iecm technical documentation: development and application of … water... · 2018. 1. 11. · from...

IECM Technical Documentation:

Development and Application of the Integrated Environmental Control Model for Power Plant Water Use Assessments

(Final Report)

April 2016

Disclaimer

This work was supported by the National Energy Technology Laboratory via a subcontract (Grant No. PO1508166-SNL-Rubin) from the Sandia National Laboratories. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference therein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed therein do not necessarily state or reflect those of the United States Government or any agency thereof.


Development and Application of the Integrated Environmental Control Model for Power Plant Water Use Assessments

(Final Report)

RES Activity No. 14-017626: Exploring Energy-Water Issues in the United States

Prepared for:

Sandia National Laboratories Albuquerque, NM 87185

www.sandia.gov/index.html

National Energy Technology Laboratory Pittsburgh, PA 15236

www.netl.doe.gov

Prepared by:

Jeffrey Anderson Kyle Borgert

Karen Kietzke Hari Chandan Mantripragada

Yang Ou Edward S. Rubin

Haibo Zhai

The Integrated Environmental Control Model Team Carnegie Mellon University

Pittsburgh, PA 15213 www.iecm-online.com

April 2016

This document brings together the five Technical Reports prepared under this contract to document enhancements made to the Integrated Environmental Control Model (IECM). Additional documentation of the model prepared under previous contracts is available at the IECM Web site: http://www.iecm-online.com.

Table of Contents Volume I: Hybrid Cooling System

Volume II: Direct Contact Cooler

Volume III: Carbon Capture Process Water Use

Volume IV: Life Cycle Water Use

Volume V: Technical, Water, and Economic Impacts of Low-Carbon Electricity Generation

IECM Technical Documentation: Volume I

Hybrid Cooling Systems for Coal- and Natural-Gas-fired Power Plants with and without Carbon Capture and Storage

April 2016

Disclaimer

This work is supported by the National Energy Technology Laboratory via the Sandia National Laboratories. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference therein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed therein do not necessarily state or reflect those of the United States Government or any agency thereof.


Hybrid Cooling Systems for Coal- and Natural-Gas-fired Power Plants with and

without Carbon Capture and Storage


Prepared for:




www.netl.doe.gov

Prepared by: The Integrated Environmental Control Model Team

Carnegie Mellon University Pittsburgh, PA 15213 www.iecm-online.com

April 2016

Integrated Environmental Control Model - Technical Documentation Table of Contents • v

Table of Contents

Hybrid Cooling Systems 1 Objectives of this Report ...................................................................................1 Introduction ........................................................................................................1 Materials and Methods .......................................................................................2 Base Case Results ..............................................................................................6 Sensitivity Analysis ...........................................................................................9 Comparisons of Different Cooling Technologies ............................................12 Conclusions ......................................................................................................13 Acknowledgements ..........................................................................................13 References ........................................................................................................14 Supporting Information ....................................................................................16

Integrated Environmental Control Model - Technical Documentation List of Figures • vi

List of Figures

Figure 1 Hybrid Cooling System ....................................................................................................... 3

Figure 2 Effects of wet unit cooling duty at coal-fired power plant without CCS (a) annual average makeup water use (b) total annualized cost of cooling system (c) cooling system LCOE (d) power plant LCOE ................................................................................................... 10

Figure 3 Effects of CO2 capture efficiency at coal-fired power plant (a) annual average makeup water use (b) cooling system LCOE ........................................................................... 11

Figure 4 Effects of capacity factor and fixed charge factor on cooling system LCOE at power plants without CCS (a) PC (b) NGCC ...................................................................................... 11

Figure 5 Comparisons of different cooling technologies (a) makeup water use (b) cooling system LCOE .............................................................................................................................. 12

Integrated Environmental Control Model - Technical Documentation List of Tables • vii

List of Tables

Table 1Major Technical and Economic Parameters and Assumptions for Base Plants .............. 6

Table 2 Ambient Air Reference Conditions ..................................................................................... 7

Table 3 Performance and Costs of Hybrid Cooling Systems for Coal- and Natural-Gas-Fired Power Plants with and without CCS .......................................................................................... 8

Table S-4 Capital Cost Components of Wet Cooling Unit ............................................................ 16

Table S-5 Operating and Maintenance Cost Components of Wet Cooling Unit ........................ 16

Table S-6 Capital Cost Components of Dry Cooling Unit ............................................................ 17

Integrated Environmental Control Model - Technical Documentation Acknowledgements • viii

Acknowledgements

This work was supported by the National Energy Technology Laboratory via a subcontract from the Sandia National Laboratories. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not reflect the views of any agency.

Integrated Environmental Control Model - Technical Documentation Hybrid Cooling Systems • 1

Hybrid Cooling Systems

Objectives of this Report This study developed and applied a new power plant modeling option for a hybrid cooling system at coal- or natural-gas-fired power plants with and without amine-based carbon capture and storage (CCS) systems.

Introduction Energy production highly depends on water. Thermoelectric power plants account for approximately 40% of U.S. freshwater withdrawals, 1 principally for cooling. Thus, thermoelectric power generation is vulnerable to water availability, which has already been affected by climate change.2 To mitigate climate change, future electricity generation will be increasingly under the pressure of controlling carbon emissions. Carbon capture and storage (CCS) is a key option for deeply cutting carbon dioxide (CO2) emissions from existing and new fossil fuel-fired power plants.3-4 However, current amine-based CCS requires a large amount of cooling water for the CO2 capture process.5 To limit CO2 emissions, the U.S. Environmental Protection Agency (EPA) has established emission performance standards for new, modified, and reconstructed power plants.6 The addition of amine-based CCS to comply with the current standard of 1400 pounds of CO2 per megawatt‐hour (lb CO2/MWh‐gross) would increase plant water use by about 17% at pulverized coal-fired plants using wet cooling towers.7 More stringent emission standards would remarkably elevate water use due to the increased CO2 removal requirements.8 Advanced cooling systems can be deployed to enhance the resilience of thermoelectric power generation systems.9

The U.S. EPA has issued regulations on cooling intake structures under Section 316(b) of the Clean Water Act,10 promoting a shift from once-through cooling to wet closed-loop cooling systems. This shift would significantly decrease national water withdrawals but increase national water consumption in the future, especially when CCS is deployed.9 In regions suffering from severe drought, limited water availability can constrain the plant operations.2 Dry cooling systems can be employed to significantly reduce water use. However, compared to wet cooling systems, dry cooling systems are much more cost-intensive, and their cooling efficiency can pronouncedly drop in hot periods.11 Thus, deployment of a hybrid cooling system that provides a compromise between the cooling efficiency and the water conservation appears attractive.12 Currently, however, there are only five hybrid cooling systems installed in the United States.13


From recent review of water use at U.S. thermoelectric power plants, Badr et al found that water usage information for hybrid cooling systems is inadequate.12 Rare engineering-economic studies on hybrid cooling are available.12,14-16 The effects of hybrid cooling deployment on the overall plant efficiency and cost of electricity generation have not been investigated thoroughly. Furthermore, bare studies have been conducted to examine hybrid cooling systems for fossil fuel-fired power plants under carbon constraints. Thus, a recent review study emphasized that efforts are needed to investigate performance and cost penalties of hybrid cooling systems for fossil fuel-fired power plants using CCS.15 Although some studies made comparisons of water use and cost among wet, dry and hybrid cooling systems, they did not address the discrepancy among collected data because of the differences in ambient air conditions, power plant and cooling system designs as well as economic assumptions.12,14-16 Such comparisons should be made under a common framework.

The major objectives of this paper, therefore, are to: (1) evaluate the plant-level performance and cost of hybrid cooling systems for water conservation at pulverized coal (PC) or natural gas combined cycle (NGCC) power plants with and without CCS; (2) examine the effects of key parameters on hybrid cooling systems; and (3) compare performance and cost among different cooling technologies under a common platform. The key performance metric considered is the annual average makeup water use for hybrid cooling systems, while the major cost metrics considered are the total annualized cost and total levelized cost of electricity. Commercially available amine-based CCS is deployed for CO2 capture when applicable.

Materials and Methods In this study, performance and cost models of hybrid cooling systems are developed and incorporated in a power plant modeling tool. Then, the newly enhanced power plant tool is employed for plant-level case studies. Integrated Environmental Control Model for Power Plant Assessments

The Integrated Environmental Control Model (IECM) is a computer tool developed by Carnegie Mellon University to perform systematic estimates of the performance, resource use, emissions, and costs for fossil-fuel-fired power plants with and without CCS.17 The IECM has an array of major cooling technologies including once-through cooling, wet towers, and dry cooling.5,11 The performance and cost models outlined below for hybrid cooling systems are incorporated in the IECM. The costing method and nomenclature employed in this study are based on the Electric Power Research Institute’s Technical Assessment Guide.18 This study employed the 2015 release of IECM Version 9.1 for assessments.17

Performance and Cost Models of Hybrid Cooling System

This study investigates a hybrid cooling system that uses both closed-loop dry and wet units, which is shown in Figure 1. Dry and wet cooling units are arranged in parallel that splits the steam flow between air-cooled condensers (ACC) and a surface condenser coupled with a wet tower unit. The dry cooling unit employs ACC and is primarily used to serve the steam cycle.


When the ambient air temperature reaches higher levels than the design, and the dry cooling unit cannot maintain a low turbine exhaust pressure, part of the exhaust steam is routed to the supplemental wet unit. Thus, this design allows ACC to reject less exhaust heat and operate at a smaller initial temperature difference (ITD) between condensing steam temperature and inlet air temperature.19-20

Figure 1 Hybrid Cooling System

Wet Cooling Unit The required amount of cooling water is determined in terms of the plant size, steam cycle

heat rate, and cooling water temperature drop across the tower.5 Given that only part of the exhaust steam is routed to the wet unit in summer, the total amount of cooling water in the wet unit for a PC power plant without CCS is then estimated as:5

( ) ( ){ }1000

110003600⋅∆⋅

+⋅⋅⋅−⋅=

wp

auxc Tc

MWgHrm ηα (1a)

where α is the fraction of the total cooling duty assigned for wet cooling unit in summer (fraction); Hr is the steam cycle heat rate (kJ/kWh) (3600 = units conversion factor); cm is the total recirculation cooling water (tonnes/hr); MWg is the plant gross size (MWe); wT∆ is the cooling water temperature drop range (oC); auxη is the auxiliary cooling load (%); and pc is the water specific heat capacity (kJ/kg∙ P

oC). When the CO2 is captured by CCS, an amount of cooling water is needed for the capture

process. So, the wet cooling unit has to serve in both summer and non-summer seasons. However, the heat rejected around the primary condenser excludes the thermal energy of the steam extracted for solvent regeneration. The larger cooling duty in summer relative to other seasons is adopted to size the wet cooling unit. The total amount of cooling water is then estimated for a PC plant with CCS as:5

( ) ( ){ } CCSc

wp

CCSraux

c mTc

qMWgHrm +

⋅∆⋅−+⋅⋅⋅−⋅

=1000

110003600 ηα

(1b)

where CCScm is the amount of cooling water required for the capture system (tonnes/hr); MWg is

the plant gross size (MWe); and CCSrq is the extracted steam heat for solvent regeneration (kJ/hr).

Condensate

Exhaust Steam


In the wet unit, the cooling water is cooled by contact with ambient air and then recirculated back to the condenser to cool the exhaust steam. The wet cooling unit relies mainly on the latent heat of water evaporation for transferring exhaust heat to the atmosphere. Water is used to make up evaporation, drift and blowdown losses. A mass and energy balance model developed in the previous study is adopted to estimate various water losses around the tower.5, 21

The engineering-economic model of a wet cooling system in the IECM is adopted to estimate the capital and O&M costs for the wet cooling unit.5,17, 21 The total capital requirement includes the direct costs plus indirect costs such as the general facilities cost, engineering and home office fees, contingency costs, and owner’s costs. The indirect costs such as general facilities cost, engineering and home office fees, and process contingency cost are empirically estimated as a percent of process facilities capital cost (PFC), whereas the project contingency cost is estimated as a percent of the sum of PFC, engineering and home office fees, and process contingency cost.18 The major direct cost components include the cooling tower structure, circulation pumps, auxiliary systems, piping, makeup water system, component cooling water system, foundation & structures, and tower structure. The variable operating and maintenance (O&M) costs are considered only when the wet cooling unit is in operation. The capital and O&M cost components are detailed in Tables S-1 and S-2 of the Supporting Information (SI), respectively.

Dry Cooling Unit

The dry cooling unit utilizes the sensible heating of atmospheric air passed across finned-tube heat exchangers to reject exhaust heat. There is no water used in the cooling process. In the non-summer seasons, the amount (kJ/hr) of exhaust heat rejected by the dry cooling unit is estimated as:

( ) ( )auxMWgHrQ η+⋅⋅⋅−= 110003600 (2a)

In summer, a large portion of the exhaust steam is routed to the dry unit, while the rest is delivered to the wet unit. The corresponding amount (kJ/hr) of exhaust heat rejected by the dry cooling unit is estimated as:

( ) ( ){ }auxMWgHrQ ηα +⋅⋅⋅−⋅−= 110003600)1(' (2b)

When the CO2 is captured by amine-based CCS, the heat rejected around the dry cooling unit does not include the thermal energy of the steam extracted for solvent regeneration. The wet cooling unit provides the cooling water required for the CO2 capture process.

The performance of dry cooling unit highly depends on the ITD between condensing steam temperature and inlet air temperature.11 The ACC plot area and fan shaft power normalized by the rejected heat are estimated using a reduced-order model:5

7401.00344.1

3.10159346.462

−−

×

×= ambientPITDa

(3)


2709.00227.1

3.10159024.848

×

×=

−ambientPITDe

(4)

Where a is the ACC plot area normalized by rejected heat (m2/MWt); e is the required fan shaft power normalized by rejected heat (kW/MWt); ITD is the initial temperature difference (oC); and

ambientP is the ambient pressure (kPa). ITD is a key parameter affecting the dry unit size and power use and varies with season. The ACC plot area and power use on an absolute basis are estimated as the product of the normalized performance parameter ( a or e ) and the cooling duty, respectively. Eq. 2 shows that the dry unit's cooling duties are different in summer versus non-summer seasons. To determine the dry unit size, the plot area is first estimated based on the cooling duty and air temperature in summer and fall or spring, respectively. Then, the larger plot area is chosen to size the dry cooling unit. A similar analysis is conducted to estimate the power use of dry unit in summer and non-summer seasons.

The engineering-economic model of a dry cooling system in the IECM is adopted to estimate the capital and O&M costs for the dry cooling unit.5,17, 22 The ACC equipment cost is estimated as a function of cooling duty and ITD:5

( )

=

−

oequip Q

QITDMC0055.1

594.663$

(5)

Where equipC is the ACC equipment capital cost (2003 US dollars); Q is the actual cooling duty

(MWt); and oQ is the reference cooling duty (288 MWt). In addition to the equipment cost, the equipment erection cost is approximately 30% of the sum of the equipment and erection costs.5 The major direct cost components include the condenser structure, steam duct support, electrical & control equipment, auxiliary cooling, and cleaning system, which are summarized in SI Table S-3. Except for water use, the dry cooling unit has O&M cost components similar to those summarized in SI Table S-2 for the wet cooling unit. However, the variable O&M costs are always considered for the dry cooling unit since it provides the primary cooling service.

Overall Cooling System Water Use and Costs

The water use intensity of a hybrid cooling system is estimated as the total annual water use divided by the annual net electricity generation of a power plant, whereas the cooling system's capital, O&M, and LCOE are estimated as the sum of those outlined above for both wet and dry cooling units. The LCOE is calculated as:23

LCOE=TCR∙FCF+FOM(CF∙Hrs)∙MW

+VOM+HR∙FC (6)

Where LCOE is the levelized cost of electricity generation ($/MWh); TCR is the total capital requirement ($); CF is the capacity factor (%); FCF is the fixed charge factor (fraction/yr); FOM is fixed O&M costs ($/yr); VOM is the variable non-fuel O&M costs ($/yr); HR is the net plant heat rate (GJ/MWh); FC is the unit fuel cost ($/GJ); MW is the net power output (MW); and Hrs is the total annual hours (hrs/yr).


Base Case Results Table 1Major Technical and Economic Parameters and Assumptions for Base Plants

Parameter Type Parameter Value

Technical Plant type Supercritical pulverized coal or GE-7FB NGCC

Plant capacity factor (%) 75%

Ambient pressure (MPa) 0.10

Environmental control systems (if applicable) Nitrogen oxide Selective catalytic reduction

Particulates Electrostatic precipitator

Sulfur dioxide Flue gas desulfurization

Carbon dioxide CCS

Hybrid cooling: dry unit Air-cooled condenser plot area per cell (m2) 110

Configuration of air-cooled heat exchanger Multiple-row

Turbine backpressure (bar) 0.14

Fan efficiency (%) 80

Auxiliary cooling load (%) 5

Hybrid cooling: wet unit Cooling duty in summer (fraction of the total) 0.3

Duration of service in summer (month) 3

Cooling water temperature drop range (oC) 11.1

Cycle of concentration (ratio) 4

Auxiliary cooling load (%) 5 Economic Dollar type 2012 Constant

Fixed charge factor 0.113

Plant book lifetime (years) 30

Water cost ($/m3) 0.3

Coal price ($/tonne) 42

Natural gas price ($/GJ) 6.9

Electricity price ($/MWh) 50

Hybrid cooling system (dry/wet unit if noted) General facilities capital (% of PFC) 10

Engineering & office fees (E) (% of PFC) 10

Process contingency cost (% of PFC) 5/0

Project contingency cost

(% of PFC+ E + process contingency) 20/10

Number of operating jobs (#) 2

Number of operating shifts (#) 4.75

Labor rate ($/hr) 34.65 Total maintenance cost (% of total plant cost) 1.5


The enhanced IECM was applied to evaluate hybrid cooling systems at baseload supercritical pulverized coal and NGCC plants. The base plants were configured using the IECM to comply with the new source performance standards for traditional air pollution controls, in which hybrid cooling systems were assumed to be installed. Major parameters and assumptions for the base plants and cooling systems are summarized in Table 1.

Coal and natural gas properties are summarized in SI Table S-4. Otherwise, the IECM default values are adopted. In this study, all costs are reported in 2012 constant dollars; all PC plants are evaluated on a basis of 550 MW net power output; and the gross power output of NGCC plants is determined by the gas turbine type and the number of gas turbines. Thus, the plant fuel requirement is 4246 GJ/hr for both the base NGCC plants with and without CCS. It was assumed that the thermal energy for solvent regeneration is extracted from the steam cycle, which results in a larger fuel requirement for the PC plant with CCS and an amount of losses in gross generating capacity for the NGCC plant with CCS, compared to the corresponding non-capture cases.

Eq. 1(a), 1(b), and 2(b) indicate that the allocation of the plant cooling duty between the dry and wet units in summer is a key factor that affects the size and performance of individual cooling units. In one previous study on geothermal power plants, the wet unit was sized to handle about 30% of the overall cooling duty.18 Besides, the total annual water use of wet cooling unit also depends on its duration of service within a year. In the base cases, the wet cooling unit is designed to share 30% of the overall cooling duty during the summer period from June to August.

The performance of wet and dry cooling units is affected by ambient air conditions.11 This study refers to the monthly meteorological data in Texas from 1981 to 2010 for designing and evaluating hybrid cooling systems.24 Table 2 summarizes the ambient air conditions adopted for assessments. The monthly average air temperature and relative humidity (RH) in summer fall within ranges from 26.6 oC to 28.1oC (in July) and 45.2% to 47.8%, respectively. The average air temperature and RH in summer were used to evaluate the cooling system performance during that period. In non-summer seasons, the corresponding average air temperature was used to evaluate the dry unit performance as well as the wet unit performance when CCS is implemented. However, the highest seasonal air temperature was referred to in sizing the dry cooling unit.

Table 2 Ambient Air Reference Conditions

Time Scale Average Temperature (oC) Average Relative Humidity (%) Annual 18.7 44.6 Summer 27.6 46.7 Non-Summer 15.7 43.9

Fall 19.3 45.7 Spring 18.7 40.8 Winter 9.2 45.2

The major performance parameters of CCS systems are summarized in SI Table S-5.25-26 As Eq. 1(b) indicates, the implementation of CCS affects power plant performance in two major


areas: the steam cycle and the cooling system. For 90% CO2 capture, the thermal energy extracted from the steam cycle for solvent regeneration is approximately 3500 kJ/kg CO2 captured for the PC plant and 3950 kJ/kg CO2 captured for the NGCC plant; and the total cooling duty for the CO2 absorption and stripping processes and CO2 product compression is 91 ton H2O/ton CO2 for the PC plant and 123 ton H2O/ton CO2 for the NGCC plant. Thus, the overall cooling duty of wet cooling unit includes the additional cooling requirement for CCS but excludes the extracted thermal energy from the steam cycle.

Table 3 Performance and Costs of Hybrid Cooling Systems for Coal- and Natural-Gas-Fired Power Plants with and without CCS

Parameter SC PC NGCC CCS (Yes or No) No Yes No Yes Gross power output (MWg) 595 686 600 544 Net power output (MWnet) 550 550 581 502 Net plant efficiency (HHV,%) 36.9 26.4 49.3 42.5 Plant CO2 emission rate (kg/kW-net) 0.85 0.12 0.37 0.04 Hybrid cooling system Annual average makeup water use (L/kWh) 0.19 1.73 0.09 0.86

Parasitic load (% of MWg) 2 3 1 1 Number of ACC cells (#) 46 39 23 24

Annual fixed cost (M$/yr) 3.0 3.4 2.0 2.4 Annual variable cost (M$/yr) 3.4 7.9 1.7 2.8

Annualized capital cost (M$/yr) 14.2 17.3 7.1 9.8 Total annualized cost (M$/yr) 20.5 28.6 10.7 14.9

Cooling system LCOE ($/MWh) 5.6 7.8 2.8 4.5 Power plant LCOE ($/MWh) 65.3 110 67.8 93.7

Table 3 summarizes the major results of the base power plants and hybrid cooling systems.

For the two base cases without CCS, the makeup water use of hybrid cooling system, equivalent to the water withdrawal, is 0.19 L/kWh for the PC case and 0.09 L/kWh for the NGCC case. This mainly offsets evaporation and blowdown losses in the wet unit. For the given cycle of concentration of the wet cooling unit, the evaporation loss accounts for 75% of the total makeup water use. This means that the cooling system's water consumption is 0.14 L/kWh for the PC case and 0.07 L/kWh for the NGCC case. Given the large amounts of power and thermal energy required for CCS operation, the implementation of amine-based CCS for 90% CO2 capture would decrease the net plant efficiency by about ten percentage points for the PC case and about seven percentage points for the NGCC case. When CCS is added to power plants, the resulting makeup water use increases by a factor of 8–9 at both the base PC and NGCC plants, compared to the non-capture cases, because the wet cooling unit primarily provides the CO2 capture process with cooling service throughout the year. In comparison between PC and NGCC plants, the makeup water use of hybrid cooling systems at NGCC plants with and without CCS is just about half that of PC plants, mainly because an NGCC plant has a much higher plant efficiency than a PC plant, and its cooling system only needs to serve the steam generation loop of the combined power cycle.8


For the given economic assumptions in Table 1, for plants without CCS, the wet cooling unit accounts for only 14% of the total capital requirement of the cooling system for the PC case and 15% for the NGCC case. However, with CCS, the wet cooling unit's capital cost share increases to 42% and 35% for the PC and NGCC cases, respectively. As shown in Table 3, the annualized capital costs of hybrid cooling systems account for roughly 60–70% of the total annualized cost of the cooling systems for the four base plants. However, the addition of CCS to the PC plant increases the plant and cooling system LCOE by roughly 70% and 40%, respectively. In contrast, the addition of CCS to the NGCC power plant increases the plant and cooling system LCOE by roughly 40% and more than 60%, respectively. These results clearly indicate that CCS deployment will have a big effect on hybrid cooling system cost. In addition, the cooling system levelized cost accounts for only 7–9% of the total plant LCOE for the PC plants with and without CCS, and only 4–5% for the NGCC plants with and without CCS. These results indicate that a hybrid cooling system is not a major contributor to overall plant cost, and suggest that retrofitting a hybrid cooling system to existing plants using wet cooling towers may not lead to substantial increases in the overall plant LCOE.

Sensitivity Analysis The cooling duty assigned to the wet unit and the CO2 capture are the key factors affecting hybrid cooling systems. Besides, as Eq. (6) shows, capacity factor and fixed charge factor also are the key parameters affecting the LCOE.3, 27-28 Thus, a sensitivity analysis was conducted for these key parameters. In each parametric analysis, other parameters were kept at the base case values given in Table 1, unless otherwise noted. Effects of Wet Unit Cooling Duty Assignment

The fraction of total plant cooling duty assigned to the wet unit and its duration of service in summer directly affect the makeup water use and size of the wet cooling unit, which in turn affect the overall plant and cooling system. The cooling duty fraction for the wet cooling unit was varied from 0.1 to 0.9, while the duration of summer service was varied from three months to one month. For the ambient air conditions given in Table 2, the monthly air temperature reaches the highest level in July. For the one-month service case (on wet cooling unit), the dry cooling unit size was determined by first estimating the required plot area based on the cooling duty and air temperature in July, then comparing it to the plot area based on the cooling duty and average air temperatures in June and August, respectively. The difference between the design temperatures is less than 2 oC, whereas the dry unit cooling duty is larger in June and August than in July. The resulting larger plot area was chosen to size the dry cooling unit.

Using the PC plant without CCS as an illustrative example, Figure 2 shows how these two parameters affect the cooling system's makeup water use, capital cost and levelized cost of electricity generation, as well as the plant-level LCOE. For either service duration design, the makeup water use monotonically increases with the cooling duty fraction of the wet cooling unit. However, the costs show a different trend. When the wet cooling unit is designed for three-month cooling service, the lowest costs occur when the wet cooling unit duty fraction is 0.3. In contrast, when the wet cooling unit is designed for one-month service all costs increase monotonically with the cooling duty fraction . This is because, although the shorter duration of


service results in a smaller makeup water requirement, it also requires a larger dry cooling unit, which leads to larger costs for both the overall cooling system and the overall plant. For example, when the wet unit cooling duty fraction is designed to be 0.3, the number of air cooled condenser cells is 46 for the three-months summer service case and 62 for the one-month service case. This finding implies that designing the wet cooling unit for three months of summer service, rather than one month, would reduce both the total capital cost and total levelized cost of a hybrid cooling system if high air temperatures occurred throughout the entire summer.

(a) (b)

(c) (d) Figure 2 Effects of wet unit cooling duty at coal-fired power plant without CCS (a) annual average makeup water use (b) total annualized cost of cooling system (c) cooling system LCOE (d) power plant LCOE

Effects of CO2 Capture Efficiency

To comply with the U.S.EPA's current emission standards, a supercritical coal-fired plant needs to remove about 20% of its carbon pollution.7 However, there is no CO2 capture needed for modern NGCC plants. Thus, we examined the effects on the hybrid cooling system of CO2 capture efficiency starting from 20% at PC plants. The bypass design was adopted for amine-based CCS because it is a cost-effective option for partial CO2 capture.29 Figure 3 shows that the makeup water use of hybrid cooling system increases by a factor of nearly 3 due to the increased cooling water use for CCS when the CO2 capture efficiency is elevated from 20% to 90%.

0.00

0.25

0.50

0.75

1.00

0.0 0.2 0.4 0.6 0.8 1.0

Ann

ual A

vera

ge M

akeu

p W

ater

U

se (L

/kW

h)

Cooling Duty of Wet Unit in Summer (fraction)

June to August

July

15

18

21

24

27

30

0.0 0.2 0.4 0.6 0.8 1.0To

tal A

nnua

lized

Cos

t of C

oolin

g Sy

stem

(201

2M$/

yr)


June to AugustJuly

2

4

6

8

10

0.0 0.2 0.4 0.6 0.8 1.0

Coo

ling

Syst

em L

CO

E (2

012$

/MW

h)


June to AugustJuly

62

64

66

68

70

0.0 0.2 0.4 0.6 0.8 1.0Pow

er P

lant

LC

OE

(201

2$/M

Wh)


June to AugustJuly


Meanwhile, the overall system LCOE increases by about $1/MWh, though the addition of CCS significantly increases the overall plant LCOE.

(a) (b) Figure 3 Effects of CO2 capture efficiency at coal-fired power plant (a) annual average makeup water use (b) cooling system LCOE

Effects of Capacity Factor and Fixed Charge Factor

Further parametric analysis was conducted to evaluate the impacts of capacity factor (CF) and FCF assumptions. Figure 4 shows the resulting effects on the cooling system LCOE at the base PC and NGCC plants without CCS. The ranges of parameter values shown in Figure 4 cover the values that were often adopted in literature.3, 28 For a given CF, the cooling system LCOE increases by 21–22% when the FCF is increased from the base value (given in Table 1) to 0.15. For a given FCF, the cooling system LCOE decreases by about 20% when the plant CF is increased from 65% to 85%. These two parameters have pronounced effects on the levelized costs of cooling systems.

(a) (b) Figure 4 Effects of capacity factor and fixed charge factor on cooling system LCOE at power plants without CCS (a) PC (b) NGCC

0.00

0.50

1.00

1.50

2.00

10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Ann

ual A

vera

ge M

akeu

p W

ater

U

se o

f Coo

ling

Syst

em (L

/kW

h)

CO2 Capture Efficiency

2

4

6

8

10

10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Coo

ling

Syst

em L

CO

E (2

012$

/MW

h)

CO2 Capture Efficiency

2

4

6

8

0.11 0.12 0.13 0.14 0.15

Coo

ling

Syst

em L

CO

E in

PC

Pl

ant (

2012

$/M

Wh)

Fixed Charge Factor (fraction)

CF=65%CF=75%CF=85%

2

4

6

8

0.11 0.12 0.13 0.14 0.15

Coo

ling

Syst

em L

CO

E in

NG

CC

Pl

ant (

2012

$/M

Wh)

Fixed Charge Factor (fraction)

CF=65%CF=75%CF=85%


Comparisons of Different Cooling Technologies A switch from once-through cooling to advanced cooling systems has been considered widely as a major strategy to reduce plant water use.9 To examine the implications, we quantify the tradeoffs in water use and cost among different cooling technologies. For the illustrative purposes, the IECM was applied to evaluate and compare the performance and costs of wet, dry, and hybrid cooling systems at the base PC and NGCC plants without CCS. The annual average ambient air conditions in Table 2 were used for assessing the wet tower systems, while the average summer conditions were used to size the dry cooling systems. Other major parameters of wet and dry cooling systems were assumed to be the same as those given in Table 1.

The PC plant using a wet tower system was found to have a net plant efficiency of 38.3%, compared to 36.9% for the PC plants using dry or hybrid cooling systems (which have similar parasitic loads). Similarly, the net plant efficiency of the NGCC plant using a wet tower system is 0.7 percentage point higher than that of the NGCC plants using hybrid or dry cooling systems. Figure 5 compares the makeup water use and LCOE of the three cooling technologies. Compared to the wet cooling system, the hybrid cooling system has much lower water use but 70% higher LCOE for the PC case and 60% higher LCOE for the NGCC case. This results in a $2–4/MWh increase in the plant LCOE for the given assumptions in Table 1. Although dry cooling has no water requirement, it has the highest LCOE among the three cooling systems. However, as shown in Figure 5(b), the difference in the LCOE between dry and hybrid cooling systems is small, which is about $1/MWh for the PC case and $0.4/MWh for the NGCC case.

A further comparison shows that the makeup water use of wet tower cooling at the PC plant without CCS is similar to that of hybrid cooling for the PC plant with 90% CO2 capture. However, as shown in Figure 3(a), hybrid cooling for cases with partial CO2 capture requires less makeup water. This result implies that for PC plants subject to the U.S. EPA's current CO2 emission standards, a switch from wet cooling to hybrid cooling can significantly decrease plant water use. Given the tradeoffs and factors discussed above, therefore, the choice of an appropriate cooling system should be made based on a full consideration of performance, costs, regulations, and resource availability.

(a) (b) Figure 5 Comparisons of different cooling technologies (a) makeup water use (b) cooling system LCOE

0.0

0.5

1.0

1.5

2.0

2.5

Wet Cooling Hybrid Cooling Dry Cooling

Coo

ling

Syst

em M

akeu

p W

ater

(L

/kW

h)

PCNGCC

0.0

2.0

4.0

6.0

8.0

Wet Cooling Hybrid Cooling Dry Cooling

Coo

ling

Syst

em L

CO

E (2

012$

/MW

h)

PCNGCC


Conclusions This study provides systematic estimates of the performance and cost of hybrid cooling systems at PC and NGCC plants. Their cost and performance depend on a range of environmental, technical and economic parameters. In general, however, hybrid cooling systems were found to significantly reduce water use, compared to prevailing wet cooling systems (e.g., reductions of roughly 91% for PC and NGCC cases without CCS). These reductions in water use, however, come with some increase in overall plant LCOE (e.g., increases of 3–5% for PC and NGCC plants without CCS). Furthermore, hybrid systems still require some makeup water in summer, and their operational control is generally more complex.

In order to minimize overall system cost in regions where hot periods last the entire summer, the wet unit of a hybrid cooling system needs to handle about 30% of the overall plant cooling load during summer months. To lower capital investments on hybrid systems, more research and development (R&D) efforts are needed to improve the dry cooling unit performance, so as to decrease the size, footprint and cost of the required air cooled condenser.

The addition of CCS systems to comply with current and future CO2 emission regulations can lead to a significant increase in power plant water use. A previous analysis for plants employing a wet cooling system showed that 90% CO2 capture using amine-based CCS increases water use by more than 80%, relative to a plant without CCS producing the same net power.5 With hybrid cooling systems, the annual makeup water requirement with CCS would still be greater than without CCS. But, compared to systems with wet cooling,5 the hybrid system water use would be substantially smaller with or without CCS (e.g., 90% less without CCS and 52% less with CCS).

Because limitations on water availability due to drought, population growth and other factors may become increasingly common in the future, water use metrics also need to be considered in R&D programs for advancing carbon capture technologies and planning water resources for energy production, especially under carbon constraints. This paper has shown that hybrid cooling technology can be an adaptive option to improve the resilience of fossil-fuel-based electricity generation, especially in the face of CO2 emission regulations. For both coal- and gas-fired power plants, hybrid cooling systems can substantially reduce water use in the electric power sector with only moderate impacts on the overall plant-level cost of electricity generation. Hybrid cooling, as an alternative to the conventional wet cooling technology, thus offers a promising option for future power plant designs.

Acknowledgements This work was supported by the National Energy Technology Laboratory via a subcontract from the Sandia National Laboratories. The authors acknowledge Karen Kietzke for assistance with the IECM computer code and Shuchi Talati for assistance with the ambient air data. Any


opinions, findings, conclusions or recommendations expressed in this material are those of the authors alone and do not reflect the views of any agency.

References (1) Kenny, J.F.; Barber, N.L.; Hutson, S.S.; Linsey, K.S.; Lovelace, J.K.; Maupin, M.A.

Estimated Use of Water in the United States in 2005. Reston, VA: U.S. Geological Survey, 2009.

(2) U.S. Department of Energy. The Water-Energy Nexus: Challenges and Opportunities. June 2014.

(3) Zhai, H.; Rubin, E. S. Comparative performance and cost assessments of coal-and natural-gas-fired power plants under a CO2 emission performance standard regulation. Energy Fuels 2013, 27(8), 4290–4301.

(4) Zhai, H.; Ou, Y.; Rubin, E. S. Opportunities for decarbonizing existing U.S. coal-fired power plants via CO2 capture, utilization and storage. Environ. Sci. Technol. 2015, 49 (13), 7571–7579.

(5) Zhai, H.; Rubin, E. S.; Versteeg, P. L. Water use at pulverized coal power plants with postcombustion carbon capture and storage. Environ. Sci. Technol. 2011, 45(6), 2479–2485.

(6) U.S. Environmental Protection Agency. 40 CFR Parts 60, 70, 71, and 98 [EPA–HQ–OAR–2013–0495; EPA–HQ–OAR–2013–0603; FRL–9930–66–OAR] RIN 2060–AQ91 Standards of Performance for Greenhouse Gas Emissions From New, Modified, and Reconstructed Stationary Sources: Electric Utility Generating Units. Federal Register 2015, Vol. 80, No. 205, October 23.

(7) Ou, Y.; Zhai, H.; Rubin E. S. Life cycle water use of coal- and natural-gas-fired power plants with and without carbon capture and storage. Int. J. Greenhouse Gas Control 2016, 44, 249–261.

(8) Talati, S.; Zhai, H.; Morgan, M. G. Water impacts of CO2 emission performance standards for fossil fuel-fired power plants. Environ. Sci. Technol. 2014, 48(20): 11769–11776.

(9) Zhai, H.; Rubin, E. S. Water impacts of a low-carbon electric power future: assessment methodology and status. Current Sustainable/Renewable Energy Reports 2015, 2(1), 1–9.

(10) U.S. Environmental Protection Agency. 40 CFR Parts 122 and 125 [EPA–HQ–OW–2008–0667, FRL–9817–3] RIN 2040–AE95 National Pollutant Discharge Elimination System—Final Regulations To Establish Requirements for Cooling Water Intake Structures at Existing Facilities and Amend Requirements at Phase I Facilities. Federal Register 2014, Vol.79, No. 158, August 15.

(11) Zhai, H.; Rubin, E. S. Performance and cost of wet and dry cooling systems for pulverized coal power plants with and without carbon capture and storage. Energy Policy 2010, 38(10), 5653–5660.

(12) Badr, L.; Boardman, G.; Bigger, J. Review of water use in US thermoelectric power plants. J. Energ. Eng. 2012, 138(4), 246–257.

(13) U.S. Energy Information Administration. Today in Energy: Many newer power plants have cooling systems that reuse water. February 11, 2014. https://www.eia.gov/todayinenergy/detail.cfm?id=14971. Accessed in December 2015.


(14) Electric Power Research Institute. Comparison of Alternate Cooling Technologies for U.S. Power Plants: Economic, Environmental, and Other Tradeoffs. Report No. 1005358. Electric Power Research Institute, Palo Alto, CA, August 2004.

(15) Macknick, J.; Newmark, R.; Heath, G.; Hallett, K. C. Operational water consumption and withdrawal factors for electricity generating technologies: a review of existing literature. Environ. Res. Lett. 2012, 7(4), 045802.

(16) Webster, M.; Donohoo, P.; Palmintier, B. Water-CO2 trade-offs in electricity generation planning. Nat. Clim. Change 2013, 3(12), 1029–1032.

(17) Rubin, E.S.; Zhai, H.; Kietze, K. Integrated Environmental Control Model, Version 9.1, Public Release, Carnegie Mellon University, Pittsburgh, PA, 2015. available at: www.iecm-online.com.

(18) Electric Power Research Institute. TAGTM –Technical Assessment Guide Vol.1:Electricity Supply, Rev.7; ReportNo.TR-102276-VIR7; Electric Power Research Institute, Palo Alto, CA, June 1993.

(19) Ashwood, A.; Bharathan, D. Hybrid Cooling Systems for Low-Temperature Geothermal Power Production. Technical Report NREL/TP-5500-48765, March 2011, National Renewable Energy Laboratory, Golden, Colorado.

(20) U.S. Department of Energy. Concentrating Solar Power Commercial Application Study: Reducing Water Consumption of Concentrating Solar Power Electricity Generation. Report to Congress. 2009. Available at https://www1.eere.energy.gov/solar/pdfs/csp_water_study.pdf. Accessed in May 2015.

(21) Zhai, H.; Berkenpas, M.; Rubin, E.S. Integrated Environmental Control Model Technical Documentation Updates Final Report Volume (I): Wet Cooling Tower Model. Final Report of Contract No. DE-AC26-04NT41917, Prepared by Carnegie Mellon University for the National Energy Technology Laboratory, Pittsburgh, PA, November 2009.

(22) Zhai, H.; Berkenpas, M.; Rubin, E.S. Integrated Environmental Control Model Technical Documentation Updates Final Report Volume (II): Air-cooled Condenser Model. Final Report of Contract No. DE-AC26-04NT41917, Prepared by Carnegie Mellon University for the National Energy Technology Laboratory, Pittsburgh, PA, November 2009.

(23) Rubin E.S. Understanding the pitfalls of CCS cost estimates. Int. J. Greenhouse Gas Control 2012, 10, 181–90.

(24) U.S. Monthly Climate Normals (1981-2010), available at https://catalog.data.gov/dataset/u-s-monthly-climate-normals-1981-2010.

(25) Rao, A. B.; Rubin, E. S. A technical, economic, and environmental assessment of amine-based CO2 capture technology for power plant greenhouse gas control. Environ. Sci. Technol. 2002, 36(20), 4467–4475.

(26) Berkenpas, M. B.; Kietzke, K.; Mantripragada, H.; McCoy, S.; Rubin, E. S.; Versteeg, P. L.; Zhai, H. Integrated Environmental Control Model Technical Documentation Updates Final Report Volume (IV): Updates to PC and IGCC Plant Models. Final Report of Contract No. DE-AC26-04NT41917, Prepared by Carnegie Mellon University for the National Energy Technology Laboratory, Pittsburgh, PA, November 2009.

(27) Rubin, E.S.; Chen, C.; Rao A.B. Cost and performance of fossil fuel power plants with CO2 capture and storage. Energy Policy 2007, 35(9), 4444–4454.

(28) Rubin, E.S.; Zhai, H. The cost of carbon capture and storage for natural gas combined cycle power plants. Environ. Sci. Technol.2012, 46(6), 3076–3084.


(29) Rao, A. B.; Rubin, E. S. Identifying cost-effective CO2 control levels for amine-based CO2 capture systems. Ind. Eng. Chem. Res. 2006, 45 (8), 2421−2429.

Supporting Information Table S-4 Capital Cost Components of Wet Cooling Unit Process Area Costs Wet Cooling Unit Costs Cooling tower structure Process facilities capital Circulation pumps General facilities capital Auxiliary systems Engineering. & home office fees Piping Process contingency cost Makeup water system Project contingency cost Component cooling water system Interest charges Foundation & structures Royalty fees Cooling tower structure Preproduction (startup) cost

Inventory capital

Process Facilities Capital (sum above) Total Capital Requirement (sum above)

Table S-5 Operating and Maintenance Cost Components of Wet Cooling Unit Variable Cost Component Fixed Cost Component Electricity Operating labor Water Maintenance labor

Maintenance material

Admin. & support labor

Total Variable Cost (sum above) Total O&M Cost (sum above)


Table S-6 Capital Cost Components of Dry Cooling Unit Process Area Costs Dry Cooling Unit Costs Condenser structure Process facilities capital Steam duct support General facilities capital Electrical & control equipment Engineering. & home office fees Auxiliary cooling Process contingency cost Cleaning system Project contingency cost

Interest charges

Royalty fees

Preproduction (startup) cost

Inventory capital

Process Facilities Capital (sum above) Total Capital Requirement (sum above) Table S-4 Fuel Properties in IECM Database Coal Property* Value Natural Gas Property Value Higher heating value (kJ/kg) 2.71E+04 Higher heating value (kJ/kg) 5.23E+04 Composition (wt.%) Composition (vol.%)

carbon 63.75 methane 93.1 hydrogen 4.5 ethane 3.2

oxygen 6.88 propane 1.1 chlorine 0.29 carbon dioxide 1.0

sulfur 2.51 oxygen 0.0 nitrogen 1.25 nitrogen 1.6

ash 9.7 hydrogen sulfide 0.0 moisture 11.12

* The Illinois #6 coal in the IECM fuel database is used as the surrogate fuel for PC plants. Table S-5 Major Performance Parameters of Amine-based Carbon Capture Systems in IECM Parameter Value CO2 removal efficiency (%) 90 Sorbent concentration (wt%) 30 Lean CO2 loading (mol.CO2/mol. solv.) 0.19 Liquid-to-gas ratio 3.06 (PC)/1.18(NGCC) Regeneration heat requirement ( kJ/kg CO2) 3524 (PC)/3954(NGCC) Heat-to-electricity efficiency (%) 18.7 (PC)/19.7(NGCC) CO2 product pressure (MPa) 13.8 CO2 compression power use (kWh/tonne CO2) 93 Cooling duty (t H2O/t CO2) 91(PC)/123(NGCC)

IECM Technical Documentation: Volume II

Direct Contact Cooler and Polishing Scrubber

April 2016

Disclaimer

Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference therein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed therein do not necessarily state or reflect those of the United States Government or any agency thereof.




Prepared for:




www.netl.doe.gov



April 2016


Table of Contents

Direct Contact Cooler and Polishing Scrubber 1

Objectives of this Report ........................................................................................................... 1 Introduction ............................................................................................................................... 1 Oxyfuel Application .................................................................................................................. 2 Modeling Approach ................................................................................................................... 3 Saturation Pressure of Water ..................................................................................................... 5 Latent Heat of Water Condensation ........................................................................................... 5 Sulfur Removal Calculation ....................................................................................................... 7 Summary of Input and Output Parameters ................................................................................. 7 Mass and Energy Balance Calculations ..................................................................................... 8

Gas Stream Flow Rate Calculations ............................................................................ 9 Reagent Stream Flow Rate Calculations ................................................................... 10 Precipitant Stream Flow Rate Calculations ............................................................... 10 Calculating the Required Cooling Load .................................................................... 11 Latent Heat of Water ................................................................................................. 13

Cooling Water Requirement Calculation ................................................................................. 13 Water Balance ........................................................................................................... 14

Auxiliary Electrical Load Requirements.................................................................................. 14 Pneumatic Head ......................................................................................................... 14 Cooling Water Pumping ............................................................................................ 14

Cost Model .............................................................................................................................. 14 Capital Cost ............................................................................................................... 15 Fixed and Variable Operations and Maintenance Cost ............................................. 16

Case Studies ............................................................................................................................. 17 References ............................................................................................................................... 18


List of Figures

Figure 1 Illustration of the physical working process utilized in the direct contact cooler polishing scrubber. ....... 2

Figure 2. Layout of oxy-combustion power plant. DCCPS unit is highlighted in blue. ................................................ 3

Figure 3 Block flow diagram representing the stream flows accounted for in the ...................................................... 4

Figure 4 Saturation pressure of water vapor in air as a function of temperature between 0 and 300 degrees centigrade. ................................................................................................................................................................... 5

Figure 5 The latent heat of vaporization of water is temperature dependent (2). The decline in LHOV magnitude is very linear over the temperature window expected during operation of the DCCPS ...................................... 6

Figure 6 Mass flow rates around DCCPS at a 650 MWg oxyfuel plant in default. .................................................... 17


List of Tables

Table 1 The constants used in the Antoine Equation are temperature dependent across the range of expected operation of the DCCPS. ............................................................................................................................................ 8

Table 2.Default O&M Parameters for the DCCPS Model ............................................................................................ 16


Acknowledgements


Integrated Environmental Control Model - Technical DocumentationDirect Contact Cooler and Polishing Scrubber • 1


Objectives of this Report This report presents the performance and cost models for the direct contact cooler and sulfur polishing scrubber used in oxy-combustion power plant model in IECM.

Introduction Direct contact coolers and polishing scrubbers are an important component of modern electricity generation units. Their application ranges from traditional pulverized coal facilities which need to decrease the relative humidity of the flue gas exiting the stack to conform to opacity limits to amine-based CO2 scrubbing systems which need ultra-low sulfur concentration flue gas to avoid heat stable salt formation. Their value is tied to the ability to accomplish three operations simultaneously:

• trace sulfur removal, • bulk flue gas cooling, and • reducing the concentration of water in the exiting flue gas.

Furthermore, the latter two operations are physically linked; the concentration of water in the exiting flue gas being a function of the exiting flue gas temperature. Direct contact coolers utilize the saturation properties of water to condense out any liquid water that is formed as the gaseous water in the flue gas stream is cooled and changes phase. This process is illustrated in Figure 1 as the entering flue gas (red diamond) is gradually cooled until the water saturation pressure curve is encountered. As further cooling of the flue gas occurs, the maximum partial pressure of water vapor in the flue gas is reduced. This means that the flue gas water, no longer able to stay in gaseous form, condenses out of the mixture as liquid water. This simultaneous reduction of temperature and gaseous water in the flue gas is continued until the desired exiting concentration of water is reached (yellow diamond).


Figure 1 Illustration of the physical working process utilized in the direct contact cooler polishing scrubber.

Trace sulfur removal can be performed simultaneously via a chemical process by adding sodium hydroxide (commonly referred to as caustic soda) which reacts with the residual sulfur dioxide in the flue gas to form sodium sulfite. This combination of a chemical and physical process allows the direct contact cooler and polishing systems (DCCPS) to accomplish their three tasks of bulk flue gas cooling, trace sulfur removal, and water concentration reduction in the exiting gas stream.

Oxyfuel Application The current generation of oxyfuel systems require the use of flue gas recirculation (FGR) to moderate temperature inside the boiler and to ensure that the heat transfer mechanisms are maintained closely to the air-fired conditions for which today’s boilers were designed. To that end, the flue gas which is recycled to the boiler must have an acceptable temperature and water concentration to ensure proper thermal regulation and to allow uninterrupted performance of the downstream traditional pollution control equipment. This last consideration is especially important for oxyfuel systems utilizing either a sub-bituminous or lignite coal with a high moisture content. Such coal types typically have a low enough sulfur content to permit the use of a spray-dry absorption (SDA) system for sulfur removal in lieu of a wet system. This is beneficial for plant heat rate. However, a direct contact cooler must then be used to reduce FGR water content to ensure that a sufficient approach temperature is maintained so that the SDA may continue to function at the desired level of sulfur removal. Figure 2 shows the configuration of a typical oxy-fuel power plant, with DCCPS unit highlighted.

0.1

1

10

100

1000

0 50 100 150 200 250 300

ABSO

LUTE

VAP

OR

PRES

SURE

[KPA

]

TEMPERATURE [°C]

DCCPS OPERATION ILLUSTRATION

Water Saturation Pressure Flue Gas Entry

Flue Gas Exit


Figure 2. Layout of oxy-combustion power plant. DCCPS unit is highlighted in blue.

Modeling Approach The direct contact cooler and polishing scrubber model uses or calculates the nine inlet and outlet streams depicted in Figure 3. These process flow streams are delineated for ease of mass and energy accounting and calculation but are not necessarily reflective of real-world DCCPS operation. The entering flue gas stream is passed to the model from the IECM fully defined: meaning that the stream is fully defined with composition, temperature, and pressure data along with the mass flow rate. There are then two main parameters which need to be specified prior to the first calculation for cooling and condensing: an anticipated pressure drop across the contacting column and either a desired exit temperature for the flue gas exiting the DCCPS or a desired water concentration in the exiting flue gas. If sulfur polishing is desired, the concentration of sulfur dioxide exiting the DCCPS must also be specified in units of parts per million. The model then steps through a series of calculations involving the saturation pressure curve of water to determine the non-specified value of either exit flue gas temperature or exit water concentration in the flue gas. From there, the model calculates the mass balance of all streams in the model along with the composition of gases, liquids, and/or solids in each stream. The energy balance is then completed in a two-step calculation process. First, changes in latent and sensible heat for each stream are calculated using a combination of the Shomate relations, heat capacity data, and the latent heat of condensation model. Secondly, the amount of cooling load required to offset the total latent and sensible heat increase is calculated treating the DCCPS as an adiabatic heat exchanger. Lastly, the amount of cooling water can be calculated from the required cooling load and reported along with the rest of the fully defined process flow streams.


Figure 3 Block flow diagram representing the stream flows accounted for in the

direct contact cooler and polishing scrubber model. There were a number of simplifying assumptions made in the creation of this model. Specifically, the separate reagent and cooling water inlet streams would in practice be a premixed solution entering the top of the contacting column. Similarly, the cooling water out, precipitate water, and spent reagent streams would in practice be a single admixture at the bottom of the contacting column. Furthermore, many studies involving a DCCPS of the size required for a coal plant equipped with carbon capture(1) (3) have indicated that it is desirable from both a cost and simplicity standpoint to construct a dedicated cooling system and water handling services for the contacting tower(s). This is in part due to the issues previously raised about the reagent stream and cooling water being combined in practice. Thus having a dedicated system to handle the caustic-doped water and precipitated solids (sodium sulfite) would be desirable. There are also balance of plant and layout considerations from the volume of cooling water required which bolster the case for a dedicated cooling water system. Treating the DCCPS system as an adiabatic heat exchanger for purposes of calculating the heat balance is another simplification. Weather conditions (including ambient water and air temperature, and associated maximum cooling water delta) will affect the quantity of cooling water required and associated parasitic load of pumping and processing that cooling water. However, absent very detailed weather data, anticipating the effect of the weather is beyond the capabilities of this analysis. Instead, a heat transfer efficiency factor 𝛼𝐶𝑜𝑜𝑙𝑖𝑛𝑔 has been provided to allow the user to enter what amounts to a cooling water safety factor into an analysis to account for non-adiabatic conditions.


Saturation Pressure of Water The maximum concentration of water vapor which can be contained in a gas mixture is a function of the gas temperature and pressure. This relationship is illustrated by the water vapor saturation pressure curve presented in Figure 4. There are numerous scientific methods for describing the relationship between vapor pressure and temperature for pure components. For this work we have chosen to use the Antoine Equation, which is derived from the Clausius-Clapeyron relation.

Figure 4 Saturation pressure of water vapor in air as a function of temperature between 0 and 300 degrees centigrade.

The units of pressure of the Antoine coefficients (A, B, and C) are reported in millimeters of mercury [mmHG]. A conversion factor (𝛾) allows for the Antoine equation to report pressures in units of kilopascals rather than millimeters of mercury using the relation.

𝛾 = 0.133322368

𝑃𝑠𝑎𝑡𝑢𝑟𝑎𝑡𝑖𝑜𝑛 = 𝑝𝐻2𝑂 = 10^ �𝐴 −𝐵

(𝐶 + 𝑇)�

The Antoine equation reports a saturation pressure which is here interpreted as the maximum partial pressure of water vapor in the flue gas mixture. The partial pressure of any given component of a gas mixture can be calculated directly given the total pressure of the gas stream and the molar fraction using Dalton’s Law of Partial Pressures.

Latent Heat of Water Condensation A reduction in the concentration of water in the flue gas is accomplished by causing the water vapor to change phase and condense out as liquid water. The amount of heat required to cause a

0.1

1

10

100

1000

10000

0 50 100 150 200 250 300

Max

imum

Par

tial P

ress

ure

of

Wat

er in

Air

[kPa

]

Temperature [C]

Water Vapor Saturation Pressure


liquid to vaporize is known as the latent heat of vaporization (LHOV). The latent heat of condensation (LHOC) is equivalent in magnitude to the (LHOV) for a pure component, but of the opposite sign. In the case of water vapor in the DCCPS, we are concerned with the amount of heat which must be removed in order to induce a change of phase from vapor to liquid. The LHOC is temperature dependent for water. More specifically, the LHOC of water displays a gradual decline in magnitude as temperature is increased (Fig. 5). However, as the triple point is approached, the LHOC rapidly converges toward zero.

Figure 5 The latent heat of vaporization of water is temperature dependent (2). The decline in LHOV magnitude is very

linear over the temperature window expected during operation of the DCCPS

Flue gas temperatures entering the DCCPS rarely exceed 170°C and the temperature zone where water condensation occurs from these entry conditions will typically be below 120°C. This operational temperature window allows us to simplify the LHOV relation of water to the linear region. A linear regression taken from 20 - 120°C provides a nearly perfect fit (R2 > 0.99) over our condensation regime.

𝐿𝑎𝑡𝑒𝑛𝑡 𝐻𝑒𝑎𝑡 𝑜𝑓 𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛 �𝑘𝐽𝑚𝑜𝑙

� = −45.161 + 0.0452 ∗ 𝑇𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛 In practice, the phase change to liquid water would not occur at a single temperature, but across a range defined by the interplay between maximum partial pressure and temperature. In our model we use a mean temperature for 𝑇𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛 in the above equation for calculating the LHOC for water. This mean temperature is the average between the DCCPS exit temperature �𝑇𝑠𝑎𝑡_𝑒𝑥𝑖𝑡� and the temperature at which the inlet flue gas first encounters the water vapor saturation curve during cooling �𝑇𝑠𝑎𝑡_𝑒𝑛𝑡𝑒𝑟�.

y = -0.0452x + 45.161 R² = 0.9994

0

10

20

30

40

50

0 100 200 300 400

Hvap

[kJ/

mol

]

Temperature [°C]

Temperature Dependance of the Latent Heat of Vaporization for Water

Latent Heat of Vaporization LHOV for DCC/PS Linear (LHOV for DCC/PS)


Sulfur Removal Calculation Deep sulfur scrubbing with sodium hydroxide (NaOH) can be enabled as a concurrent process to flue gas cooling in the updated DCCPS model. The amount of sodium hydroxide, also referred to as caustic soda or caustic, required for polishing is determined on a strictly mass balance basis according to the mass of sulfur to be removed from the flue gas. The kinetics are assumed to be sufficiently fast as to not preempt the completion of the following reaction:

𝑆𝑂2 + 2𝑁𝑎𝑂𝐻 𝑦𝑖𝑒𝑙𝑑𝑠�⎯⎯⎯� 𝑁𝑎2𝑆𝑂3 + 𝐻2𝑂

The heat of reaction from the formation of solid sodium sulfite (Na2SO3) and water is also neglected in determining the required cooling load of the DCCPS unit. However, the water formed has been assumed to be in the gaseous state and is therefore accounted for in the latent heat removal. A desired exiting sulfur concentration, in parts per million (ppm), may be specified when using the DCCPS model. This volumetric concentration, along with the entering flue gas composition, is then used to determine the mass of sulfur which must be removed from the flue gas stream per unit time. Because the quantity of sulfur in the DCCPS exit stream is much less than the total quantity of gas in the exit stream, the number of sulfur moles can be calculated with sufficient accuracy based upon the balance of gas moles in the exit stream.

Summary of Input and Output Parameters Input Parameters The key input parameters defining the performance of the DCCPS are as follows: �̇�𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 Mass flow rate of flue gas entering the contacting tower [kgmol/hr] 𝑥𝑖,𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 Mole fraction of all species (i) in the entering flue gas 𝑃𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 Absolute pressure [kPa] of the entering flue gas 𝑃𝑑𝑟𝑜𝑝_𝐷𝐶𝐶𝑃𝑆 Absolute pressure drop across the contacting tower [kPa] 𝑥𝑆𝑂2,𝑓𝑔_𝑒𝑥𝑖𝑡 Molar concentration of sulfur dioxide in exiting flue gas [ppm] 𝑃𝑀𝐶 Moisture content in the sodium sulfite slurry [mass fraction] 𝛼𝐶𝑜𝑜𝑙𝑖𝑛𝑔 Heat transfer efficiency of the DCCPS system (unity being ideal heat transfer) ∆𝑇𝑚𝑎𝑥,𝐶𝑜𝑜𝑙𝑖𝑛𝑔 𝑤𝑎𝑡𝑒𝑟 Largest acceptable cooling water temperature increase [°C] Additionally, one of the following two parameters must be specified about the exit flue gas stream: 𝑇𝑓𝑔_𝑒𝑥𝑖𝑡 Temperature of exiting flue gas [°C] 𝑥𝐻2𝑂,𝑓𝑔_𝑒𝑥𝑖𝑡 Mole fraction of water vapor in exiting flue gas


Output Parameters The model will then calculate or report the following key output parameters: �̇�𝑓𝑔_𝑒𝑥𝑖𝑡 Mass flow rate of flue gas exiting the contacting tower [kgmol/hr] 𝑥𝑖,𝑓𝑔_𝑒𝑥𝑖𝑡 Mole fraction of all species (i) in the exiting flue gas 𝑃𝑓𝑔_𝑒𝑥𝑖𝑡 Absolute pressure [kPa] of the exiting flue gas 𝑇𝑓𝑔_𝑒𝑥𝑖𝑡 Temperature of exiting flue gas [°C] 𝜑𝐻2𝑂_𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑒𝑑 Molar flow rate of water condensed out of flue gas [mol/sec] 𝜑𝑁𝑎2𝑆𝑂3,𝑝𝑟𝑒𝑐𝑖𝑝𝑖𝑡𝑎𝑛𝑡 Molar flow rate of sodium sulfite produced from sulfur treatment [mol/sec] 𝜑𝐻2𝑂,𝑝𝑟𝑒𝑐𝑖𝑝𝑖𝑡𝑎𝑛𝑡 Molar flow rate of water produced for sulfur treatment [mol/sec] �̇�𝑁𝑎𝑂𝐻 Mass flow rate of sodium hydroxide required for sulfur treatment [kg/sec] �̇�𝑆𝑙𝑢𝑟𝑟𝑦 Mass flow rate of sodium sulfite slurry produced by sulfur treatment [kg/sec] 𝐶𝑜𝑜𝑙𝑖𝑛𝑔 𝐿𝑜𝑎𝑑𝑡𝑜𝑡𝑎𝑙 Total cooling load requirement of DCCPS [kJ/sec] �̇�𝐶𝑜𝑜𝑙𝑖𝑛𝑔 𝑤𝑎𝑡𝑒𝑟 Mass flow rate of cooling water required for DCCPS cooling [kg/sec] �̇�𝑤𝑎𝑡𝑒𝑟_𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑒𝑑 Mass flow rate of water generated during DCCPS operation [kg/sec]

Mass and Energy Balance Calculations The DCCPS model has been designed so that either a desired water concentration of the exiting flue gas or a desired exiting flue gas temperature may be specified by the user. These two variables are co-dependent and cannot be specified independently. Regardless of starting information, the Antoine Equation is then utilized to determine either the exit concentration or exit temperature of the flue gas leaving the DCCPS. Table 1 The constants used in the Antoine Equation are temperature dependent across the range of expected operation of

the DCCPS.

Antoine Constants for Water [°C and mmHG] 1 – 100 °C 99 – 374 °C

A 8.07131 8.14019 B 1730.63 1810.94 C 233.426 244.485

Starting Equations Given: fully defined inlet (pressure, temperature, composition, mass flow rate), pressure drop across DCCPS, and desired flue gas exit composition.

𝑝𝐻2𝑂,𝑓𝑔_𝑒𝑥𝑖𝑡 = 𝑥𝐻20,𝑓𝑔_𝑒𝑥𝑖𝑡 ∗ �𝑃𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 − 𝑃𝑑𝑟𝑜𝑝_𝐷𝐶𝐶𝑃𝑆�

𝑇𝑓𝑔_𝑒𝑥𝑖𝑡 = 𝑇𝑠𝑎𝑡_𝑒𝑥𝑖𝑡 = 𝐵

𝐴 − �log10�𝑝𝐻2𝑂,𝑓𝑔_𝑒𝑥𝑖𝑡/𝛾��− 𝐶

Starting equations given: fully defined inlet, pressure drop across DCCPS, desired flue gas exit temperature.


𝑝𝐻2𝑂,𝑓𝑔_𝑒𝑥𝑖𝑡 = 𝛾 �10^�𝐴 −𝐵

�𝐶 + 𝑇𝑓𝑔_𝑒𝑥𝑖𝑡��

𝑥𝐻2𝑂,𝑓𝑔_𝑒𝑥𝑖𝑡 = 𝑝𝐻2𝑂,𝑓𝑔_𝑒𝑥𝑖𝑡

�𝑃𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 − 𝑃𝑑𝑟𝑜𝑝_𝐷𝐶𝐶𝑃𝑆�

Once one of the above sequences has been followed, the temperature at which the incoming flue gas reaches the water vapor saturation curve during cooling must be calculated:

𝑇𝑠𝑎𝑡_𝑒𝑛𝑡𝑒𝑟 = 𝐵

𝐴 − �log10�𝑝𝐻2𝑂,𝑓𝑔_𝑒𝑛𝑡𝑒𝑟/𝛾��− 𝐶

Gas Stream Flow Rate Calculations At this point, the temperature, pressure, and composition of the flue gas has been either calculated or specified at the three important states of the DCCPS. The mass flows through the system must now be balanced: Inlet stream needs to be in [mol/sec] for each compound. This conversion can be accomplished utilizing the following if total mass flow is in [kgmol/hr]:

𝜑𝑖,𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 = 𝑥𝑖,𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 ∗�̇�𝑓𝑔_𝑒𝑛𝑡𝑒𝑟

3.6

Where: 𝜑𝑖,𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 is the molar flow rate [mol/sec] of species i in the entering flue gas 𝑥𝑖,𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 is the mole fraction of species (i) in the entering flue gas �̇�𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 is the total mass flow rate [kgmol/hr] of flue gas into the DCCPS The flue gas exit molar flow rate can then be calculated using the below for all gaseous species other than water and sulfur dioxide:

𝜑𝑖,𝑓𝑔_𝑒𝑥𝑖𝑡 = 𝜑𝑖,𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 For water the following is used to determine the exit flue gas molar flow rate:

𝜑𝐻2𝑂,𝑓𝑔_𝑒𝑥𝑖𝑡 = 𝑥𝐻2𝑂,𝑓𝑔_𝑒𝑥𝑖𝑡 ∗ ∑ 𝜑𝑖,𝑓𝑔_𝑒𝑥𝑖𝑡𝑗

1 − 𝑥𝐻2𝑂,𝑓𝑔_𝑒𝑥𝑖𝑡

Where: ∑ 𝜑𝑖,𝑓𝑔_𝑒𝑥𝑖𝑡𝑗 is the total flow rate of all non-water gas species in the exiting flue gas

(neglecting sulfur) The molar flow rate of water condensed out of the DCCPS is then calculated by subtracting the outlet flue gas flow rate from the inlet flow rate

𝜑𝐻2𝑂_𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑒𝑑 = 𝜑𝐻2𝑂,𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 − 𝜑𝐻2𝑂,𝑓𝑔_𝑒𝑥𝑖𝑡


The sulfur dioxide molar flow rate in the exiting flue gas is then defined by:

𝜑𝑆𝑂2,𝑓𝑔_𝑒𝑥𝑖𝑡 = 𝑥𝑆𝑂2,𝑓𝑔_𝑒𝑥𝑖𝑡[𝑝𝑝𝑚] ∗ ∑ 𝜑𝑖,𝑓𝑔_𝑒𝑥𝑖𝑡𝑘

1,000,000

Where: ∑ 𝜑𝑖,𝑓𝑔_𝑒𝑥𝑖𝑡𝑘 is the total flow rate of all gas species in the exiting flue gas (neglecting sulfur) 𝑥𝑆𝑂2,𝑓𝑔_𝑒𝑥𝑖𝑡[𝑝𝑝𝑚] is the desired concentration of sulfur dioxide in the exiting flue gas in ppm The mass flow rate [kgmol/hr] of the exiting flue gas stream can now be calculated by taking the sum across all component gas molar flow rates and multiplying by their respective molecular weights.

�̇�𝑓𝑔_𝑒𝑥𝑖𝑡 = 3.6 ∗ ��𝜑𝑖,𝑓𝑔_𝑒𝑥𝑖𝑡 ∗ 𝑀𝑊𝑖� Where: 𝜑𝑖,𝑓𝑔_𝑒𝑥𝑖𝑡 is the molar flow rate [mol/sec] of species (i) in the exiting flue gas 𝑀𝑊𝑖 is the molecular weight of species (i) [g/mol] �̇�𝑓𝑔_𝑒𝑥𝑖𝑡 is the total mass flow rate [kgmol/hr] of flue gas exiting the DCCPS

Reagent Stream Flow Rate Calculations The amount of sodium hydroxide (NaOH) added for sulfur removal in the DCCPS model is assumed to be equal to the stoichiometric requirement. In practice, a surplus quantity of caustic would be supplied to the DCCPS in the recycled cooling water to ensure sufficient availability to achieve the stipulated exiting sulfur dioxide concentration. For ease of calculation however, we have assumed that the stoichiometric quantity of caustic closely approximates steady state behavior for caustic consumption in the DCCPS model.

𝜑𝑁𝑎𝑂𝐻 = 2 ∗ �𝜑𝑆𝑂2,𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 − 𝜑𝑆𝑂2,𝑓𝑔_𝑒𝑥𝑖𝑡� The mass flow rate of required sodium hydroxide added as reagent is then

�̇�𝑁𝑎𝑂𝐻 �𝑘𝑔𝑠𝑒𝑐

� = 0.04 �𝑘𝑔𝑚𝑜𝑙

� ∗ 𝜑𝑁𝑎𝑂𝐻

Precipitant Stream Flow Rate Calculations The precipitant stream is assumed to be comprised of only water and the sodium sulfite solid created by the removal of sulfur from the incoming flue gas. The quantity of sodium sulfite can be calculated based upon the difference in sulfur dioxide flow rate between the entering and exiting flue gas. This is true because SO2 and Na2SO3 are equimolar in reaction (A). Additionally, an equivalent number of moles of water are generated in the production of sodium sulfite which must be added to the molar flow rate of precipitate water.

𝜑𝑁𝑎2𝑆𝑂3,𝑝𝑟𝑒𝑐𝑖𝑝𝑖𝑡𝑎𝑛𝑡 = 𝜑𝑆𝑂2,𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 − 𝜑𝑆𝑂2,𝑓𝑔_𝑒𝑥𝑖𝑡 = 𝜑𝐻2𝑂_𝑆𝑂2𝑟𝑒𝑚𝑜𝑣𝑎𝑙


The molar flow rate of all precipitate water can then be calculated by adding the water generated by the creation of sodium sulfite to the condensed water calculated previously.

𝜑𝐻2𝑂,𝑝𝑟𝑒𝑐𝑖𝑝𝑖𝑡𝑎𝑛𝑡 = 𝜑𝐻2𝑂_𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑒𝑑 + 𝜑𝐻2𝑂_𝑆𝑂2𝑟𝑒𝑚𝑜𝑣𝑎𝑙 For reporting mass flow rates from the model, it is necessary to convert the molar flow rates of the above streams. In practice, a fraction of the water produced in the formation of sodium sulfite remains with the solid to form a slurry. This fraction, denoted as Precipitant Moisture Content [mass fraction], may be specified by the user but carries a default value of 25%. A lower limit of 14.3% is stipulated for Precipitant Moisture Content (PMC) because this represents the equimolar mixture of sodium sulfite and water which would be produced simultaneously when sodium hydroxide reacts with sulfur. Therefore, absent drying, a PMC of less than 14.3% is not possible. For PMC’s greater than 14.3% additional water from the DCCPS is entrained with the precipitant slurry. For a generic PMC the resulting precipitant slurry mass flow rate is defined by:

�̇�𝑆𝑙𝑢𝑟𝑟𝑦 �𝑘𝑔𝑠𝑒𝑐

� = 0.126 �𝑘𝑔𝑚𝑜𝑙

� ∗ 𝜑𝑁𝑎2𝑆𝑂3,𝑝𝑟𝑒𝑐𝑖𝑝𝑖𝑡𝑎𝑛𝑡 + 0.018 �𝑘𝑔𝑚𝑜𝑙

� ∗ �𝑃𝑀𝐶14.3�

Because the sodium sulfite is produced in a slurry, rather than as a dry product, there is no excess water created from the removal of sulfur which can be returned to the DCCPS. In fact, for all PMC’s greater than the minimum value, the sulfur removal process is water negative. The required slurry water must be subtracted from the overall water balance and is calculated as follows:

�̇�𝑆𝑙𝑢𝑟𝑟𝑦_𝑤𝑎𝑡𝑒𝑟 �𝑘𝑔𝑠𝑒𝑐

� = 0.018 �𝑘𝑔𝑚𝑜𝑙

� ∗ 𝜑𝐻2𝑂_𝑆𝑂2𝑟𝑒𝑚𝑜𝑣𝑎𝑙 ∗ �𝑃𝑀𝐶14.3

− 1�

Calculating the Required Cooling Load To determine the amount of cooling which must be provided to the DCCPS, we define the system as an adiabatic heat exchanger. This allows us to neglect any second order effects of environmental temperature fluctuations and focus on the primary bulk fluid heat transfer required. The cooling load is made up of a sensible heat component (temperature change) and a latent heat component (phase change). The sensible heat change of each of the non-reactive gas species is calculated using the Shomate equation (Section 4.5.1) to determine the enthalpy of each component at the entering and exiting states of the DCCPS. To then calculate the change in sensible heat of all the non-reactive gas (NRG) species we assume ideal gas behavior and apply the Gibbs-Dalton law for calculating the combined enthalpy of a gas stream.

𝐻 = 𝑚ℎ = 𝑚1ℎ1 + 𝑚2ℎ2 + ⋯+ 𝑚𝑘ℎ𝑘 = �𝑚𝑖ℎ𝑖

𝑘

𝑖=1

The difference in enthalpy of each component of the entering and exiting flue gas can then be multiplied by their respective mass flow rates to calculate the total sensible heat which must be removed from the non-reactive gas species while in the DCCPS.

∆ 𝑆𝑒𝑛𝑠𝑖𝑏𝑙𝑒 𝐻𝑒𝑎𝑡𝑁𝑅𝐺 = 𝐻𝑒𝑛𝑡𝑒𝑟_𝑁𝑅𝐺 − 𝐻𝑒𝑥𝑖𝑡_𝑁𝑅𝐺


The sensible heat delta calculation for sulfur dioxide is calculated using the same formula as the non-reactive gases. Theoretically, there should be some accounting for the reduction in moles of SO2 gas as the flue gas passes through the DCCPS which would result in a slightly lower value for the sensible heat of SO2 than calculated using the non-reactive gases methodology. However, due to the diminutive mass flow rate of SO2 even before removal, the reduction in the system heat balance through precise mass accounting is negligible (<0.1%). We therefore chose to use the following relation to calculate the sensible heat delta of the sulfur dioxide gas in the flue gas stream.

∆ 𝑆𝑒𝑛𝑠𝑖𝑏𝑙𝑒 𝐻𝑒𝑎𝑡𝑆𝑂2 = 𝜑𝑆𝑂2,𝑓𝑔_𝑒𝑛𝑡𝑒𝑟 ∗ [ℎ𝑆𝑂2�𝑇𝑓𝑔_𝑒𝑛𝑡𝑒𝑟� − ℎ𝑆𝑂2(𝑇𝑓𝑔_𝑒𝑥𝑖𝑡)] The sensible heat delta from water in the flue gas is calculated in three parts. Two of which correspond to the vapor and liquid phase of the condensing water, while the third accounts for the bulk cooling of the non-condensing water vapor (NCV) in the flue gas. The third part is the most straightforward and is calculated in identical fashion to the sensible heat of sulfur dioxide save that the molar flow rate used is the exiting, rather than entering, gas flow rate of water vapor from the DCCPS.

∆ 𝑆𝑒𝑛𝑠𝑖𝑏𝑙𝑒 𝐻𝑒𝑎𝑡𝐻2𝑂_𝑁𝐶𝑉 = 𝜑𝐻2𝑂_𝑁𝐶𝑉,𝑓𝑔_𝑒𝑥𝑖𝑡 ∗ [ℎ𝐻2𝑂�𝑇𝑓𝑔_𝑒𝑛𝑡𝑒𝑟� − ℎ𝐻2𝑂(𝑇𝑓𝑔_𝑒𝑥𝑖𝑡)] The remaining two sensible heat components for water relate to the vapor (CV) and then liquid water (CL) which is condensed out of the incoming flue gas. The vapor phase of the condensate water sensible heat change is calculated in a similar fashion to the non-reactive gaseous components using the Shomate relations. The specific calculation varies in that the final temperature of the water vapor is not assumed to be the exit temperature of the DCCPS, but rather the average condensation temperature as defined below:

𝑇𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛=

𝑇𝑠𝑎𝑡_𝑒𝑛𝑡𝑒𝑟+𝑇𝑠𝑎𝑡_𝑒𝑥𝑖𝑡2

∆ 𝑆𝑒𝑛𝑠𝑖𝑏𝑙𝑒 𝐻𝑒𝑎𝑡𝐻2𝑂_𝐶𝑉 = 𝜑𝐻2𝑂_𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑒𝑑 ∗ [ℎ𝐻2𝑂�𝑇𝑓𝑔_𝑒𝑛𝑡𝑒𝑟� − ℎ𝐻2𝑂(𝑇𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛)] The liquid phase of the condensate sensible heat change is calculated using the heat capacity of liquid water.

𝐶𝑝𝐻2𝑂_𝑙𝑖𝑞𝑢𝑖𝑑 �𝑘𝐽

𝑚𝑜𝑙 ∗ 𝐾� = 0.075

The heat capacity of liquid water is close enough to constant over the range of temperature involved within the DCCPS to allow us to safely assume a fixed specific heat for liquid water. The sensible heat change of the liquid water can be calculated using the change in temperature of the condensate and the molar flow rate.

∆ 𝑆𝑒𝑛𝑠𝑖𝑏𝑙𝑒 𝐻𝑒𝑎𝑡𝐻2𝑂_𝐶𝐿 = 𝜑𝐻2𝑂_𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑒𝑑 ∗ 𝐶𝑝𝐻2𝑂_𝑙𝑖𝑞𝑢𝑖𝑑 ∗ �𝑇𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛 − 𝑇𝑓𝑔_𝑒𝑥𝑖𝑡�


Latent Heat of Water Calculating the latent heat required to be removed from a gas stream in order to have a specified fraction of a component condense is not a straight-forward calculation to obtain an exact answer. In order to simplify the calculation of the latent heat of cooling, 𝑇𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛 was assumed to be the temperature at which all water vapor was condensed out of the flue gas within the DCCPS. This assumption preempts the use of a much more computationally intensive, iterative method which could capture continuous changes in water vapor concentration as a function of temperature. It was determined that this degree of precision was not appropriate given the inherent uncertainty of the electrical generation unit as a whole and therefore a decision was made in favor of computational economy. The total latent heat of condensation for the water condensed out of the flue gas is calculated for the new DCCPS model using the molar flow rate of the precipitated water and the molar flow rate of water created by the sulfur removal process chemistry. The water created by the formation of sodium sulfite is very small in comparison (typically 3 orders of magnitude less) but is included here as the sole means of thermally accounting for the exothermic removal of sulfur in the model. The sum of these two molar flow rates is then combined with the latent heat of condensation correlation detailed earlier to calculate the required cooling load for the latent heat of water condensation. ∆ 𝐿𝑎𝑡𝑒𝑛𝑡 𝐻𝑒𝑎𝑡 = (𝜑𝐻2𝑂_𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑒𝑑 + 𝜑𝐻2𝑂_𝑆𝑂2𝑟𝑒𝑚𝑜𝑣𝑎𝑙) ∗ (45.161 − 0.0452 ∗ 𝑇𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛)

Cooling Water Requirement Calculation At this point all of the mass and energy streams have been calculated with the exception of the cooling water flow rate. The required flow rate of cooling water is a function of four parameters: 𝛼𝐶𝑜𝑜𝑙𝑖𝑛𝑔 heat transfer efficiency of the DCCPS system (unity being ideal heat transfer) 𝐶𝑜𝑜𝑙𝑖𝑛𝑔 𝐿𝑜𝑎𝑑𝑡𝑜𝑡𝑎𝑙 sum of all latent and sensible cooling loads in the DCCPS system ∆𝑇𝑚𝑎𝑥,𝐶𝑜𝑜𝑙𝑖𝑛𝑔 𝑤𝑎𝑡𝑒𝑟 largest acceptable temperature increase of cooling water 𝐶𝑝𝐶𝑜𝑜𝑙𝑖𝑛𝑔 𝑤𝑎𝑡𝑒𝑟 specific heat of cooling water, default value of 4.2 [kJ/kg K] The total cooling load is the first piece of information which is required to be calculated. It can be found by taking the sum of all the latent and sensible heat deltas calculated in the previous section to get a total cooling load in kilojoules per second.

𝐶𝑜𝑜𝑙𝑖𝑛𝑔 𝐿𝑜𝑎𝑑𝑡𝑜𝑡𝑎𝑙 �𝑘𝐽𝑠𝑒𝑐

� = ∆𝑆𝐻𝑁𝑅𝐺 + ∆𝑆𝐻𝑆𝑂2 + ∆𝑆𝐻𝐻2𝑂_𝑁𝐶𝑉 + ∆𝑆𝐻𝐻2𝑂_𝐶𝑉 + ∆𝑆𝐻𝐻2𝑂_𝐶𝐿 + ∆𝐿𝐻

The required mass flow of cooling water can then be calculated:

�̇�𝐶𝑜𝑜𝑙𝑖𝑛𝑔 𝑤𝑎𝑡𝑒𝑟 �𝑘𝑔𝑠𝑒𝑐

� = 𝐶𝑜𝑜𝑙𝑖𝑛𝑔 𝐿𝑜𝑎𝑑𝑡𝑜𝑡𝑎𝑙

𝛼𝐶𝑜𝑜𝑙𝑖𝑛𝑔 ∗ 𝐶𝑝𝐶𝑜𝑜𝑙𝑖𝑛𝑔 𝑤𝑎𝑡𝑒𝑟 ∗ ∆𝑇𝑚𝑎𝑥,𝐶𝑜𝑜𝑙𝑖𝑛𝑔 𝑤𝑎𝑡𝑒𝑟


Water Balance One of the primary functions of the DCCPS is to reduce the moisture content in the entering flue gas. It follows logically then that a substantial amount of liquid water (𝜑𝐻2𝑂_𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑒𝑑) is produced during normal operation. However, as discussed previously, the reaction to sodium sulfite can be a water consuming process if a water-rich slurry is specified. However, the creation of sodium sulfite slurry is a secondary process compared to the precipitation of flue gas moisture under typical operating conditions and has correspondingly little effect on the net water produced during operation.

�̇�𝑤𝑎𝑡𝑒𝑟_𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑒𝑑 �𝑘𝑔𝑠𝑒𝑐

� = �0.018 �𝑘𝑔𝑚𝑜𝑙

� ∗ 𝜑𝐻2𝑂_𝐶𝑜𝑛𝑑𝑒𝑛𝑠𝑒𝑑� − �̇�𝑆𝑙𝑢𝑟𝑟𝑦_𝑤𝑎𝑡𝑒𝑟 �𝑘𝑔𝑠𝑒𝑐

�

Auxiliary Electrical Load Requirements

Pneumatic Head The required pneumatic head to overcome the pressure loss across the contacting tower is provided by a combination of the main induced draft fan and the primary forced draft fan unit. The electrical load required for their operation is calculated outside this component model and is included in the base plant energy use calculations performed by the IECM.

Cooling Water Pumping The electrical load required to pump the above quantity of cooling water is calculated using a pre-existing IECM relation for pumping:

𝑀𝑊𝑓𝑔_𝑐𝑜𝑜𝑙𝑖𝑛𝑔 = 4.7 ∗ 10−5 ∗ 𝐶𝑜𝑜𝑙𝚤𝑛𝑔𝑊𝑎𝑡𝑒𝑟̇ 𝑓𝑔[𝑔𝑝𝑚] The cooling water flowrate was in units of [kg/sec], so we must use the following relation to convert from volume to mass flowrate:

1 𝑔𝑎𝑙𝑙𝑜𝑛 𝑤𝑎𝑡𝑒𝑟 = 3.79 𝑘𝑔 𝑤𝑎𝑡𝑒𝑟 With this, and correcting for the time unit disparity, the electrical use of the DCCPS pumps can be expressed as:

𝑀𝑊𝐷𝐶𝐶𝑃𝑆,𝑝𝑢𝑚𝑝 = 7.44 ∗ 10−4 ∗ �̇�𝐶𝑜𝑜𝑙𝑖𝑛𝑔 𝑤𝑎𝑡𝑒𝑟 �𝑘𝑔𝑠𝑒𝑐

�

Cost Model Available cost information for DCCPS is currently very limited. This is largely a result of the private vendors who produce cost estimates for DOE, and other publicly available information organizations, rolling the cost of the DCCPS into the cost quote for an entire system; be that


oxyfuel or post-combustion scrubbing. Consequently, it is impossible to discern the proportion of costs which should be ascribed to the DCCPS. A concerted effort was made to contact representatives of the vendors responsible for producing said system estimates. But at the time of writing, no new cost information has been garnered. The only reliable cost quote available was produced by the Alstom Corporation in 2001 for an evaluation study for the State of Ohio Department of Development. The accuracy of applying this, slightly dated, point estimate is perhaps poor. However, the overall cost contribution of the DCCPS is fairly small and any inaccuracy in this specific estimate is well within the margin of error of the overall plant cost estimate.

Capital Cost The PFC of the DCCPS is scaled on the basis of the flue gas flow rate entering the contacting tower. The reference cost of the DCCPS is $17.6 million in 2001 USD (4), corresponding to a treated flue gas flow rate of 810,000 acfm. A single train limit of 2,000,000 acfm is enforced and is adopted from flue gas desulfurization train limits (5). Actual cubic feet per minute (acfm) of gas flow can be calculated using the ideal gas law and the molar flow rate of flue gas entering the DCCPS.

�̇� �𝑚3

𝑠 �=𝜑𝑡𝑜𝑡𝑎𝑙_𝐹𝐺 ∗ 𝑅𝑇

𝑃

Where �̇� flue gas velocity [m3/sec] entering the DCCPS 𝜑𝑡𝑜𝑡𝑎𝑙_𝐹𝐺 total molar flow rate [mol/sec] of all gases entering the DCCPS 𝑅 the universal gas constant [J/mol K] (R = 8.314) 𝑇 bulk gas temperature [Kelvin] 𝑃 bulk gas pressure [Pascals] The reference cost information and train size limit are both in units of actual cubic feet per minute, so it is necessary to convert prior to calculating a PFC for the DCCPS being examined.

�̇�[𝑎𝑐𝑓𝑚] = �̇� �𝑚3

𝑠 �∗ 35.314 �

𝑓𝑡3

𝑚3� ÷ 60 �𝑠

𝑚𝑖𝑛�

The PFC can then be calculated using the following set of operations: If �̇� > 2,000,000 𝑎𝑐𝑓𝑚, divide the stream into equal parts which are each less than 2,000,000 acfm. The total PFC of the DCC will be given by:

𝑃𝐹𝐶𝐷𝐶𝐶𝑃𝑆[$𝑀] = 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑇𝑟𝑎𝑖𝑛𝑠 ∗ 𝑝𝑓𝑐𝐷𝐶𝐶𝑃𝑆_𝑠𝑖𝑛𝑔𝑙𝑒 The PFC for a single train is calculated based upon the reference cost and associated volumetric flow rate. The cost estimate must also be adjusted to the base financial year selected by the user using the chemical engineering plant cost index (PCI).


𝑝𝑓𝑐𝐷𝐶𝐶𝑃𝑆_𝑠𝑖𝑛𝑔𝑙𝑒[$𝑀] = 17.6 ∗ ��̇�

810,000�0.6

∗ �𝑃𝐶𝐼

𝑃𝐶𝐼2001�

Fixed and Variable Operations and Maintenance Cost The fixed O&M costs which are specific to the direct contact cooler and polishing scrubber are calculated using a combination of equations presented below and the values presented in Table 2. The internal cost of electricity is calculated internal to the IECM using the base plant COE as the default (case specific) but may also be specified by the user.

Table 2.Default O&M Parameters for the DCCPS Model

O&M Cost Elements Default Value

Fixed O&M Costs

A Electricity Price (Internal) ($/MWh) Case Specific

B Total Maintenance Cost (% TPC) 2

C Number of Operating Jobs (jobs/shift) 2.000

D Number of Operating Shifts (shifts/day) 4.750 E Maintenance Cost Allocated to Labor (% total) 40.00

F Admin. & Support Labor Cost (% total labor) 30.00 The electrical load required for operation of the DCCPS is limited to the pumping required to circulate the water in the contacting column. The electric load of the pump is calculated in the performance model and is debited from the gross plant output. The cost of electricity use is assessed to the DCCPS model by multiplying the internal cost of electricity [$/MWh] by the pumping load [MW].

𝑉𝑂𝑀𝐷𝐶𝐶𝑃𝑆,𝑒𝑙𝑒𝑐𝑡𝑟𝑖𝑐 �$ℎ𝑟�

= 𝑀𝑊𝐷𝐶𝐶𝑃𝑆,𝑝𝑢𝑚𝑝 ∗ 𝐶𝑂𝐸𝑖𝑛𝑡𝑒𝑟𝑛𝑎𝑙

When the ability to perform polishing scrubbing is being utilized, sodium hydroxide (NaOH), also known as caustic, is consumed. The mass consumed is calculated by the performance model and is reported to the cost model where it is multiplied by the unit cost.

𝑉𝑂𝑀𝐷𝐶𝐶𝑃𝑆,𝑐𝑎𝑢𝑠𝑡𝑖𝑐 �$ℎ𝑟�

= 𝑁𝑎𝑂𝐻 �$

𝑡𝑜𝑛𝑛𝑒�∗ 3.6 ∗ �̇�𝑁𝑎𝑂𝐻 �

𝑘𝑔𝑠𝑒𝑐

�

With the consumption of caustic, the DCCPS generates sodium sulfate (Na2SO3) as part of a slurry which must be treated and/or disposed of as solid waste. The disposal cost is assumed to be equivalent to the waste disposal cost of solids produced by the flue gas desulfurization system.


𝑉𝑂𝑀𝐷𝐶𝐶𝑃𝑆,𝑠𝑜𝑙𝑖𝑑𝑤𝑎𝑠𝑡𝑒 �$ℎ𝑟�

= 𝑊𝑎𝑠𝑡𝑒 𝐷𝑖𝑠𝑝𝑜𝑠𝑎𝑙 𝐶𝑜𝑠𝑡 �$

𝑡𝑜𝑛𝑛𝑒�∗ 3.6 ∗ �̇�𝑆𝑙𝑢𝑟𝑟𝑦 �

𝑘𝑔𝑠𝑒𝑐

�

The quantity of water produced by the DCCPS is affected by whether or not polishing scrubbing is being performed; as a fraction of the produced water is entrained in the sodium sulfate slurry. Regardless, the DCCPS performance model reports the quantity of water generated to the cost model which must be treated by the wastewater treatment model.

𝑉𝑂𝑀𝐷𝐶𝐶𝑃𝑆,𝑤𝑤𝑡 �$ℎ𝑟�

= �̇�𝑤𝑎𝑡𝑒𝑟_𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑒𝑑 �𝑘𝑔𝑠𝑒𝑐

� ∗ 3.6 ∗ 𝑊𝑊𝑇 �$

𝑡𝑜𝑛𝑛𝑒�

Case Studies The performance and cost models of DCCPS were developed for the oxyfuel combustion module in the IECM. For the illustrative purpose, Figure 6 shows the result diagram as it will be implemented in the IECM for a 650 MWg coal-fired oxyfuel plant in default.

Figure 6 Mass flow rates around DCCPS at a 650 MWg oxyfuel plant in default.


References

1. Babcock & Wilcox Power Generation Group. Engineering and Economic Evaluation of Oxy-Fired 1100F (593C) Ultra-Supercritical Pulverized Coal Power Plant with CO2 Capture. Palo Alto : EPRI, 2011. 1021782.

2. IEA-GHG. 2010/07 Oxyfuel Combustion of Pulverized Coal . s.l. : ieaghg, 20120. 3. NETL. Cost and Performance Baseline for Fossil Energy Plants, Volume 1: Bituminous

Coal and Natural Gas to Electricity. s.l. : U.S. Department of Energy, 2013. 4. Alstom, et al. Engineering feasibility and economics of CO2 capture on an existing coal-

fired power plant. 2001. Final report prepared by ALSTOM Power Inc., ABB Lummus Global Inc., ALSTOM Power Environmental Systems and American Electric Power (AEP).

5. Black & Veatch. Wisconsin Public Service: Weston Unit 4 Flue Gas Desulfurization. 2004. pp. 3-4.

IECM Technical Documentation: Volume III

CO2 Capture Process Water Use

April 2016

Disclaimer

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference therein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed therein do not necessarily state or reflect those of the United States Government or any agency thereof.




Prepared for:




www.netl.doe.gov



April 2016


Table of Contents

CO2 Capture Process Water Use 1

Objectives of this Report ........................................................................................................... 1 Introduction ............................................................................................................................... 1 Water Use Estimation in the IECM ........................................................................................... 2 Water Consumption in Pulverized Coal Plants .......................................................................... 4

Amine CO2 capture system ......................................................................................... 4 Ammonia CO2 capture system .................................................................................... 5 Membrane CO2 capture system ................................................................................... 5 Calcium looping CO2 capture system ......................................................................... 6 PC Plant Case Studies ................................................................................................. 6 Gross power output and net plant efficiency ............................................................... 7 Water consumption ...................................................................................................... 8 Sensitivity analysis ...................................................................................................... 8

Water Consumption in IGCC Plants ........................................................................................ 11 Slag handling ............................................................................................................. 11 Slurry water ............................................................................................................... 12 Quench/Wash and Sour Water Stripper (SWS) ......................................................... 12 Syngas Humidifier and Combustion Turbine Dilution .............................................. 12 Gasifier Steam ........................................................................................................... 12 Water Gas Shift Steam .............................................................................................. 12 IGCC Water Use Estimation in IECM ...................................................................... 12 IGCC Plant Case Studies ........................................................................................... 14 Gross power output and net plant efficiency ............................................................. 14 Water consumption .................................................................................................... 15 Sensitivity analysis .................................................................................................... 16

Water Consumption in Oxy-Combustion Plants ...................................................................... 17 Sensitivity analysis .................................................................................................... 17

Conclusion ............................................................................................................................... 18 Appendix – CO2 Compressor Cooling Water .......................................................................... 19 References ............................................................................................................................... 20


List of Figures

Figure 1. Schematic of a wet cooling tower used in the IECM ................................................................................................................. 2

Figure 2. Schematic of an air cooled condenser used in the IECM ........................................................................................................... 2

Figure 3. Screenshot showing plant-level and process-level make-up water requirement for a PC power plant ...................................... 3

Figure 4. Screenshot showing plant-level and process-level water consumption estimation in IECM ..................................................... 3

Figure 5. A sample IECM screenshot showing cooling water requirement for a PC power plant using an amine-based CO2 capture system. ............................................................................................................................................................................................. 4

Figure 6. Schematic of amine-based CO2 capture system ........................................................................................................................ 5

Figure 7. Schematic of ammonia-based CO2 capture system ................................................................................................................... 5

Figure 8. Schematic of membrane-based CO2 capture system ................................................................................................................. 6

Figure 9. Schematic of calcium looping CO2 capture system ................................................................................................................... 6

Figure 10. Net plant efficiency of PC power plants using different CO2 capture technologies. ............................................................... 7

Figure 11. Net plant efficiency of plants using different CO2 capture technologies ................................................................................. 7

Figure 12.Plant-level water consumption for plants with different CO2 capture technologies ................................................................. 8

Figure 13. Water consumption for the plant and increase in water consumption because of CO2 capture using FG+ solvent ................. 9

Figure 14. Power plant water consumption because of CO2 capture for different capture technologies ................................................ 10

Figure 15. Change in plant power output and CO2 capture power requirement as a function of sorbent weight fraction for FG+ ........ 10

Figure 16. Change in water consumption as a function of sorbent weight fraction for FG+ ................................................................... 11

Figure 17.IECM screen showing makeup water requirement for an IGCC power plant ......................................................................... 13

Figure 18. IECM screen showing the cooling water requirements of an IGCC power plant with CO2 capture ..................................... 13

Figure 19. Gross plant output of IGCC power plants. ............................................................................................................................. 14

Figure 20. Net plant efficiency of IGCC power plants ........................................................................................................................... 15

Figure 21.Plant-level water consumption for IGCC power plant ............................................................................................................ 15

Figure 22. Water consumption for the plant and increase in water consumption because of CO2 capture for an IGCC power plant using Shell gasifier. Water gas shift reactor H2O/CO = 0.99. ................................................................................................................. 16

Figure 23. Effect of WGS steam to inlet CO on plant water consumption and water consumption because of CO2 capture, for an IGCC plant using Shell gasifier (CO to CO2 conversion = 95%) ............................................................................................................. 16

Figure 24. Screenshot of cooling water requirement in an oxy-combustion power plant. ...................................................................... 17

Figure 25. Effect of direct contact cooler exit temperature on the cooling water requirement and net plant efficiency for an oxy-combustion plant burning Wyoming PRB coal. ............................................................................................................................. 18

Integrated Environmental Control Model - Technical Documentation Acknowledgements • vii

Acknowledgements

This work is supported by the U.S. Department of Energy under Activity No. 0004000.2.672.241.003 from the National Energy Technology Laboratory (DOE/NETL). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not reflect the views of any agency.

Integrated Environmental Control Model - Technical Documentation CO2 Capture Process Water Use • 1


Objectives of this Report This report presents a comparative analysis of plant level water consumption for power plants using different CO2 capture technologies, using the Integrated Environmental Control Model (IECM).

Introduction Water is one of the biggest resources utilized in thermoelectric power plants. Many processes in a power plant use water in one way or another. Some processes consume water while some other processes produce water. Depending on the plant design, a fraction of total water is consumed by the plant or recycled. The net water use of a power plant (called consumption) is the difference between the total plant water use and the amount of recycled water. Plant water consumption increases with the addition of more processes. Depending on the type of power plant (PC, IGCC or NGCC), water is consumed in different processes such as cooling tower, boiler feedwater, slurry preparation, ash and slag handling, syngas humidification, quench water system and FGD system. Among the processes, cooling tower consumes the most water. The addition of CO2 capture process to the power plant increases its water use significantly. Because of the energy penalty associated with CO2 capture, the gross power generation of the power plant should increase in order to produce the same amount of net power output as a plant without CO2 capture. The CO2 capture process itself consumes water. For these two reasons, the plant level water consumption goes up with the addition of CO2 capture process. In this report, the IECM is used to estimate process-level and plant-level water consumption for coal-fired power plants using different CO2 capture technologies. Post-combustion, pre-combustion and oxy-combustion CO2 capture technologies are discussed. Technologies considered in this report for the different plant types are:

1. Post-combustion CO2 capture: a. Amine-based systems:

i. FG+ (Econamine) ii. MEA

b. Chilled ammonia process


c. Membrane-based system and d. Calcium-looping process

2. Pre-combustion CO2 capture: a. GE-Quench gasifier with Selexol CO2 capture b. Shell gasifier with Selexol CO2 capture

3. Oxy-combustion CO2 capture

Water Use Estimation in the IECM The IECM models three different types of cooling water systems for power plants – once-through systems, the wet cooling tower and the air cooled condenser. The schematic representations of the wet cooling tower and the air cooled condenser are shown in figures 1 and 2 respectively.

Figure 1. Schematic of a wet cooling tower used in the IECM

Figure 2. Schematic of an air cooled condenser used in the IECM

Since a typical new plant which meets the new source performance standards uses wet cooling tower as the cooling system, only plants using a wet cooling tower are focused on in this report.


For different plant types, IECM estimates water requirement for different processes and the overall power plant. Three types of results are presented – water withdrawal (called “make-up”), water consumption and the amount of cooling water requirement. Figure 3 shows a sample screenshot of the PC plant in which the plant level make-up water is shown, along with process-level make-up water requirement.

Figure 3. Screenshot showing plant-level and process-level make-up water requirement for a PC power plant

Figure 4 shows a screenshot from IECM PC plant configuration showing water consumption at the plant level divided into different process-level requirements.

Figure 4. Screenshot showing plant-level and process-level water consumption estimation in IECM

Similarly, Fig 5 shows the screenshot showing the cooling water requirement for a PC plant. In plants with CO2 capture, cooling water is also used for intercooling in the multi-stage CO2 product compressor. Appendix A shows the performance model used for calculating the cooling requirement of CO2 compressor.


Figure 5. A sample IECM screenshot showing cooling water requirement for a PC power plant using an amine-based CO2

capture system.

Similar models are used for IGCC and NGCC plant designs also.

Water Consumption in Pulverized Coal Plants Pulverized coal power plants involve combustion of coal in air. For PC plants with CO2 capture, After the removal of criteria air pollutants (PM, SOx and NOx), the flue gas is sent to a CO2 capture process, where CO2 is separated mainly from N2. There are four post-combustion CO2 capture options in IECM – Amine-based, Ammonia-based, membrane-based and chemical (calcium) looping. Amine and ammonia systems are chemical solvent based systems in which absorption of CO2 happens at a low temperature and regeneration happens at a higher temperature by supply of heat in the form of low pressure steam. Membrane system uses pressure swing absorption and regeneration system where absorption occurs at high pressure and desorption happens by releasing the pressure. The calcium looping (CaL) process is a high temperature sorbent based system using CaO/CaCO3 as the sorbent. In the CaL process, regeneration (calcination) happens at a very high temperature. Calcination heat is supplied by oxy-combustion of coal. Brief descriptions of the processes are given below:

Amine CO2 capture system A schematic of the amine system is shown in Fig 6. There are two variations of the amine system in IECM – the conventional MEA and the advanced amine (FG+ or Econamine). The two solvents mainly differ in their heats of absorption and regeneration. MEA has a regenerator heat requirement of 4468 kJ/kg CO2 while FG+ has a regeneration heat requirement of 3517 kJ/kg CO2. Hence, FG+ needs less steam compared to MEA because of which the plant using an FG+ system has higher efficiency than the one using MEA-based system.


Figure 6. Schematic of amine-based CO2 capture system

Ammonia CO2 capture system A schematic of the ammonia system is shown in Fig 7. Compared to amine system, absorption occurs at a much lower temperature in the ammonia system, in order to minimize the vapor losses of ammonia. A refrigeration (chiller) system is needed to achieve such low temperatures. However, the regenerator heat requirement for an ammonia system (2438 kJ/kg CO2) is much lower than that of either of the amine systems.

Figure 7. Schematic of ammonia-based CO2 capture system

Membrane CO2 capture system A schematic of the membrane capture system is shown in Fig 8. Absorption occurs at a high pressure (feed pressure of about 4 bar) and regeneration occurs under vacuum conditions (about 0.2 bar). The regeneration energy needed for this system is predominantly electrical energy for the compressor and vacuum pumps.


Figure 8. Schematic of membrane-based CO2 capture system

Calcium looping CO2 capture system A schematic of the CaL system is shown in Fig X. Adsorption (carbonation) and regeneration (calcination) occur using the reversible reaction of CO2 with CaO to form CaCO3. Calcination heat is supplied through oxy-combustion of coal. The CaL process also produces electricity using heat recovery from the exothermic carbonation reaction and the cooling of hot solid and gas streams.

Figure 9. Schematic of calcium looping CO2 capture system

PC Plant Case Studies In this section, the utility of IECM in estimating plant-level water consumption for power plants using different CO2 capture technologies is illustrated using a few case studies. Only pulverized coal power plant design is presented. All the configurations are assumed to generate a net electrical output of 550MW using Illinois#6 coal. All CO2 capture technologies assume a CO2 capture rate of 90%.


Gross power output and net plant efficiency Almost all the solvent-based post-combustion CO2 capture technologies use thermal energy to regenerate the solvent. Regeneration is usually done by the supply of low pressure steam from the steam cycle. This results in loss of steam to the steam turbine and hence a decrease in the plant net power output. In order to maintain the same net power output, a bigger plant (gross plant size) needs to be built, which results in a corresponding increase in water consumption. Figure 10 shows the gross power output of the power plant in order to generate 550MW of net power. The membrane-based process needs the highest gross power to generate 550MW of net power.

Figure 10. Net plant efficiency of PC power plants using different CO2 capture technologies.

Figure 11 shows the net plant efficiency for the base plant as well for plants when different CO2 capture technologies are used. The net power output of the plant is maintained at 550MW for all these cases. Among the solvent-based CO2 capture technologies (MEA, FG+ and Ammonia), the plant with FG+ has the highest efficiency, followed by the plants with chilled ammonia and MEA.

Figure 11. Net plant efficiency of plants using different CO2 capture technologies


The chilled ammonia process consumes some electrical energy in order to meet the energy requirement for refrigeration of the solvent. Hence, not all the energy penalty is in the form of thermal energy. There is no thermal energy requirement in regeneration when a membrane-based system is used. The energy penalty is because of the electrical energy required to compress the flue gas or to create vacuum conditions. In the calcium looping process, a huge amount of electricity is generated by heat recovery. Hence, energy penalty is not as high as it is when other post-combustion capture technologies are used.

Figure 12.Plant-level water consumption for plants with different CO2 capture technologies

Water consumption Water consumption results are normalized by the net power output. Figure 12 shows the plant specific water consumption (kg/kWh-net) for plants using different CO2 capture technologies. Gross plant size is the main indicator of the water consumption of the power plant. Hence, among the solvent-based processes, the plant with chilled ammonia system consumes the most water. Though the gross plant size is much higher when a membrane-based system is used, the water consumption is lower than for an ammonia-based plant. The reason for such a high water consumption rate for the plant using an ammonia-system is that the chilled ammonia process requires a much higher cooling duty because of the use of much bigger direct contact coolers. Hence, the use of chilled ammonia process causes a significant increase in the plant water consumption. Among all the process areas, cooling tower consumes the maximum amount of water. Losses from cooling tower include losses due to evaporation and also drift losses.

Sensitivity analysis The sensitivity analysis capability of IECM is illustrated here by varying plant design and process design parameters. For all the sensitivity cases, plant gross output was fixed at 650MW (the default PC plant size in IECM).


Effect of CO2 capture rate The CO2 capture rate was varied for the different cases to understand its effect on the plant water consumption. In order to calculate the increase in water consumption because of CO2 capture, the following equation is used:

∆𝑊𝑎𝑡𝑒𝑟 =𝑊𝑎𝑡𝑒𝑟�|𝐶𝐶𝑆 � 𝑘𝑔

𝑘𝑊ℎ� − 𝑊𝑎𝑡𝑒𝑟�|𝑁𝑜 𝐶𝐶𝑆 � 𝑘𝑔𝑘𝑊ℎ�

𝐶𝑂2 �|𝑐𝑎𝑝𝑡𝑢𝑟𝑒𝑑 �𝑘𝑔𝑘𝑊ℎ�

Figure 13. Water consumption for the plant and increase in water consumption because of CO2 capture using FG+ solvent

Figure 13 shows the change in plant water consumption (kg/kWh) and increase in water consumption because of CO2 capture (kg water/kg CO2) with CO2 capture rate for a plant using FG+ CO2 capture technology. The overall water consumption increases with the capture rate owing to the decrease in plant efficiency and increase in the size of CO2 capture process. However, the specific water consumption (kg/kg CO2) decreases with increasing CO2 capture rate. Figure 14 shows the increase in water consumption because of CO2 capture, as a function of the CO2 capture technology and the CO2 capture percentage. For all technologies, plant water consumption decreases with increasing CO2 capture rate. Water consumption per unit CO2 captured is the lowest for an MEA-based plant, followed by the plants with FG+system, membrane-system and Ammonia-based system, respectively. As explained before, ammonia based system consumes the most amount of water.


Figure 14. Power plant water consumption because of CO2 capture for different capture technologies

Effect of process design conditions Amine solvents are usually a mixture of amine sorbent in water. In order to illustrate the effect of process design conditions on the performance of the plant and its water consumption, the sorbent loading of the solvent has been varied. Only FG+ solvent has been considered for this case study.

Figure 15. Change in plant power output and CO2 capture power requirement as a function of sorbent weight fraction for FG+

Figure 15 shows the effect of varying the sorbent weight fraction on the net output of the power plant and the equivalent electrical energy required for the CO2 capture process. The power required for CO2 capture decreases with increasing sorbent weight fraction and so the net power output also increases. Figure 16 shows that the plant water consumption and increase in water consumption because of CO2 capture increase slightly with increasing sorbent weight fraction. This shows that having a solvent with higher sorbent weight fraction leads to increased net plant efficiency but also increases the water consumption of the power plant, though the increase of water consumption is only marginal.


Figure 16. Change in water consumption as a function of sorbent weight fraction for FG+

Water Consumption in IGCC Plants In IGCC power plants, coal is first gasified in oxy-fired gasifiers. The produced syngas is cooled and sent to a gas cleanup process consisting of a water gas shift (WGS) reactor and a Selexol-based H2S and CO2 capture process. The clean syngas, consisting mainly of H2, is combusted in a combined cycle power plant. In an IGCC power plant, several process areas have a water demand. However, because of availability of water from several other process areas, a significant amount of water is recycled internally in the plant. The water which cannot be recycled should be supplied from external sources. The process areas with water demand in an IGCC power plant are:

• Slag handling • Slurry water • Quench or wash • Humidifier • Sour water stripper • Boiler feedwater • Gasifier steam • Water gas shift steam • Combustion turbine dilution

Slag handling Slag is a by-product of gasification reactions. Ash in coal comes out as slag, either in a dry or a wet form. For slurry-based gasifiers such as GE (quench and radiant designs), ash mixes with excess water in the slurry, forming wet slag, which is removed from the bottom of the gasifier. For dry-feed gasifiers such as Shell designs, molten slag is cooled and run down the gasifier walls to be removed at the bottom. Slag handling equipment has water demand in treating and disposing the slag.


Slurry water For slurry-based gasifiers such as GE, coal is fed into the gasifier in the form of a slurry. The amount of water in the slurry depends on the coal type. For eg, high-quality bituminous coals need a moisture content of about 33% in the coal feed to gasifier. There is water demand for the coal slurry in slurry-based systems.

Quench/Wash and Sour Water Stripper (SWS) Raw gas exiting the gasifier is sometimes cooled using quench water. This process also helps remove impurities such as chlorides, NH3, particulates etc. The cooled and clean syngas is sent for further treatment. The bottoms from this scrubber (water mixed with particulates) is sent to a sour water stripper. Both quench and SWS processes need water.

Syngas Humidifier and Combustion Turbine Dilution In IGCC plants, syngas is diluted with water vapor or nitrogen to reduce its heating value to meet the limits set by gas turbine manufacturers. When sufficient N2 is not available for dilution, water vapor is used. Hence syngas humidification has some water demand.

Gasifier Steam In dry-feed gasifiers such as Shell designs, steam is added to the gasifier to supply water for the gasification reactions.

Water Gas Shift Steam In CO2 capture cases, water gas shift (WGS) reaction is used to convert CO in the syngas to CO2 and H2, as shown in the following reaction: CO + H2O CO2 + H2 Usually, moisture in the syngas (from the gasifier and quench processes) is sufficient for the shift reaction. Additional steam might be needed to meet the WGS requirements. The molar ratio of H2O:CO is maintained around 1.8:1 for the WGS reaction.

IGCC Water Use Estimation in IECM The NETL baseline report (2015) was used to estimate the water use factors for different IGCC process areas in IECM models. The reported values of water demand, internal recycle and process discharge for different process areas were used to estimate the process level water use factors. The cases used from the NETL report were – Shell gasifier without CCS (Case B1A), Shell gasifier with CCS (Case B1B) and GE (Quench) gasifier with CCS (Case B5B-Q). Table 1 shows the factors incorporated into the IGCC models of IECM. Figure


Table 1. Water use factors (gal/MWh) added to IECM

GE Quench Shell

No CCS CCS No CCS CCS Slag handling 0 0 6.56 0 Quench/Wash 139.5 154.0 0 30.30 Humidifier 0 0 34.78 0 Gasifier steam 0 0 8.06 10.84 Shift steam (only CCS) 0 0 0 105.6 CT steam dilution 0 0 12.85 0

Figure 17.IECM screen showing makeup water requirement for an IGCC power plant

Cooling water is also used for the air separation unit (ASU) and intercooling of CO2 compressor. Figure 18 shows a screenshot of the cooling water requirements in an IGCC power plant with CO2 capture, which includes water for ASU and CO2 compressor.

Figure 18. IECM screen showing the cooling water requirements of an IGCC power plant with CO2 capture


IGCC Plant Case Studies As was done for the PC cases, in this section the utility of IECM in estimating plant-level water consumption for IGCC power plants is illustrated. All the configurations are assumed to generate use two GE-7FB gas turbines, burning Illinois#6 coal. Both GE (quench) and Shell gasifier cases are presented. A CO2 capture rate of 90% is assumed.

Gross power output and net plant efficiency In an IGCC power plant, pre-combustion CO2 capture occurs at a high pressure (> 30 bar). Hence solvent regeneration is achieved using a pressure swing. Because no additional heat (supplied by steam energy) is needed for regeneration, the efficiency loss caused by pre-combustion CO2 capture is much less compared to that in post-combustion capture. Additionally, CO2 is also released at a high pressure (~ 2 bar). As a result, energy needed for CO2 compression is also much smaller than in post-combustion CCS. Figure 19 shows the gross power and net power output of IGCC power plants using GE (Quench) and Shell gasifiers, with and without CCS. Figure 20 shows the net plant efficiency for the same plant configurations. It can be seen that plants using the dry-feed Shell gasifier are in general more efficient than the slurry-feed GE gasifier cases.

Figure 19. Gross plant output of IGCC power plants.


Figure 20. Net plant efficiency of IGCC power plants

Figure 21.Plant-level water consumption for IGCC power plant

Water consumption Water consumption results are normalized by the net power output. Figure 21 shows the plant specific water consumption (kg/kWh-net) for the different IGCC plant configurations considered here. In general, IGCC plants consume much less water compared to the PC plants. Unlike in the PC cases, gross plant size is not the only indicator of the water consumption of the power plant. As mentioned before, a significant portion of water is produced in the IGCC power plant and a major portion of it is recycled. The water consumption of plants with GE gasifier is higher because of the wet slurry-based coal feed system. Among all the process areas, cooling tower consumes the maximum amount of water. Losses from cooling tower include losses due to evaporation and also drift losses.


Sensitivity analysis The sensitivity analysis capability of IECM is illustrated here by varying the CO2 capture rate. In IGCC power plants, CO2 capture rate is varied by varying the CO to CO2 conversion efficiency in the WGS reactor. Figure 22 shows the change in plant water consumption (kg/kWh) and increase in water consumption because of CO2 capture (kg water/kg CO2) with CO2 capture rate (CO to CO2 conversion rate in the WGS reactor) for a plant using Shell gasifier. The overall water consumption increases with the capture rate owing to the decrease in plant efficiency and increase in the size of CO2 capture process. However, the specific water consumption (kg/kg CO2) decreases with increasing CO2 capture rate.

Figure 22. Water consumption for the plant and increase in water consumption because of CO2 capture for an IGCC power

plant using Shell gasifier. Water gas shift reactor H2O/CO = 0.99.

The other variable that was varied was the stoichiometric molar ratio of steam to inlet CO in the WGS reactor. Figure 23 shows the effect of this variable on the plant water consumption and water consumption because of CO2 capture.

Figure 23. Effect of WGS steam to inlet CO on plant water consumption and water consumption because of CO2 capture, for an

IGCC plant using Shell gasifier (CO to CO2 conversion = 95%)


Water Consumption in Oxy-Combustion Plants In oxy-combustion plants, coal is combusted in an oxy-fired boiler which results in a flue gas containing predominantly of CO2 and H2O. Oxidant is supplied from an air separation unit (ASU). After purification and recycling, H2O is condensed using a direct contact cooler (DCC) and the resulting flue gas is further purified in a CO2 purification unit (CPU) to obtain very high purity CO2 for pipeline transport. Both ASU and CPU consume significant amounts of water. Detailed thermodynamic models for direct contact cooler and CPU were developed and incorporated into IECM. All the water consumption is to supply cooling load to different processes. Figure 24 shows a sample screenshot of the cooling water requirement in an oxy-fired power plant.

Figure 24. Screenshot of cooling water requirement in an oxy-combustion power plant.

Table 2 shows the performance results for two oxy-combustion plants, one burning Appalachian Medium Sulfur (bituminous, high-sulfur) coal and the other burning Wyoming PRB coal (sub-bituminous, low-sulfur coal) modeled in IECM. The plant with higher sulfur coal uses a wet fluegas desulfurization (FGD) technology for SOx control, while the plant with low sulfur coal uses a lime spray dryer. As can be seen in the table, the plant burning sub-bitumimous coal (Wyoming PRB) is slightly less efficient than the one burning bituminous coal. Because of the low carbon content of sub-bituminous coal, much higher quantity of coal is required to produce the same gross power output. As a result, the CO2 emissions are also slightly higher for the Wyoming PRB plant. As mentioned before, ASU and CPU require cooling water. Higher coal flow rate requires higher oxygen supply, because of which the Wyoming PRB plant has higher cooling water requirement for ASU. Also, because of much higher coal moisture content in Wyoming PRB, more water needs to be removed in the direct contact cooler before the flue gas is sent to CPU. As a result, the DCC cooling water requirement is much higher for the PRB plant compared with the bituminous coal plant. However, the dry sulfur capture technology used in the PRB plant does not consume any water while the wet FGD used in the bituminous plant consumes water. The net make-up water (withdrawal) requirement for both the plants is similar.

Sensitivity analysis Since DCC needs a significant amount of water, it was chosen here to demonstrate the sensitivity analysis capability of IECM. The effect of varying DCC exit temperature on its cooling water requirement and plant efficiency is shown in Figure 24. Lower the DCC exit temperature, more


cooling is required. Hence as the DCC temperature increases, the cooling water requirement decreases. However, higher temperature leads to a decrease in efficiency because of its effect on the downstream CPU power requirement. Table 2. Performance results of two oxy-combustion plants modeled in IECM. The gross power output is 650 MW for both plants. Purity of oxygen is 97%.

Coal Appl Med S Wyoming PRB Coal Sulfur content (% wt, as-received) 2.13 0.37 Coal moisture content (%wt, as-received) 5.05 30.2 Coal flow (tonnes/hr) 182 304 Net power output (MW) 484 482 Net plant efficiency (%HHV) 31.0 29.5 CO2 emissions (kg/kWh) 0.10 0.11 CO2 captured (kg/kWh) 0.93 1.01 Water consumption (total withdrawal) (kg/kWh)

3.42 3.45

Cooling water for ASU (kg/kWh) 5.47 5.63 Cooling water for DCC (kg/kWh) 14.6 20.4

Figure 25. Effect of direct contact cooler exit temperature on the cooling water requirement and net plant efficiency for an oxy-

combustion plant burning Wyoming PRB coal.

Conclusion This report presented a brief overview of the IECM water consumption models. Case studies for three coal plant designs were presented – pulverized coal (PC), integrated gasification combined cycle (IGCC) and oxy-combustion. The IECM was used to estimate water consumption in plants using several CO2 capture technologies. The capability of IECM to conduct sensitivity analyses


was presented through a series of case studies. The IECM was demonstrated to be a versatile tool that can be evaluating the water requirements of a power plant.

Appendix – CO2 Compressor Cooling Water The following equations are used to calculate the cooling water requirement for multi-stage CO2 product compressor. Compressor energy requirement

∆𝐻 = 𝐶𝑝∆𝑇

𝜂 =Δ𝐻𝑖𝑠𝑒𝑛Δ𝐻𝑎𝑐𝑡

Actual outlet temperature:

𝑇2,𝑎𝑐𝑡 = 𝑇1 +𝑇2,𝑖𝑠𝑒𝑛 − 𝑇1

𝜂

Isentropic outlet temperature:

𝑇2,𝑖𝑠𝑒𝑛 = 𝑇1 �𝑝2𝑝1�𝛾−1𝛾

Actual heat requirement (single stage):

Δ𝐻𝑎𝑐𝑡(𝑘𝐽

𝑘𝑚𝑜𝑙) =

𝐶𝑃𝑇1𝜂 ��

𝑝2𝑝1�𝛾−1𝛾− 1�

Actual heat requirement (multiple stages):

Δ𝐻𝑎𝑐𝑡(𝑘𝐽

𝑘𝑚𝑜𝑙) = 𝑛

𝐶𝑃𝑇1𝜂 ��

𝑝2𝑝1�𝛾−1𝛾𝑛

− 1�

For simplicity, Cp can be assumed to be constant (42 kJ/kmol), which is fairly constant for temperatures less than 120oC. For PC and Oxy-fuel cases, initial pressure = 1 atm. For IGCC, initial pressure = 10 atm (NETL baseline 2015). Assume 6 stages (n = 6) for both cases (it is not clear from NETL baseline 2015 how many stages are there in IGCC plant).


References [1] Integrated Environmental Control Model Carbon Sequestration Edition, Carnegie Mellon University http://www.iecm-online.com.

[2] US DoE/NETL Cost and performance baseline for fossil energy plants, DOE/NETL-2010/1392, Revision 2a, 2013

[3] US DoE/NETL Cost and performance baseline for fossil energy plants, Volume 1a, DOE/NETL-2015/1723, Revision 3, 2015

http://www.iecm-online.com/

IECM Technical Documentation: Volume IV

Life Cycle Water Use of Coal- and Natural-Gas-fired Power Plants with and without Carbon Capture and Storage

April 2016

Disclaimer

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference therein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed therein do not necessarily state or reflect those of the United States Government or any agency thereof.


Life Cycle Water Use of Coal- and Natural-Gas-fired Power Plants

with and without Carbon Capture and Storage


Prepared for:




www.netl.doe.gov



April 2016


Table of Contents

Life Cycle Water Use 1

Objectives of this Report ........................................................................................................... 1 Introduction ............................................................................................................................... 1 Materials and Methods............................................................................................................... 2

Overview of Analysis Scope and Methods .................................................................. 2 Water Use by Stage ..................................................................................................... 3 Uncertainty Analysis ................................................................................................... 6

Results and Discussion .............................................................................................................. 7 Conclusions ............................................................................................................................. 18 Acknowledgements .................................................................................................................. 19 References ............................................................................................................................... 19 Appendix ................................................................................................................................. 22


List of Figures

Figure 1 Life cycle analysis boundary ............................................................................................................................... 3

Figure 2 Effects of coal type on life cycle water use of a PC plant under emission limit of 1400 lb CO2/MWh ...... 15

Figure 3 Effects of cooling type on life cycle water use of PC and NGCC power plants under CO2 emission limits .................................................................................................................................................................................... 15

Figure 4 Probability distribution of life cycle water use of a supercritical PC plant using a recirculating cooling system with and without CCS .................................................................................................................................. 16

Figure 5 Probability distribution of life cycle water use of a NGCC plant using a recirculating cooling system .... 17

Figure 6 Tornado diagrams on life cycle water withdrawals for PC plant with partial CO2 capture and shale-gas-fired NGCC plant (a) PC plant; (b) NGCC plant .................................................................................................. 18


List of Tables

Table 1 Use and consumption factors applied to Ecoinvent elementary water flowsa .................................................. 5

Table 2 Water use factors of coal production cyclea ........................................................................................................ 7

Table 3 Water use factors of natural gas production cyclea ............................................................................................ 7

Table 4 Water use for plant infrastructure based on EIO-LCAa ................................................................................... 9

Table 5 Water use for major chemical production based on Ecoinvent 2.2 ................................................................ 10

Table 6 Major technical assumptions and parameters for power plant modeling ..................................................... 11

Table 7 Water use of PC and NGCC power plants with and without CCS ................................................................. 12

Table 8 Life cycle water use of coal-fired power plants with and without CCS.......................................................... 13

Table 9 Life cycle water use of NGCC power plants with and without CCS .............................................................. 14


Acknowledgements


Integrated Environmental Control Model - Technical Documentation Life Cycle Water Use • 1

Life Cycle Water Use

Objectives of this Report This study conducts a hybrid life cycle analysis to estimate and characterize the water use of coal- and natural-gas-fired power plants with and without carbon capture and storage (CCS), including quantification of variability and uncertainty in the life cycle water use.

Introduction Low-carbon energy production has been increasingly becoming important for mitigating climate change. To cut carbon dioxide (CO2) emissions, the U.S. Environmental Protection Agency (EPA) has established CO2 emission performance standards for new fossil fuel-fired power plants. The final standard for new coal-fired power plants is an emission limit of 1400 pounds of CO2 per megawatt‐hour on a gross‐output basis (lb CO2/MWh‐gross), which is less stringent than the initially proposed standard of 1100 lb CO2/MWh‐gross (EPA, 2014 and 2015a). For new natural gas combined-cycle (NGCC) plants, the standard is an emission limit of 1000 lb CO2/MWh‐gross (EPA, 2015a). The EPA also issued the Clean Power Plan under Section 111(d) of the Clean Air Act to reduce CO2 emissions from existing power plants, which sets up state-specific emission targets to reduce nationwide carbon pollution by an average of 32% below 2005 levels in 2030 (EPA 2015b). Carbon capture and storage (CCS) is regarded as one of the best emission reduction systems in the EPA’s rules for new fossil fuel-fired electricity generating units (EGUs), whereas improved utilization of NGCC power plants is considered as one of the key mitigation measures for existing plants. However, retrofitting CCS for partial CO2 capture also is a viable option for some existing coal-fired EGUs to comply with the Clean Power Plan, depending on unit attributes and fuel prices (Zhai et al, 2015). To comply with the CO2 emission limit of 1100 lb/MWh gross, adding amine-based CCS to new pulverized coal-fired (PC) power plants would increase plant-level water use by roughly 20−50% (Talati et al, 2014). In addition, the high energy and infrastructure demands of CCS also contribute to natural resource use and chemicals used for CO2 capture. When unconventional fuels such as shale gas are used to fire NGCC power plants, the production of shale gas requires large volumes of water for drilling and hydraulic fracturing, which could be several times of those for conventional natural gas production (Clark et al, 2013; Meldrum et al, 2013) and intensify pressure on local water resources (Soeder and Kappel 2009). Depending on the availability of water resources at different supply chains for electricity generation, low-carbon energy regulations and policies


could pose complex water management challenges for fossil fuel-fired power plants, and their impacts on water resources need to be examined on a life cycle basis. Life cycle environmental impacts of CCS have received increasing attention (Koornneef et al, 2008; Korre et al, 2010; Singh et al, 2011; Zapp et al, 2012; Corsten et al, 2013; Grant et al, 2014; Zhang et al, 2014). However, less attention has been paid to the life cycle water use issue. Water withdrawal and water consumption are the two metrics that are often adopted to measure water use. Water withdrawal refers to the total amount of water taken from a source while consumption refers to the loss of water that is not returned to the source (e.g. due to evaporation) (Zhai et al, 2011). Meldrum et al (2013) present consolidated estimates of life cycle water use for various electricity generation technologies by conducting a broad review and analysis on existing references and find that the plant operation for electricity generation dominates the life cycle water use for fossil fuel-fired power plants in most cases. When power generation is considered as end use of shale gas, hydraulic fracturing only accounts for 6.2% of the life cycle water consumption (Laurenzi and Jersey 2013). In comparison, NGCC plants use much less water than PC plants (Meldrum et al, 2013; Fthenakis et al, 2010; Laurenzi and Jersey 2013). Deterministic comparisons of water footprints among 36 coal-based power generation pathways show that life cycle water use is sensitive to the choice of fuel production and transport methods and the power plant configuration (Ali and Kumar 2015). Adding CCS to PC and NGCC power plants would roughly double the life cycle water use (Fthenakis and Kim 2010; Meldrum et al, 2013). However, existing life cycle studies on CCS mainly focus on 90% CO2 capture and highly depend on water use data collected from various sources without sufficient access to power plant or system designs and quantification of variability and uncertainty. Besides, water used in plant infrastructure and chemicals production is not often included in the analysis.

The major objectives of this paper, therefore, are to (1) estimate and characterize the life cycle water use of coal- and natural-gas-fired power plants under the constraint of different CO2 emission limits, especially the U.S. EPA's newly issued new source performance standards (NSPS) for limiting CO2 emissions; (2) examine the variability in life cycle water use by key factors including fuel type and supply approach, and power plant and CCS designs (e.g. wet vs. dry cooling; partial vs. full CO2 capture); and (3) quantify the uncertainties in major stages or supply chains across the life cycle to provide probabilistic water use estimates. When CO2 capture is needed, commercially available amine-based CCS is employed for power plants. This study provides a systematic inventory and water implications of low-carbon electricity generation across the life cycle. The term of water use includes both water withdrawal and water consumption. Given that the U.S. EPA’s rules present the standards in the English units, the variables of this study are expressed in the English unit system. However, unit conversion factors from English to metric unit systems are provided for international readers in the appendix.

Materials and Methods

Overview of Analysis Scope and Methods In this analysis, the major stages of the life cycle for electric power generation include fuel acquisition, processing and transport, power plant operation, production of chemicals used in power plants, and power plant infrastructure. Figure 1 shows the life cycle analysis scope, in which the fuel supply stage includes fuel extraction, processing and transport; the power plant


operation stage covers water use for electricity generation; ammonia, limestone, and monoethanolamine (MEA) used in environmental control systems are included in the water analysis for chemical production; and the plant infrastructure stage mainly takes into account upstream component manufacturing and power plant construction. A process-based hybrid life cycle assessment (LCA) is conducted by incorporating relevant information from a power plant model with inventory databases and an Economic Input-Output -LCA tool for water use estimates. The life cycle water use measured in gallons per megawatt hour (gal/MWh) is estimated as: 𝑳𝑪𝑾𝑼 = 𝒘𝒖𝒇𝑭𝑺 ∗ 𝑭𝑼

𝑬𝑴𝑾+ 𝒘𝒖𝒇𝐎𝐏 + 𝒘𝒖𝒇𝐂𝐇𝐄 ∗ 𝑪𝑯𝑬

𝑬𝑴𝑾+ 𝒘𝒖𝐏𝐈 ∗ 𝟏

𝑬𝑴𝑾𝒉 (1) where 𝑳𝑪𝑾𝑼 is the life cycle water use (gal/MWh); 𝒘𝒖𝒇𝑭𝑺 is the water use factor of fuel supply (gal/ton for coal, gal/MMscf for natural gas); ); 𝒘𝒖𝒇𝐎𝐏 is the water use factor of plant operation for electricity generation (gal/MWh); ); 𝒘𝒖𝒇𝐂𝐇𝐄 is the water use factor of chemical production (gal/ton); 𝒘𝒖𝐏𝐈 is the total water use for plant infrastructure (gal); 𝑪𝑯𝑬 is the amount of chemical used in a power plant (ton/hr); 𝑭𝑼 is the amount of fuel used in a power plant (ton/hr for coal; MMscf/hr for natural gas); 𝑬𝑴𝑾is the net plant power output (MWnet); 𝑬𝑴𝑾𝒉 is the total electricity generation of power plant over the lifetime (MWh). The Integrated Environmental Control Model, a power plant modeling tool, serves as the basis for the process-based LCA and estimates a variety of mass and water streams at coal- and natural gas-fired power plants (IECM, 2012). The water use factors of fuel supply are estimated based on the existing data from literature. A life cycle inventory database is applied to estimate the water use factors of chemical production (Goedkoop et al, 2013), while an Economic Input-Output Life Cycle Assessment (EIO-LCA) tool is employed to estimate the water use factors of plant infrastructure (CMUGDI, 2015).

Figure 1 Life cycle analysis boundary

Water Use by Stage Fuel Supply Coal can be extracted by surface mining or underground mining and transported mainly by trains or by slurry pipelines in occasional cases. Dust control inside the mine leads to more direct water use for underground mining than surface mining. Furthermore, extensive use of mine equipment

Extraction Processing Transport

Fuel Supply

Plant Infrastructure

Plant Operation Plant w/o CCSPlant w/ CCS

Chemical Production

Electricity


for constructing the shaft and operation of ventilation fan also result in more indirect water use for underground mining (Fthenakis et al, 2010). Natural gas can be extracted by conventional drilling or hydraulic fracturing (mainly for shale gas) and transported mainly by pipelines (Meldrum et al, 2013). Water use may vary significantly with extraction site. Meldrum et al (2013) conducted a comprehensive review that collects and screens a wide range of references regarding water use for fuel production and electricity generation. To address the discrepancy among collected data because of the differences in production pathways, analysis boundaries, and performance parameters, their review study first conducts a harmonization analysis based on such key performance parameters as net plant efficiency and fuel heating value and then presents the minimum, median, mean and maximum estimates of fuel supply water use intensities in the form of gallons per MWh for both coal and natural gas (Meldrum et al, 2013). Thus, their water use intensity estimates are used in our analysis, including uncertainty analysis. Given that the water use factor required in Equation (1) is on the basis of either mass or volume, the fuel supply water use factors in the form of gallons per ton (gal/ton) for coal or gallons per million standard cubic feet (gal/MMscf) for natural gas are derived or back-calculated from the reported water use intensities in terms of the common harmonization performance metrics. Plant Infrastructure The plant infrastructure stage involves manufacturing of many components (e.g. coal and sorbent handling systems, and combustion turbines) and plant construction (Ruether et al, 2004). An EIO-LCA tool developed by Carnegie Mellon University (CMU)'s Green Design Institute (CMUGDI) is employed to estimate the water use in this stage (CMUGDI, 2015). The EIO-LCA tool consists of economic input-output models, and resource use and emissions data and provides estimates of the materials and energy resources required for economy-related activities and their resource use and emissions throughout the supply chain (CMUGDI, 2015). The assessment is based on proper classification of different predefined economic activities. The EIO-LCA model requires use of the costs in particular economic sectors (CMUGDI, 2015). The classification method developed by Ruether et al (2004) who applied the EIO-LCA model to estimate greenhouse emissions associated with the construction of IGCC plants is adopted to formulate 16 economic sectors for power plants. New PC and NGCC plants presented in the Nation Energy Technology Laboratory (NETL)'s baseline report are used as the surrogate plants for infrastructure-related water assessments, including both cases with and without CCS (NETL 2013). For the illustrative purposes, Appendix Tables A-1 and A-2 present examples of the economic sector classification and cost by plant section for PC and NGCC plants without CCS. The 2002 purchaser price model of EIO-LCA is employed to estimate water use for each economic sector and then the total water use for the entire plant infrastructure. However, the 2002 purchaser model only reports water withdrawal requirements. Although the 1992 producer price model of EIO-LCA reports both water withdrawal and consumption information, the 2002 purchaser model has updated water withdrawal information. The total water consumption for plant infrastructure is estimated as the product of the total water withdrawals multiplied by a water consumption adjustment factor. The water consumption adjustment factor is estimated as the ratio of the total water withdrawals versus water consumption derived from the 1992 EIO-LCA producer model for a given power plant. Chemical Production


There are several major chemicals used in environmental control systems, including ammonia for hot-side selective catalytic reduction (SCR), limestone for wet flue gas desulfurization (FGD), and monoethanolamine (MEA) for amine-based CCS. Ecoinvent 2.2, a widely-used life cycle inventory (LCI) database, is used to estimate the water use for chemical production. Ecoinvent contains inventories of more than 10,000 standard processes in terms of materials, construction and so on (Goedkoop et al, 2013). For each process, there are four dataset options at most: allocation default/unit process, allocation default/ system process, consequential/ unit process, and consequential/system process. “Allocation default” is applied for attributional modeling to measure an environmental footprint from cradle to grave, whereas “consequential” modeling is applied to measure the consequence of a change to the baseline situation (Goedkoop et al, 2013). The “unit process” only contains inputs and emissions from one process step plus upstream input from other unit processes, whereas the “system process” summaries the detailed life cycle inventories (LCI) results of all these related unit processes (Goedkoop et al, 2013). The allocation default/system process dataset is selected to estimate the water use for chemical production. Detailed information of chemical production processes is presented in Appendix Table A-3. The selection of system processes used for inventory estimation for individual chemicals is described in Appendix Table A-4 (Goedkoop et al, 2013). Ecoinvent 2.2 reports a range of elementary water intake flows to a production system based on the type of water sources such as fresh water and salt water. For an elementary water flow, the water intake is equal to water withdrawal. However, the information of water discharge flows is neither completely qualified nor quantified completely (Flury et al, 2012). A simplified approach developed by Flury et al (2012) is adopted to estimate water consumption by applying a set of consumption fraction factors to the elementary water flows. The total water withdrawals and consumption are estimated as the sum of individual elementary water flows weighted by withdrawal factor and consumption factors reported in Table 1 (Flury et al, 2012), respectively. Table 1 Use and consumption factors applied to Ecoinvent elementary water flowsa

Elementary water flow Withdrawal factor

Consumption factor

Water, cooling, unspecified origin 1b 0.05c Water, lake 1b 0.1d Water, river 1b 0.1d Water, well, in ground 1b 0.1d Water, rain 0e 0e Water, unspecific natural origin 1b 0.1d Water, turbine use,

0f

0f unspecified natural origin Water, salt, ocean

0g

0g Water, salt, sole a Source of information: Flury et al, 2012. b For an elementary water flow, the water intake is equal to water withdrawal. So, the water withdrawal factor is one. c 95% of water withdrawn for industrial cooling is returned based on statistical data (Munoz et al, 2010). Thus, 0.05 is the consumptive use factor for cooling water. d According to the water use categories provided by Shaffer (2008), these elementary flows are categorized as “industrial water use” with a water consumption coefficient of 0.1.


e A zero value is applied to rain water because it is not reported in Ecoinvent 2.2 (Frischknecht and Jungbluth 2007). f Turbine-use water is the intake water from hydroelectric power plants. According to the U.S. Geological Survey, hydroelectric power generation does not withdraw water or divert water flow, it is categorized as "in-stream" water use (Fthenakis and Kim 2010). g Because traditional water footprint studies only concern three types of water category: blue water (fresh water surface/underground), green water (rain) and gray water (for pollutant dilution). Salt water (ocean/sole) is out of water footprint studies and is just related to “offshore works, overseas ship transport” (Frischknecht and Jungbluth 2007). Thus, the contribution to water withdrawal/consumption calculation also is 0. Power Plant Operation

The Integrated Environmental Control Model (IECM) developed by CMU is applied to conduct plant-level modeling and analysis for coal- and natural gas-fired power plants (IECM, 2012). The IECM has a range of power plant systems and configurations that can employ a variety of fuels, air pollution control systems, and cooling systems, including wet towers and air-cooled condensers for dry cooling. Plant and process performance models are developed mainly based on fundamental mass and energy balances and further link to their economic models that estimate capital and operation and maintenance (O&M) costs, and annual levelized costs.

The IECM has a water systems module that can estimate water use for the cooling system, steam cycle, and environmental control systems if applicable (Zhai and Rubin 2010; Zhai et al, 2011). We apply the IECM to configure a new PC plant with a supercritical (SC) boiler and a new NGCC plant with two GE 7FB gas turbines and then estimate the amounts of fuel use, chemical use, and water use for both cases with and without CCS, depending on fuel and cooling system types. To evaluate the life cycle water impacts of controlling CO2 emissions, we first determine the CO2 capture level required for CCS to comply with the given emission rate limit and then estimate the amounts of fuel use, chemical use, and water use (Zhai and Rubin 2013). Given that the bypass design is cost-effective for non-full CO2 control by amine-based CCS (Rao and Rubin 2006), it is adopted for partial CO2 capture.

Uncertainty Analysis Uncertainties are quantified for major stages across the life cycle. Measured data and literature data along with our own judgment are used to formulate probability distribution functions of key uncertain parameters. For example, uniform or triangular distributions assigned to uncertain parameters of the fuel supply based on the minimum, median and maximum water use factors reported by a review study (Meldrum et al, 2013). To assess the uncertainty of water use in the plant operation stage, we first assign probability distribution functions to key uncertain variables such as ambient air conditions, auxiliary cooling, and CCS cooling duty and then conduct a stochastic simulation in the IECM to get the probabilistic distribution of plant water use intensity. With the minimum, median and maximum values of the probabilistic results, a triangular distribution is formulated to specify the uncertainty in the plant operation water use. It is assumed that there is no correlation between the stages. With probability distribution functions assigned to uncertain parameters of major stages, Monte Carlo simulations with 1000 samples are conducted using the uncertainty analysis tool @Risk 6.3 (Palisade 2015) to obtain the probabilistic distribution of life cycle water use, which allows the collective effects of uncertainties in many different parameters to be considered simultaneously and quantify the likelihood of various outcomes.


Results and Discussion Water use factors by stage

Fuel Supply. The amount of water use for fuel production is estimated as the product of mass- or volume-based water use factors and the mass or volume flow rate of fuel consumed for electricity generation, which is estimated using the IECM for given power plant and CCS designs. As discussed in Section 2.2.1, the water use factors (gal/ton or gal/MMscf) are derived from the water use intensities in gallons of water per MWh based on the common harmonization metrics reported by Meldrum et al (2013): the assumed net plant efficiency is 35.4% for PC plants and 51% for NGCC plants, while the assumed fuel heating value is 21 MMBtu/ton for coal and 1031 MMBtu/MMscf for natural gas. Tables 2 and 3 report the minimum, nominal, and maximum values of fuel supply water use for coal and natural gas production, respectively.

Table 2 Water use factors of coal production cyclea

Stage Water Withdrawals (gal/ton) (min., nominal, max.)

Water Consumption (gal/ton) (min., nominal, max.)

Extraction Surface miningb (1.09, 6.54, 28.3) (1.09, 6.54, 28.3)

Underground miningb (17.4, 58.8, 392) (14.7, 58.8, 392) Processing (19.6, 39.2, 2179.2) (19.6, 39.2, 2179.2) Transport

Train (1.09, 2.18, 4.36) (0.22, 1.09, 2.18) Slurryb (218, 240, 893) (218, 240, 893)

a The water use factors are derived from the water use intensities in gallons of water per MWh based on the common harmonization metrics reported by Meldrum et al (2013).

b Water consumption is equal to water withdrawal due to lack of data and nature of water use. Table 3 Water use factors of natural gas production cyclea

Stage Water withdrawals (gal/MMscf)

(min., nominal, max.)

Water consumption (gal/MMscf) (min., nominal, max.)

Extraction conventional drillingb (15.4, 154, 2927) (15.4, 154, 2927) hydraulic fracturingb (154, 1849, 28655) (154, 1849, 28655)

Processingb (77.0, 77.0, 77.0)c (77.0, 77.0, 77.0)c

Transport pipeline (616, 616, 2003)d (154, 462, 924)

a The water use factors are derived from the water use intensities in gallons of water per MWh based on the common harmonization metrics reported by Meldrum et al (2013).

b Water consumption is equal to water withdrawal due to lack of data and nature of water use. c There is only one value reported in this category. d The minimum value and median (nominal) value are equal in Meldrum et al (2013).


Tables 2 and 3 show that the fuel supply water use varies significantly with the fuel

extraction and transport approaches. Underground coal mining uses roughly eight times more water than surface coal mining, mainly for underground dust control. Transporting coal by slurry pipelines results in a much greater amount of water use than train transport. Hydraulic fracturing for shale gas extraction uses roughly eleven times more water than conventional natural gas production because it requires large amounts of fracture fluid (contains proppants and chemical additives) to create fracture pathways through shale, while conventional geological formations contain natural gas deposits (Clark et al, 2013). Besides, there is a big uncertainty in water use for the shale gas extraction, mainly driven by the amount of fracturing liquid used, times of hydraulic fracturing work conducted, and the estimated ultimate recovery of the well (Clark et al, 2013). Plant Infrastructure. The water use for component manufacturing and plant construction mainly depends on the plant size, plant configuration and process designs, which determine the capital costs of individual plant sectors. In this assessment, the reference PC plant has a capacity of 550 MW(net) for both cases with and without CCS, whereas the reference NGCC plant have a capacity of 555 MW-net for the case without CCS and 474 MW-net for the case with CCS (NETL 2013). The addition of CCS for 90% CO2 capture would decrease the net plant efficiency from 39.3% to 28.4% for the PC plant and 50.2% to 42.8% for the NGCC plant (NETL 2013). Their total plant purchaser prices are reported in Table 4. Given the plant capacity factor of 85% and the assumed plant lifetime of 60 years, the water use factors of this stage are estimated as the product of the total water use divided by the total electricity generation over the plant lifespan. Using the same approach discussed in Section 2.2.2, the water use of plant infrastructure is also estimated for coal-fired plants under the emission limits of 1100 and 1400 lb CO2/MWh. As shown in Table 4, adding CCS systems to PC and NGCC power plants for partial and full CO2 capture would significantly increase the infrastructure water use. In comparison of water use factors between PC and NGCC plants, the manufacturing and construction of NGCC plants use roughly 35% of water required for PC plants for the case without CCS and about 45% for the case with 90% capture.


Table 4 Water use for plant infrastructure based on EIO-LCAa

Parameter Supercritical PC NGCC

CO2 capture design w/o CCS

20% CCSb

40% CCSb

Full CCSb

w/o CCS

Full CCSb

Plant Size (MWnet) 550 550 550 550 555 474 Total plant purchaser price (2002M$) 682 884c 965c 1206 244 437

Total water consumption (billion gal)d 0.70 1.07 1.10 1.37 0.26 0.55

Total water withdrawals (billion gal) 4.00 5.81 6.08 7.60 1.40 2.98

Water consumption factor (gal/MWh) 2.85 4.34 4.47 5.55 1.03 2.59

Water withdrawal factor (gal/MWh) 16.3 23.6 24.7 30.9 5.62 14.1 a There are 14 cost categories defined for a coal- or gas-power plant in the NETL's baseline report. To match with economic activities defined in the EIO-LCA model, the mapping rule used in Ruther et al (2004) is applied to formulate 16 categories: (a) The equipment costs of the first 13 sections in the NETL's baseline report; (b) the equipment installation cost as the sum of material and installation costs for the first 13 sections; (c) contingencies as the sum of process and project contingencies for the first 13 sections and (d) the engineering contract management, home office, and fee as the sum of all the costs. The purchaser price is adjusted from 2007 to 2002 dollars using the chemical engineering plant cost index and is used as an input for the EIO-LCA model to obtain water withdrawal information for each category. b The 20% CCS and 40% CCS refer to the cases under the constraint of 1400 lb CO2/MWh and 1100 lb CO2/MWh, respectively; the full CCS refers to the case with 90% CO2 capture. c Since the cost information of PC with CCS case in NETL (2013) is based on 90% capture rate, an adjustment factor is applied to estimate the cost for the case with 20% and 40% CO2 capture. The adjustment factor is the ratio of total capital requirements for plants with 20% or 40% versus 90% CO2 capture simulated in the IECM. d Water consumption is calculated by applying the water consumption adjustment factors to total water withdrawals from the 2002 purchaser model. The consumption adjustment factors are the ratio of total water consumption and total water withdrawal reported in EIO-LCA 1992 Producer Model, using the same data source and methodology. The average ratios of water withdrawal versus consumption are 5.6 and 5.4 for the PC and NGCC cases, respectively. Chemical Production. The Ecoinvent database (v2.2) reports the flow rates of various elementary water intake flows for the production of one kilogram chemical, which are presented in Appendix Table A-5 in detail. The water use factors are estimated for the three chemicals using the method discussed in Section 2.2.3 and are summarized in Table 5. Given the production of MEA uses ethylene oxide and ammonia, the MEA production uses more water than the ammonia production. Different from limestone and ammonia, pure MEA is often diluted to have a concentration of 30% in weight before it is used as a solvent for CO2 capture (Zhai et al, 2011).


Table 5 Water use for major chemical production based on Ecoinvent 2.2

Parameter Ammonia Limestone Amine Solvent (30-wt% MEA)

Total water withdrawals (gal H2O/lb chem.) 1.3 0.06 3.0 Total water consumption (gal H2O/lb chem.) 0.09 0.004 0.24

Performance of Power Plants with and without CCS. IECM v8.0.2 was employed to model

a supercritical PC plant fired by bituminous coal and a NGCC plant. Given that Section 316(b) of the Clean Water Act requires the EPA to regulate cooling intake structures to minimize adverse environmental impacts, wet cooling towers were assumed for new PC and NGCC plants. Air pollution control systems such as ESP, SCR, and FGD were used to comply with the federal NSPS for traditional air pollutants. When needed, an amine-based CCS system with bypass was used for partial CO2 capture. Table 6 summarizes the major technical assumptions and parameters of power plants and environmental and cooling systems, including CCS.


Table 6 Major technical assumptions and parameters for power plant modeling

Variable Value Plant type PC or NGCC

Plant capacity (MW net) 550 (PC) 542 (NGCC)

Plant capacity factor (%) 85% Ambient Air Conditions

Temperature (oF) 56 Relative humidity (%) 59

Traditional Air Pollution Controls (if applicable)

Nitrogen oxides Hot-side selective catalytic reduction(SCR)

Particulates Electrostatic precipitator (ESP)

Sulfur oxides Wet flue gas desulphurization (FGD)

Partial CO2 capture Bypass Carbon Capture and Storage (if applicable)

Capture system type Econamine FG+ CO2 removal efficiency (%) 90

Sorbent concentration (wt.%) 30 CO2 product pressure (psia) 2000

Heat-to-electricity equivalent eff. of extracted steam (%) 18.7 Regeneration heat requirement (Btu/lb CO2) 1516 Makeup water for washing (% of flue gases) 0.8

Process cooling duty (t H2O/t CO2) 92.8 Wet Cooling System

Water temperature drop across the tower (oF) 20 Cycles of concentration (ratio) 4

Auxiliary cooling duty (% of primary cooling) 1.4


Table 7 Water use of PC and NGCC power plants with and without CCS

Parameter PCa NGCC

w/o CCS Partial CCS Full CCS w/o CCS Full CCS

Gross output (MW) 590 622 636 681 557 505 Net power output (MW) 550 550 550 550 542 469 Net plant efficiency (%) 38.2 34.8 32.8 27.5 50.1 43.3 CO2 capture rate (%) 20.3 39.8 90.0 90.0 CO2 emission intensity

(lb/MWh gross) 1686 1400 1100 204 781 91.0 (lb/MWh net) 1808 1584 1272 252 802 92.7

Fuel (ton/hr or MMscf/hr) 210 231 245 293 3.57 3.57 Ammonia makeup (ton/hr) 0.22 0.24 0.26 0.31 Limestone makeup (ton/hr) 15.1 19.0 20.2 24.1 Amine solvent makeup (lb/hr) 0.0 22.5 46.9 127 0.00 39.7 Water consumption by unit (x103gal/hr) 238 279 309 409 95.5 136

FGD 34.2 37.6 39.9 47.6 Water tower 203 238 266 357 95.5 136

CCS 0.00 3.0 3.2 3.8 Water withdrawals by unit (x103gal/hr) 339 401 443 582 127 181

Boiler 33.5 36.7 39.0 46.6 Cooling 271 318 355 476 127 181

SCR 0.34 0.38 0.40 0.48 FGD 34.2 37.6 39.9 47.6 CCS 0.00 8.76 9.36 11.3

Total water consumption (gal/MWh) 432 507 563 744 176 290 Total water withdrawals (gal/MWh) 617 729 806 1059 235 387

a The bituminous Illinois #6 coal in the IECM database is used as the surrogate fuel for power plant assessment.

Based on the IECM modeling results for the given designs, Table 7 reports the flow rates of

fuel, chemical makeup, and water streams at PC and NGCC plants with and without CCS. The cooling system is the largest water user at power plants. Compliance with the EPA's CO2 emission performance standard of 1400 lb/MWh would require about 20% carbon capture at the PC plant. The addition of CCS for partial CO2 capture decreases the net plant efficiency by more than 3% (absolute basis) and then increases the plant fuel use by 10%, the total plant water use by about 17-18%, and the limestone and ammonia use by 26% and 9%, respectively. To meet the more stringent CO2 emission limit of 1100 lb/MWh, the implementation of CCS at the PC plant decreases the net plant efficiency by more than 5% (absolute basis) and then increases the plant fuel use by 17%, the total plant water use by about 30%, and the limestone and ammonia use by 33% and 18%, respectively. There is no CO2 capture need for the NGCC plant under the regulation of the EPA's CO2 emission performance standard (1000 lb/MWh). As shown in Table


7, enhancing the CO2 capture level to 90% further significantly decreases the plant efficiency and increases the resource use. Similar to the PC plant, the addition of CCS for 90% CO2 capture at the NGCC plant decreases the net plant efficiency by about 7% (absolute basis) and then increases the total plant water use by 65%. Significant increases in plant water use mainly come from the large amount of cooling water (93 ton water/ton CO2) required for the amine-based capture process (Zhai et al, 2011). In comparison between the two plant types, the NGCC plant with CCS even requires 33% and 37% less water consumption and withdrawals than the PC plant without CCS, respectively. Life Cycle Water Use The total life cycle water use was estimated using Equation (1) based on the nominal water use factors reported in Tables 2, 3, 4, 5 and 7, and the fuel and chemicals requirements reported in Table 7 for the given power plant designs. Tables 8 and 9 summarize the life cycle water use of PC and NGCC plants with and without CCS, respectively.

Table 8 Life cycle water use of coal-fired power plants with and without CCS

Stage

Water Consumption (gal/MWh)

Water Withdrawals (gal/MWh)

w/o CCS

20% CCS

40% CCS

Full CCS

w/o CCS

20% CCS

40% CCS

Full CCS

Fuel supply Surface mining+train 18 20 21 25 18 20 21 26

Underground mining+train 38 42 44 53 38 42 45 53

Surface mining+slurry pipeline 109 120 127 152 109 120 127 152

Underground mining+ slurry pipeline 129 142 151 180 129 142 151 180

Plant infrastructure 2.9 4.3 4.5 5.6 16 24 25 31 Makeup chemicals

Ammonia 0.07 0.08 0.08 0.1 1.0 1.1 1.2 1.5 Limestone 0.2 0.3 0.3 0.3 3.5 4.4 4.7 5.6

Amine 0.0 0.01 0.02 0.06 0.00 0.12 0.25 0.68 Plant operation 432 507 563 744 617 729 806 1059 Total life cycle

Surface mining 453 531 588 775 656 778 858 1123 Underground mining 473 553 612 802 676 800 882 1151

Surface mining + slurry pipeline 544 631 695 901 747 878 964 1249

Underground mining+ slurry pipeline 564 653 718 929 767 900 988 1277


Table 9 Life cycle water use of NGCC power plants with and without CCS

Stage Water Consumption (gal/MWh) Water Withdrawals (gal/MWh) w/o CCS Full CCS w/o CCS Full CCS

Fuel supply via pipeline Conventional gas 4.6 5.3 5.6 6.5

shale gas 16 18 17 19 Plant infrastructure 1.0 2.6 5.6 14.1 Makeup Chemicals

Amine 0.00 0.02 0.0 0.3 Plant operation 176 290 235 387 Total life cycle

Conventional gas 182 298 246 407 shale gas 193 311 257 420

The results show that in the total life cycle water consumption and withdrawals, the plant operation accounts for 77% to 96% and 80% to 94% for PC plants and 91% to 97% and 91% to 95% for NGCC plants, respectively, whereas other stages contribute much less to the life cycle water use. The life cycle water use does not significantly vary with the coal mining and transport methods. Although the water use of hydraulic fracturing is numerous times than that of conventional drilling (Meldrum et al, 2013), the life cycle water use difference between conventional gas and shale gas is not significant. The total water use for the production of three chemicals is much smaller. Because the large amount of cooling duty required for CO2 capture process increases the plant water use and the energy penalty of CCS increases the fuel and chemicals use in power plants, the addition of CCS would significantly increase the life cycle water use for both PC and NGCC plants, depending on the CO2 capture level. The results given in Tables 8 and 9 imply that a shift from coal to natural gas for electricity generation would lower water use on both plant and life cycle bases, including unconventional natural gas for electricity generation, mainly because of the NGCC plant's larger plant efficiency and smaller CO2 emission intensity. Considering that the plant operation dominates the life cycle water use for electricity generation, we further examine the variability by fuel type and cooling type, under the constraint of the EPA's CO2 emission performance standards. Coal type is the key factor that affects the required CO2 removal level and water use of PC power plants under the CO2 emission regulation because carbon content and heating value have large effects on the plant efficiency and CO2 emission intensity (Talati et al, 2014). Figure 2 compares the life cycle water use of the PC plants fired by three coals, including bituminous Illinois No.6 coal, sub-bituminous Wyoming Powder River Basin (WY PRB) and North Dakota lignite (ND LIG), in which coal is assumed to be extracted by surface mining and transported by train. High coal quality lowers the life cycle water use for the regulated power plants. Compared with Ill. No.6 coal, burning ND LIG coal in the PC plant subject to the given emission limit would increase the total life cycle water withdrawals and consumption by 18% and 20%, respectively.


Figure 2 Effects of coal type on life cycle water use of a PC plant under emission limit of 1400 lb CO2/MWh

A shift from wet cooling to dry cooling can significantly reduce plant water use. Air-cooled condensers for dry cooling are used as the primary cooling system for power plants. When CCS is needed, an auxiliary wet cooling system is still used for CCS due to the cooling water requirement for the capture process (Zhai et al, 2011). Surface mining and train transportation were assumed for the PC plant, while conventional drilling and pipeline transportation were assumed for the NGCC plant. Figure 3 compares the resulting life cycle water use of power plants using different cooling systems under the CO2 emission regulation. For the PC plant with CCS, hybrid cooling saves 60% of the life cycle water withdrawals and 66% of the life cycle water consumption, compared to wet cooling. For the NGCC plant without CCS, the deployment of dry cooling almost eliminates water use.

(a) PC plant under 1400 lb CO2/MWh (b) NGCC plant under 1000 lb/CO2/MWh Figure 3 Effects of cooling type on life cycle water use of PC and NGCC power plants under CO2 emission limits


Uncertainty in Life Cycle Water Use The deterministic results presented in Tables 8 and 9 indicate that the plant operation and the fuel supply account for the majority of life cycle water use. Thus, uncertainties in these stages are characterized and quantified to provide probabilistic estimates of life cycle water use for power plants, especially under the regulation of the EPA's CO2 emission performance standards. Specifically, the key uncertain parameters considered in the fuel supply include the water use factors associated with fuel extraction or drilling and transport. Uniform or triangular distribution functions are assigned to the uncertain water use factors based on the minimum, nominal and maximum values reported in Tables 2 and 3 and are summarized in Appendix Tables A-6 and A-7. The key uncertain parameters considered in the plant operation stage include ambient air conditions and those that affect water use around the steam cycle, FGD, cooling, and amine-based CCS systems. The assumptions about the distributions for the uncertain parameters are based primarily on the previous studies (Zhai et al, 2011; Talati et al, 2014; Meldrum et al, 2013) and are summarized in Appendix Tables A-8 and A-9. Please note that the choice of different probability distribution functions might affect the uncertainty analysis results. Surface mining and train transport were assumed for coal production, while hydraulic fracturing and pipeline transport were assumed for natural gas. The resulting probability distributions of life cycle water use address only the influences of uncertainties for the given cooling system. Power plants using different cooling systems, such as cooling ponds or dry cooling, could have different distributions. Figure 4 shows the cumulative distribution functions (CDFs) of life cycle water use for the PC plant with different CO2 capture levels. Given the assumed probability distribution functions assigned to uncertain parameters, the resulting distribution for the PC plant without CCS has a 95-percentile range from 583 to 732 gal/MWh for water withdrawals and 427 to 495 gal/MWh for water consumption; for the PC plant subject to the 1400 lb CO2/MWh emission limit, the resulting distribution has a 95-percentile range from 756 to 996 gal/MWh for water withdrawals and 542 to 672 gal/MWh for water consumption. The implementation of CCS for 90% CO2 capture would result in a life cycle water use range from 953 to 1502 gal/MWh for water withdrawals and from 666 to 1035 gal/MWh for water consumption.

(a) (b) Figure 4 Probability distribution of life cycle water use of a supercritical PC plant using a recirculating cooling system with and without CCS


Figure 5 shows the CDFs of life cycle water use for the NGCC plant fired by conventional natural gas and shale gas. The 95-percentile distribution for the NGCC plant fired by conventional natural gas ranges from 234 to 283 gal/MWh for water withdrawals and from 172 to 210 gal/MWh for water consumption, while for the NGCC plant fired by shale gas, the probability distribution ranges from 251 to 406 gal/MWh for water withdrawals and from 183 to 339 gal/MWh for water consumption. In contrast, the life cycle water use distribution for the shale gas case has a broader range than the conventional gas case, mainly due to the larger uncertainty of water used for hydraulic fracturing.

(a) (b) Figure 5 Probability distribution of life cycle water use of a NGCC plant using a recirculating cooling system

Tornado diagrams in @ Risk were further employed to compare the relative importance of different uncertain variables and identify the variable that contributes the most to the uncertainty of life cycle water use for coal-fired and shale-gas-fired NGCC power plants subject to the EPA's CO2 emission performance standards. Figures 6 shows the tornado diagrams for PC and NGCC plants, in which the variable with highest impact on the life cycle water use appears at the top of the diagram, followed by other variables in descending impact order. The Spearman's rank correlation coefficient is almost 1.0 between plant operation and life cycle water use for the PC case and between hydraulic fracturing and life cycle water use for the NGCC case. The plant operation is the dominant chain influencing the uncertainty of life cycle water withdrawals for the PC plant, whereas the hydraulic fracturing is the dominant chain of uncertainty for the shale-gas-fired NGCC plant. There are similar findings regarding water consumption.


(a) (b) Figure 6 Tornado diagrams on life cycle water withdrawals for PC plant with partial CO2 capture and shale-gas-fired NGCC plant (a) PC plant; (b) NGCC plant The baseline value is the overall mean of the simulation outputs.

Conclusions This study performs a process-based hybrid analysis to provide systematic estimates of the water use of electricity generation through supply chains across the life cycle. The life cycle water use is characterized for PC and NGCC plants under the constraint of different CO2 emission limits. Among multiple supply chains for electricity generation, the plant operation dominates the life cycle water use for both types of power plants. The addition of CCS to power plants would remarkably increase the plant and life cycle water use, depending on the CO2 capture level. The resulting influence of increased water use also depends on the water availability at different supply chains, such as fuel mining and power plant sites. The compliance with the U.S. EPA's current CO2 emission performance standards would sizably increase the life cycle water use of coal-fired power plants, but does not affect the water use of NGCC plants because of no need for CO2 capture. In the context of regulating CO2 emissions from power plants, burning high-quality coal for electricity generation would improve the net plant efficiency and lower the fuel, CO2 capture, and water requirements. Advancing cooling technologies and carbon capture systems with heat integration or recovery that lowers the required cooling duty would help lower the life cycle water use. Given the current gas price and potential vast availability of shale gas, the U.S. EPA's low-carbon regulations on new and existing power plants would promote a shift from coal to natural gas for electricity generation, and in turn, pronouncedly reduce the life cycle water use. However, special attention should be paid to the local impacts of extracting natural gas from vast shale deposits on water resources and the environment. The results of uncertainty analysis demonstrate that there is a larger probabilistic distribution in the life cycle water use of PC plants with CCS than those without CCS, mainly due to the uncertainty associated with the capture process. NGCC plants fired by shale gas have a larger probabilistic distribution in the life cycle water use than those fired by conventional natural gas. The plant operation and hydraulic fracturing are the major sources contributing the most to the uncertainty of life cycle water use for PC and shale-gas-fired NGCC plants, respectively.


Acknowledgements This work was supported by the National Energy Technology Laboratory via the Sandia National Laboratories under Grant No. PO 1508166-SNL-Rubin. The authors acknowledge Dr. Michael Griffin and Dr. Paulina Jaramillo for the comments on the draft manuscript, Lucy Xiao for the assistance at the early stage of this effort, Dr. H. Scott Matthews for the assistance with the ecoinvent database, and Dr. Niels Jungbluth for the discussion about water use estimation based on the ecoinvent database. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors alone and do not reflect the views of any agency.

References Ali, B., & Kumar, A. (2015). Development of life cycle water-demand coefficients for coal-

based power generation technologies. Energy Conversion and Management, 90, 247−260. Carnegie Mellon University Green Design Institute (2015). Economic Input-Output Life Cycle

Assessment (EIO-LCA), U.S. 2002 (428 sectors) Purchaser model [Internet], Available from: www.eiolca.net, Accessed 28 July, 2015.

Clark, C. E., Horner, R. M., & Harto, C. B. (2013). Life cycle water consumption for shale gas and conventional natural gas. Environmental Science & Technology, 47(20), 11829−11836.

Corsten, M., Ramírez, A., Shen, L., Koornneef, J., & Faajj, Andre. (2013). Environmental impact assessment of CCS chains–lessons learned and limitations from LCA literature. International Journal of Greenhouse Gas Control, 13, 59−71.

U.S. Environmental Protection Agency (2014). 40 CFR Part 60 Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility Generating Units; Proposed Rule. Fegeral Register, Vol. 79, No. 117, June 18.

U.S. Environmental Protection Agency (2015a). 40 CFR Parts 60, 70, 71, and 98 Standards of Performance for Greenhouse Gas Emissions from New, Modified, and Reconstructed Stationary Sources: Electric Utility Generating Units, Final Rule. August 3. Available at http://www.epa.gov/airquality/cpp/cps-final-rule.pdf.

U.S. Environmental Protection Agency (2015b). 40 CFR Part 60 Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility Generating Units, Final Rule. August 3. Available at http://www.epa.gov/airquality/cpp/cpp-final-rule.pdf.

Flury, K., Jungbluth, N., Frischknecht, R., & Munoz, I. (2012). Recommendation for life cycle inventory analysis for water use and consumption, Working paper, ESU Services, http://www. esu-services. ch/fileadmin/download/flury-2012-water-LCI-recommendations.pdf. Accessed May 2015.

Frischknecht, R., & Jungbluth, N. (2007). Overview and methodology. Data v2. 0. Ecoinvent report No. 1. Ecoinvent Centre, Swiss Federal Laboratories for Materials Testing and Research (EMPA), Duebendorf (Switzerland).

Fthenakis, V., & Kim, H. C. (2010). Life-cycle uses of water in US electricity generation. Renewable and Sustainable Energy Reviews, 14(7), 2039−2048.

Goedkoop M., Oele M., Leijting J., Ponsioen, T., & Meijer, E. (2013). Introduction to LCA with SimaPro. Version 5.1. PRé, Nov. 2013.

Grant, T., Anderson, C., & Hooper, B. (2014). Comparative life cycle assessment of potassium carbonate and monoethanolamine solvents for CO2 capture from post combustion flue gases. International Journal of Greenhouse Gas Control, 28, 35−44.

http://www.cmu.edu/epp/people/faculty/michael-griffin.html

http://www.cmu.edu/epp/people/faculty/michael-griffin.html

http://www.cmu.edu/epp/people/faculty/h-scott-matthews.html

http://www.eiolca.net/


Integrated Environmental Control Model (IECM). Carnegie Mellon University’s Integrated Environmental Control Model Version 8.0; Carnegie Mellon University: Pittsburgh, PA, 2012; http://www.cmu.edu/epp/iecm/index.html (accessed May 2015).

Koornneef, J., van Keulen, T., Faaij, A., & Turkenburg, W. (2008). Life cycle assessment of a pulverized coal power plant with post-combustion capture, transport and storage of CO2. International Journal of Greenhouse Gas Control, 2(4), 448−467.

Korre, A., Nie, Z., & Durucan, S. (2010). Life cycle modelling of fossil fuel power generation with post-combustion CO 2 capture. International Journal of Greenhouse Gas Control, 4(2), 289−300.

Laurenzi, I. J., & Jersey, G. R. (2013). Life Cycle Greenhouse Gas Emission and Freshwater Consumption of Marcellus Shale Gas. Environmental Science & Technology, 47(9), 4896−4903.

Meldrum, J., Nettles-Anderson, S., Heath, G., & Macknick, J. (2013). Life cycle water use for electricity generation: a review and harmonization of literature estimates. Environmental Research Letters, 8(1), 015031.

Muñoz, I., Milà-i-Canals, L., & Fernández-Alba, A. R. (2010). Life cycle assessment of water supply plans in Mediterranean Spain. Journal of Industrial Ecology, 14(6): 902−918.

U.S. National Energy Technology Laboratory. (2013). Cost and Performance Baseline for Fossil Energy Plants, Volume 1: Bituminous Coal and Natural Gas to Electricity, Revision 2a, DOE/NETL-2010/1397, Pittsburgh, PA.

Palisade Corporation (2015). @Risk 6.3 Trial Version, http://www.palisade.com/risk/ (Accessed in May 2015).

Rao, A. B., & Rubin, E. S. (2006). Identifying cost-effective CO2 control levels for amine-based CO2 capture systems. Industrial & Engineering Chemistry Research, 45(8), 2421−2429.

Ruether, J. A., Ramezan, M., & Balash, P. C. (2004). Greenhouse Gas Emission from Coal Gasification Power Generation Systems. Journal of Infrastructure Systems, 10, 111−119.

Shaffer, K. (2008). Consumptive water use in the Great Lakes Basin, US Department of the Interior, US Geological Survey.

Singh, B., Strømman, A. H., & Hertwich, E. (2011). Life cycle assessment of natural gas combined cycle power plant with post-combustion carbon capture, transport and storage. International Journal of Greenhouse Gas Control, 5(3), 457−466.

Soeder, D. J., & Kappel, W. M. (2009). Water resources and natural gas production from the Marcellus Shale (pp. 1-6). Reston, Virginia: US Department of the Interior, US Geological Survey.

Talati, S., Zhai, H., & Morgan, M. G. (2014). Water impacts of CO2 emission performance standards for fossil fuel-fired power plants. Environmental Science & Technology, 48(20), 11769−11776.

Zapp, P., Schreiber, A., Marx, J., Haines, M., Hake, J. F., & Gale, J. (2012). Overall environmental impacts of CCS technologies—a life cycle approach. International Journal of Greenhouse Gas Control, 8, 12−21.

Zhai, H., & Rubin, E. S. (2010). Performance and cost of wet and dry cooling systems for pulverized coal power plants with and without carbon capture and storage. Energy Policy, 38(10), 5653−5660.

Zhai, H., Rubin, E. S., & Versteeg, P. L. (2011). Water use at pulverized coal power plants with postcombustion carbon capture and storage. Environmental Science & Technology, 45(6), 2479–2485.

http://www.cmu.edu/epp/iecm/index.html

http://www.palisade.com/risk/


Zhai, H., & Rubin, E. S. (2013). Comparative performance and cost assessments of coal-and natural-gas-fired power plants under a CO2 emission performance standard regulation. Energy & Fuels, 27(8), 4290-4301.

Zhai, H., Ou, Y., & Rubin, E.S. (2015). Opportunities for decarbonizing existing U.S. coal-fired power plants via CO2 capture, utilization and storage. Environmental Science & Technology, 49(13), 7571−7579.

Zhang, X., Singh, B., He, X., Gundersen, T., Deng, L., & Zhang, S. (2014). Post-combustion carbon capture technologies: Energetic analysis and life cycle assessment. International Journal of Greenhouse Gas Control 27, 289−298.


Appendix Conversion Factors from English Units to Metric System of Units 1 gal = 3.79 E-3 m3

1 MMscf = 2.83 E+4 m3

1 gal/ton = 4.17 E-6 m3 /kg 1 gal/MMscf = 1.34 E-7 m3 /m3

1 lb = 4.54 E-1 kg

1 gal/lb = 8.34 E-3 m3/kg

1 oF = 2.56 E+2 K

∆1oF = ∆0.56oK

1 Btu/lb = 2.33 E+0 kJ/kg 1 psi = 6.89E+3 Pa 1 lb/MWh = 4.54 E-1 kg/MWh 1 ton/hr = 9.07 E-1 t/hr 1 MMscf/hr = 2.83 E+4 m3/hr 1 gal/MWh = 3.79 E-3 m3/MWh


Table A-1 Economic Sector Classification and Cost by Plant Section for an illustrative PC plant without CCS

Categorya 2007 Purchaser Price (K$)b

2002 Purchaser Price (K$)c

I-O Category in 2002 Purchaser Modeld

Total water withdrawals (Billion Gallon)e

Coal & Sorbent Handing 16381 12334 Machinery and Engines/Material

Handling equipment manufacturing 0.099

Coal & Sorbent prep & feed 11008 8288 Machinery and Engines/Other

industrial machinery manufacturing 0.058

Feedwater & Misc. BOP systems 42453 31965

Other Metal Hardware and ordnance manufacturing/Valve and fittings other than plumbing

0.226

PC Boiler(Gasifier & Accessories) 157253 118404

Machinery and Engines/Industrial Process furnace and oven manufacturing

0.774

Flue Gas Cleanup 79643 59967 Other Metal Hardware and ordnance manufacturing /Other fabricated metal manufacturing

0.557

Combustion Turbine & Accessories 0 0

Machinery and Engines/Turbine and turbine generator set units manufacturing

0.000

HRSG Ducting & Stack 17397 13099

Machinery and Engines/Heating equipment (except warm air furnaces) manufacturing

0.119

Steam Turbine Generator 74579 56154


0.262

Cooling Water System 12303 9264 Mining and Utilities/water, sewage and other system 0.194

Ash/Spent Sorbent Handling System 4409 3320 Machinery and Engines/Material


Accessory Electric Plant 17541 13207

Lighting, electrical components, batteries/Electric power and specially transformer manufacturing

0.105

Instrumentation & Control 8739 6580

Semiconductors, electronic equipment and media reproduction/industrial process variable instruments

0.040

Improvements to site 2969 2236 Construction/nonresidential maintenance and repair 0.013

Equipment Installation 295084 222184 Construction/nonresidential manufacturing structure 0.859

Eng'g CM H.O. & Fee 69584 52393 Management, administrative, and waste services/facilities support service

0.242

Contingencies 96558 72703 Construction/nonresidential maintenance and repair 0.423

a The cost category is based on the NETL report (NETL, 2013). b The cost values come from the NETL report (NETL2013). c The cost values are converted from 2007 Purchaser Price based on the Chemical Engineering Plant Cost Index. d The I-O sector is defined in EIO-LCA 2002 Purchaser Model (CMUGDI 2015). e Water withdrawal requirements reported in the EIO-LCA 2002 Purchaser Model (CMUGDI 2015).


Table A-2 Economic Sector Classification and Cost by Plant Section for an illustrative NGCC plant without CCS

Itema 2007 Purchaser Price (K$)b

2002 Purchaser Price (K$)c

I-O Category in 2002 Purchaser Modeld Total water withdrawals (Billion Gallon)e

Coal & Sorbent Handing 0 0 Machinery and Engines/Material


Coal & Sorbent prep & feed 0 0 Machinery and Engines/Other industrial

machinery manufacturing 0.000

Feedwater & Misc. BOP systems 22444 16899

Other Metal Hardware and ordnance manufacturing/Valve and fittings other than plumbing

0.120

Gasifier & Accessories 0 0

Machinery and Engines/Industrial Process furnace and oven manufacturing

0.000

Gas Cleanup and Piping 0 0

Other Metal Hardware and ordnance manufacturing/Other fabricated metal manufacturing

0.000

Combustion Turbine & Accessories 0 0

Other Metal Hardware and ordnance manufacturing/Other fabricated metal manufacturing

0.264

HRSG Ducting & Stack 75294 56693


0.234

Steam Turbine Generator 34200 25751

Machinery and Engines/Heating equipment (except warm air furnaces) manufacturing

0.113

Cooling Water System 32078 24153


0.087

Ash/Spent Sorbent Handling System 5524 4159 Mining and Utilities/water, sewage and

other system 0.000

Accessory Electric Plant 0 0 Machinery and Engines/Material


Instrumentation & Control 16639 12528

Lighting, electrical components, batteries/Electric power and specially transformer manufacturing

0.027

Improvements to site 5778 4351 Semiconductors, electronic equipment and media reproduction/industrial process variable instruments

0.008

Equipment Installation 1722 1297 Construction/nonresidential

maintenance and repair 0.215

Eng'g CM H.O. & Fee 73920 55658 Construction/nonresidential

manufacturing structure 0.078

Contingencies 22400 16866 Management, administrative, and waste services/facilities support service 0.150

a The cost category is based on the NETL report (NETL, 2013). b The cost values come from the NETL report (NETL2013). c The cost values are converted from 2007 Purchaser Price based on the Chemical Engineering Plant Cost Index. d The I-O sector is defined in EIO-LCA 2002 Purchaser Model (CMUGDI 2015). e Water withdrawal requirements reported in the EIO-LCA 2002 Purchaser Model (CMUGDI 2015).


Table A-3 Documentation of Chemical Production Processes in Ecoinvent 2.2 Process Ammonia, liquid, at

regional storehouse/RER WITH US ELECTRICITY S

Limestone, milled, loose, at plant/CH WITH US ELECTRICITY S

Monoethanolamine, at plant/RER WITH US ELECTRICITY S

Included processa Mostly present state of the art technology used in European ammonia production plants.

Milling, sieving, filtering and storing. One part of the total heating energy for "production" and "administration" is included. Equipment included in the infrastructure: 1 crusher, 1 sieve, and 2 small silos and 1 filter.

Raw materials and chemicals used for production, transport of materials to manufacturing plant, estimated emissions to air and water from production(incomplete), estimation of energy demand and infrastructure of the plant(approximation). Solid wastes omitted.

Remarka Average of numbers for stream reforming (85%) and partial oxidation of heavy fuel oil (15%)

Infrastructure data are estimated based on a tour and sketches of process, whose value is normalized with an annual production capacity of about 6'000 tons of product per year.

Large uncertainty of the process data due to weak data on the production process and missing data on process emissions; Geography: Data used has no specific geographical origin. Average European processes for raw materials, transport requirements and electricity mix used.

Technologya Infrastructure data are estimated based on a tour and sketches of process, whose value is normalized with an annual production capacity of about 6'000 tons of product per year.

Heavy machines (exd. Building machines) are operated electrically; air is recirculated in closed loop to avoid dust emissions

Production from ethylene oxide and ammonia with a process yield of 95%. Inventory bases on stoichiometric calculations.

Time perioda Values based on reports from 1995 and 2000, and on literature with data from unknown data

No documentation Date of published literature

Category Processes - Material - Chemicals - Inorganic

Processes - Material - Chemicals - Inorganic

Processes - Material - Chemicals - Organic

Data libraryb US-EI 2.2 US-EI 2.2 US-EI 2.2 a The documentation of processes comes from Ecoinvent 2.2. b US-EI is an amalgamated life cycle inventory(LCI) dataset of EarthShift developed data with proper expansion to North American region. All processes using electricity from European regions were indirectly adapted to instead use US electricity by rerouting electricity production/distribution to US electricity production/distribution in a series of processes.


Table A-4 Selection of System Processes Used for Inventory Estimation in Ecoinvent Pollutant Description of Process Selection NOx In the Ecoinvent database, the operation of hot-side SCR is defined in the “processing”, “power plants”

category, and in the “NOx retained, in SCR/GLO WITH ELECTRICITY U” (global, US electricity utilized, unit model). The production of ammonia used in SCR is specified by the unit process of “ammonia, liquid, at regional storehouse/RER WITH US ELECTRICITY U” (Europe, US electricity utilized, unit model).

SO2 Limestone is a common alkaline sorbent to remove sulfur dioxide (SO2) from flue gas. In the Ecoinvent database, the operation of wet FGD is defined in the “processing”, “power plants” category, and in the “SOx retained, in hard coal flue gas desulfurization/RER WITH ELECTRICITY U” (European, US electricity utilized, unit model). The limestone production is specified by the unit process of “Limestone, milled, loose, at plant/CH WITH US ELECTRICITY U” (Switzerland, US electricity utilized, unit model).

CO2 Water is used for producing pure MEA and diluting the MEA concentration to 30% in weight for application to CO2 capture. Since the amine-based CCS system is an emerging technology without widely commercial deployment for power plants, there is no specific process entry in the processing category. Thus, the only MEA-related production model is selected for assessment: “Monoethanolamine, at plant/RER WITH US ELECTRICITY S” (European, US electricity utilized, system model).

Table A-5 Water use per kilogram of chemical produced Elementary water flow Ammonia Limestone MEA Water, cooling, unspecified natural origin (m3/kg) 7.23E-03 4.49E-04 7.21E-02

Water, lake (m3/kg) 9.05E-06 4.32E-07 3.10E-05 Water, river (m3/kg) 7.02E-04 3.20E-05 2.61E-03 Water, salt, ocean (m3/kg) 5.60E-04 9.01E-07 6.43E-04 Water, salt, sole (m3/kg) 2.66E-04 2.38E-06 2.34E-04 Water, turbine use, unspecified natural origin (m3/kg) 1.14E+00 7.86E-01 4.41E+00

Water, unspecified natural origin (m3/kg) 2.66E-03 1.54E-05 4.83E-03 Water, well, in ground (m3/kg) 1.74E-04 3.49E-05 4.85E-04

Total water withdrawals (m3/kg) 1.08E-02 5.31E-04 8.00E-02 Total water consumption (m3/kg) 3.90E-03 3.07E-05 4.40E-03 Table A-6 Distribution functions assigned to uncertain parameters for coal supply* Water Withdrawals Water Consumption

Parameter Unit Nominal Value

Distribution Function

Nominal Value

Distribution Function

Extraction - surface gal/ton 6.54 Uniform (1.09, 28.33) 6.54 Uniform (1.09, 28.33) Extraction - underground gal/ton 58.8 Uniform (17.4, 392.3) 58.84 Uniform (17.4,392.3) Transport - train gal/ton 2.18 Uniform (1.09, 4.36) 1.09 Uniform (0.22, 2.18) Transport - slurry gal/ton 239.7 Uniform (217.9,893.5) 239.71 Uniform (217.9,893.5) *Reference: Meldrum et al, 2013


Table A-7 Distribution functions assigned to uncertain parameters for natural gas supply* Withdrawal Consumption

Parameter Units Nominal Value Distribution Function Nominal

Value Distribution Function

Drilling gal/MMscf 154.1 Triangular

(15.4,154.1,2927.2 ) 154.06 Triangular (15.4,154.1,2927.2 )

Fracturing (shale gas)

gal/MMscf 1848.7 Triangular

(154.1,1848.7,28655.3 ) 1848.73 Triangular (154.1,1848.7,28655.3 )

Transport - pipeline

gal/MMscf 616.2 Triangular (616.2, 616.2,

2002.8) 462.18 Triangular (154.1, 154.1, 924.4)

*Reference: Meldrum et al, 2013. Table A-8 Distribution functions assigned to uncertainty parameters for PC power plant* Category Parameter Units Nominal Value Distribution Function

Ambient Air

Ambient air temperature oF 55.9 Uniform (42.1, 70.3) Relative humidity % 59 Uniform (50,68)

Base plant (steam cycle)

Boiler blowdown % 6 Uniform (0,10)

Miscellaneous stream losses % 0.4 Uniform (0,1.0) Demineralizer underflow % 8.5 Uniform (0,17) Auxiliary cooling duty % 1.4 Uniform (0,2.8)

FGD

Total pressure drop across FGD cm H2O gauge 25.4 Uniform (0, 50.8) Temperature rise across ID fan oF 14 Uniform (0,25)

CCS

Makeup water for wash section % 0.8 Uniform (0,1.6) Capture system cooling duty t H2O/ t CO2 92.8 Triangular (67, 92.8, 162) Regeneration heat requirement Btu/lb CO2 1516 Triangular (1137, 1516, 1895) Heat-to-electricity efficiency % 18.7 Triangular (14,18.7, 22)

* Reference: Talati et al, 2014. Table A-9 Distribution functions assigned to uncertain parameters for NGCC power plant* Category Parameter Units Nominal Value Distribution Function

Ambient Air

Ambient air temperature oF 55.9 Uniform (42.1, 70.3) Relative humidity % 59 Uniform (50,68)

Base plant Auxiliary cooling duty % 1.4 Uniform (0,2.8)

* Reference: Talati et al, 2014.

IECM Technical Documentation: Volume V

Technical, Water, and Economic Impacts of Low-Carbon Electricity Futures under the Clean Power Plan: A Case Study in

New Mexico

April 2016

Disclaimer

This work is supported by the National Energy Technology Laboratory via the Sandia National Laboratories. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference therein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed therein do not necessarily state or reflect those of the United States Government or any agency thereof.


Technical, Water, and Economic Impacts of Low-Carbon Electricity Futures under the Clean Power Plan: A Case Study in New Mexico


Prepared for:




www.netl.doe.gov



April 2016


Table of Contents

Technical, Water, and Economic Impacts of Low-Carbon Electricity Futures under the Clean Power Plan: A Case Study in New Mexico 14

Objectives of this Report ......................................................................................................... 14 Introduction ............................................................................................................................. 14 Research Methods .................................................................................................................... 16

Databases ................................................................................................................... 17 Power Plant Profiles 17 Ambient Temperature, Humidity and Pressure 18 Water Use and Alternative Water Sources 19

Data Analysis ............................................................................................................ 20 Database Calculations ............................................................................................... 20 The IECM .................................................................................................................. 20 Financial Calculations ............................................................................................... 22 Emission Compliance ................................................................................................ 23 Fuel Price ................................................................................................................... 23 EPA CO2 Mitigation Measures ................................................................................. 24

Improving Plant Heat Rate 24 Generation Shifting 25 Power Plant Retirement 25 Emission Reduction Credits 25 Mass Allowance 27 Renewable Energy Sources 27

Techno-economic Analysis ....................................................................................... 34 The IECM Parameters 34 Mitigation Technologies 36

State-Level Mitigation Analysis .............................................................................................. 47 Historical Plant-Level Results ................................................................................... 47 2030 Plant-Level Results ........................................................................................... 49

Scenario Results 50 Unit VOM 54 Plant Makeup Water 54 Alternative Plant Cooling Water Availability and Cost 57

Discussion ................................................................................................................................ 58 Emission Intensity Reduction .................................................................................... 58

Scenario CO2 Avoidance Cost and Increase in LCOE 58 CCS Mitigation Option Sensitivity to ERC/MA and Natural Gas Price 64 Plant VOM and Dispatch Order 67 Uncertainty Analysis 69 Mass-based Approach 74

Water Use .................................................................................................................. 76 Plant Water Consumption Comparison to USGS Data 76 Plant Makeup Water 78 Plant Makeup Water VOM for Alternative Water Sources 80

Policy ......................................................................................................................... 82 Conclusions ............................................................................................................................. 84 References ............................................................................................................................... 86 Appendix A: Tables ................................................................................................................. 91

Integrated Environmental Control Model - Technical Documentation Table of Contents • vi

Appendix B: Figures .............................................................................................................. 110 Appendix C: Scaling the IECM Simulation Model for NGCC Plant Profile from NEEDS and eGRID Database: LCOE, Fuel, Plant Makeup Water ............................................................ 124

LCOE and Fuel ........................................................................................................ 127 Plant Makeup Water ................................................................................................ 133

Appendix D: Panhandle A and B Equations and Constants .................................................. 134 Panhandle A ............................................................................................................ 135 Panhandle B ............................................................................................................. 136

Integrated Environmental Control Model - Technical Documentation List of Figures • vii

List of Figures

Figure 1 Schematic diagram of model construction and implementation for the existing coal-fired EGUs and NGCC plants to be compliant for different scenarios under the CPP. ................................................................ 17

Figure 2 Fleet CO2 emission intensity and associated avoidance cost for the different CPP scenarios. The fleet emission intensity is the summation of the pounds of CO2 emitted by all endogenous sources divided by all endogenous net generation, inclusive of renewable sources existing prior to 2030. Implementation of the technology option for the rate-based approach that allows for retirement of the San Juan #1 EGU and maintains net generation at the 2012 level results in the lowest emission intensity and the third lowest cost of avoidance. ERC and MA prices have no cost in this analysis. ............................................................................. 60

Figure 3 CO2 avoidance cost for the different scenarios from implementing the CPP in New Mexico as measured by the reduction in the CO2 emission intensity for the fleet. The fleet emission intensity is the summation of the pounds of CO2 emitted by all endogenous sources divided by all endogenous net generation, inclusive of renewable sources existing prior to 2030. As all of the scenarios are compliant with the CPP, scenario S2 and S3 dominate all choices as that with the least avoidance cost. T5 results in a larger reduction in emission intensity but at a slightly higher avoidance cost. ERC and MA prices have no cost in this analysis. .............. 60

Figure 4 CO2 avoidance cost for the different scenarios from implementing the CPP in New Mexico as measured by the reduction in the fossil fuel emission intensity. As all of the scenarios are compliant with the CPP, scenario T1 dominates all choices as that with the least avoidance cost. ERC and MA prices have no cost in this analysis. .............................................................................................................................................................. 61

Figure 5 Increase in fleet LCOE for the different scenarios from implementing the CPP in New Mexico as measured by the reduction in the emission intensity for the fleet. The fleet emission intensity is the summation of the pounds of CO2 emitted by all endogenous sources divided by all endogenous net generation, inclusive of renewable sources existing prior to 2030. As all of the scenarios are compliant with the CPP, scenario M1 dominates all choices as the ones with the least avoidance cost. ERC and MA prices have no cost in this analysis. .................................................................................................................................... 62

Figure 6 Increase in fleet LCOE for the different scenarios from implementing the CPP in New Mexico as measured by the reduction in the fossil fuel emission intensity. As all of the scenarios are compliant with the CPP, scenario M1 dominates all choices as that with the least avoidance cost and greatest reduction. ERC and MA prices have no cost in this analysis. .......................................................................................................... 63

Figure 7 Increase in percent of endogenous net generation coming from renewable sources in relation to the overall increase in net generation under the different scenarios from the implementation of the CPP in New Mexico for 2030. The initial assumption was that the CPP would result in a renewable source penetration of 30%. Of the scenario 1 implementations, the mass-based approach has the lowest increase in net generation. .................................................................................................................................................................................... 64

Figure 8 Sensitivity of avoidance cost to ERC price for San Juan EGU #4 with and without 40% capture CCS system. When fitted with CCS, an auxiliary natural gas boiler provides the steam and electricity for the system. The two lines intersect when the ERC price is $34.9/MWh and result in an avoidance cost of $54.7/ton of CO2. ...................................................................................................................................................... 65

Figure 9 Sensitivity of avoidance cost for the fossil fuel fleet to percent increase in natural gas price relative to 2012. Negative avoidance costs occur when the price of natural gas is low enough to cause the LCOE of the NGCC plants to be lower than that for the PC EGUs. .......................................................................................... 66

Figure 10 Sensitivity of the avoidance cost for the fossil fuel fleet to ERC price and percent increase in natural gas price relative to 2012 for San Juan EGU #4 with a 40% capture CCS system. An auxiliary natural gas boiler provides the steam and electricity for the CCS system. Areas shaded red represent the combinations

Integrated Environmental Control Model - Technical Documentation List of Figures • viii

of ERC price and natural gas price increases that have an avoidance cost of $54.7/ton of CO2 or greater. In these regions, CCS with an auxiliary natural gas boiler has a cost effective avoidance cost. ............................ 66

Figure 11 Variation in generation-weighted average VOM due to changes in the ERC or MA price. In the implementation of the state option for the rate-based approach on the New Mexico PC EGUs and NGCC plants, the higher NGCC VOM (due to the projected natural gas price in 2030) can be compensated for with an ERC price greater than $11.1/MWh. In the mass-based approach, the MA price must be greater than $389.4/ton of CO2 to compensate for the natural gas price. The high MA price relates to the mass allowance allocation process. ..................................................................................................................................................... 68

Figure 12 Cumulative distribution function of the uncertainty in the 2030 CO2 avoidance cost from implementing the S1 scenario for the CPP in New Mexico, as measured by the reduction in the fossil fuel emission intensity. Ten thousand simulations of the S1 stochastic model were run. ......................................... 70

Figure 13 Sensitivity analysis to input parameter uncertainty in the 2030 CO2 avoidance cost from implementing the S1 scenario for the CPP in New Mexico, as measured by the reduction in the fossil fuel emission intensity. The natural gas price dominates this cost, followed by the ERC price, the capital cost component of the LCOE for the wind turbines, the price for the coal, and the capital costs component of the LCOE for the solar power. ............................................................................................................................................................... 70

Figure 14 Cumulative distribution function of the uncertainty in the 2030 increase in fleet LCOE, relative to 2012, from implementing the S1 scenario for the CPP in New Mexico. Ten thousand simulations of the S1 stochastic model were run. ....................................................................................................................................... 71

Figure 15 Sensitivity analysis to input parameter uncertainty in the 2030 increase in fleet LCOE, relative to 2012, from implementing the S1 scenario for the CPP in New Mexico. The natural gas price dominates this cost, followed by the capital cost component of the LCOE for the wind turbines, the price for the coal, and the capital costs component of the LCOE for the solar power. The ERC price is not significant because ERCs purchased by the any plant are offset by a reduction in LCOE from the selling plant. ..................................... 72

Figure 16 Sensitivity of selected metrics to CCS retrofit factor for San Juan EGU #4 in the mass-based approach, when fitted with a CCS system with an auxiliary natural gas boiler. The baseline retrofit factor for all components in the CCS system used in this study was 1.2, which indicates that all related capital cost factors are increased by 20%. .............................................................................................................................................. 73

Figure 17 Sensitivity of selected metrics to CCS retrofit factor for San Juan EGU #4 in the mass-based approach, when fitted with a CCS system without an auxiliary natural gas boiler. The baseline retrofit factor for all components in the CCS system used in this study was 1.2, which indicates that all related capital cost factors are increased by 20%. .............................................................................................................................................. 73

Figure 18 Fleet CO2 mass emissions and associated avoidance cost for the different CPP scenarios. The fleet mass emission is the summation of the tons of CO2 emitted by all endogenous sources. Implementation of the technology options for the rate-based approach and that maintain net generation at the 2012 level results in the lowest carbon emissions and the third lowest cost of avoidance. ERC and MA prices have no cost in this analysis. ..................................................................................................................................................................... 74

Figure 19 Reduction in fleet CO2 emission intensity and associated fleet CO2 mass emissions for the different CPP scenarios. The fleet emission intensity is the summation of the pounds of CO2 emitted by all endogenous sources divided by all endogenous net generation, inclusive of renewable sources existing prior to 2030. The fleet mass emission is the summation of the tons of CO2 emitted by all endogenous sources. Scenarios S1, S2, S3, and T1 show a nonlinear relationship for the reduction in emission intensity and mass, while the other scenarios may show a biased, linear behavior. ............................................................................ 75

Figure 20 Reduction in fossil fuel CO2 emission intensity and associated fossil fuel CO2 mass emissions for the different CPP scenarios. The fossil fuel emission intensity is the summation of the pounds of CO2 emitted by

Integrated Environmental Control Model - Technical Documentation List of Figures • ix

all endogenous sources divided by all endogenous net generation, exclusive of renewable sources existing prior to 2030. The mass emission is the summation of the tons of CO2 emitted by all endogenous sources. Scenarios S1, S2, S3, and T1 show a nonlinear relationship for the reduction in emission intensity and mass, while the other scenarios may show a linear behavior. Scenario M1 has a lower fossil fuel intensity reduction because the PC EGUs have a great intensity since CCS is not used in this scenario. ......................................... 76

Figure 21 Reduction in makeup water use and fossil fuel emission intensity relative to 2012 for the implementation of the CPP scenarios on the New Mexico PC EGUs and NGCC plants. Fossil fuel emission intensity for the rate-based approach scenarios includes the offset in intensity due to the inclusion of ERCs. The largest reduction in water use corresponds to the implementation of the two scenarios that cap net generation from the fossil fuel sources at the 2012 levels. ..................................................................................... 79

Figure 22 Reduction in makeup water use and increase in total net generation relative to 2012 for the implementation of the CPP scenarios on the New Mexico PC EGUs and NGCC plants. Extra generation from the CCS sorbent generation process is not available for sale on the grid. The reduction in net generation for scenarios S2 and S3 relates to the decreased requirements for ERCs due to the lower emission intensity for the CCS plant. The largest reduction in water use corresponds to the implementation of the two scenarios that cap net generation from the fossil fuel sources at the 2012 levels. ............................................... 80

Figure 23 Impact of alternative makeup cooling water availability and cost on the levelized water cost and the water system VOM in the implementation of the state option for the rate-based approach on the New Mexico PC EGUs and NGCC plants. ................................................................................................................................... 81

Figure 24 Impact of alternative makeup cooling water availability and cost on the levelized water cost and the water system VOM in the implementation of mass-based approach on the New Mexico PC EGUs and NGCC plants. ......................................................................................................................................................................... 82

Integrated Environmental Control Model - Technical Documentation List of Tables • x

List of Tables

Table 1 Site-specific coal-fired EGU emission compliance configuration. .................................................................. 23

Table 2 CCS specific parameters for the IECM. ........................................................................................................... 33

Table 3 Site-specific coal-fired EGU solids management configuration. ..................................................................... 35

Table 4 Financial parameters for the IECM. ................................................................................................................. 35

Table 5 Description of scenarios analyzed in study. All scenarios apply to the state option for the rate-based approach. Scenarios 1, 2, and 3 apply to the mass-based approach. Scenarios 4 and 5 apply to the technology option for the rate-based approach. ..................................................................................................... 38

Table 6 State option required and excess rate-based ERCs for coal-fired EGUs and NGCC plants for 2030 generation in scenarios 1 through 3, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state emission rate goal is 1,146 lbs/MWh. ............................................................. 41

Table 7 Technology option required and excess rate-based ERCs for coal-fired EGUs and NGCC plants for 2030 generation in scenario 1, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The technology option emission rate goals are 1,305 lbs/MWh for steam generation and 771 lbs/MWh for NGCC generation. The GS-Shift ERCs are calculated with Equations [3] and [4]. ................... 41

Table 8 State option required and excess rate-based ERCs for coal-fired EGUs and NGCC plants for 2030 generation in scenario 4, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state emission rate goal is 1,146 lbs/MWh. Fleet net generation is maintained at 2012 levels. .................................................................................................................................................................................... 42

Table 9 Technology option required and excess rate-based ERCs for coal-fired EGUs and NGCC plants for 2030 generation in scenario 4, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The technology option emission rate goals are 1,305 lbs/MWh for steam generation and 771 lbs/MWh for NGCC generation. Fleet net generation is maintained at 2012 levels. The GS-Shift ERCs are calculated with Equations [3] and [4]. .................................................................................................................... 42

Table 10 State option required and excess rate-based ERCs for coal-fired EGUs and NGCC plants for 2030 generation in scenario 5, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state emission rate goal is 1,146 lbs/MWh Fleet net generation is maintained at 2012 levels. .................................................................................................................................................................................... 43

Table 11 Technology option required and excess rate-based ERCs for coal-fired EGUs and NGCC plants for 2030 generation in scenario 5, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The technology option emission rate goals are 1,305 lbs/MWh for steam generation and 771 lbs/MWh for NGCC generation. Fleet net generation is maintained at 2012 levels. The GS-Shift ERCs are calculated with Equations [3] and [4]. .................................................................................................................... 43

Table 12 Distribution of mass-based allowances for fossil fuel based upon 2012 generation. ................................... 45

Table 13 Scenario 1 net generation and required/excess mass-based allowances for coal-fired EGUs and NGCC plants for 2030 generation, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. Net generation is maintained at 2012 levels. The mass-based goal is 12,412,602 short tons of CO2. ....................................................................................................................................................................... 45

Table 14 Scenario 2 net generation and required/excess mass-based allowances for coal-fired EGUs and NGCC plants for 2030 generation, after the generation shift and the 2.1% heat rate improvement is applied to the

Integrated Environmental Control Model - Technical Documentation List of Tables • xi

coal-fired EGUs. Fleet net generation is maintained at 2012 levels. NGCC plants remain unchanged from scenario 1 (Table 13) and are omitted. The mass-based goal is 12,412,602 short tons of CO2. ........................ 45

Table 15 Scenario 3 net generation and required/excess mass-based allowances for coal-fired EGUs and NGCC plants for 2030 generation, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. Fleet net generation is maintained at 2012 levels. NGCC plants remain unchanged from scenario 1 (Table 13) and are omitted. The mass-based goal is 12,412,602 short tons of CO2. ........................ 46

Table 16 CCS system and pipeline transport parameters for the IECM. ................................................................... 46

Table 17 Historical net generation and summertime peak capacity, and IECM simulated LCOE, VOM, and makeup water for New Mexico generation sources in 2010. ................................................................................. 48

Table 18 Historical net generation and IECM simulated LCOE, VOM, and makeup water for New Mexico generation sources in 2012, excluding hydropower. .............................................................................................. 49

Table 19 Historical net generation and CO2 emission intensity, and IECM simulated LCOE and makeup water for New Mexico generation sources in 2010 and 2012 by plant type, excluding hydropower. ........................... 49

Table 20 Scenario 1 net generation and CO2 emission intensity, and IECM simulated LCOE for New Mexico generation sources under the rate and mass-based CPP approaches applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state and technology options are used for in rate-based approach. The mass-based approach does not include new generation sources. ......... 51

Table 21 Scenario 2 net generation and CO2 emission intensity, and IECM simulated LCOE and makeup water for New Mexico generation sources under the rate and mass-based CPP approaches applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. In this scenario, carbon capture and storage technology with an auxiliary natural gas boiler is applied to San Juan EGU #4 to capture 40% of coal relate CO2 emissions. The state option is used for the rate-based approach. The mass-based approach does not include new generation sources. ................................................................................... 52

Table 22 Scenario 3 net generation and CO2 emission intensity, and IECM simulated LCOE and makeup water for New Mexico generation sources under the rate and mass-based CPP approaches applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. In this scenario, carbon capture and storage technology without an auxiliary natural gas boiler is applied to San Juan EGU #4 to capture 40% of coal relate CO2 emissions. The state option is used for the rate-based approach. The mass-based approach does not include new generation sources. .......................................................................... 52

Table 23 Scenario 4 net generation and CO2 emission intensity, and IECM simulated LCOE and makeup water for New Mexico generation sources under the rate-based CPP approach applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state and technology options are used for the rate-based approach to simulate state profile, when the coal-fired capacity factor in scenario 1 is reduced to maintain total 2012 net generation levels and achieve CPP compliance. .................... 52

Table 24 Scenario 5 net generation and CO2 emission intensity, and IECM simulated LCOE and makeup water for New Mexico generation sources under the rate-based CPP approach applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state and technology options are used for the rate-based approach to simulate state profile, when the San Juan EGU #1 in scenario 1 is retired to maintain total 2012 net generation levels and achieve CPP compliance. ..................................... 53

Table 25 Scenario 1 IECM simulated VOM for New Mexico generation sources under the mass-based approach and the state option for the rate-based approach of the CPP applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. ERC and MA prices are set to $0/MWh and $0/ton of CO2, respectively. .............................................................................................................................. 54

Integrated Environmental Control Model - Technical Documentation List of Tables • xii

Table 26 Scenario 1, IECM simulated plant makeup water needs (gigagallons/year) for New Mexico generation sources under the rate and mass-based CPP approaches applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state and technology options are used for in rate-based approach. The mass-based approach does not include new generation sources. ........................ 55

Table 27 Scenario 1, IECM simulated cooling system makeup water needs (tons/hour) for specific New Mexico generation plants under the rate and mass-based CPP approaches applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state and technology options are used for in rate-based approach. The mass-based approach does not include new generation sources. ......... 56

Table 28 Scenario 1 makeup water 95% CDF ranges for specific New Mexico generation plants for state option in rate-based approach and the mass-based approach in 2030. The technology option for the rate-based approach yields water use metrics identical to those for the state option. Estimations are based upon an uncertainty simulation in the IECM that only accounts for the uncertainty in the average annual temperature and humidity. ...................................................................................................................................... 56

Table 29 Scenario 1 makeup water 95% CDF ranges for New Mexico generation sources for state option in rate-based approach and the mass-based approach in 2030. The technology option for the rate-based approach yields water use metrics identical to those for the state option. Estimations are based upon an uncertainty simulation in the IECM that only accounts for the uncertainty in the average annual temperature and humidity..................................................................................................................................................................... 57

Table 30 Comparison of dispatch orders for New Mexico generation sources in 2012 and 2030, based upon VOM. 2030 projections are for scenario 1 IECM simulated VOM under the mass-based approach and the state option for the rate-based approach of the CPP applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. ERC and MA prices are set to $0/MWh and $0/ton of CO2, respectively. ..................................................................................................................................................... 68

Table 31 Comparison of USGS estimation of 2010 site water consumption for New Mexico modeled fossil fuel plants to the results for water consumption from simulating plant performance for 2010 in the IECM. Previously reported PC EGU plant water consumption simulation results are summed by site and converted to millions of gallons per day. .................................................................................................................................. 78

Table 32 Comparison of USGS estimation of 2010 site normalized water consumption for New Mexico modeled fossil fuel plants and the results for normalized water consumption from simulating plant performance for 2010 in the IECM to the normalized water consumption ranges by technology type reported Macknick (2012) and by the plant owners. Previously reported PC EGU plant water consumption simulation results are summed by site. The IECM estimation method yields normalized values that are typically closer to the owner reported values than does the USGS estimation. ........................................................................................ 78

Integrated Environmental Control Model - Technical Documentation Acknowledgements • xiii

Acknowledgements

This work is supported by the National Energy Technology Laboratory via the Sandia National Laboratories under Work Package 14-017626. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not reflect the views of any agency.

Integrated Environmental Control Model - Technical Documentation Exploring Energy-Water Issues in the United States • 14

Technical, Water, and Economic Impacts of Low-Carbon Electricity Futures under the Clean Power Plan: A Case Study in New Mexico

Objectives of this Report This report presents the application of the compliance procedures outlined in the Clean Power Plan final regulation to the affected fossil fuel power plants in New Mexico, as modeled with the IECM.

Introduction Water use by electric power plants is becoming an increasingly important issue in many regions of the country, especially in the southwestern United States. To mitigate carbon dioxide (CO2) emissions from fossil-fueled power plants, carbon capture and storage (CCS) technologies are receiving considerable attention from the U.S. Department of Energy (DOE) and others. In the recently announced Clean Power Plan (CPP),1 the U.S. Environmental Protection Agency (EPA) indicated that CCS could be a viable, cost-effective option for states to achieve the required reduction in CO2 emissions from resident power plants. Therefore, the main objectives of our project are to extend and apply the analytical capabilities of the Integrated Environmental Control Model (IECM), developed by Carnegie Mellon University, to simulate possible CO2 mitigation measures for coal and natural gas power plants in New Mexico to comply with the CPP regulations 1 “Federal Plan Requirements for Greenhouse Gas Emissions From Electric Utility Generating Units Constructed on or Before January 8, 2014; Model Trading Rules; Amendments to Framework Regulations; Proposed Rule,” Title 40 Code of Federal Regulations, Parts 60, 62, and 78. 2015 ed.


concerning existing power plants. In particular, we examine the impact of the CPP on the levelized cost of electricity (LCOE) for the state, the resulting cost of CO2 avoided for the different implementations, assess the subsequent makeup water requirements for the state fossil fuel-fired electric power plants using a variety of cooling technologies and processes for CCS, and determine the availability and resulting cost of using alternative water sources for cooling makeup water at these sites In August 2015, the EPA released the final regulation for the CPP.2 The backbone of this plan consists of three building blocks to guide state stakeholders as how to reduce CO2 emissions from fossil fuel power plants. The guidance in the first block (B1) is to improve the heat rate efficiency of the coal-fired power plants affected by this plan by 2.1% to 4.3%. The degree of this improvement depends upon the interconnection region in which the plant resides, which acts as a proxy for the age of the plant. The suggestion in the second building block (B2) is to re-dispatch electrical generation from the coal-fired plants to lower CO2-emitting natural gas and combined cycle plants (NGCC). This is a one-for-one shift between the two sources such that the capacity for the NGCC may increase to 75% of the summertime peak capacity for the plant. The recommendation in the third building block (B3) is that new no- and low-carbon generation sources be built to offset the emissions from the existing fossil fuel fleet and to provide additional electrical generation. This is particularly relevant because the EPA contemplates creation of certain mass allowances and emission reduction credits associated with these no/low-carbon generators that can be sold to the fossil fuel plants.

Whether mass allowances or emission reduction credits are sold depends upon which of the two recommended approaches (mass-based or rate-based) is taken by the state in which the plant resides. The mass-based approach sets a limit on the quantity of short tons of CO2 that can be emitted from the fossil fuel plants in a given year. Here, there are two limits—one when new fossil fuel generating sources are included and one that considers only the existing fossil fuel sources. Of these two options, the new source limit for a state permits more CO2 emissions. In the rate-based approach, there are also two options. One option (subcategory) considers that the CO2 emission intensity depends upon the technology of the generation source. Therefore, separate emission intensity standards are used for generation from NGCC and steam turbine plants. The second option is to create a unique emission intensity standard for the state. This standard is the 2012 generation source-weighted average of the of the two aforementioned standard emission intensities. For each approach there is also a backstop that allows the state to devise its own plan. This backstop may consist of existing plans such as the California cap-and-trade carbon market, the Regional Greenhouse Gas Initiative, state and renewable portfolio standards.

2 On 9 February 2016, the United States Supreme Court issued a stay order concerning the CPP. On the date of issuance of this report on the CPP implementation in New Mexico, a ruling has yet to be given by the lower court or by the Supreme Court. It is possible that the forthcoming ruling will require the EPA to modify some or all of the structure of the CPP. If the CPP is struck down or requires modification, this research is still useful in that it provides insights into how a policy concerning CO2 reduction in fossil fuel power plants can be modeled. Furthermore, the analysis of the current CPP


Research Methods Assessing a possible scenario under which New Mexico can comply with the CPP requires site-specific and generic modeling of multiple power plant technologies: coal-fired electric generation units (EGUs), NGCC plants, geothermal plants, utility photovoltaic (PV) projects, and wind turbine farms. Attributes concerning the plant location, configuration, year online, capacity, heat rate, fuel type and cost, operational hours or capacity factor, and annual generation are needed to improve the accuracy of either model type. Therefore, the operational and chronological characteristics of the individual, existing EGUs in the fleet are gathered from various public databases. With this information, the existing EGUs can be profiled in different models to determine the change in CO2 emissions, water consumption, and LCOE that arise from applying the EPA recommended compliance methods for the fossil-fuel fleet, and applying CCS as an alternative CO2 mitigation measure for coal-fired EGUs. The assessment framework of the model construction and implementation for the existing coal-fired EGUs and NGCC plants is shown in Figure 1. Three tools are constructed to evaluate implementation of the CPP: the Power Plant Simulation Tool (PPST), the State Generation Parameter Tool (SGPaT), and the State Generation Portfolio Tool (SGPoT). The site-specific power plant database provides the inputs of the characteristics of the different power plants in New Mexico for the PPST with which the unique, site-specific simulation of plant performance attributes and costs are constructed. The performance attributes of this simulation are then modified according to the CPP building blocks for improving the heat rate of the coal-fired EGUs (B1) and redistributing the mix of electrical power generation (B2). The simulation parameters are further modified dependent upon the CPP approach and options scenarios that are run. Some of these scenarios may include retrofitting a coal-fired EGU with CCS to mitigate the emissions and emission intensity. The site-specific power plant database also provides inputs for the SGPaT, which is a spreadsheet. This tool catalogues the characteristics and historical performance of the fossil fuel and renewable energy sources and calculates the estimated historical LCOE of the renewable energy sources. The outputs from the PPST and SGPaT tools are combined to form the SGPoT tool. In this spreadsheet, the implications of the required, new renewable energy sources (B3) on the state power plant fleet in the different scenarios are determined. These implications include the additional capacity of new utility solar and wind farms and the associated LCOE. The makeup water use, LCOE, and cost of CO2 avoidance metrics for the fossil fuel and overall fleet in the given scenarios are then calculated with the prices associated with the water, fuel, and emission reduction credits and mass allowances described by the EPA. Compliance also shapes the future EGU generation needs. Here, required attributes (primarily concerning capital and operation and maintenance (O&M) costs, and capacity factors) are needed to create models to estimate the LCOE and annual generation from


the new renewable technology EGUs that are necessary to provide the emission reduction credits and mass allowances specified by the EPA for compliance. These attributes are gathered from scientific literature on the respective technologies. Insights into CPP compliance scenarios can be garnered by combining the outputs from the above models into deterministic and stochastic models that determine the costs of compliance and assess the uncertainty around the underlying input parameters that determine these costs.

Figure 1 Schematic diagram of model construction and implementation for the existing coal-fired EGUs and

NGCC plants to be compliant for different scenarios under the CPP.

Databases Power Plant Profiles Databases from several U.S. agencies are used to profile the New Mexico power plants in the PPST (Table A-A1). This study uses the EPA’s two emissions and electricity generation databases that provide operations information of power plant EGUs and plants in 2010: National Electronic Energy Data Systems (NEEDS)3 version 5.13 and the ninth edition Emissions and Generation Resources Database (eGRID)4 version 1.0. The United States Energy Information Administration (EIA) supplies year-specific datasets on fuel (Form 923)5, as well as descriptions of power plant cooling systems and water use (Form 860).6 Operational data are also obtained from the Federal Energy Regulatory

3 “EPA Power Sector Modeling Platform v.5.13.” www.epa.gov. 25 June. 2014. n.p. Web. 20 October.

2014. <http://www.epa.gov/airmarket/progsregs/epa-ipm/BaseCasev513.html>. 4 “eGRID.” http://www.epa.gov. 5 August. 2014. n.p. Web. 20 October. 2014. <http://www.epa.gov/cleanenergy/energy-resources/egrid/>. 5 “Form EIA-923 detailed data” www.eia.gov. n.d. n.p. Web. 20 August. 2015. <http://www.eia.gov/electricity/data/eia923/>. 6 “Form EIA-860 detailed data” www.eia.gov. n.d. n.p. Web. 20 August. 2015. <http://www.eia.gov/electricity/data/eia860/>.

http://www.nrel.gov/

http://www.epa.gov/airmarket/progsregs/epa-ipm/BaseCasev513.html

http://www.epa.gov/

http://www.epa.gov/cleanenergy/energy-resources/egrid/

http://www.eia.gov/

http://www.eia.gov/electricity/data/eia923/

http://www.eia.gov/



Commission (FERC) Form 1.7 In particular, the NEEDS database provides unit-level information about the configuration, capacity, heat rate and type of fuel used, as well as the location, boiler online time, and associated emission control devices and their on-line time. The eGRID database provides additional characteristics of EGUs including the emission and power generation characteristics, such as the unit operating hours, net power generation. Form 923 supplies information concerning the quantity and price of the fuels consumed by the power plants in the various years. Form 860 provides plant-specific information concerning the water-cooling system type, operational date, and the water source and use.8 The FERC database supplies the EGU gross generation, with which the parasitic load can be calculated. The EPA supplies other data in various documentation9 to directly support of the 111(d) proposal and final rule. State-level data on historical coal-fired power generation and associated CO2 emission intensity for the option 1 building blocks come from Appendix 1 of the technical support documentation for the proposal goal computation. Data concerning the characteristics of the fossil fuel sources that are classified by the EPA as “affected” sources by the CPP are from the data file concerning 2012 unit-level data.10

Ambient Temperature, Humidity and Pressure Temperature, humidity and pressure will affect the makeup water use at a water-cooled fossil fuel plant (Zhai et al., 2011), while steam cycle capacity and plant efficiency will be affected at a dry-cooled plant (Electric Power Research Institute, 2005). As such, site-specific annual temperature, humidity and pressure data are required to improve accuracy when simulating the performance parameters of individual coal-fired EGUs and NGCC plants. For this study, these data were obtained by using monthly, average, meteorological data from the Quality Controlled Local Climatological Data (QCLCD) from the National Climatic Data Center (NCDC)11 weather stations nearest to the individual power plants. Site-specific data for 2010 and 2012 are used to determine the average annual temperature, humidity and pressure12 for those years, respectively. In some instances, complete datasets for the nearest weather station are not available; in this case, data from the next closet weather station are used (Table A-A2). Humidity data 7 “Form 1- Electric Annual Utility Report.” http://www.ferc.gov. 25 April. 2012. n.p. Web. 20 October. 2014. <http://www.ferc.gov/docs-filing/forms/form-1/viewer-instruct.asp>. 8 The United States Geological Survey (USGS) also compiles the power plant-specific EIA source and withdrawal information for 2010 and compares these figures to estimates made by the USGS. “Scientific Investigations Report 2014-5184.” www.usgs.gov. 10 November. 2014. n.p. Web. 20 August. 2015. <http://pubs.usgs.gov/sir/2014/5184/>. 9 “Clean Power Plan Final Rule Technical Documents.” http://www2.epa.gov. 13 August. 2015. n.p. Web. 27 August. 2014. < http://www2.epa.gov/cleanpowerplan/clean-power-plan-final-rule-technical-documents>. 10 “Clean Power Plan State Goal Visualizer.” http://www2.epa.gov. 26 August. 2015. n.p. Web. 26 August. 2015. <http://www2.epa.gov/cleanpowerplantoolbox>. 11 “NOAA National Centers for Environmental Information.” http://gis.ncdc.noaa.gov. n.d. n.p. Web. 27 August. 2015. <http://gis.ncdc.noaa.gov/map/viewer/#app=cdo>. 12 QCLCD temperatures are reported in degrees Celsius and converted to degrees Fahrenheit for the IECM. The QCLCD reports dew point rather than relative humidity; therefore, the dew point data is converted to relative humidity with the August-Roche-Magnus equation. Atmospheric pressures are reported in inches of mercury and converted to pounds per square inch absolute (psia), though corrections are not made for variations in mean sea level pressure.

http://www.ferc.gov/

http://www.ferc.gov/docs-filing/forms/form-1/viewer-instruct.asp

http://www.usgs.gov/

http://pubs.usgs.gov/sir/2014/5184/

http://www2.epa.gov/

http://www2.epa.gov/cleanpowerplan/clean-power-plan-final-rule-technical-documents



http://www2.epa.gov/cleanpowerplantoolbox

http://gis.ncdc.noaa.gov/

http://gis.ncdc.noaa.gov/map/viewer/#app=cdo


from the next closest weather station are used as a proxy, when only humidity data are missing from the nearest station (Table A-A3). Here, we assume that the air mass around the stations is stationary so that the dew point will be the same regardless of the air temperature.13 The month-specific average dew point is not recorded in eight of the 400 station samples taken between 2008 and 2012. In these cases, the month and site-specific average for the remaining years is used as the proxy relative humidity for that month. Site-specific, mean temperature and humidity for 2030 are estimated by using the mean values for these parameters that result from the best-fit cases when continuous distribution functions are fitted to the average of the annual values for each parameter from 2008 through 2012.14 15 For all sites, the best-fit distribution for temperature data is a uniform distribution. For three of the four weather stations, the uniform distribution is also the best-fit distribution for humidity data; the uniform distribution is the second best fit for the fourth station. Therefore, the uniform distribution is used for humidity in all cases, for simplicity. These fitted distributions (Table A-A4) also bound the probable minimum and maximum values for the metrics at each site so that a probabilistic analysis of the uncertainty in the metrics on other power generation parameters can be examined. Hence, the mean temperature and humidity values are used for deterministic predictions, while the temperature and humidity distributions are used for probabilistic simulation in the PPST.

Water Use and Alternative Water Sources In addition to the information obtained on the plant-specific cooling system parameters from FERC Form 860, this study uses data obtained by the United States Geological Service (USGS) for comparison with plant-level water consumption estimates from the IECM model.16 The USGS report data include the water source and the estimated, plant-specific annual water consumption per day from 2010 (Table A-A5). As such, the water consumption is not compiled at the unit-level and requires the model to aggregate the unit-level consumption at the site. Tidwell et al. (2014) provides the database for alternative water sources that can be used as makeup water for cooling (Table A-A6).17 These sources comprise of unappropriated groundwater, appropriated surface water, brackish groundwater, and wastewater that is available in the basin in which the power plant is located. The costs associated with these water sources include the capital and operation and management costs associated purchase rights, transportation, and any subsequent treatment (Table A-A7).

13 “Unit 4: Temperature-Moisture Relationship.” http://ocw.usu.edu. n.d. n.p. Web. 27 August. 2015. <http://ocw.usu.edu/Forest__Range__and_Wildlife_Sciences/Wildland_Fire_Management_and_Planning/Unit_4__Temperature-Moisture_Relationship_4.html>. 14 This topic will be discussed in the description of the IECM. 15 The mean, site-specific atmospheric pressure is taken as the mean of the annual average pressure for each of these five years. Uncertainty for the 2030 atmospheric pressure is not considered because the historical standard deviations over five years for each site are less than one percent of the mean values. 16 United States Geological Survey. Scientific Investigations Reports 2014-5184. http://www.usgs.gov. 10 November. 2014. n.p. Web. 2 February. 2016. <http://pubs.usgs.gov/sir/2014/5184/>. 17 Replacement water for cooling water makeup is only considered in this study to minimize the required water treatment cost.

http://ocw.usu.edu/Forest__Range__and_Wildlife_Sciences/Wildland_Fire_Management_and_Planning/Unit_4__Temperature-Moisture_Relationship_4.html

http://ocw.usu.edu/Forest__Range__and_Wildlife_Sciences/Wildland_Fire_Management_and_Planning/Unit_4__Temperature-Moisture_Relationship_4.html

http://www.usgs.gov/

http://pubs.usgs.gov/sir/2014/5184/


Data Analysis The NEEDS version 5.13 database lists ninety-three EGUs for New Mexico. Of these, the EPA requires only twenty-five EGUs to comply with the CPP (Table A-A1). Sixteen of these units are classified as coal-steam or NGCC plants that are modeled in this study.18 The remaining nine units are classified as oil/gas steam (OGST) units and are not modeled in full in this study. While the LCOE for these units is ignored in the fossil fuel and overall fleet LCOE estimation, the water consumption for these units is calculated based upon Macknick et al. (2012)19, and the CO2 emission intensities are considered and mitigated for to achieve the state CPP goal.20

Database Calculations To study the change in fleet CO2 emissions, net CO2 emission intensity, LCOE, and water use due to implementation of the CPP mitigation technologies, we require the database to contain all information necessary to simulate the generating units and the associated environmental control systems in the IECM. This includes the summer nameplate capacity/plant size, annual operational hours, coal type, the gross and net power production, and the parasitic load. The NEEDS, eGRID, and FERC databases do not include the parasitic load, but this parameter can be estimated based on gross and net power outputs. One must determine the capital and operating and maintenance costs for the EGU to calculate the LCOE. The capital portion of this cost depends upon the percent of the cost already amortized. This amortization is taken as a direct percent of the age of the component, estimated to be equal to the referred year (e.g. 2012 or 2015) minus the on-line year for the component, to the component booklife.

The IECM The Power Plant Simulation tool used in this study is the Integrated Environmental Control Model21 (IECM, version 8.02). The IECM is a publically available computer-modeling tool developed by Carnegie Mellon University for the U.S. Department of Energy’s National Energy Technology Laboratory (DOE/NETL). The model is based on mass and energy balances along with empirical data to estimate the plant-level performance, emissions, and costs of pulverized coal (PC), NGCC, and integrated gasification combined cycle (IGCC) plants. With the model, the user is able to specify the plant features, fuel properties, EGU and generator characteristics, and emission 18 This accounts for 89% of the 2012 state net electrical generation and 92% of the CO2 emissions. 19 This study uses the median value (826 gals/MWh) for the water consumption for a natural gas steam plant as a proxy for all OGST plants. 20 The net generation for the OGST plants is held constant at 2012 values in 2010 and 2030 for illustrative purposes. 21 “Integrated Environmental Control Model Documentation.” http://www.cmu.edu/epp/iecm/iecm_doc.html. n.d. n.p. Web. 22 December. 2013.

http://www.cmu.edu/epp/iecm/iecm_doc.html


control characteristics to simulate plant performance under these different designs and configurations. In particular, the water system module estimates the cooling water, water consumption, and makeup water required for the steam cycle, cooling system, and any fitted emission control devices. Cost estimates are also made concerning the associated capital cost, O&M, and total annual levelized cost of electricity generation in the absence of production tax credits or other incentives, defined as:

𝐿𝐶𝑂𝐸 = 𝑝𝑙𝑎𝑛𝑡 𝑐𝑎𝑝𝑖𝑡𝑎𝑙 𝑐𝑜𝑠𝑡 ×𝑓𝑖𝑥𝑒𝑑 𝑐ℎ𝑎𝑟𝑔𝑒 𝑓𝑎𝑐𝑡𝑜𝑟+𝑜𝑝𝑒𝑟𝑎𝑡𝑖𝑜𝑛 𝑎𝑛𝑑 𝑚𝑎𝑖𝑛𝑡𝑒𝑛𝑎𝑛𝑐𝑒 𝑐𝑜𝑠𝑡𝑛𝑒𝑡 𝑝𝑜𝑤𝑒𝑟 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑖𝑜𝑛

, [1]

where LCOE is measured in dollars per megawatt hour ($/MWh) on the net generation basis and also is a function of fixed charge factor. The IECM output for the various performance metrics can be estimated deterministically or stochastically. 22 In the stochastic method, multiple, quasi-independent input parameters are modeled as continuous distributions23 from which samples are drawn, using Monte Carlo or similar techniques (Zhai et al., 2012), to construct the set of inputs for evaluation in the model. These randomly drawn sets are then sequential evaluated to construct a probability distribution function (PDF), or cumulative distribution function (CDF), of the performance metrics that are used to assign probabilistic uncertainty bounds for the output metrics given the uncertainty in and the interaction of the model input parameters. From our study dataset of New Mexico, the plant characteristics of the selected, individual EGUs are used as the IECM inputs to configure and model the unique EGUs. These aspects include the installed cooling system and emission control devices, the coal type and price, capacity factor (here defined as the operating hours divided by the hours in one year), gross power output, net power output, parasitic load, and net heat rate. The EGU-specific inputs also include average temperature, humidity, and atmospheric pressure for modeled years. The IECM defaults are used for all other parameters, including the finance, performance, and supply prices.24 Once the EGU model is established, the resulting performance, emissions, water use, and LCOE from the associated costs are used to determine the state fleet metrics under the CPP. The IECM modeling procedure details specific to EGU types are discussed in the mitigation section. The CO2 emission intensity as calculated from eGRID data (Table A-A8) will not exactly match that derived from the IECM model. This inaccuracy relates to various uncertainties in the model assumptions, such as the composition of the fuel, as well as in the eGRID data. For the coal-fired EGUs, the ratio of the IECM simulated CO2

22 Ibid. 23 The IECM uses the following distribution for uncertainty analysis: lognormal, normal, triangular, uniform, half normal, and negative half normal. Therefore, only these distributions were considered for fitting the site-specific temperature, relative humidity, and atmospheric data. 24 No cost is associated with historical or future water use from current sources. The site-specific costs used for alternative water sources in 2030 are discussed in another section.


emission rate from mitigation improvements and that from the simulated 2010 baseline rate is multiplied by the eGRID calculated 2010 value to compensate for this difference. Modeling NGCC plants in the IECM has two limitations. Firstly, only two models of gas turbines can be simulated—GE 7FA and GE 7FB—that may differ from the gas turbines used in the existing NGCC plants. The GE 7FA is an older model found in many already constructed NGCC plants. The GE 7FB is a newer, advanced turbine that has a higher inlet temperature, pressure ratio, and adiabatic efficiency, which result in a lower net heat rate. This turbine is appropriate for newly constructed NGCC plants. Therefore, with these defined performance standards that yield net generation based substantially on the turbine type, the number of turbines and the capacity factor, this study uses the GE-7FA model, with the number of gas turbines described in eGRID, and varies the hours of operation to achieve the documented net generation.25 The second limitation is that the steam cycle heat rate cannot be adjusted; therefore the plant net heat rate cannot be adjusted to simulate the net plant heat rate in the NEEDS and eGRID databases. These limitations result in the LCOE, makeup water use, and emissions rate for the simulated plant to be lower than those for the real plant (Appendix C). To overcome these biases, the LCOE for the modeled plant is determined by multiplying the IECM simulated LCOE by the ratio of the capacity for the modeled plant and that for the simulation raised to a power for the capital cost, and then adding the fuel and water VOM components scaled according to the ratio of the net heat rate for the modeled plant and the IECM value for the net heat rate for O&M costs (Equation [A-C4]). Similarly, the makeup water use (tons/hr) is adjusted by multiplying the simulated makeup water use by the product of the ratios of the capacities raised to a power, which accounts for differences due to size, and the ratio of the actual heat rate to the simulated heat rate, which accounts for the impact of net heat rate on water use (Equation [A-C6]). Since the emission rate is a function of heat rate rather than capacity factor, the emissions rate calculated from the eGRID data is reported.

Financial Calculations The fixed charge factor (FCF) for calculating the LCOE is estimated based upon the financial schedule outlined in the IECM.26 The FCF for all past, non-wind turbine construction is the default 11.3%, which is based upon a book life of thirty years, and a twenty-five year depreciation schedule. Current and future projects are assessed at the same FCF, given the same book life and depreciation schedule. Wind turbines have a twenty-year life (Tidball et al., 2010); therefore, the default FCF for all wind projects is 13%. As the characteristics concerning CO2 emissions (operating hours, net power generated, emissions estimate) derive from 2010 data, all fiscal calculations are in 2010

25 Afton (Groves, 2009) and Luna (PCL, 2016) NGCC plants use GE 7FA gas turbines, while the Hobbs facility uses Mitsubishi F series turbines (Colorado Energy, 2012). The type of turbine used at the Bluffview facility is a GE LM6000 PD turbine (Farmington, 2015). All simulated NGCC plants went online between 2002 and 2008; therefore, we assume that all facilities use the older GE 7FA models. 26 “Modeling of Integrated Environmental Control Systems for Coal-Fired Power Plants.” www.cmu.edu. n.d. n.p. Web. 20 October. 2014. <http://www.cmu.edu/epp/iecm/iecm_doc.html>.

http://www.cmu.edu/



dollars. For instances where capital and O&M costs are estimated external to the IECM, the Chemical Engineering Plant Cost Index (CEPCI) is used to convert the costs to 2010 dollars. When water27 and fuel prices are taken from historical data or estimates are made for future years, the Consumer Price Index (CPI) is used to convert the costs to 2010 dollars.28

Emission Compliance As coal-fired EGUs will require emissions upgrades to comply with current Mercury and Air Toxics Standards (MATS) and National Ambient Air Quality Standards (NAAQS) from 2015, all coal-fired EGUs without every emission control device are upgraded to include the missing devices. Here, wet scrubbers for sulfur oxides (SOx), low-nitrogen oxide burners (LNB) for in-furnace nitrogen oxide (NOx) control, hot-side SCR for post-combustion NOx control, and carbon injection for mercury (Hg) control are added, per Table 1. For New Mexico, all affected EGUs are already equipped with wet scrubbers and LNB, and two EGUs have Hg control devices. Other devices, per below, are added in the model where required.

Table 1 Site-specific coal-fired EGU emission compliance configuration.

Device Type Parameter IECM Default Setting

Combustion Controls NOx Low NOx burner (LNB)

Post Combustion Controls NOx Hot-side Selective Catalytic

Reduction (SCR)

SOx Wet Flue-Gas Desulfurization

(FDG)

Hg Carbon Injection

Fuel Price The prices per MMBtu based upon quantity-weighted averages from Form 923 were used to determine the historical coal and natural gas prices for specific EGUs and NGCC plants for 2010 and 2012 generation. If no information was available for a plant in either year, the MMBtu quantity-weighted average price for other plants is used as a proxy. Future fuel prices (Figure A-B1) are based upon the percentage change of the base case 2012 price relative to the EIA estimate of the 2030 price due to implementation of the

27 We assume that values given by Tidwell et al. (2014) are in 2012 dollars. 28 “Consumer Price Index, 1913-.” http://www.minneapolisfed.org. n.d. n.p. Web. 20 August. 2015. <https://www.minneapolisfed.org/community/teaching-aids/cpi-calculator-information/consumer-price-index-and-inflation-rates-1913>.

http://www.minneapolisfed.org/

https://www.minneapolisfed.org/community/teaching-aids/cpi-calculator-information/consumer-price-index-and-inflation-rates-1913

https://www.minneapolisfed.org/community/teaching-aids/cpi-calculator-information/consumer-price-index-and-inflation-rates-1913


CPP proposal (EIA, 2015).29 Uncertainty in the forecast is accounted for by applying to the EIA forecasted price a nominal offset percent error that is normally distributed. Values for the nominal offset and the standard deviation for the normal distribution are taken from studies done by Newcomer and Apt (2007) for EIA coal related forecasts and by Rode and Fischbeck (2006) for EIA natural gas related forecasts (Figure A-B2). The nominal values for the coal and natural gas forecast error are 2.5% and -0.7%, respectively. The standard deviation values for the coal and natural gas forecast error distributions are 5% and 34.2%, respectively. While the statistics in the aforementioned studies may not include the effect of policies such as the CPP on fuel prices, the study durations do cover system shocks such as hurricanes, electricity market deregulation, the California energy crisis and the second Gulf War. Hence, the variation statistics may be useful to measure fuel price uncertainty in the low-carbon CPP world.

EPA CO2 Mitigation Measures The CPP utilizes two explicit methods to mitigate CO2 emissions and emission intensity—improving the heat rate of coal-fired EGUs and shifting generation from coal-fired EGUs to NGCC plants. These emissions are also implicitly mitigated through the retirement of coal-fired power plants and through the reliance on generation from new zero- and low-carbon emitting generation sources that enable the emission intensity rates for the fossil fuel plants to be reduced through the use of emission reduction credits, or allow the fossil fuel plants to emit more CO2 through the purchase of mass allowances from these zero- and low-carbon emitting facilities. The CPP also discusses the use of carbon capture and sequestration as a possible method to reduce emission from fossil-fuel plants. As such, this mitigation technology is applied with a 40% capture rate to one coal-fired EGU in a case study. All of the coal-fired EGUs, NGCC plants in this study are located in New Mexico and are classified by the EPA as being affected by the CPP. This does not mean that a single corporation that operates solely in New Mexico owns all of the plants, or that the electricity is explicitly generated for the New Mexico market. However, this study treats all the generators and the resulting electricity as such for simplicity.

Improving Plant Heat Rate In the CPP, the EPA proposes that the average heat rate for the western fleet of coal-fired EGUs can be improved by 2.1%, based upon the EPA’s regional analysis of an eleven-year study of EGU gross heat rate variations.30

29 “Analysis of the Impacts of the Clean Power Plan.” http://www.eia.gov. n.d. 22 May. 2015. Web. 27 August. 2015. <http://www.eia.gov/analysis/requests/powerplants/cleanplan/>. Data from the table for energy prices by sector and source was used for natural gas and coal prices 2012 base case and 2030 policy case for the electric power sector (http://www.eia.gov/beta/aeo/#/?id=3-CPP2015&cases=ref_cpp2015~rf15_111_all). 30 “Federal Plan Requirements for Greenhouse Gas Emissions From Electric Utility Generating Units Constructed on or Before January 8, 2014; Model Trading Rules; Amendments to Framework Regulations; Proposed Rule,” Title 40 Code of Federal Regulations, Parts 60, 62, and 78. 2015 ed.


http://www.eia.gov/analysis/requests/powerplants/cleanplan/

http://www.eia.gov/beta/aeo/#/?id=3-CPP2015&cases=ref_cpp2015~rf15_111_all

http://www.eia.gov/beta/aeo/#/?id=3-CPP2015&cases=ref_cpp2015~rf15_111_all


Generation Shifting In New Mexico, the 2012 emission intensity of the NGCC fleet is substantially lower than that for the coal-fired fleet (907 lbs/MWh and 2,341 lbs/MWh, respectively).31 Therefore, shifting electricity generation from existing coal-fired EGUs to existing NGCC plants will decrease the overall mass of CO2 emitted and the average rate at which it is emitted. Here, the EPA goal is for all NGCC plants to run at 75% capacity factor and to reduce the capacity factor for the coal-fired EGUs by a comparable amount to offset the increased generation from the NGCC plants.

Power Plant Retirement Retiring fossil fuel plants that have high CO2 emission intensity rates or produce significant quantities of CO2 relative to other fossil fuel power sources in the state, and replacing that electrical generation with generation from NGCC or renewable sources is an appropriate method to lower emissions and emission rates. In New Mexico, two of the five CPP affected coal-fired EGUs are schedule to retire in 2017—San Juan unit 2 and 3, thereby decreasing the available nameplate capacity by 924 MW. However, these units are not particularly high CO2 emitting units, as the heat rates for these EGUs are similar to those for the other three coal-fired EGUs in the state. Emission Reduction Credits The EPA outlines two approaches with which the states can comply with the CPP in 2030. One approach is rate-based, and has two options to meet the emission rate goal. The first of these options is for the individual fossil fuel plants to meet the state emission goal—1,146 lbs/MWh for New Mexico. The other rate-based option entails that the individual plants meet the nation-wide emission rate goal based upon the generation technology used; the fossil steam technology goal is 1,305 lbs/MWh and the NGCC technology goal is 771 lbs/MWh. This technology standard also serves as the basis for the state emission goal, since the latter is based upon the weighted-average of the percent of overall affected fossil fuel generation derived from each technology option standard for the state in 2012. Each option is similar in nature, with the main difference being the level of the goal for the emission intensity. When the objective is for each coal-fired EGU and NGCC plant to meet the state emissions rate goal, the NGCC plants that already have a lower emission rate than the goal are able to claim emission rate credits (ERCs) for a portion of the difference between the goal and the plant emission rate. The formula for the total credits from an affected EGU is:

𝐸𝑅𝐶𝑠 =

�𝑀𝑊ℎ 𝑓𝑜𝑟 𝑎𝑓𝑓𝑒𝑐𝑡𝑒𝑑 𝐸𝐺𝑈∗(𝑠𝑡𝑎𝑡𝑒 𝐶𝑂2 𝑒𝑚𝑖𝑠𝑠𝑖𝑜𝑛 𝑟𝑎𝑡𝑒 𝑔𝑜𝑎𝑙−𝐶𝑂2 𝑒𝑚𝑖𝑠𝑠𝑖𝑜𝑛 𝑟𝑎𝑡𝑒 𝑓𝑜𝑟 𝐸𝐺𝑈)𝑠𝑡𝑎𝑡𝑒 𝐶𝑂2 𝑒𝑚𝑖𝑠𝑠𝑖𝑜𝑛 𝑟𝑎𝑡𝑒 𝑔𝑜𝑎𝑙

� . [2]

31 “Clean Power Plan State Goal Visualizer.” http://www2.epa.gov. 26 August. 2015. n.p. Web. 26 August. 2015. <http://www2.epa.gov/cleanpowerplantoolbox>.




Conversely, coal-fired EGUs that exceed the goal can purchase credits to meet the goal. These credits have units of zero CO2 emissions per MWh of generation; therefore, the EGUs must purchase enough of these credits from new renewable energy sources, or from generation related to increased capacity factors at nuclear and hydropower plants, to increase the site actual and credited generation enough to offset the carbon emissions for the actual site generation. Equation [2] also indicates that fossil fuel generating sources can produce ERCs for sale, when the emission rate for the source is less than that for the state goal or the technology. As such, many NGCC plants or coal-fired EGUs fitted with CCS systems capturing carbon at sufficiently high rates can produce ERCs. Similarly, fitting a CCS system to a fossil fuel generating source can decrease the need to purchase ERCs. When the rate-based option related to the generation technology is employed, each coal-fired EGU and NGCC plant must meet the intensity goal appropriate for the generation technology used. Meeting the goal still entails purchasing (or selling) necessary ERCs to increase the site actual and credited generation. An important difference between the two options is that the NGCC plants receive credits, called Gas-Shift ERCs (GS-ERCs), to compensate the site for increasing the capacity factor from that in 2012 to 75%. The GS-ERC credits are calculated based upon the product of the NGCC generation, the Incremental Generation Factor and the GS-ERC Emission Factor:

𝐺𝑆 − 𝐸𝑅𝐶𝑠 = 𝑁𝐺𝐶𝐶 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑖𝑜𝑛 ∗ 𝐼𝑛𝑐𝑟𝑒𝑚𝑒𝑛𝑡𝑎𝑙 𝐺𝑒𝑛𝑒𝑟𝑎𝑡𝑖𝑜𝑛 𝐹𝑎𝑐𝑡𝑜𝑟 ∗ 𝐺𝑆

− 𝐸𝑅𝐶 𝐸𝑚𝑖𝑠𝑠𝑖𝑜𝑛 𝐹𝑎𝑐𝑡𝑜𝑟. [3]

Here, the EPA defines the Incremental Generation Factor for 2030 as a constant (0.26), and calculates the GS-ERC Emission Factor as the percent difference between the emission rates of the NGCC plant and the technology option emission rate for steam generation. This relationship is expressed as

𝐺𝑆 − 𝐸𝑅𝐶 𝐸𝑚𝑖𝑠𝑠𝑖𝑜𝑛 𝐹𝑎𝑐𝑡𝑜𝑟 = 1 − �𝑁𝐺𝐶𝐶 𝑒𝑚𝑖𝑠𝑠𝑖𝑜𝑛 𝑟𝑎𝑡𝑒𝑆𝑡𝑒𝑎𝑚 𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 � . [4]

These GS-ERC credits may not be used directly by the NGCC plant that produces them; the credits must be sold to other fossil fuel sources, and as such offset any credits that the NGCC plant needs to purchase to be compliant or further decreases the plant marginal cost. While these ERC credits for both options can theoretically be purchased from any source32 within the contiguous U.S. power plant fleet, this study limits such trading to New Mexico. As such, endogenous market forces will determine the price for these credits. For simplicity, the price of these credits is treated in two manners for the 32 These credits can derive from new renewable, NGCC, hydropower sources, or from extra generation due to increased capacity factor from nuclear and hydropower sources. The CPP also discusses other available credits, but these are not considered in this study. One restriction on this trading is that sources that trade must be in states that have the mass or rate based plan.


deterministic study. The first is as if the individual generating sources involved in the trade are commonly owned so that the credits are internal transfers that do not affect the LCOE. This is equivalent to a $0/MWh price for the credits. The second is with the credits having a non-zero price for external transfers. Here, the LCOE for the technologies will increase through trades between different generating plants, but the overall LCOE for the state will not.33 The stochastic model uses a range of ERC prices, with a uniform distribution in prices from $0 to $50/MWh, to examine the implications of the ERC price on the cost of CO2 avoided for the fossil fleet and the breakeven point for CCS implementation at 40% capture.

Mass Allowance The second approach that the EPA outlines is mass-based and also has two options. Here, each option limits the tonnage of CO2 from fossil fuel sources that can be emitted within the state during the year. The first option imposes an emission limit for New Mexico of 12,412,602 short tons that does not allow generation from new fossil-fuel sources. The second option does allow for new source generation and is limited to 13,229,925 short tons. This study examines only the first option. While it is up to the state to outline a plan to allocate these allowances amongst the power generation sources within the state, the EPA does propose that 5% of the allowances be given to new renewable energy sources as a set-aside, and that another set-aside (called the output-based set-aside) be given to NGCC plants to compensate for additional CO2 emissions from the increased generation related to increasing the capacity factor to 75%. The out-based set-aside for New Mexico is 627,085 short tons. These allowance may also be purchased and sold and are modeled as outlined in the rate-based approach.

Renewable Energy Sources The EPA supporting data indicate that 2.56 terawatt hours (TWh) of New Mexico generation were derived from solar and wind sources in 2012, with an additional 0.22 TWh from hydropower generation (Table A-A9). These sources represent 14.4% of the total generation from fossil fuel sources that will be regulated in the CPP. According to this plan, the reduction in the emissions and the emission intensity from these fossil sources will come in part by investing in new renewable energy sources; this study examines three types of renewable energy as future sources: utility solar photovoltaic, onshore wind, and geothermal. The National Renewable Energy Laboratory (NREL) projections of generation mix for 2030, with 30% renewable energy penetration and incremental technology improvements (30% RE-ITI), are used to set the relative percentage of 2030 generation requirements from each type of renewable source.34 With this relationship, one can determine how much of the required, new renewable generation in 2030 should be allocated to wind and to utility PV. To do this, the projected 2030 percentages of relative generation from onshore wind and from utility PV energy are determined by dividing the NREL projected 33 The increase in LCOE for the coal-fired fleet related to the purchase of the credit is offset by the decrease in LCOE for the NGCC and renewable fleet by the sale of the credit. Here we assume that the credit is not taken as profit by the provider. 34 “Renewable Electricity Futures Scenario Viewer.” http://www.nrel.gov. n.d. n.p. Web. 20 August. 2015. <http://www.nrel.gov/analysis/re_futures/data_viewer/#>.



generation from each of these sources by the sum of the NREL projected state generation from geothermal, hydropower, concentrated solar power (CSP), utility PV, and onshore wind.35 This results in 64.8% of the renewable energy in 2030 coming from wind generation. The renewable generation produced by new wind capacity is then determined by multiplying the sum of the generation from existing wind sources and the additional new renewable sources required to meet the CPP requirement minus the existing and planned geothermal generation by this percentage. The remaining new renewable generation, minus the existing and planned geothermal generation, is attributed to new utility PV capacity. The number of new solar and wind farms is calculated from the required capacity for each type of new renewable generation (based upon the required ERCs and mass allowances for compliance, this percentage, and the IPM associated capacity factors36 for New Mexico), and by limiting the capacity of these farms to the largest existing capacity for each technology in the state.37 The LCOE for each new wind and solar farm is calculated according to methods outlined in the proceeding sections.38 Several types of new renewable energy sources are excluded from this study. The relative percentage of contribution of geothermal generation to the 2030 renewable generation mix is limited to the future generation from this source to those plants already constructed or under construction by 2016. Future hydropower and CSP generation sources are also not considered. Hydropower is not considered because the EPA Integrated Planning Model (IPM) v.5.15 cites the potential new capacity for the New Mexico as only 75 MW. 39 CSP is also not considered a source of future generation because no facilities are currently listed as operational, under construction, or under development in New Mexico by the Solar Industries Association.40 Biomass is not considered, as the EPA believes that this renewable source may not be economically viable on a broad scale and the allowance of ERCs for this source will depend upon the EPA approving that the biogenic feedstock for the source is carbon neutral.41 The solar and wind capacity factors for existing facilities are taken as the empirical values from eGRID and the EPA supporting data for the appropriate year. For new solar and wind facilities, this study uses the capacity factors used in IPM v.5.15 for New Mexico and the particular generation source.42 The capacity factor for geothermal

35 In the final CPP regulation, the EPA states that biomass may not be economically viable on a broad scale and that the CO2 emissions may not be carbon neutral for all biogenic sources. Therefore, new generation from biomass is not considered in this study. 36 The IPM capacity factors for New Mexico wind turbines and utility PV are 34% and 26% respectively. 37 According to the EPA supporting documentation for 2012 generation and emissions, the largest wind farm capacity is 200 MW and he largest utility PV capacity is 50 MW. 38 The estimated LCOE for new wind and solar generation is $82.03/MWh and $116.7/MWh, respectively. 39 “EPA’s Power Sector Modeling Platform v.5.15.” http://www.epa.gov. 3 August. 2015. n.p. Web. 27 August. 2015. <http://www.epa.gov/powersectormodeling/psmodel515.html>. 40 “SEIA Major Solar Projects List Update.” http://www.seia.org. n.d. n.p. Web. 20 August. 2015. <http://www.seia.org/research-resources/major-solar-projects-list/seia-major-solar-projects-list-update>. 41 “Federal Plan Requirements for Greenhouse Gas Emissions From Electric Utility Generating Units Constructed on or Before January 8, 2014; Model Trading Rules; Amendments to Framework Regulations; Proposed Rule,” Title 40 Code of Federal Regulations, Parts 60, 62, and 78. 2015 ed. 42 “EPA’s Power Sector Modeling Platform v.5.15.” http://www.epa.gov. 3 August. 2015. n.p. Web. 27 August. 2015. <http://www.epa.gov/powersectormodeling/psmodel515.html>.

http://www.epa.gov/

http://www.seia.org/

http://www.epa.gov/


generation is the monthly average capacity factor for utility geothermal sources, as reported by the EIA.43 Future capacity values for an existing facility are taken as the maximum value for the facility from 2010 generation, 2012 generation or the IPM model.

Solar Energy The capital cost requirement for existing solar capacity is based upon historical data for the annual, capacity-weighted average cost for individual projects built between 2007 and 2013 compiled by Barbose et al. (2014, Figure A-B3).44 When applied to the summertime capacity for individual farms, a scaling factor (Table A-A10), based upon a regression from system installed price data as a function of capacity (Feldman et al., 2014), is applied to the capacity-weighted average capital costs to determine the estimated capital cost. Capital costs for future capacity are based upon an extrapolation of a regression of the Barbose et al. (2014) historical cost data (Table A11) that are also scaled to the future capacity, per Equation [5]:

𝑃𝑟𝑜𝑗𝑒𝑐𝑡𝑒𝑑 𝑐𝑎𝑝𝑖𝑡𝑎𝑙 𝑐𝑜𝑠𝑡= ((4.7994 ∗ (𝑝𝑟𝑜𝑗𝑒𝑐𝑡𝑒𝑑 𝑦𝑒𝑎𝑟 − 𝑏𝑎𝑠𝑒 𝑦𝑒𝑎𝑟)−0.3321)∗ 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦−0.03157) ∗ 1000 ∗ 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦, [5]

where the base year is 2009 and the capacity is the summertime peak capacity (MW). When compared to the range of analyst expectations for utility PV capital costs (Feldman et al., 2014, Figure B-B4), the scaled value in 2015 for a 100 MW-AC capacity farm ($2.02$/W-AC in 2010$) is $1.89/W-DC, (in 2010$) using an invertor load ratio of 1.27. This value is near the elicited $2/W-DC (in 2010$) for 2015, and is similar to the $2.26/W-

AC (in 2010$) value cited in the IPM v5.15 model for a 150 MW plant available in 2016. Uncertainty is expected in this value. The stochastic component for this variable was determined by measuring the year-specific range in installed prices for all capacity projects, as catalogued by Barbose et al. (2014) and using one-sixth of the range to define the positive and negative deviations from the capacity-weighted mean.45 The distribution for these deviations is taken as uniform (Table A-A21), as the capacities associated with the installed prices are not published. The costs for total operation and maintenance charges are derived in a similar fashion used for the capital costs. Bolinger et al. (2015) catalogues the annual generation-weighted O&M costs ($/kWAC-year), based upon historical data from 2010 through 2013 (Figure A-B5). A nonlinear regression (Table A-A12), based upon the year of operation, is used to extrapolate costs beyond 2013 (Equation [6]). The resulting value is not scaled for instances when the facility capacity factor deviates from the expected capacity factor. As the generation-weighted data are based upon an empirical population, actual values 43 Ibid. 44 Study costs utilize the capacity-weighted average costs based upon $/MW-AC. 45 This range is chosen to introduce an amount of variation that is considered suitable, given that the annual capital costs for these projects are not know as a function of capacity. Assuming a standard deviation of one-fourth or one-sixth of the range and then applying this in a normal or other type of distribution function to generate the capacity-weighted capital cost in a stochastic model can produce costs that are clearly not appropriate for the given capacity installation.


for O&M costs fall between $20-$40/kWAC-year (Bolinger et al., 2015). Therefore, the deviation around the mean value for any year is taken as a uniform distribution of one-sixth of this range (Table A-A21).

𝑂&𝑀 𝑐𝑜𝑠𝑡 = 39.7383 ∗ (𝑝𝑟𝑜𝑗𝑒𝑐𝑡𝑒𝑑 𝑦𝑒𝑎𝑟 − 𝑏𝑎𝑠𝑒 𝑦𝑒𝑎𝑟)−0.2815 , [6]

where the base year is 2009.

Wind Energy The capital cost requirement for existing wind capacity is derived from historical data for the annual, capacity-weighted average cost for individual projects built between 1982 and 2013 (Wiser and Bolinger, 2014, Figure A-B6). Capital costs for future capacity are based upon the total overnight cost given in IPM v5.15 for onshore wind at $1.59/W (in 2010$). Though the IPM projection is for online service in 2016, independent projections from NREL (Chapman et al., 2012) and Black & Vetch (2012) indicate that these projections may remain constant for many years.46 While the future capital cost may be constant, uncertainty in the value is still present in existing and future capital cost; Wiser and Bolinger (2014) suggest that this uncertainty may be dependent upon the capacity of the facility. In their study, Wiser and Bolinger (2014) document 136 individual project costs with capacities from less than 5 MW to greater than 200 MW, and categorized these costs into six capacity buckets. The range for the individual project costs diminishes as the capacity of the category increases. To quantify the uncertainty for any project, regardless of the online service year, we can then take one-half of the range for each capacity category to define the deviation around the capacity-weighted mean capital cost for the particular year (Table A-A21).

Wiser and Bolinger (2014) also empirically document the total O&M costs for wind turbines as a function of installed year and years since installation (Figure A-B7). These data suggest that the initial O&M decreases with more recent commercial installation dates, and is not constant with years in service. To accommodate this model in this study, we supplement the matrix created by Wiser and Bolinger (2014) by using the O&M cost for the initial year to normalize that for subsequent years. The normalized O&M for years not defined in their study are determined by repeating the normalized pattern for years that contain such data. The normalized O&M is then held constant for all installation years after the eleventh year of service because data from the wind turbines with more than ten years of service is greater than that for any previous year (Table A-A13). For wind turbines built after 2012, the AEO 2014 value is used for the initial year O&M; the same schedule as that for 2012 installation is used with subsequent years. The range in the empirical data from 2003 to 2013, which is the period during which the existing wind turbines in 2012 were constructed, defines uncertainty in these values. The measured range is approximately $26/MWh (in 2013$); therefore, the deviation around the mean value for any year is taken as a uniform distribution of one-

46 Furthermore, the projected capital costs given by the EIA, Black and Veatch, and shown for the EPA in Chapman et al. are all similar. However, the recently released IPM v5.15 estimates a 2016 decreased capital cost of 1.6 2010$/W.


sixth of this range, per the method used for the solar capital cost uncertainty (Table A-A21).

Geothermal Energy There is one, 4 MW geothermal plant currently in operation in New Mexico, and one, 6 MW plant under construction that is due to be operational in 2016. According to a media report,47 the 4 MW facility cost 43 million dollars to construct;48 therefore, the estimated capital cost is $10,437/kW (in 2010$).49 The capital cost contribution to LCOE for the 6 MW facility is determined in part by scaling the aforementioned levelized capital cost with economies of scale data per Sanyal (2005)50 that have been converted to 2010 dollars. The result is then annualized with the FCF, according to the remaining years for amortization, and levelized with the expected net generation (Equation [7]):

𝐿𝐶𝑂𝐸𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝐶𝑜𝑠𝑡

=10,437 ∗ 𝑒−0.002498∗(𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦−4) ∗ 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦 ∗ �1 − 𝑎𝑔𝑒

𝑔𝑒𝑜𝑡ℎ𝑒𝑟𝑚𝑎𝑙 𝑙𝑖𝑓𝑒 � ∗ 𝐹𝐶𝐹

𝑛𝑒𝑡 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑖𝑜𝑛,

[7]

where capacity is the built capacity (MW), age is the current age (years) of the facility from the year on-line, geothermal life is the expected booklife (years) for a geothermal plant as specified by Tidball et al. (2010), FCF is fixed charge factor for non-wind turbine sources (11.3%), and net generation is the expected net generation taken as the product of the source capacity factor and the facility capacity.

The O&M expenditure for this facility is not published; therefore, the cost is determined by scaling the O&M cost estimate cited by Sanyal (2005) with the plant capacity, in a similar manner as used for the capital cost (Table A-A15). Doing so yields an O&M cost of $22.4/MWh for the 4 MW facility:51

𝐿𝐶𝑂𝐸𝑂&𝑀 𝐶𝑜𝑠𝑡 = 22.36 ∗ 𝑒−0.002514∗𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦, [8]

where capacity is the summertime peak capacity (MW). Uncertainties for these costs are not evaluated because the total generation for the two plants represents only 0.25% of the expected compliant generation.

47 Danko, Pete. ”New Mexico Joins The Geothermal Power Ranks." The New York Times, 16 Jan., 2014. Web. 27 Aug. 2015. 48 This study assumes that this cost is in 2013 dollars. 49 Black and Veatch (2012) estimate the capital cost at $10,448/kW. 50 A nonlinear regression (Table A14) is fitted to the economies of scale data to derive the scaling factor. 51 Black and Veatch (2012) estimate the O&M cost at $32.7/MWh.


Retrofitting CCS In this study, a coal-fired EGU is retrofitted with a post-combustion CCS subsystem, with which the flue gas can bypass the subsystem to enable capture of any percent of CO2 emissions required to meet the reduction goal (Rao and Rubin, 2006).52 This system requires a large amount of low-quality steam to regenerate the advanced monoethanolamine-based (MEA) sorbent after it strips CO2 from the flue gas, and the equipment that maintains the process operation consumes large amounts of electricity. Decreasing the net generation of the PC EGU can provide the required additional steam and power to meet the targeted CO2 capture rate. Alternatively, the capacity factor of the EGU can be increased, up to the limits of the nameplate capacity or the available hours, to provide the original net generation and still meet the target capture rate. In lieu of using steam and electricity generated by the EGU, an auxiliary gas-fired power generation system can be added to provide the steam and power requirements. Two of these scenarios are examined in this study. In the first scenario, the net generation for the EGU is held constant, as the re-dispatch process lowers the summertime peak capacity sufficiently to allow for required extra capacity to generate the necessary CCS steam and power without exceeding the overall boiler capacity. In the second scenario, the auxiliary natural gas boiler is employed, and it is sized so that no additional net generation is produced for sale on the grid. Details for these scenarios are shown in Table 2.

52 Land and additional water are also required, each of which are approximately equal to that used in the existing system. Currently, this study assumes that there is enough land and water at the site to accommodate these increased needs.


Table 2 CCS specific parameters for the IECM.

Variable Increased Capacity

Type Post-combustion

Capture method Fluor, FG+ amine

Retrofit factor for CCS components 1.253

Capture efficiency (%) 90

Flue bypass control Bypass

Power and Steam Source Increased Capacity/Auxiliary gas-fired boiler

Thermal efficiency of auxiliary gas power system (%)

NA/35

SOx polisher Yes

CO2 purity (%) 99.5

CO2 transportation method Pipeline

CO2 storage method Geological

This mitigation option will incur capital and O&M costs for the CCS plant and auxiliary boiler, the required retrofits to existing subsystems, the transportation system for the CO2 to the saline sequester site, and the CO2 storage. The sequestration expenditures are dependent upon the EGU and storage site-specific parameters relating to the transportation and storage of the captured CO2 product. The NETL storage reservoir database54 lists CO2 sequester sites within the U.S., from which this study uses the five nearest sequestration sites55 to the San Juan coal power plant to determine the lowest overall cost from the combination of transportation and storage costs from the EGU to the site.56 To calculate the transportation cost, we first find the distance from the EGU to the center of the sequestration site.57 The additional capital and O&M for transporting and storing the CO2 at different flow rates resulting from the various capture rates are determined with the CO2 pipe diameter and compressor station equations described in the IECM support documentation,58 the pipeline cost equations used in the NG pipeline

53 NETL, 2013. 54 “NATCARB/ATLAS” www.netl.doe.gov. n.d. n.p. Web. 1 December. 2014. <http://www.netl.doe.gov/research/coal/carbon-storage/natcarb-atlas>. 55 All of these sites have adequate volume to store 100 years worth of annual CO2 production at 90% capture. 56 The storage costs for national sites range from $1.90 to $24.0 per ton of CO2. 57 This may over estimate the pipeline costs in some instances. 58 “Development And Application of Optimal Design Capability For Coal Gasification Systems.” www.cmu.edu. n.d. n.p. Web. 20 October. 2014. <http://www.cmu.edu/epp/iecm/IECM_Publications/2008ra%20McCoy%20et%20al,%20IECM%20Trans%20&%20Storage%20Tech.pdf>.

http://www.netl.doe.gov/research/coal/carbon-storage/natcarb-atlas

http://www.cmu.edu/

http://www.cmu.edu/epp/iecm/IECM_Publications/2008ra%20McCoy%20et%20al,%20IECM%20Trans%20&%20Storage%20Tech.pdf



calculations (as described later), and the site specific storage costs. As the transportation and storage costs vary between sites, a site with a low transportation cost but high storage cost may be economical for low capture rates and costly at high storage rate. Therefore, we calculate the transport and storage cost at specific capture rates and use the site with the lowest overall cost at the desired capture rate to evaluate this mitigation technology. Further capital and O&M cost will be incurred that are related to bringing NG to the site to power the auxiliary boiler. The pipeline capital costs will be a function of the pipe diameter, which is determined in terms of the required flow rate for the given percent capture, and the distance from the current site to the nearest site with NG or to the nearest pipeline large enough to deliver the required flow. This study uses the IECM calculation for the required EGU natural gas consumption (tons/hour) as an input to the Panhandle A equation that calculates the pipe diameter (See Equation A-D1 and Table A-D). Pipeline cost is based upon regressions on material, labor, miscellaneous costs, and right of way, as a function of pipe diameter and pipeline length derived by Parker (2004).59 Additional capital cost may be required for electrically powered compressor stations to boost the line pressure that are added at 50-mile intervals, if fitted.60 The O&M cost for the NG pipeline is taken as $4,965 per mile, which is the IECM defaults for the CCS pipeline.61

Techno-economic Analysis

The IECM Parameters The plant configuration and financial parameters are required to calculate the LCOE, emissions, and makeup water use in the IECM. In addition to the emission control configurations shown in Table 1, Table 3 shows the solids management configuration for the coal-fired EGUs in the study. The age of the emission control devices, and the other subsystems, will affect the LCOE due to the amortization of the capital costs. For this study, we assume that all subsystems have a 30-year book life; the percent of amortization for each year of interest (2010, 2012, and 2015) is then determined uniquely at the EGU subsystem-level as the difference of the year of interest and the installation year divided by the booklife. Other financial parameters for constructing these subsystems are in Table 4. As all of the coal-fired EGUs already exist in 2010, we assume that these parameters represent the actual conditions at the time of each subsystem construction. Hence, we model historical performance of the coal-fired EGUs and the NGCC plants in 2010 and 2012 and future performance in 2030 from the data provided in the 2010 databases, with the previously outlined adjustments. To create the performance models, the coal-fired EGU or NGCC plant is first configured in the IECM as it is described in the 2010 databases. The 2010 site-average annual 59 The IECM uses a similar model for calculating the capital cost of CCS pipeline installations, but differentiates costs by geographic region. 60 “Technical Documentation: The Economics of CO2 Transport by Pipeline and Storage in Saline Aquifers and Oil Reservoirs.” www.cmu.edu. n.d. n.p. Web. 20 October. 2014. <http://www.cmu.edu/epp/iecm/IECM_Publications/2008ra%20McCoy%20et%20al,%20IECM%20Trans%20&%20Storage%20Tech.pdf >. 61 Ibid.

http://www.cmu.edu/


pressure, temperature, and humidity are entered in the IECM overall plant parameters, and the simulation capacity factor is adjusted so that the simulation operating hours are the same as those in the 2010 databases. For a coal-fired EGU, the parasitic loads for various base plant components and the steam cycle heat rate are then adjusted to achieve the overall parasitic load and net heat rate for the unit, as reported in the databases. The simulated gross capacity for the coal-fired EGU is then adjusted so the resulting net generation is the same as the reported value. The net capacity that is then lower than the documented summertime peak capacity. This difference is important because it provides a margin with which the capacity of the coal-fired EGU can be increased for future peak and baseload demand. With the fuel price adjusted according to the 2010 receipts, this model serves as the baseline model for the unique coal-fired EGU. To determine the 2012 performance parameters and costs, the atmospheric conditions, annual operating hours, gross capacity, receipts, and amortization for the subsystems are adjusted to those appropriate for 2012 as the reference year. Future performance and costs are determined with the appropriate aforementioned adjustments and the inclusion of any required, additional emission control subsystems for future generation. With this model, the CPP building blocks to improve heat rate and reduce capacity factor are simulated.

Table 3 Site-specific coal-fired EGU solids management configuration.

Device Type Parameter IECM Default Setting

Solids Management Wastewater Ash Pond

Fly ash Disposal No Mixing

Table 4 Financial parameters for the IECM.

Variable Value

Year costs reported 2010

Costs basis (dollar type) Constant

Discount rate (%) 7.09

Fixed charge factor (fraction) 0.113

Book life (years) 30

Construction costs Overnight

To simulate a NGCC plant in the IECM, several modifications must be made to the 2010 simulation procedure previously described for the coal-fired EGU. The GE-7FA gas turbine model, with the number of gas turbines described in eGRID, is selected for the power block and the documented net generation is achieved by further varying the hours of operation. The appropriate cooling system and related performance parameters must also be chosen to simulate the makeup water use in the plant (Table A-A1). With the


configuration for the NGCC plant determined and the financial parameters in Table 4, the historical and future performance and costs are simulated in the same fashion as used for a coal-fired EGU. The simulation for future NGCC plant performance and costs under the CPP building blocks only necessitates increasing the operating hours for the plant to produce a net generation equivalent to that from maintaining a 75% capacity factor. The Luna and Bluffview NGCC plants use recirculating water towers; therefore, the wet cooling tower defaults are selected in the IECM and the makeup water use is adjusted for the heat rate, as previously noted. To simulate the parasitic load for the dry cooling system at the Hobbs NGCC plant, the default settings for the air-cooled condenser system are used in the IECM in conjunction with the site-specific, mean atmospheric conditions. The capacity of the cooling system is sized for peak temperature conditions rather than for the mean conditions, however. Therefore, the capacity of the system and the associated capital cost used in this study is based upon the average of the three hottest months for each year from 2008 through 2012 (Table A-A16). This modification necessitates a correction in the LCOE calculation for the cooling system, for which an addition is made to the portion of the LCOE that is the annualized and levelized difference between the capital cost of the dry-cooling system as modeled for the mean atmospheric conditions and that for the peak temperature conditions. Currently, the hybrid cooling system at the Afton NGCC plant cannot be directly modeled in the IECM version 8.02. The capital and O&M costs for this system are taken as those for a dry-cool system, as modeled for the Hobbs NGCC plant. The makeup water use is determined per Webster et al. (2013) as the product of the rate of fuel input (millions of British thermal units per hour, MMBtu/hr) and the assumed water withdrawal rate for a new NGCC plant with a hybrid cooling system (19 gal/MMBtu). This result is also adjusted for the current heat rate. The fuel is also a key parameter to determine plant performance, emissions, and costs. The coal-fired EGUs in the study use the correct coal type, but the properties of the coal are not adjusted to be site-specific. Rather, Illinois #6 is used as a proxy for the bituminous coal and Wyoming Powder River Basin is used as a proxy for the sub-bituminous coal. The properties for these proxy coals are the IECM defaults (Table A-A17). The other fuel type is natural gas, for which we use the IECM default to define the properties (Table A-A18). The historical, site-specific prices for these commodities that are used in the IECM are taken from form 923 (Tables A-A19 and A-A20). Adjusting these values according to the EIA 2030 predictions derives the 2030 prices (EIA, 2015). Here, the EIA expects the 2030 coal price to be 3.3% lower than the 2012 prices, while the natural gas price will increase by 81.3% relative to the 2012 value.

Mitigation Technologies The EPA outlines two approaches with which a state can meet CO2 emission compliance—rate-based and mass-based. While this study examines multiple scenarios for each approach (Table 5), it does not examine all possible options. In particular, it does not examine the mass-based approach with new source generation and applying all possible CO2 mitigation measures (such as co-firing with biomass or natural gas, upgrading coal type, and upgrading steam generator) to all fossil fuel plants. Scenarios


explored for the rate-based approach for the state emission rate target include those in which the net generation for the fossil fuel plants is maintained at 2012 levels and sufficient renewable resource generation is added to the fleet to ensure compliance (scenario 1). This is inclusive of adding 40% capture CCS, with and without an auxiliary gas boiler, to San Juan unit 4 (scenarios 2 and 3, respectively). Alternative scenarios for this approach include maintaining the net fleet generation at 2012 levels by reducing only the coal-fired generation from 2012 levels and adding the minimum amount of renewable energy to be complaint (scenario 4); and using the same procedure as in scenario 4 to maintain the net fleet generation at 2012 levels, but allowing for the retirement of one coal-fired plant (scenario 5). The rate-based approach using the technology standard examines scenarios 1, 4, and 5, while the mass-based approach examines scenarios 1 through 3. For each of these scenarios, the first two steps involve improving the heat rate for the coal-fired EGUs and shifting generation for the coal-fired plants to the NGCC plants. Subsequent steps entail determining the required renewable energy generation to obtain necessary ERCs or mass allowances to be compliant and to meet generation requirements.


Table 5 Description of scenarios analyzed in study. All scenarios apply to the state option for the rate-based

approach. Scenarios 1, 2, and 3 apply to the mass-based approach. Scenarios 4 and 5 apply to the technology option for the rate-based approach.

Compliance Type

Scenario Description for 2030 Net Generation

Mass-based M1 Fossil fuel EGU generation is maintained at 2012 levels. M2 Fossil fuel EGU generation is maintained at 2012 levels and

40% CCS with an auxiliary natural gas boiler is applied to San Juan #4 EGU.

M3 Fossil fuel EGU generation is maintained at 2012 levels and 40% CCS without an auxiliary natural gas boiler applied to San Juan #4 EGU.

Rate-based S1 Fossil fuel EGU generation is maintained at 2012 levels. S2 Fossil fuel EGU generation is maintained at 2012 levels and

40% CCS with an auxiliary natural gas boiler is applied to San Juan #4 EGU.

S3 State option where fossil fuel EGU generation is maintained at 2012 levels and 40% CCS without an auxiliary natural gas boiler applied to San Juan #4 EGU.

S4 State option where total net generation is maintained at 2012 total net generation level, by reducing capacity factors for coal-fired EGUs.

S5 State option where total net generation is maintained at 2012 total net generation level, by retiring San Juan #1 EGU.

T1 Technology option where fossil fuel EGU generation is maintained at 2012 levels.

T4 Technology option where total net generation is maintained at 2012 total net generation level, by reducing capacity factors for coal-fired EGUs.

T5 Technology option where total net generation is maintained at 2012 total net generation level, by retiring San Juan #1 EGU.

Improving Plant Heat Rate The technical support documents for the Clean Power Plan62 outline the efforts to establish the best systems of emission reduction (BSER) to reduce CO2 emissions from coal-fired EGUs. As part of this effort, the EPA describes the results of an eleven-year study that uses unit-level, historical, hourly data on heat input and gross generation to assess the possible heat rate improvement for the fleet-wide coal-fired EGUs. The database for this study is 884 subcritical and supercritical coal- and petroleum coke-fired

62 “Clean Power Plan Proposed Rule: GHG Abatement Measures.” http://www2.epa.gov. 13 August. 2015. 3 August. 2015. Web. 27 August. 2015. < http://www2.epa.gov/cleanpowerplan/clean-power-plan-final-rule-technical-documents>.





EGUs of various ages that use different coal ranks, are of multiple configurations, and have capacities greater than or equal to 25 MW. With this database, three analytical methodologies are examined to determine a reasonable means to estimate the potential heat rate improvements for each interconnection region. The method chosen as the most conservative estimate of the potential improvement is based upon the percent difference between the best gross heat rate from a two-year rolling average between 2002 and 2012, and the 2012 value. This is done while taking into account the hourly capacity factor and ambient temperature. The result of this analysis is that the heat rate for the coal-fired EGUs in New Mexico can be improved by 2.1%. This reduction is directly applied to those EGUs to obtain a corresponding 2.1% reduction in CO2 emission intensity rate. The overall reduction in emissions rate may not be 2.1%; however, because of the addition of the parasitic load from new emission control devices. In the CPP proposal, the EPA cited a Sargent & Lundy (2009) report that presents several best practices and mechanical upgrades that may be useful in realizing the heat rate improvement. The improvement cost of $100/kW-net from that report and the EPA proposal are used in this study to cost the 2.1% heat rate improvement. The impact of the heat rate improvement on the CO2 emission intensity is modeled in the IECM through a 2.1% improvement in the steam cycle heat rate for the boiler, prior to the addition of any new emission control devices. As such, when new emission controls are later added, the effect of the parasitic load is accounted for in the IECM model. The cost associated with the heat rate improvement is added external to the IECM model, as an additional annualized and levelized LCOE component for the specific EGU.

Generation Shifting Building block 2 of the CPP outlines increasing the capacity factor of the NGCC plants to 75% and reducing the coal-fired generation by the difference between the generation at 75% capacity and that in 2012. For the New Mexico plants, the coal-fired EGUs generation-weighted average capacity factor was 73.6%, and that for the NGCC plants was 53.9%. Increasing the NGCC capacity factor to 75% produces an extra 3.37 TWh to be offset by a reduction in the capacity factor of the coal-fired EGUs. Two of these EGUs63 are retiring in 2017, however. Therefore, the average capacity factor for the remaining coal-fired EGUs can increase to 83.9% to obtain an overall reduction in coal-fired generation of 3.37 TWh.

Emission Reduction Credits The generation needed from new renewable sources for the rate-based state and technology option scenarios in Table 5 is the summation of the credits produced or the credits required from each fossil fuel plant, as calculated with Equation [2]. In scenarios 1 through 3 (Table 6), the resulting number of credits or required credit purchases to maintain the fossil fuel generation at 2012 levels with the state option goal for the OGST

63 The San Juan units 2 and 3 produced 4.73 TWh of electricity in 2012.


EGUs and for each coal-fired EGU and NGCC plant only varies with the implementation of CCS on San Juan EGU #4.64 Whereas the NGCC plants in the state option have a surplus of ERCs in the three scenarios, all power generation sources require ERCs with the technology option (Table 7). Scenarios 4 and 5 require that the net fleet generation be maintained at 2012 levels. Here, the NGCC fossil fuel generation, which has lower CO2 emission rates, is held constant at the original generation shift levels, and some of the coal-fired generation is replaced with new renewable generation to minimize the overall required new source generation. In scenario 4 to achieve the necessary emission intensity reduction, the capacity factors for the coal-fired EGUs are decreased uniformly to approximately 50% for the state option and to approximately 41% for the technology option. While this may not be the overall optimal reduction for the individual EGUs, it does achieve the 2012 net generation in a manner that also yields compliance (Tables 8 and 9, respectively). The resulting required new source generation is 3.2 TWh and 4.1 TWh, respectively.

Operating the three coal-fired EGUs will increase the PC LCOE because of the lower capacity factor, however. Therefore, a more cost effective measure is to retire the oldest and more expensive EGU (San Juan 1) to increase the overall capacity factors and decrease the LCOE (scenario 5). Following the same procedures as in scenario 4, the required new renewable generation for compliance at 2012 generations levels is similar to that without retirement: New renewable generation for the state goal is 3.3 TWh (Table 10), while 4.1 TWh is required for the technology goal (Table 11).

64 Any credits that are available from the OGST EGUs are ignored in the credit calculation because the LCOE for the individual EGUs is not known and these EGUs are not used in the resulting $/ton of CO2 avoided calculations. The required credits for the lumped OGST intensity are 215,789. The emission intensity for these units is taken into account for meeting the state emission rate goal.


Table 6 State option required and excess rate-based ERCs for coal-fired EGUs and NGCC plants for 2030

generation in scenarios 1 through 3, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state emission rate goal is 1,146 lbs/MWh.

Plant name Unit number

Emission intensity

(lbs/MWh)

Net Generation (MWh)

Emission Reduction

Credits (requirements)

OGST NA 1,313 1,476,861 (215,789) San Juan 1 2,299 2,366,901 (2,381,789) San Juan 4 2,314 3,719,416 (3,790,812)

San Juan with CCS auxiliary gas boiler

(40% capture)

4 1,642 3,719,416 (1,611,418)

San Juan with CCS and no auxiliary gas boiler (40%

capture)

4 1,721 3,719,416 (1,865,860)

Escalante NA 2,406 1,892,950 (2,081,229) Afton NGCC NA 973 1,552,000 234,709

Bluffview NGCC NA 937 427,300 77,800 Hobbs NGCC NA 897 3,458,000 750,200

Luna NGCC NA 905 3,669,000 770,193

Table 7 Technology option required and excess rate-based ERCs for coal-fired EGUs and NGCC plants for 2030

generation in scenario 1, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The technology option emission rate goals are 1,305 lbs/MWh for steam generation and 771

lbs/MWh for NGCC generation. The GS-Shift ERCs are calculated with Equations [3] and [4]. Plant name Unit

number Emission intensity

(lbs/MWh)

Net Generation

(MWh)

GS-Shift Emission Reduction

Credits

Emission Reduction


OGST NA 1,313 1,476,861 NA (9,558) San Juan 1 2,299 2,366,901 NA (1,803,351) San Juan 4 2,314 3,719,416 NA (2,875,773) Escalante NA 2,406 1,892,950 NA (1,597,018)

Afton NGCC NA 973 1,552,000 102,757 (405,996) Bluffview

NGCC NA 937 427,300 31,300 (92,190)

Hobbs NGCC

NA 897 3,458,000 280,830 (556,823)

Luna NGCC NA 905 3,669,000 292,079 (639,773)


Table 8 State option required and excess rate-based ERCs for coal-fired EGUs and NGCC plants for 2030 generation in scenario 4, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state emission rate goal is 1,146 lbs/MWh. Fleet net generation is maintained at 2012 levels.


Emission intensity

(lbs/MWh)


Emission Reduction


OGST NA 1,313 1,476,861 (215,789) San Juan 1 2,299 1,399,352 (1,408,246) San Juan 4 2,314 2,198,981 (2,241,192) Escalante NA 2,406 1,119,144 (1,230,457)

Afton NGCC NA 973 1,552,000 234,709 Bluffview NGCC NA 937 427,300 77,800

Hobbs NGCC NA 897 3,458,000 750,200 Luna NGCC NA 905 3,669,000 770,193

Table 9 Technology option required and excess rate-based ERCs for coal-fired EGUs and NGCC plants for 2030

generation in scenario 4, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The technology option emission rate goals are 1,305 lbs/MWh for steam generation and 771

lbs/MWh for NGCC generation. Fleet net generation is maintained at 2012 levels. The GS-Shift ERCs are calculated with Equations [3] and [4].


Emission intensity

(lbs/MWh)

Net Generation

(MWh)


Credits

Emission Reduction


OGST NA 1,313 1,476,861 NA (9,558) San Juan 1 2,299 1,157,751 NA (882,095) San Juan 4 2,314 1,819,323 NA (1,406,662) Escalante NA 2,406 925,922 NA (781,169)


NGCC NA 937 427,300 31,300 (92,190)

Hobbs NGCC NA 897 3,458,000 280,830 (556,823) Luna NGCC NA 905 3,669,000 292,079 (639,773)


Table 10 State option required and excess rate-based ERCs for coal-fired EGUs and NGCC plants for 2030

generation in scenario 5, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state emission rate goal is 1,146 lbs/MWh Fleet net generation is maintained at 2012 levels.


Emission intensity

(lbs/MWh)


Emission Reduction


OGST NA 1,313 1,476,861 (215,789) San Juan 1 NA 0 0 San Juan 4 2,314 3,108,269 (3,167,934) Escalante NA 2,406 1,581,915 (1,739,257)

Afton NGCC NA 973 1,552,000 234,709 Bluffview NGCC NA 937 427,300 77,800

Hobbs NGCC NA 897 3,458,000 750,200

Table 11 Technology option required and excess rate-based ERCs for coal-fired EGUs and NGCC plants for 2030 generation in scenario 5, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The technology option emission rate goals are 1,305 lbs/MWh for steam generation and 771

lbs/MWh for NGCC generation. Fleet net generation is maintained at 2012 levels. The GS-Shift ERCs are calculated with Equations [3] and [4].


Emission intensity

(lbs/MWh)

Net Generation

(MWh)


Credits

Emission Reduction


OGST NA 1,313 1,476,861 NA (9,558) San Juan 1 NA 0 NA 0 San Juan 4 2,314 2,571,620 NA (1,988,322) Escalante NA 2,406 1,308,794 NA (1,104,185)


NGCC NA 937 427,300 31,300 (92,190)

Hobbs NGCC NA 897 3,458,000 280,830 (556,823) Luna NGCC NA 905 3,669,000 292,079 (639,773)

Mass Allowances The mass-based approach for the CPP allows New Mexico to emit 12,412,602 short tons of CO2 per year; five percent of these allowances are set-aside for new renewable resource capacity, and 627,085 short tons are set-aside for increased generation from existing NGCC plants. The EPA does not stipulate how the remaining allowance should be allocated, but the agency does suggest that the allowances can be allocated to the individual plants according to the associated average percent contribution to the overall fossil fuel generation between 2010 through 2012 (40 CFR parts 60, 62, and 78). As this study only utilizes generation in 2012, these remaining allowances are allocated


proportionately to the EGUs and NGCC plants that are scheduled to functioning in 2030, according to 2012 generation. The allowances that would have been given to the retiring San Juan EGUs are proportionately allocated to the three remaining coal-fired EGUs.65 The resulting allowance distribution is shown in Table 12. The allowance distribution enables one to calculate the allowable generation from any fossil fuel source as the difference between the total allowances allocated and the product of the proposed net generation66 and emission intensity rate for the source (Equation [9]). If this difference is negative, than the source must buy surplus allowances from another fossil source or from the set-aside for a new renewable source; otherwise, the source must curtail generation to comply. 𝐴𝑣𝑎𝑖𝑙𝑎𝑏𝑙𝑒 𝐴𝑙𝑙𝑜𝑤𝑎𝑛𝑐𝑒𝑖

= 𝑡𝑜𝑡𝑎𝑙 𝑎𝑙𝑙𝑜𝑤𝑎𝑛𝑐𝑒𝑖 −𝑛𝑒𝑡 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑖𝑜𝑛𝑖 ∗ 𝑒𝑚𝑖𝑠𝑠𝑖𝑜𝑛 𝑖𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦 𝑖

2,000 , [9]

where the subscript i is the unique generation source. Summing these available allowances between sources to achieve zero net allowances and adding enough new renewable resources to maintain 2012 net generation levels, to limit and overbuilding capacity, result in a different distribution and corresponding net generation for each scenario. This is because the emission intensity for San Juan #4 varies according to the steam and power source used to operate the retrofitted CCS. The related decrease in intensity can have several effects on the fleet: (1) The allowable generation from San Juan #4 can increase and result in a corresponding decrease in generation from the other PC EGUs; (2) The lower overall intensity for the PC fleet can result in an increase in capacity factors for all PC EGUs; (3) The required new renewable generation can decrease. This study only considers that the addition of CCS mitigation permits the capacity factor for all PC EGUs to increase with a resulting decrease in required new renewable resource (Tables 13-15).67 65 The allowances for San Juan 2 and 3 are not reserved for the remaining San Juan EGUs because units 2 and 3 retire thirteen years before the study date. As such, the San Juan plant may lose the right to these allowances. These allowances could also be given to the new renewable sources as additional incentives. 66 The generation for OGST is held constant at 2012 levels. Generation from the NGCC plants is limited to the 75% capacity factor target. Generation from the coal-fired EGUs is limited by the total available mass allowances and the employed CO2 mitigation options. 67 When the original coal-fired boiler supplies steam and electricity for the CSS, the summertime peak capacity is increased to the 540 MW and the capacity factor is increased to 85%.


Table 12 Distribution of mass-based allowances for fossil fuel based upon 2012 generation. Plant name Unit

number 2012 Net

Generation (MWh)

Mass Allowances (short tons)

Set-Asides (short tons)

OGST NA 1,476,861 969,889 0 San Juan 1 2,072,236 2,135,831 0 San Juan 4 3,431,787 3,537,106 0 Escalante NA 1,122,027 1,156,461 0

Afton NGCC NA 468,702 333,209 51,286 Bluffview NGCC NA 463,606 329,585 50,728

Hobbs NGCC NA 2,987,812 2,124,091 326,928 Luna NGCC NA 1,810,838 1,287,359 198,143

New Renewable capacity

NA 0 0 620,630

Table 13 Scenario 1 net generation and required/excess mass-based allowances for coal-fired EGUs and NGCC plants for 2030 generation, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. Net generation is maintained at 2012 levels. The mass-based goal is 12,412,602 short tons of CO2.


Emission intensity

(lbs/MWh)



OGST NA 1,313 1,476,861 (81,560) San Juan 1 2,299 1,851,212 (278,720) San Juan 4 2,314 2,909,048 171,340 Escalante NA 2,406 1,480,524 (624,598)

Afton NGCC NA 973 1,552,000 (421,599) Bluffview NGCC NA 937 427,300 129,322

Hobbs NGCC NA 897 3,458,000 572,522 Luna NGCC NA 905 3,669,000 (373,658)

New Renewable NA 0 2,457,076 620,630 Table 14 Scenario 2 net generation and required/excess mass-based allowances for coal-fired EGUs and NGCC plants for 2030 generation, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. Fleet net generation is maintained at 2012 levels. NGCC plants remain unchanged from scenario 1

(Table 13) and are omitted. The mass-based goal is 12,412,602 short tons of CO2. Plant name Unit


(lbs/MWh)



OGST NA 1,313 1,476,861 (81,560) San Juan 1 2,299 2,138,288 (322,435)

San Juan with CCS auxiliary gas boiler

(40% capture)

4 1,384 3,360,167 777,571

Escalante NA 2,406 1,710,115 (900,794) New Renewable NA 0 1,489,290 620,630


Table 15 Scenario 3 net generation and required/excess mass-based allowances for coal-fired EGUs and NGCC plants for 2030 generation, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. Fleet net generation is maintained at 2012 levels. NGCC plants remain unchanged from scenario 1

(Table 13) and are omitted. The mass-based goal is 12,412,602 short tons of CO2. Plant name Unit


(lbs/MWh)



OGST NA 1,313 1,476,861 (81,560) San Juan 1 2,299 2,100,263 (278,720)

San Juan with CCS w/o aux. gas boiler

(40% capture)

4 1,721 3,300,414 697,272

Escalante NA 2,406 1,679,705 (864,210) New Renewable NA 0 1,617,479 620,630

CCS Capture, Transportation and Storage Cost Analysis from the nearby sequestration areas indicates that the Hermosa 1a site is the closest and least expensive site to the San Juan unit 4 EGU. The center of this storage field is forty-two miles from the EGU and has an associated storage cost of $6.38/ton of CO2.68 To determine the transportation cost when capturing 40% of the coal related CO2, the additional parameters are used in the IECM, CCS retrofit scenario (Table 16). This simulation yields a transportation cost to the sequestration site of $0.7/MWh.

Table 16 CCS system and pipeline transport parameters for the IECM. Variable Value

Percent capture (%) 40 Pipeline region Southwest

Pipeline length (miles) 42 Net change in elevation (feet) 0

Number booster stations 0 Fixed O&M cost ($/mile-yr) 4,989

Natural Gas Pipeline Cost The addition LCOE for adding CCS with NG auxiliary boilers to the EGU depends upon the availability of NG at the EGU. For the San Juan unit 4 EGU, the nearest NG source is within approximately 1 mile. The cost to bring NG to the individual EGUs is estimated in the study, and is considered an expense added to the cost of the LCOE for the book life of the required pipeline. The construction cost is estimated with national average costs for the right of way, materials, labor, and miscellaneous charges as a function of the length and diameter of the pipeline (Parker, 2004). The pipe diameter is

68 “NATCARB/ATLAS” www.netl.doe.gov. n.d. n.p. Web. 1 December. 2014. <http://www.netl.doe.gov/research/coal/carbon-storage/natcarb-atlas>.



determined according to the Panhandle A equation (Equation A-E1), which is suitable for approximation in turbulent flow for Reynolds numbers from 5 to 11 million.69 70 The constants given in Table A-D for this calculation are based upon the natural gas properties and plausible natural gas distribution pressure ranges;71 however, the pipe properties and the temperatures are estimations of feasible values. The capital cost portion of the LCOE for the required pipeline then becomes the levelized product of the capital cost for these additions and the FCF that is levelized by the net generation for the EGU:

𝐿𝐶𝑂𝐸𝑐𝑎𝑝𝑖𝑡𝑎𝑙 𝑐𝑜𝑠𝑡 𝑁𝐺 𝑝𝑖𝑝𝑒𝑙𝑖𝑛𝑒 = 𝐹𝐶𝐹∗𝑐𝑎𝑝𝑖𝑡𝑎𝑙 𝑐𝑜𝑠𝑡𝑝𝑖𝑝𝑒 𝑛𝑒𝑡 𝑔𝑒𝑛𝑒𝑟𝑎𝑡𝑖𝑜𝑛

. [10]

State-Level Mitigation Analysis

Historical Plant-Level Results Applying the information in the aforementioned databases as inputs to the IECM model allows us to simulate the past and future performance of the coal-fired and NGCC plants. With the inclusion of the cost model for the renewable resource generation, one can construct the unit-level profile for net generation, LCOE and water use for any period (Table 17 and Table 18). Doing so for 2010 indicates that the modeled generators produced 20.7 TWh of electricity at a fleet-wide average cost of $52.8/MWh72 and required 11.99 billion gallons of water (Ggal) (Table 19). The net generation for 2012 increased to 21.1 TWh, requiring an increase in makeup water to 11.85 billion gallons, while the LCOE decreased to $50.5/MWh. (Table 19). For these years the generation weighted-average emission intensity decreased slightly from 1,795 lbs/MWh to 1,710 lbs/MWh, due in part to the increase in renewable generation. The similarity in the aggregate of the performance metrics for these two years illustrates the overall stability of the fleet operation and the sensitivity of LCOE to changes in natural gas price (Table A20). The sensitivity of the cost of NGCC generated electricity to the natural gas price is particularly evident in the VOM cost. In 2010, the cost of natural gas resulted in VOM 69 L. Carroll and R. Weston Hudkins, Advanced Pipeline Design, [Online]. Available: http://www.ou.edu/class/che-design/a-design/projects-2009/Pipeline%20Design.pdf. 70 The Panhandle B equation (Equation D2) can also be used to calculate the pipe diameter for Reynolds numbers from 4 to 40 million (Carroll and Hudkins). In a study comparing various simplified NG pipeline equations to a simulation for a pipe network, Carroll and Hudkins find that the Panhandle A and B equations produce a 3.5% to 10% error in pressure drop from the simplified theoretical optimal pipe diameter relative to the computationally accurate optimal pipe diameter. 71 Pressure ranges given for interstate transmission and within state distribution range from 200 to 1,500 psi. “The Transportation of Natural Gas.” http://www.naturalgas.org. 20 September. 2013. n.p. Web. 20 December. 2014. <http://naturalgas.org/naturalgas/transport/>. 72 The generation weighted-average LCOE always excludes the OGST generation for which we do not estimate the cost.

http://www.ou.edu/class/che-design/a-design/projects-2009/Pipeline%20Design.pdf

http://www.naturalgas.org/

http://naturalgas.org/naturalgas/transport/


costs for these plants to be greater than that for the PC EGUs. In 2012, the lower natural gas price led to VOMs for the NGCC plants that were lower than or roughly equivalent to those for the PC EGUs, however. While the fossil fuel capacity and performance was stable, those for the renewable energy sources were not. As such, the generation from and LCOE for these sources, when brought forward from 2012 to 2030, can vary greatly due to the dependence of the capacity factor on when the source came on-line in 2012. This is evident in 19.5% of the renewable source capacity available in 2012—sources that had capacity factors less than or equal to one-half of the IPM capacity factor.73 The resulting difference between the historical generation from the existing renewable sources in 2012 and the projected generation from these sources in 2030 results in an additional generation of 0.72 TWh. Consequently, calculating metrics that use the LCOE from 2012 may be biased because of the lower capacity factor for some of the new renewable sources and the lower natural gas price.

Table 17 Historical net generation and summertime peak capacity, and IECM simulated LCOE, VOM, and makeup water for New Mexico generation sources in 2010.


Summertime Peak

Capacity (MW)


LCOE ($/MWh)

VOM ($/MWh)

Makeup Water

(Ggal/year)

OGST NA NA 1,476,861 NA NA 1.22 San Juan 1 322 2,339,308 44.8 29.6 2.03 San Juan 2 320 1,936,519 47.9 29.8 1.69 San Juan 3 495 2,585,338 45 29.4 1.57 San Juan 4 506 3,272,251 41.5 28.8 2.74 Escalante NA 247 1,664,059 44.7 25.8 1.47

Afton NGCC

NA 236 293,055 136.2 47.9 0.08

Bluffview NGCC

NA 65 421,331 86.7 38.2 0.16

Hobbs NGCC

NA 526 2,994,894 53.3 36.7 NA

Luna NGCC NA 558 1,907,383 58.7 36.3 1.03 Wind NA 700 1,832,182 72.3 0.0 0.0 Solar NA 30 3,817 1,347 0.0 0.0

73 “EPA’s Power Sector Modeling Platform v.5.15.” http://www.epa.gov. 3 August. 2015. n.p. Web. 27 August. 2015. <http://www.epa.gov/powersectormodeling/psmodel515.html>.

http://www.epa.gov/


Table 18 Historical net generation and IECM simulated LCOE, VOM, and makeup water for New Mexico

generation sources in 2012, excluding hydropower. Plant name Unit

number Net

Generation (MWh)

LCOE ($/MWh)

VOM ($/MWh)

Makeup Water (Ggal/year)

OGST NA 1,476,861 NA NA 1.22 San Juan 1 2,072,236 45.8 30 1.73 San Juan 2 1,969,293 46.6 29.2 1.74 San Juan 3 2,758,644 43.1 28.9 2.39 San Juan 4 3,431,787 40.1 29.2 2.58 Escalante NA 1,122,027 50.3 26.4 1.51

Afton NGCC NA 468,702 82.4 27.7 0.06 Bluffview NGCC NA 463,605 76.2 28.5 0.15

Hobbs NGCC NA 2,987,812 42.9 22.6 0.0 Luna NGCC NA 1,810,838 48 24.8 0.48

Wind NA 2,225,788 65.3 0.0 0.0 Solar NA 333,797 175.3 0.0 0.0

Table 19 Historical net generation and CO2 emission intensity, and IECM simulated LCOE and makeup water

for New Mexico generation sources in 2010 and 2012 by plant type, excluding hydropower. Net

Generation (TWh)

LCOE ($/MWh)

CO2 Emission Intensity

(lbs/MWh)

Annual Makeup Water

(Ggal/year) Plant

Type/Year 2010 2012 2010 2012 2010 2012 2010 2012

PC 11.8 11.35 44.4 44.0 2,334 2,329 9.5 9.9 NGCC 5.62 5.73 61.9 50.4 907 909 1.3 0.7 OGST 1.48 1.48 NA NA 1,313 1,313 1.2 1.2

Renewable 1.83 2.56 78.5 79.6 0.0 0.0 0.0 0.0 Total 20.7 21.1 NA NA NA NA 11.99 11.85 Fleet

Generation Weighted-

Average

NA NA 52.8 50.5 1,795 1,710 NA NA

2030 Plant-Level Results The predicted future performance for the plants and the associated state generation profile is dependent upon the CPP approach and scenario taken. As this study only considers endogenous credits, allowances and net generation, the capacity of the state renewable sources must increase enough to provide any required credits and allowances to enable the state compliance and meet generation needs. The decrease in emission intensity from the heat rate improvement for each coal-fired EGU and the dispatch change associated


with the generation shift between the coal-fired and gas-fired EGUs define the amount of credits or allowances needed for each scenario, as previously outlined. Balancing these credits and allowances with the fossil fuel generation then provides the inputs for the IECM to model the costs and performance of the coal and natural gas plants in 2030. Those for the new renewable sources are derived from the renewable energy cost models, according to the percent of new utility solar and wind capacity projected in the NREL 30% RE-ITI scenario74 and the appropriate IPM capacity factors. The results for the individual generating plants and the required new renewable generation from following this procedure, for the two approaches and the five scenarios, indicate a wide range of outcomes for the net generation, LCOE, CO2 emission intensity and plant makeup water (Tables A-A22 to A-A32). Aggregating these results by scenario and fuel type facilitates result analysis (Tables 15-19).

Scenario Results In scenario 1 (Table 20), the metric results for the rate-based options are similar with an increase in generation from new renewables of 6.5 TWh for the state goal and 7.4 TWh for the technology option goal. This difference in renewable generation accounts for the greater LCOE and lower fleet emission intensity for the technology option. In contrast, the mass-based approach decreases the PC generation by 1.8 TWh, thereby reducing the need for more renewable generation; only 1.4 TWh of new renewable generation is added. Therefore, the overall generation is approximately the same as in 2012, though at a greater cost due to the lower capacity factor for the PC generation and the increased renewable cost. The lower increase in renewables also increases the fleet emission intensity. Scenario 2 examines the use of CCS at 40% capture with an auxiliary natural gas boiler on San Juan #4 EGU for the state option in the rate-based and the mass-based approaches (Table 21). While this scenario does not allow for more net generation from that EGU in the state option, more generation is permitted in the mass-based approach so that all mass allowances from PC can be used; therefore, there is 1 TWh more net generation from PC in this instance relative to that for the M1 scenario. Conversely, the net generation from renewable resources decreases by 2.2 TWh for the state option because the emission intensity for the PC EGUs decreases. The net effect is an overall 2.2 TWh reduction in generation for the state option and no change in generation for the mass-based approach. As the resulting additional LCOE for the CCS plant is lower than that for the required new renewable resources, the LCOE for the state option is almost $1/MWh lower than that in scenario 1 (Tables A-A22 and A-A23). Conversely, the mass-based LCOE is almost $1.5/MWh greater in scenario 2 than in scenario1 because less additional renewable energy is required to compensate for the additional cost of the CCS plant. In scenario 2, there is a corresponding decrease in emission intensity associated with the reduction in fossil fuel intensity for the state option; however, this does not occur for the mass-based approach because of the reduction in renewable generation.

74 “Renewable Electricity Futures Scenario Viewer.” http://www.nrel.gov. n.d. n.p. Web. 20 August. 2015. <http://www.nrel.gov/analysis/re_futures/data_viewer/#>.



In scenario 3, the CCS plant steam and electricity requirements come from the coal-fired EGU instead of an auxiliary power system. As such, the levelized fuel cost for San Juan #4 does not increase as much as when natural gas is used. This then lowers the LCOE for each approach, relative to scenario 2, while having only a small effect on the fleet net generation and emission intensity (Table 22).

Scenarios 4 and 5 maintain the fleet generation at the 2012 level by decreasing the PC generation to meeting the CPP requirements for the rate-based approaches (Tables 23 and 24). In scenario 4, the associated reduced capacity factor for all coal-fired EGUs increases the PC LCOE for the state option by $10/MWh and for the technology option by almost $13/MWh, relative to scenario 1. These increases are substantially offset by the lower requirements for new renewable resources; as such, the overall LCOE relative to scenario 1 increases by $1.5/MWh for the state option and by almost $3/MWh for the technology option. However, the large reduction in generation from coal shifts the generation emission further away from the higher carbon content fuel to further lower the overall emission intensity for each option. Allowing San Juan #1 EGU to retire, scenario 5, maintains these lower fleet emission intensities, while greatly decreasing the PC LCOE, relative to those in scenario 4, through increasing the capacity factors of the remaining coal-fired boilers. The resulting fleet LCOE is approximately equivalent to that for the state option and roughly $1/MWh higher for the technology option, respectively in scenario 1.

Table 20 Scenario 1 net generation and CO2 emission intensity, and IECM simulated LCOE for New Mexico generation sources under the rate and mass-based CPP approaches applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state and technology options are

used for in rate-based approach. The mass-based approach does not include new generation sources. Net Generation

(TWh) LCOE

($/MWh) CO2 Emission Intensity

(lbs/MWh) Plant

Type/Year State

option Tech

option Mass-based

State option

Tech option

Mass-based

State option

Tech option

Mass-based

PC 8 8 6.2 42.3 42.6 46.1 2,331 2,331 2,331 NGCC 9.1 9.1 9.1 60.4 60.4 60.4 915 915 915 OGST 1.48 1.48 1.48 NA NA NA 1,313 1,313 1,313

Renewable 9.9 10.6 5.7 90.0 90.0 87 0.0 0.0 0.0 Total 28.5 29.1 22.6 NA NA NA NA NA NA Fleet


Average

NA NA NA 65.9 66.7 63.4 998 974 1,086


Table 21 Scenario 2 net generation and CO2 emission intensity, and IECM simulated LCOE and makeup water for New Mexico generation sources under the rate and mass-based CPP approaches applied in 2030, after the

generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. In this scenario, carbon capture and storage technology with an auxiliary natural gas boiler is applied to San Juan EGU #4 to capture

40% of coal relate CO2 emissions. The state option is used for the rate-based approach. The mass-based approach does not include new generation sources.

Net Generation (TWh)

LCOE ($/MWh)


(lbs/MWh)

Annual Makeup Water

(Ggal/year) Plant

Type/Year State

option Mass-based

State option

Mass-based

State option

Mass-based

State option

Mass-based

PC 8 7.2 51.8 57.2 2,018 2,018 6.9 6.4 NGCC 9.1 9.1 60.4 60.4 915 915 1.5 1.5 OGST 1.48 1.48 NA NA 1,313 1,313 1.2 1.2

Renewable 7.7 4.8 83.7 85.5 0.0 0.0 0.0 0.0 Total 26.3 22.6 NA NA NA NA 9.7 9.2

Fleet


Average

NA NA 64.9 65.0 985 1086 NA NA

Table 22 Scenario 3 net generation and CO2 emission intensity, and IECM simulated LCOE and makeup water for New Mexico generation sources under the rate and mass-based CPP approaches applied in 2030, after the

generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. In this scenario, carbon capture and storage technology without an auxiliary natural gas boiler is applied to San Juan EGU #4 to

capture 40% of coal relate CO2 emissions. The state option is used for the rate-based approach. The mass-based approach does not include new generation sources.


LCOE ($/MWh)


(lbs/MWh)

Annual Makeup Water

(Ggal/year) Plant

Type/Year State

option Mass-based

State option

Mass-based

State option

Mass-based

State option

Mass-based

PC 8 7.1 50.3 53.3 2,055 2,055 7.1 6.4 NGCC 9.1 9.1 60.4 60.4 915 915 1.5 1.5 OGST 1.48 1.48 NA NA 1,313 1,313 1.2 1.2

Renewable 8 4.9 82.4 85.7 0.0 0.0 0.0 0.0 Total 26.6 22.6 NA NA NA NA 9.82 9.11

Fleet


Average

NA NA 64.2 63.9 986 1086 NA NA

Table 23 Scenario 4 net generation and CO2 emission intensity, and IECM simulated LCOE and makeup water


for New Mexico generation sources under the rate-based CPP approach applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state and technology options

are used for the rate-based approach to simulate state profile, when the coal-fired capacity factor in scenario 1 is reduced to maintain total 2012 net generation levels and achieve CPP compliance.


LCOE ($/MWh)


(lbs/MWh)

Annual Makeup Water

(Ggal/year) Plant

Type/Year State

option Tech

option State

option Tech

option State

option Tech

option State

option Tech

option PC 4.7 3.9 52.4 55.2 2,331 2,331 4.1 3.1

NGCC 9.1 9.1 60.4 60.4 915 915 1.5 1.5 OGST 1.48 1.48 NA NA 1,313 1,313 1.2 1.2


Fleet


Average

NA NA 67.4 69.6 949 856 NA NA

Table 24 Scenario 5 net generation and CO2 emission intensity, and IECM simulated LCOE and makeup water

for New Mexico generation sources under the rate-based CPP approach applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state and technology options

are used for the rate-based approach to simulate state profile, when the San Juan EGU #1 in scenario 1 is retired to maintain total 2012 net generation levels and achieve CPP compliance.


LCOE ($/MWh)


(lbs/MWh)

Annual Makeup Water

(Ggal/year) Plant

Type/Year State

option Tech

option State

option Tech

option State

option Tech

option State

option Tech

option PC 4.7 3.9 43.7 45.7 2,345 2,345 3.8 3

NGCC 9.1 9.1 60.4 60.4 915 915 1.5 1.5 OGST 1.48 1.48 NA NA 1,313 1,313 1.2 1.2


Fleet


Average

NA NA 65.4 67.8 949 856 NA NA


Unit VOM By 2030, there is a predicted 3% decrease in coal prices and an 81% increase in natural gas prices, relative to those for 2012. These changes result in a significant increase in the NGCC plant VOM, such that all plants may have a greater VOM than those for the PC EGUs (Table 25). This result is independent of the CPP approach taken, when the associated credits and allowances have no price. Table 25 Scenario 1 IECM simulated VOM for New Mexico generation sources under the mass-based approach and the state option for the rate-based approach of the CPP applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. ERC and MA prices are set to $0/MWh and

$0/ton of CO2, respectively. Plant name Unit

number M1 VOM ($/MWh)

S1 VOM ($/MWh)

OGST NA NA NA San Juan 1 29.4 29.4 San Juan 2 NA NA San Juan 3 NA NA San Juan 4 28.6 28.6 Escalante NA 26.4 26.4

Afton NGCC NA 39.3 39.3 Bluffview NGCC NA 55.6 55.6

Hobbs NGCC NA 48.3 48.3 Luna NGCC NA 55.6 55.6

Wind NA 0.0 0.0 Solar NA 0.0 0.0

Plant Makeup Water The annual quantity of plant makeup water also varies with most scenarios, approaches and options. In scenario 1, while the rate-based options have the same annual water use, the mass-based water use is almost 1.5 billion gallons per year lower because the net generation from coal is lower (Table 26). The addition of CCS in scenarios 2 and 3 lowers the difference between the state option and the mass-based approach to 0.5 billion gallons in scenario 2 due to the increased generation from the coal in the mass-based approach (Table 21). The state option in scenario 3 uses 0.2 billion gallons more than in scenario 2 because the higher net generation from San Juan #4 EGU requires more water use when the coal boiler is used to create steam for the CCS sorbent regeneration process (Table 22).75 Hence, there is an overall increase in annual water use of 1.3 to 1.5 billion gallons over that in scenario 1, when CCS is used on San Juan #4 EGU in either scenario for one of the approaches. The reductions in generation in scenarios 4 and 5 cause a reduction in water use for the rate-based options. The scenario 4 reduction results in a 2.5 and 3.5 billion gallon reduction in annual water use for the state and technology

75 This is also true for San Juan #4 in the mass-based approach; however, the generation is lower and the deviation in water use is lost during round off.


options, respectively (Table 23). The retirement of San Juan #1 EGU slightly increases these water reductions because of more efficient water use (Table 24). One component of the plant makeup water is the cooling system water, which is quantified for possible replacement from alternative sources.76 A comparison of the water system cooling makeup water rate of use for each generation source under scenario 1 shows that the cooling water requirements for the two rate-based approaches are identical, and that for the mass-based approach is slightly lower (Tables 27 and 28). Given the same trend for the plant makeup water use rate (Tables A-A22, A-A27, and A-A30), this study need only model the mass-based approach and the state option for the rate-based approach in the stochastic model to determine the sensitivity of the plant and cooling system makeup water to the atmospheric conditions. The results from this model indicate that 95% of the variation in annual makeup and cooling water caused by the uncertainty in the forecasted temperature and relative humidity for each approach is within 2% of the respective median value for each generation source (Table 28).77 The summation of these sources then gives a non-probabilistic upper and lower bounds for the annual water use under each approach (Table 29).78 Table 26 Scenario 1, IECM simulated plant makeup water needs (gigagallons/year) for New Mexico generation sources under the rate and mass-based CPP approaches applied in 2030, after the generation shift and the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state and technology options are used for in rate-

based approach. The mass-based approach does not include new generation sources. Plant name State option

(Ggal/year) Technology option

(Ggal/year) Mass-based (Ggal/year)

PC 6.6 6.6 5.1 NGCC 1.5 1.5 1.5 OGST 1.2 1.2 1.2

Renewable 0.0 0.0 0.0 Total 9.33 9.33 7.86

76 For the NGCC plants that use water or hybrid cooling, all of the plant makeup water is considered as cooling system water. 77 The CDFs for the plant makeup and cooling water rates for each plant is shown in Figures B8-B11. 78 These bounds do not take into account uncertainties other than temperature and relative humidity. Other operational parameters of the units, such as capacity factor, can have a significant impact on water needs.


Table 27 Scenario 1, IECM simulated cooling system makeup water needs (tons/hour) for specific New Mexico generation plants under the rate and mass-based CPP approaches applied in 2030, after the generation shift and

the 2.1% heat rate improvement is applied to the coal-fired EGUs. The state and technology options are used for in rate-based approach. The mass-based approach does not include new generation sources.


State option (tons/hour)

Technology option

(tons/hour)

Mass-based (tons/hour)

San Juan 1 940.8 940.8 934.6 San Juan 4 1421 1421 1,413 Escalante NA 684.5 684.5 680.5

Afton NGCC NA 163.3 163.3 163.3 Bluffview NGCC NA 87.3 87.3 87.3

Hobbs NGCC NA 0.0 0.0 0.0 Luna NGCC NA 705.6 705.6 705.6

Table 28 Scenario 1 makeup water 95% CDF ranges for specific New Mexico generation plants for state option in rate-based approach and the mass-based approach in 2030. The technology option for the rate-based approach yields water use metrics identical to those for the state option. Estimations are based upon an

uncertainty simulation in the IECM that only accounts for the uncertainty in the average annual temperature and humidity.

Plant Water (Ggal/year)

Cooling System Water (Ggal/year)


Level State option

Mass-based

State option

Mass-based

San Juan 1

2.5 1.99 1.55 1.64 1.28 50 2.01 1.56 1.66 1.29

97.5 2.03 1.58 1.68 1.30

San Juan 4

2.5 3.02 2.35 2.48 1.93 50 3.05 2.37 1.51 1.95

97.5 3.07 2.39 2.53 1.97

Escalante

NA 2.5 1.53 1.19 1.25 0.97 50 1.54 1.20 1.26 0.98

97.5 1.55 1.20 1.27 0.99

Afton NGCC

NA 2.5 0.26 0.26 0.26 0.26 50 0.26 0.26 0.26 0.26

97.5 0.26 0.26 0.26 0.26

Bluffview NGCC

NA

2.5 0.14 0.14 0.14 0.14 50 0.14 0.14 0.14 0.14

97.5 0.14 0.14 0.14 0.14

Hobbs NGCC

NA

2.5 0.0 0.0 0.0 0.0 50 0.0 0.0 0.0 0.0

97.5 0.0 0.0 0.0 0.0

Luna NGCC

NA

2.5 1.11 1.11 1.11 1.11 50 1.13 1.13 1.13 1.13

97.5 1.14 1.14 1.14 1.14


Table 29 Scenario 1 makeup water 95% CDF ranges for New Mexico generation sources for state option in rate-

based approach and the mass-based approach in 2030. The technology option for the rate-based approach yields water use metrics identical to those for the state option. Estimations are based upon an uncertainty

simulation in the IECM that only accounts for the uncertainty in the average annual temperature and humidity. Plant Water

(Ggal/year) Cooling System Water

(Ggal/year) Level State option

Mass-based

State option

Mass-based

2.5 8.06 6.6 6.88 5.69 50 8.13 6.66 6.95 5.75

97.5 8.19 6.71 7.02 5.8

Alternative Plant Cooling Water Availability and Cost The makeup water for the wet cooling tower system that is used to replenish the consumed water used to cool the steam generator water for the boiler in a PC EGU, or the heat recovery steam generator in a NGCC plant, requires less purification than that for steam generation. Given that the alternative water sources examined in this study may require more treatment for use as steam than the current treatment system is able to provide, this study only considers using alternative water sources for makeup water in the wet cooling system. Therefore, the amount of the current cooling water that can be replaced is dependent upon the cooling system requirements for the specific EGU, or plant, for the given scenario and the availability of the alternative water sources. The overall makeup water cost (dollars per thousand gallons ($/kgal)) is then the gallon-weighted average of the cooling system water to the total plant makeup water use (Equation [11]). In the case when the current water source has no associated costs, the equation reduces to the first term and is then related solely to the proportion of cooling water that can be replaced at the given cost for the alternative source to the total cooling requirements.79 𝑤𝑎𝑡𝑒𝑟 𝑐𝑜𝑠𝑡𝑡𝑜𝑡𝑎𝑙

= 𝑤𝑎𝑡𝑒𝑟%𝑐𝑜𝑜𝑙𝑖𝑛𝑔

𝑎𝑙𝑡 ∗ 𝑤𝑎𝑡𝑒𝑟𝑐𝑜𝑜𝑙𝑖𝑛𝑔 ∗ 𝑤𝑎𝑡𝑒𝑟 𝑐𝑜𝑠𝑡𝑎𝑙𝑡 𝑤𝑎𝑡𝑒𝑟𝑝𝑙𝑎𝑛𝑡

+�1 − 𝑤𝑎𝑡𝑒𝑟%𝑐𝑜𝑜𝑙𝑖𝑛𝑔

𝑎𝑙𝑡 � ∗ 𝑤𝑎𝑡𝑒𝑟𝑐𝑜𝑜𝑙𝑖𝑛𝑔 ∗ 𝑤𝑎𝑡𝑒𝑟 𝑐𝑜𝑠𝑡𝑐𝑢𝑟𝑟𝑒𝑛𝑡𝑤𝑎𝑡𝑒𝑟𝑝𝑙𝑎𝑛𝑡

+(𝑤𝑎𝑡𝑒𝑟𝑝𝑙𝑎𝑛𝑡 − 𝑤𝑎𝑡𝑒𝑟𝑐𝑜𝑜𝑙𝑖𝑛𝑔) ∗ 𝑤𝑎𝑡𝑒𝑟 𝑐𝑜𝑠𝑡𝑐𝑢𝑟𝑟𝑒𝑛𝑡)

𝑤𝑎𝑡𝑒𝑟𝑝𝑙𝑎𝑛𝑡 , [11]

where 𝑤𝑎𝑡𝑒𝑟 𝑐𝑜𝑠𝑡𝑡𝑜𝑡𝑎𝑙 is the total cost for the plant makeup water ($/kgal), 𝑤𝑎𝑡𝑒𝑟%𝑐𝑜𝑜𝑙𝑖𝑛𝑔

𝑎𝑙𝑡 is the percent of makeup cooling water for the generator that can be replaced by the alternative water source,

79 This study assumes that the current water source has no cost.


𝑤𝑎𝑡𝑒𝑟𝑐𝑜𝑜𝑙𝑖𝑛𝑔 is the annual amount of makeup cooling water used at the generator (kgal), 𝑤𝑎𝑡𝑒𝑟 𝑐𝑜𝑠𝑡𝑎𝑙𝑡 is the cost for the alternative water source ($/kgal), 𝑤𝑎𝑡𝑒𝑟 𝑐𝑜𝑠𝑡𝑐𝑢𝑟𝑟𝑒𝑛𝑡 is the cost for the current cooling water ($/kgal), and 𝑤𝑎𝑡𝑒𝑟𝑝𝑙𝑎𝑛𝑡 is the annual quantity of makeup water used by the generator (kgal).

If one only considers the scenario 1 values for the cooling makeup water simulation for each generator (Table 27) and compares this to the estimates of alternative water sources available at each site (Table A-A6), one finds that a greater percentage of the cooling water for the NGCC plants can be replaced than for the PC EGUs (Tables A-A33 to A-A35).80 This is due in part to the lower water requirements for the NGCC plants. An exception to this is the San Juan EGUs that can have 100% of the required cooling water replaced with unappropriated ground water. For Scenario 1, the percent availability and cost for the plant makeup water is substantially independent of the approach taken because each approach and option uses similar amounts of cooling water. As the cost for the alternative water is greater than that for the water currently used on site, the range in makeup water cost for each approach is from $0/kgal, when no alternative water is available, to $2.78/kgal, when the current source is supplemented with brackish water.

Discussion

Emission Intensity Reduction

Scenario CO2 Avoidance Cost and Increase in LCOE As the names suggest, the rate-based and mass-based approaches use different standards to show compliance. This study looks at each with respect to emission intensity as a simple, common denominator to analyze the cost-effectiveness of these approaches and the different scenarios. For cost effectiveness, two metrics are used—CO2 avoidance cost and increase in fleet LCOE.81 The CO2 avoidance cost is examined from the perspective of the overall fleet and the fossil fuel generation sources. These two perspectives are relevant because the former relates to what is emitted into the atmosphere and the latter relates to the intent of the CPP to reduce emissions intensity of the fossil fuel plants by augmenting the fossil fuel generation with ERCs to reduce the perceived emission intensity of these plants. Thus for these perspectives, the emission intensity can be defined as: (1) the overall fleet net emission intensity is the total pounds of CO2 emitted by the fleet divided by the total fleet net generation, inclusive of net generation from renewable sources existing prior to 2012; (2) the fossil fuel net emission

80 The uncertainty in the generator requirements expressed in the CDF is not used because the uncertainty in the availability of the alternative water source is not known. 81 OGST emissions, net generation and LCOE are excluded from the calculations for both metrics. ERC and MA prices are set to zero, unless otherwise indicated.


intensity is restricted to the pounds of CO2 emissions divided by the net generation from these fossil fuel sources and any generation related to emission reduction credits that are purchased by the fossil fuel sources. For each case, the expression for the avoidance cost is taken as:

𝑎𝑣𝑜𝑖𝑑𝑎𝑛𝑐𝑒 𝑐𝑜𝑠𝑡𝑖 = �𝐿𝐶𝑂𝐸𝑖2030−𝐿𝐶𝑂𝐸𝑖

2012��𝐼𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦𝑖

2012−𝐼𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦𝑖2030�

, [12]

where i is the measured condition (fleet or fossil fuel), 𝑎𝑣𝑜𝑖𝑑𝑒𝑛𝑐𝑒 𝑐𝑜𝑠𝑡𝑖 is the CO2 avoidance cost ($/ton CO2) for condition i, 𝐿𝐶𝑂𝐸𝑖2012 is the LCOE in 2012 for condition i, 𝐿𝐶𝑂𝐸𝑖2030 is the LCOE in 2030 for condition i, 𝑖𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦𝑖2012 is the net CO2 emission intensity in 2012 for condition i (lbs/MWh), and 𝑖𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦𝑖2030 is the net CO2 emission intensity in 2012 for condition i (lbs/MWh).

Implementation of the various scenarios reduces the fleet emission intensity from 1,611 lbs/MWh in 2012 to between the range from 1,086 to 856 lbs/MWh, at an avoidance cost of $44-$55/ton of CO2 (Figure 2). The highest intensities are those for the mass-based approach, while the lowest intensities are those for the technology option for the rate-based approach that maintains net generation at the 2012 levels (T4 and T5). When these data are plotted as a percent reduction in fleet emission intensity rather than the absolute emission intensity after implementation of the scenario (Figure 3), one might be indifferent between choosing between scenarios S2, S3, S5, and T5 because of the similarity in avoidance cost and each assures compliance with the CPP. While each scenario may be compliant, the range for reductions in fleet emission intensity is large (39% to 47%).


Figure 2 Fleet CO2 emission intensity and associated avoidance cost for the different CPP scenarios. The fleet

emission intensity is the summation of the pounds of CO2 emitted by all endogenous sources divided by all endogenous net generation, inclusive of renewable sources existing prior to 2030. Implementation of the

technology option for the rate-based approach that allows for retirement of the San Juan #1 EGU and maintains net generation at the 2012 level results in the lowest emission intensity and the third lowest cost of avoidance.

ERC and MA prices have no cost in this analysis.

Figure 3 CO2 avoidance cost for the different scenarios from implementing the CPP in New Mexico as measured by the reduction in the CO2 emission intensity for the fleet. The fleet emission intensity is the

summation of the pounds of CO2 emitted by all endogenous sources divided by all endogenous net generation, inclusive of renewable sources existing prior to 2030. As all of the scenarios are compliant with the CPP,

scenario S2 and S3 dominate all choices as that with the least avoidance cost. T5 results in a larger reduction in emission intensity but at a slightly higher avoidance cost. ERC and MA prices have no cost in this analysis.


When these same scenarios are viewed from the perspective of fossil fuel emission intensity, the overall avoidance cost diminishes and more pronounced clustering is observed in the approaches (Figure 4). The overall range for avoidance cost increases from approximately $14/ton of CO2 in the fleet emission intensity analysis to approximately $47/ton of CO2 in the fossil fuel emission intensity analysis, however. The technology option for the rate-based approach produces the greatest reductions in emission intensity at the lowest avoidance costs, with T1 being the dominant scenario that results in a 45% reduction in fossil fuel emission intensity at a cost of $14.3/ton of CO2. Conversely, the mass-based approaches yield the least reduction at the highest cost. Therefore, from the perspective of CPP compliance with a minimized avoidance cost, it is preferable to use the technology option of the rate-based approach, but the scenario chosen depends upon the emission intensity that is considered for the avoidance cost.

Figure 4 CO2 avoidance cost for the different scenarios from implementing the CPP in New Mexico as measured by the reduction in the fossil fuel emission intensity. As all of the scenarios are compliant with the

CPP, scenario T1 dominates all choices as that with the least avoidance cost. ERC and MA prices have no cost in this analysis.

One way to reconcile this choice may be to consider the increase in fleet LCOE. Doing so shows a different trend, however (Figure 5). In regards to the reduction in the fleet emission intensity, the technology based scenarios result in the largest increase in fleet LCOE, while the mass-based scenarios result in the lowest increase in LCOE. In particular, M1 dominates all other scenarios with a 33% reduction in emission intensity at a 25% increase in LCOE. The clustering of scenarios, based upon option and approach, is more evident when the increase in LCOE is examined relative to the reduction in fossil fuel intensity (Figure 6). The mass-based scenarios still result in the lowest increase in


LCOE, with M1 dominating all choices with a 20% reduction in fossil fuel emission intensity and the 25% increase in LCOE. Thus, the consistency of the scenario dominance is more pronounced when the increase in LCOE is the cost effectiveness metric.

Figure 5 Increase in fleet LCOE for the different scenarios from implementing the CPP in New Mexico as measured by the reduction in the emission intensity for the fleet. The fleet emission intensity is the summation of the pounds of CO2 emitted by all endogenous sources divided by all endogenous net generation, inclusive of

renewable sources existing prior to 2030. As all of the scenarios are compliant with the CPP, scenario M1 dominates all choices as the ones with the least avoidance cost. ERC and MA prices have no cost in this analysis.


Figure 6 Increase in fleet LCOE for the different scenarios from implementing the CPP in New Mexico as

measured by the reduction in the fossil fuel emission intensity. As all of the scenarios are compliant with the CPP, scenario M1 dominates all choices as that with the least avoidance cost and greatest reduction. ERC and

MA prices have no cost in this analysis.

It should be noted that the LCOE calculations do not include ancillary services. This is relevant because the increased use of renewable sources requires more fees for ancillary services to compensate for the variability of the renewable sources (Apt and Jaramillo, 2014, chapters 7&8). Since each scenario has the same proportion of new solar and wind energy, scenarios with higher renewable energy penetration may have an increased likelihood of incurring these higher costs. The mass-based scenarios have the lowest renewable energy penetration because only 5% of the mass allowances are set-aside for new renewable resources, however (Figure 7). Additional costs may include the capital and operational costs for the new transmission lines to connect new renewable sources to the main grid.


Figure 7 Increase in percent of endogenous net generation coming from renewable sources in relation to the overall increase in net generation under the different scenarios from the implementation of the CPP in New Mexico for 2030. The initial assumption was that the CPP would result in a renewable source penetration of 30%. Of the scenario 1 implementations, the mass-based approach has the lowest increase in net generation.

CCS Mitigation Option Sensitivity to ERC/MA and Natural Gas Price The CCS options for CO2 mitigation in scenarios 2 and 3 have similar cost effectiveness as the respective scenario 1 rate and mass-based approaches outlined in the CPP, with regards to the related percent increase in fleet LCOE for the percent reduction in fossil fuel CO2 emission intensity. The similarity is not apparent when the metric is the CO2 avoidance cost with regard to the reduction in the fossil fuel fleet emission intensity, when there is no price for the ERCs and MAs. Since implementation of CCS decreases the tonnage of CO2 emitted for a constant net generation, one expects that the avoidance cost for a PC EGU without CCS that must purchase ERCs or MAs from new renewable sources or NGCC plants to be compliant with the CPP will at a particular ERC/MA price exceed that for a similar PC EGU that has CCS and must also purchase these credits to be compliant. The point of intersection of the avoidance cost loci for the ERC/MA prices for these two EGUs represents the point at which one is indifferent between choosing CCS or purchasing the credits or allowances. If the price for these credits or allowances is higher, it is more economical to use CCS, with regard to the avoidance cost. This point will also depend upon the percent capture and the steam and electricity source for the CCS system. When a 40% capture system with an auxiliary gas boiler is modeled on San Juan EGU #4 for the rate-based approach (S2), then the breakeven point is $34.9/MWh for an avoidance cost of $54.7/ton of CO2 (Figure 8). Allowing the EGU to supply the steam and electricity for the CCS system (S3) increases the slope on the CCS line so that the intersection decreases to $33.3/MWh and the avoidance cost is $53/ton of CO2 (Figure


A-B12). For this EGU and CCS system in the mass-based approach, the MA price must be $221.9/ton of CO2 for an avoidance cost of $89.7/ton of CO2 for the CSS system with an auxiliary boiler (M2, Figure A-B13). The same system without the auxiliary boiler (M3) has a MA price of $96.8/ton for an avoidance cost of $65.4/ton of CO2 (Figure A-B14). In each of these approaches, the avoidance cost for the CCS system without the auxiliary boiler is less than that with the auxiliary boiler because the cost for the natural gas to run the boiler is greater than that for the cost for the coal to do so and the emissions from the auxiliary boiler are not captured. The reason for the high threshold for the MA is discussed in the following section on dispatch order.

Figure 8 Sensitivity of avoidance cost to ERC price for San Juan EGU #4 with and without 40% capture CCS

system. When fitted with CCS, an auxiliary natural gas boiler provides the steam and electricity for the system. The two lines intersect when the ERC price is $34.9/MWh and result in an avoidance cost of $54.7/ton of CO2.

Block 2 of the CPP relates to shifting generation from the PC EGUs to the NGCC plants. Though these NGCC plants have a lower CO2 emission intensity than that for the PC EGUs, the LCOE for the NGCC plants is higher than that for the PC EGUs because the 2030 price for the natural gas is predicted to be higher than it was in 2012. This results in a $16.3/ton CO2 avoidance cost for the fossil fuel fleet in S1, without any ERC price. Lower natural gas prices will result in lower avoidance costs that can be negative, because the price for the natural gas is low enough that the LCOE for the NGCC plants is lower than that for the PC EGUs (Figure 9). Thus, if the natural gas price is no longer held fixed as it was in the previous ERC/MA analysis, then the ERC/MA price that is required to make CCS feasible in each scenario will depend upon the price of the natural gas. Here, the required ERC/MA price to meet the above avoidance cost threshold is inversely related to the natural gas price. This relationship is shown for S1 (Figure 10), where an ERC price of $30/MWh is required for the avoidance cost for the fossil fuel


fleet to be greater than the threshold of $54.7/ton of CO2, when the natural gas price in 2030 is 100% that for 2012. This ERC price increases to $50/MWh, when the increase in natural gas price is only 40%. (Similar trends for the other scenarios are shown in Figures A-B15 to A-B17.)

Figure 9 Sensitivity of avoidance cost for the fossil fuel fleet to percent increase in natural gas price relative to

2012. Negative avoidance costs occur when the price of natural gas is low enough to cause the LCOE of the NGCC plants to be lower than that for the PC EGUs.

Figure 10 Sensitivity of the avoidance cost for the fossil fuel fleet to ERC price and percent increase in natural

gas price relative to 2012 for San Juan EGU #4 with a 40% capture CCS system. An auxiliary natural gas boiler provides the steam and electricity for the CCS system. Areas shaded red represent the combinations of ERC

price and natural gas price increases that have an avoidance cost of $54.7/ton of CO2 or greater. In these regions, CCS with an auxiliary natural gas boiler has a cost effective avoidance cost.


Plant VOM and Dispatch Order While the dispatch order in a state with a regulated electricity market may be affected by contractual commitments in additional to variable costs, the dispatch for states with deregulated electricity markets is generally based upon bidding the electricity at the lowest cost for the generation source. In this case, the lowest cost is the marginal cost of the electricity—the VOM. Generating sources such as renewable or hydropower have zero or near zero marginal costs and are typically the first to be dispatched in this system. Those that use nuclear and fossil fuel, respectively, usually follow these sources, though nuclear is regularly taken out-of-merit. Under the CPP with regard to dispatch order, it is desirable to dispatch NGCC plants before PC EGUs, since the former have a lower CO2 emission intensity and the intent of the regulation is to achieve a 75% capacity factor with these plants. This may also be the premise behind the proposed markets setup for the emission reduction credits and the mass allowances that favor lower carbon emitting sources by having the higher emitting sources pay for the credits and allowances that can offset the lower emitting sources. The fuel price is the main component of the VOM for the fossil fuel sources; therefore, the projected fuel price will have a significant effect on dispatch order in market-based systems. In 2012, the natural gas price was low enough that most of the NGCC plants were dispatched before the PC EGUs. This order is almost reversed in 2030 due to the projected higher cost of natural gas (Table 30); however, the future dispatch order can be changed through market intervention associated with the price of the ERCs and MAs. A sensitivity analysis of the generation-weighted VOM for each technology to the ERC or MA price indicates that at an ERC price of greater than $11.1/MWh, the VOM for the PC EGUs will be greater than that for the NGCC plants under the rate-based approach (Figure 11). The VOM sensitivity for the mass-based approach does not converge until the MA price is $389.4/ton of CO2, however. This apparent disparity in convergence points is related to the mass allocation process in which most generating sources already have all the allowances required to meet the respective electricity generation targets. Consequently, the MA price must be more than an order of magnitude larger than current market-based carbon prices82 to compensate for the projected higher natural gas price. Hence, while the ERC price may be effective in changing an undesirable dispatch order related to high natural gas price, the MA allowance may not be useful for this purpose unless the allocation process is designed to require PC EGUs to purchase a sufficient number of allowances from NGCC plants. 82 The current price for carbon in the California and Regional Greenhouse Gas Initiative (RGGI) markets is less than $12/ton of CO2, as of December 31, 2015.


Table 30 Comparison of dispatch orders for New Mexico generation sources in 2012 and 2030, based upon VOM. 2030 projections are for scenario 1 IECM simulated VOM under the mass-based approach and the state option for the rate-based approach of the CPP applied in 2030, after the generation shift and the 2.1% heat rate

improvement is applied to the coal-fired EGUs. ERC and MA prices are set to $0/MWh and $0/ton of CO2, respectively.

VOM ($/MWh) Dispatch Order Plant name Unit

number 2012 2030

2012 2030

OGST NA NA NA NA NA San Juan 1 30 29.4 9 3 San Juan 2 29.2 NA 7 NA San Juan 3 28.9 NA 6 NA San Juan 4 29.2 28.6 7 2 Escalante NA 26.4 26.4 3 1

Afton NGCC NA 27.7 39.3 4 4 Bluffview NGCC NA 28.5 55.6 5 6

Hobbs NGCC NA 22.6 48.3 1 5 Luna NGCC NA 24.8 55.6 2 6

Figure 11 Variation in generation-weighted average VOM due to changes in the ERC or MA price. In the

implementation of the state option for the rate-based approach on the New Mexico PC EGUs and NGCC plants, the higher NGCC VOM (due to the projected natural gas price in 2030) can be compensated for with an ERC price greater than $11.1/MWh. In the mass-based approach, the MA price must be greater than $389.4/ton of

CO2 to compensate for the natural gas price. The high MA price relates to the mass allowance allocation process.


Uncertainty Analysis

Stochastic Model Some of the inputs used for the deterministic model are based upon averages that may not be historically accurate for the specific generation sites, are projections of future costs and expenses that are based upon estimates that are likely to change, or are future costs that are defined but not yet quantitatively understood. Among these uncertainties are the capital costs and O&M expenses of existing plants that are used to calculate the historical LCOEs, the future capital and O&M costs for plants yet to be built and the prices for the fuel to generate the electricity for those that are not renewable sources, and the price that the market will set for the ERCs or the MAs. One can determine the probable effect of parameters such as these, and determine which might be important in driving the avoidance cost and LCOE metrics, with a stochastic model based upon distributions around those parameters with uncertainty—a model that is similar to that used to determine the effect of future site-specific temperature and relative humidity variations on plant makeup water. Here, the Monte Carlo analysis run is in the Excel model using @RISK, rather than in the IECM, with the inputs of relevance now the fuel prices for the future fossil fuel plants, the capital and O&M cost for the solar and wind farms, and the price for the ERCs (Table A-A21). For existing solar and wind farms, the farm capital cost is determined stochastically for the year the farm went online; the capital cost component of the LCOE is determined deterministically from that value for all subsequent periods. All O&M values for existing and new solar and wind farms are determined stochastically, however. Only the state option for the rate-based approach is modeled stochastically. The deterministic, projected cost for compliance with the CPP under this option with no ERC price is $16.3/ton of CO2 avoided for the fossil fuel fleet, which is less than the EPA’s national average estimate of $30/ton of CO2 avoided. Results from the simulation of 10,000 runs of the stochastic model indicate that the mean projected cost is $31.4 ± 0.4 per ton of CO2 (Figure 12). A sensitivity analysis of the avoidance cost (Figure 13) shows that the price of natural gas, where the proxy for the price is the predicted percent increase in the price over the 2012 price, is a larger driver than the ERC price.83 The avoidance cost is also sensitive to the capital cost component of the LCOE for the wind turbines, the price for the coal, and the capital costs component of the LCOE for the solar power—to a lesser extent.

83 The larger expected cost is due in part to modeling the price for the ERCs as a uniform distribution with an expected value of $25/MWh.


Figure 12 Cumulative distribution function of the uncertainty in the 2030 CO2 avoidance cost from

implementing the S1 scenario for the CPP in New Mexico, as measured by the reduction in the fossil fuel emission intensity. Ten thousand simulations of the S1 stochastic model were run.

Figure 13 Sensitivity analysis to input parameter uncertainty in the 2030 CO2 avoidance cost from

implementing the S1 scenario for the CPP in New Mexico, as measured by the reduction in the fossil fuel emission intensity. The natural gas price dominates this cost, followed by the ERC price, the capital cost

component of the LCOE for the wind turbines, the price for the coal, and the capital costs component of the LCOE for the solar power.


The simulations show that there is a low probability that the avoidance cost will be negative or near zero. These occasions relate in part to 1% of the simulations having a lower natural gas price in 2030 than that in 2012 (Figure A-B18). On these occasions, the LCOE for the NGCC plants will also be lower than that for 2012 (Figure A-B19) to produce an overall fossil fuel fleet cost in 2030 that is lower than that from 2012. This then results in a negative avoidance cost. It is this sensitivity of the avoidance cost to the natural gas price, and to the ERC price as demonstrated in Figure 10, that illustrates the importance of the simulation distributions for the significant input parameters. While a CDF for the natural gas price can be produced from peer-reviewed methodology, this study does not have a similar CDF for the ERC to improve the prediction of the cost of avoidance.

The EPA documentation on the CPP does not cite a specific U.S. fleet LCOE for 2030 or the projected increase in LCOE from 2012. The estimation from the stochastic model is that the LCOE for the state option will increase by an average of 30.7 ± 0.2 percent, with a 95% probable range of 16.4% to 44.7% (Figure 14). The factors that affect this metric are the increase in the natural gas price, the capital cost component of the LCOE for the wind turbines, the variation in the coal price, and the capital costs component of the LCOE for the solar power (Figure 15). Here, the ERC price is not significant because the resulting increase in LCOE from the purchase of ERCs by any plant is offset by a reduction in LCOE from the sale of those ERCs by another plant.

Figure 14 Cumulative distribution function of the uncertainty in the 2030 increase in fleet LCOE, relative to

2012, from implementing the S1 scenario for the CPP in New Mexico. Ten thousand simulations of the S1 stochastic model were run.


Figure 15 Sensitivity analysis to input parameter uncertainty in the 2030 increase in fleet LCOE, relative to

2012, from implementing the S1 scenario for the CPP in New Mexico. The natural gas price dominates this cost, followed by the capital cost component of the LCOE for the wind turbines, the price for the coal, and the capital costs component of the LCOE for the solar power. The ERC price is not significant because ERCs purchased by

the any plant are offset by a reduction in LCOE from the selling plant.

CCS Retrofit Factor Few commercial, PC CCS facilities are currently in operation: none of these are facilities with retrofitted CCS systems. As such, the CCS retrofit factors used to determine the increase in LCOE for the San Juan EGU #4, when fitted with this system, are estimates with associated uncertainty. The effects of the uncertainty can be quantified by running the IECM model for the plant at different levels for this factor on all associated components; in this manner, a lower and upper bounds of the effect of varying this factor can be determined for the various metrics of interest. The nominal retrofit factor is varied between 1 and 1.4 to examine the sensitivity of the various economic output metrics to the retrofit factor. When these factors are used in the IECM model to determine the LCOE for this EGU with an auxiliary natural gas boiler as part of the CCS system (Figure 16), the nominal LCOE varies linearly by approximately ± 3%. This variation in LCOE is also seen in a similar increase in the fossil fuel avoidance cost. The variation is lessened as the LCOE changes for a single EGU are diluted by the generation-weighted average LCOEs from more renewable energy sources. The resulting impact on the fleet LCOE is only ± 0.5%. When the auxiliary boiler is not used, the same trends are observed with smaller resulting variations, because the absence of the auxiliary boiler decreases the retrofit capital costs (Figure 17).


Figure 16 Sensitivity of selected metrics to CCS retrofit factor for San Juan EGU #4 in the mass-based

approach, when fitted with a CCS system with an auxiliary natural gas boiler. The baseline retrofit factor for all components in the CCS system used in this study was 1.2, which indicates that all related capital cost factors

are increased by 20%.

Figure 17 Sensitivity of selected metrics to CCS retrofit factor for San Juan EGU #4 in the mass-based

approach, when fitted with a CCS system without an auxiliary natural gas boiler. The baseline retrofit factor for all components in the CCS system used in this study was 1.2, which indicates that all related capital cost

factors are increased by 20%.


Mass-based Approach Scenario 1 for the state and technology options for the rate-based approach set the goal for the emission intensity from the fossil fleet. Similarly, scenario 1 for the mass-based approach sets the goal for the total mass emissions from the fossil fuel fleet. Implementing the different scenarios resulted in different overall fleet emission intensities, such that the emission intensity for the mass-based approaches is greater than those for the other approaches (Figure 2). When one considers the emitted mass for the different scenarios, one finds that the rate-based approaches for scenarios one through three emit approximately more than one to two million tons of CO2 annually than do the mass-based scenarios (Figure 18). Here, only scenarios four and five emit less CO2 than the targeted amount, due to the process to obtain the limited generation in these scenarios.

Figure 18 Fleet CO2 mass emissions and associated avoidance cost for the different CPP scenarios. The fleet mass emission is the summation of the tons of CO2 emitted by all endogenous sources. Implementation of the

technology options for the rate-based approach and that maintain net generation at the 2012 level results in the lowest carbon emissions and the third lowest cost of avoidance. ERC and MA prices have no cost in this

analysis.

The discrepancy in reduction in mass and emission intensity observed in Figures 2 and 18, for which all scenarios are compliant with the CPP, suggests that the scenarios exhibit a nonlinear behavior. When plotted, this behavior does appear to be nonlinear for some of the scenarios (S1, S2, S3, T1) because the percent reduction in the emitted mass is almost a one-half that for the percent reduction in the fleet CO2 emission intensity, and is substantially independent of the emission intensity (Figure 19). Additionally, the clustering of the scenarios predominately by type for each approach appears to be biased, for the remaining scenarios, in that the percent reduction in the emitted mass is 10% to 25% less than that for the percent reduction in the fleet CO2 emission intensity and


appears to have a linear dependence. This linearity is more apparent when existing renewable sources are removed from the emission intensity calculation (Figure 20). This behavior still results in the same dominant scenarios for the avoidance cost and increase in LCOE metrics when the objective is mass reduction, though the data are clustered by scenario for the rate-based approaches (Figures A-B20 and A-B21).

Figure 19 Reduction in fleet CO2 emission intensity and associated fleet CO2 mass emissions for the different CPP scenarios. The fleet emission intensity is the summation of the pounds of CO2 emitted by all endogenous sources divided by all endogenous net generation, inclusive of renewable sources existing prior to 2030. The

fleet mass emission is the summation of the tons of CO2 emitted by all endogenous sources. Scenarios S1, S2, S3, and T1 show a nonlinear relationship for the reduction in emission intensity and mass, while the other scenarios

may show a biased, linear behavior.


Figure 20 Reduction in fossil fuel CO2 emission intensity and associated fossil fuel CO2 mass emissions for the

different CPP scenarios. The fossil fuel emission intensity is the summation of the pounds of CO2 emitted by all endogenous sources divided by all endogenous net generation, exclusive of renewable sources existing prior to

2030. The mass emission is the summation of the tons of CO2 emitted by all endogenous sources. Scenarios S1, S2, S3, and T1 show a nonlinear relationship for the reduction in emission intensity and mass, while the other scenarios may show a linear behavior. Scenario M1 has a lower fossil fuel intensity reduction because the PC

EGUs have a great intensity since CCS is not used in this scenario.

Water Use Plant Water Consumption Comparison to USGS Data Comparing the results of the plant water consumption from the IECM simulation of the historical performance of the plants to the historical water consumption estimates given by the USGS serves as a benchmark for the accuracy of the simulation. As the simulation and the USGS data are both estimates of use, this benchmark is not to be taken as a high degree of accuracy, but rather to be “in the ballpark” of an order of magnitude. Here, water consumption in the IECM is measured as the evaporation loss from the wet flue-gas desulfurization unit and the wet cooling tower for the coal-fired EGUs. For the NGCC plants, the IECM water consumption is comprised of drift and evaporation losses from the wet cooling tower. Since the USGS data are taken at the plant-level, the unit-level data from the simulation are summed at the plant-level and converted to millions of gallons of water consumed per day. Doing so for the 2010 historical performance shows that all of the simulation estimates for both PC and NGCC plants are within a factor of two of the USGS estimates (Table 31). Another method to verify the water consumption is to normalize the usage with the net generation. This normalized value (gallons per megawatt-hour) can then be compared to


the normalized ranges compiled by Macknick (2012), and the site-unique data from the integrated resource planning documentation provided by the power plant owners (Table 32): Public Service Company of New Mexico84 (PNM) are the primary owners for Afton, Luna and the San Juan power plants; Tri-State Generation and Transmission Associate, Inc.85 is the primary owner for the Escalante power plant. In this comparison, the USGS values for the power plants fall within the ranges given by Macknick (2012). This is also true of the IECM estimates, with the exception of that for the Luna NGCC plant. The data provided by the plant owner indicates that the IECM estimation is typically closer to the reported value than that from the USGS estimate, however. Therefore, the IECM estimate method may provide a simulation of water use that is accurate enough to show the trends in water use due to implementation of the CPP scenarios. This is particularly true with estimating the water use for the Afton and Luna NGCC plants. As Afton uses a hybrid cooling system, the water use per MWh should be much lower than that for Luna, though the yearly operating hours for the Afton dry-cooling system are not given. According to PNM, Afton has a water efficiency that is 50% better than that for Luna. This improved efficiency is not indicated in the USGS data, though the trend is apparent in the IECM estimate.

The simulation estimate for the Afton plant is almost two times greater than the value disclosed by the plant owner. While this is a large deviation, it may not be significant, since the Afton plant had the lowest water use of any of the modeled fossil-fuel plants. This low use may relate to the low capacity factor (12%) for 2010. The discrepancy in IECM estimation and the owner reported value may relate in part to the aforementioned modeling limitations for adjusting the capacity and heat rate of the NGCC plants. Other factors may relate to the differences in wet cooling tower performance parameters between the site and the model, such as the cooling water inlet temperature and the atmospheric conditions. Additionally, this plant may be more difficult to simulate than the other plants because the water use for the hybrid system is based upon an averaged empirical relationship between the input fuel Btu and the corresponding water use (Webster et al., 2013). This factor may differ from that observed at Afton. Furthermore, for Afton we do not know the annual ratio of overall water tower cooling use to dry cooling use.

84 “PNM Integrated Resource Plan: 2014 - 2033.” https://www.pnm.com. n.d. n.p. Web. 20 March. 2016. <https://www.pnm.com/irp>. 85 “Integrated Resource Plan/Electric Resource Plan.” http://www.tristate.coop. 12 November. 2015. Tri-State Generation and Transmission Association, Inc. Web. 20 March. 2016. <http://www.tristategt.org/resourceplanning/documents/Tri-State_IRP-ERP_Final.pdf>.

https://www.pnm.com/


Table 31 Comparison of USGS estimation of 2010 site water consumption for New Mexico modeled fossil fuel

plants to the results for water consumption from simulating plant performance for 2010 in the IECM. Previously reported PC EGU plant water consumption simulation results are summed by site and converted to

millions of gallons per day. Plant name USGS

(Mgal/day) Estimation (Mgal/day)

Deviation (𝐼𝐸𝐶𝑀𝑈𝑆𝐺𝑆

-1) San Juan 11.2 16.8 0.51

Escalante 2.2 2.9 0.32

Afton NGCC 0.2 0. 2 -0.11

Bluffview NGCC 0.3 0.3 0.08

Hobbs NGCC NA NA NA

Luna NGCC 1.2 2.1 0.77

Table 32 Comparison of USGS estimation of 2010 site normalized water consumption for New Mexico modeled fossil fuel plants and the results for normalized water consumption from simulating plant performance for 2010 in the IECM to the normalized water consumption ranges by technology type reported Macknick (2012) and by the plant owners. Previously reported PC EGU plant water consumption simulation results are summed by site.

The IECM estimation method yields normalized values that are typically closer to the owner reported values than does the USGS estimation.

Plant name USGS (gal/MWh)

Estimation (gal/MWh)

Macknick (gal/MWh)

PMN (gal/MWh)

Tri-State (gal/MWh)

San Juan 404 609 394-664 594 NA Escalante 474 626 394-664 NA 667

Afton NGCC 224 200 NA 122 NA Bluffview

NGCC 226 254 130-300 NA NA

Hobbs NGCC NA NA NA NA NA Luna NGCC 234 400 130-300 333 NA

Plant Makeup Water All of the scenarios investigated meet the criteria and goals for the CPP, and all result in different performance characteristics for the fossil fuel fleet. To understand which scenario may be preferable, one needs to look at the interaction of the scenario with the various metrics (Figure A-B22). For plant makeup water, each scenario results in a reduction in water use of at least 20% relative to the quantity used in 2012, due primarily to the retirement of the two San Juan EGUs and the generation shift in CPP block two. Associated with these reductions is a reduction in the fossil fuel emission intensity of at least 20% (Figure 21). In most cases, the largest reductions in water use correlates to the largest reductions in fossil fuel intensity that are related to the fourth and fifth scenarios that maintain net generation at the 2012 levels through decreasing the overall capacity factors for the PC


EGUs (Figure 22). The scenarios with CCS result in a small absolute decrease or no decrease in emission intensity, relative to the corresponding scenario 1 approaches, and increase makeup water use due to the steam generation process for amine sorbent.

Figure 21 Reduction in makeup water use and fossil fuel emission intensity relative to 2012 for the

implementation of the CPP scenarios on the New Mexico PC EGUs and NGCC plants. Fossil fuel emission intensity for the rate-based approach scenarios includes the offset in intensity due to the inclusion of ERCs. The

largest reduction in water use corresponds to the implementation of the two scenarios that cap net generation from the fossil fuel sources at the 2012 levels.


Figure 22 Reduction in makeup water use and increase in total net generation relative to 2012 for the

implementation of the CPP scenarios on the New Mexico PC EGUs and NGCC plants. Extra generation from the CCS sorbent generation process is not available for sale on the grid. The reduction in net generation for

scenarios S2 and S3 relates to the decreased requirements for ERCs due to the lower emission intensity for the CCS plant. The largest reduction in water use corresponds to the implementation of the two scenarios that cap

net generation from the fossil fuel sources at the 2012 levels.

Plant Makeup Water VOM for Alternative Water Sources The impact of the increased cost of the makeup cooling water from alternative water sources upon the LCOE depends upon the rate at which the water is used and the amount of electricity that is generated. If there are no disposal charges for the treatment of the makeup water and any required discharge, then the main components of the water system VOM are the electricity cost required for pumping and treatment and the costs incurred for the water. The electricity cost is dependent only upon the net generation and is thus a constant for the various alternative water sources. This leaves the levelized water cost as the only variable for the water system VOM. While there is no difference in the water use and associated costs for the two options in the rate-based approach, there is a difference for the PC EGU costs between the rate-based and the mass-based approach that is related to the difference in net generation for the EGUs in each case (Tables A-A35 to A-A37). In the state option for the rate-based approach, the more expensive groundwater and brackish water sources have a larger impact on the water cost and the water VOM for the PC EGUs than does the less expensive surface water, (Figure 23). Here, the increase in the levelized water cost for the two San Juan EGUs differ because of the difference in makeup water use rates, but the overall impact on the percent increase in water VOM is the same. The impact of the alternative water source cost on the NGCC plant LCOE is less than that for the PC EGUs because of the lower water use for the NGCC plants. However, the differences in levelized cost for the wastewater and the related increase in VOM between the Afton and


Luna NGCC plants are due to the difference in cost for this source at the different sites. For the mass-based approach, the same trends are observed, but the magnitudes differ due to the differences in net generation between the two approaches, (Figure 24). Therefore in either approach, the LCOE for each generation plant will increase with the use of alternative cooling water sources, but this increase will depend upon the cost of the alternative source and the cooling water requirements of the plant.

Figure 23 Impact of alternative makeup cooling water availability and cost on the levelized water cost and the

water system VOM in the implementation of the state option for the rate-based approach on the New Mexico PC EGUs and NGCC plants.


Figure 24 Impact of alternative makeup cooling water availability and cost on the levelized water cost and the water system VOM in the implementation of mass-based approach on the New Mexico PC EGUs and NGCC

plants.

Policy Most of the EPA guidelines for reducing the CO2 emissions from existing power plants are now in place, and it is up to the state policy makers as to how the state plan should be outlined. In doing so, there are many choices that must be made: the metric used, the approach and option taken, how to deal with the interactions and uncertainties that may affect the execution of the plan, and whether or not to consider future water requirements and availability. If the cost-effectiveness of implementing the CPP is the objective of the policymaker, there are at least two possible metrics to use—CO2 avoidance cost and increase in LCOE. In economic terms, one might consider the CO2 avoidance metric to be the marginal cost of reducing emissions—a metric that might be of most interest to utility owners, environmental groups, and the EPA (as the avoidance cost is often used in EPA documentation to compare mitigation options). This metric can be considered for two classifications: (1) relative to the change in emissions and cost for the entire fleet (inclusive of all renewable resources), and (2) relative to the change in emission and cost for the fossil fuel fleet (inclusive of only new renewable resources to offset emissions). In either case, one wants to use this metric to determine which scenario for emissions reduction results in the lowest marginal cost to reach the emissions goal. Analysis of the simulated scenarios for both emission classifications shows that the fossil fuel avoidance cost is more sensitive to the percent reduction in emissions from the scenarios than is the


fleet classification (Figures 3 and 4). Therefore since all scenarios are CPP compliant, the policy maker should consider the fossil fuel classification for this metric.

The increase in LCOE metric is one that consumers, and utility commissions may be more concerned with. Because consumers are rate takers, they seek to minimize their monthly bills. The utility commissioners in regulated states (such as New Mexico) may also be explicitly mandated to share this view, while still wanting to give the utility providers a fair profit when implementing the CPP. Regardless, the scenario with the lowest increase in LCOE will satisfy both parties. Analysis of the simulated scenarios for both emission classifications shows that the increase in LCOE from the scenarios is independent of the classification, though the sensitivity varies due to the reduction of the fleet emission intensity from existing renewable energy sources (Figures 4 and 5). Therefore with this metric, the policy maker need only consider which classification to use to communicate the scenario choice to the stakeholders.

By design, each scenario that a policy maker examines will be compliant with the CPP. The choice that must be made is whether to use the mass-based or the rate-based approach, and if the rate-based approach is chosen, which option. The main difference in scenarios for these approaches and options may be the availability of resources for a given state, the ability to purchase/sell electricity and ERC/MAs with other states, or the future conditions relating to fuel prices, plant retirement schedules, and installation costs for new energy sources. Given the endogenous scenarios examined in this study, the approach taken may depend upon the metric choice. Here, the most cost-effective approach for fossil fuel avoidance cost results from implementation of the rate-based approach with the technology option (Figure 4), while this choice results in one of the least cost-effective approach with regard to increase in LCOE (Figure 6). These diametrically opposed yet compliant outcomes emphasize the importance of understanding the intent of the policy maker with regards to taking the most cost-effective measure for some stakeholders or seeking to avoid costs for other stakeholders. The policy maker should carefully consider the metric choice before choosing the approach.

Factors other than the stakeholder perspective may be relevant. This study examined some of these factors, particularly concerning uncertainties. Here, the most significant factor may be the price for natural gas, as this expense will impact both the avoidance cost and the LCOE because the second building block for the CPP is increasing the generation from the NGCC plants. If the price does increase as much as is projected, this may result in an unintended change in the dispatch order in unregulated electricity markets that may need to be corrected through an adjustment in the price of the ERCs or MAs. While the increase in price for the ERCs to compensate for the natural gas price may be reasonable, that for the MAs is much greater than the current carbon price. However, the MA price is also dependent upon the allocation scheme, of which the policy maker has some control. If the policy maker chooses the mass-based approach, it may be possible to design the MA allocation process to lower the inherent price such that fluctuations in natural gas price can still be compensated for to maintain NGCC plants dispatching before coal-fired plants.


The projected ERC/MA and natural gas prices may also affect the choice for mitigation options taken in either approach. Analysis of the scenarios for the LCOE metric suggested that the mass-based approach with CCS (M2) is the one of the most cost effective method to achieve CPP compliance; this is not true for the avoidance cost metric under the same conditions, however. To achieve parity between the metrics for the approaches at the projected natural gas price, the ERC price needs to be greater than $46/MWh, and have no cost associated with the MA.

Both metrics are also impacted by uncertainties concerning the renewable energy that must be considered. The annualized capital cost of the future renewable sources will directly impact the LCOE components of each metric. As more facilities of each type are built, the capital costs for facilities constructed nearer to 2030 may decrease more than the regressions and standard deviations in this study suggest. This may make these alternatives more economically attractive than the current NGCC plants or the PC EGUs retrofitted with CCS. Partially offsetting these improvements in capital costs are the inclusion of additional ancillary services to improve reliability of the system and additional transmission lines to bring the power to the main grid. The need for both of these LCOE components may be dependent upon the approach and scenario chosen, as the penetration of renewable energy generation to meet the CPP compliance requirements is different for each. Therefore, the policy maker should consider the existing infrastructure and future electricity demand.

The current and future need for water for thermoelectric power generation may also be a concern for the policy maker. Besides minimizing the future need for water, the availability and cost of alternative water sources may influence the choice of CPP scenarios. Consequently, if the objective of the policy maker is to concurrently decrease water use and fossil fuel emission intensity, the policy maker might implement the rate-based approach in which the net generation levels are maintained at the 2012 levels. Here, demand requirements above this level might be met through additional renewable resources without an increase in water use. Alternatively, the scenario 1 mass-based approach may be implemented to achieve a reduction in water use of more than 35% and still meet the CPP requirements. In either scenario, minimizing the need for future makeup water increases the percent of this water that can be replaced by alternative sources. This tradeoff may be impacted by the location of the fossil fuel plant, the type of power plant, and the source of the alternative water. Furthermore, the associated increase in LCOE or avoidance cost due to the increase in water VOM should be integrated into the policy makers CPP considerations.

Conclusions This study provides an integrated analysis to support water resources planning for low-carbon electricity generation. The preliminary results indicate that while adding CCS to a coal-fired power plant would significantly increase plant-level water use (with


magnitudes depending on the power plant, cooling system, and CCS designs), the overall increase in water use could be small in comparison to fleet-level water use in the CPP rate implementation. In contrast, a shift from coal to natural gas and renewable energy for low-carbon electricity generation would reduce the state-level water use by 24% from the current makeup water level. Modeling the implementation of the mass-based and rate-based CPP in New Mexico in deterministic and stochastic forms demonstrates seven points. Firstly, the choice of approaches taken to meet the CPP will produce different results for the overall CO2 mass emitted or the intensity at which it is emitted for the same scenario; there is not necessarily a one-to-one correspondence between the emitted mass when the rate-based approach is taken to comply with the CPP and the emission intensity when the mass-based approach is taken to comply with the CPP. As such, the most cost-effective solution may be dependent of the approach choice. Secondly, this cost-effective scenario depends upon the metric chosen to do so. The compliance solution that leads to the least increase in LCOE may not be the result in the lowest cost of avoidance. Thirdly, building new renewables generation sources to produce ERCs may lead to over generation, when all electricity sources are considered. This over generation may be minimized in the mass-based approach. Fourthly, it may be possible for the states to achieve the required CO2 mitigation at the avoidance cost that the EPA suggests. Simulations of avoidance cost for the rate-based approach with the state option suggest that this cost may be similar to estimated $30/ton of CO2, when uncertainty in the natural gas price and non-zero ERC prices are considered. Fifthly, the projected higher natural gas price may lead to a change in dispatch order. With a sufficiently high ERC price, the dispatch order may be corrected so that NGCC plants are dispatched before coal-fired EGUs; however, it is unlikely that the dispatch order can be corrected with the MA price, unless the MA allocation process is designed to allow this. Sixthly, CCS with an auxiliary natural gas boiler at 40% capture in the mass-based approach may be an economically viable mitigation option in the CPP when the cost-effective metric is the percent increase in the fleet LCOE. When the metric is the CO2 avoidance cost for the fossil fuel fleet, the choice of this scenario depends upon the ERC price for the rate-based approach. For New Mexico, ERC prices above $46/MWh make CCS economically viable when compared to the methodology outlined by the EPA to meet the rate-based standard. Therefore, further studies to determine the distribution of possible ERC/MA prices and using this information to model the uncertainty in the ERC/MA and natural gas prices to determine the most cost-effective mitigation choices for the CPP are warranted.


References Apt, Jay, and Paulina Jaramillo. Variable Renewable Energy and the Electricity Grid. Routledge, 2014. Barbose, G., S. Weaver, and N. Darghouth. "Tracking the Sun VII: An Historical Summary of the Installed Price of Photovoltaics in the United States from 1998 to 2013." Lawrence Berkeley National Laboratory, Berkeley, CA (2014). Black, Veatch. "Cost and Performance Data for Power Generation Technologies." National Renewable Energy Laboratory, Golden, CO (2012). Bolinger, M., Weaver, S., & Zuboy, J. (2015). Is $50/MWh Solar for Real? Falling Project Prices and Rising Capacity Factors Drive Utility-Scale PV Toward Economic Competitiveness. Progress in Photovoltaics: Research and Applications. doi:10.1002/pip.2630 - Report Number: LBNL-183129 “Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility Generating Units; Final Rule,” Title 40 Code of Federal Regulations, Part 60. 2015 ed. Chapman, J.; Lantz, E.; Denholm, P.; Felker, F.; Heath, G.; Mai, T.; Tegen, S. (2012). "Wind Energy Technologies," Chapter 11. National Renewable Energy Laboratory. Renewable Electricity Futures Study, Vol. 2, Golden, CO: National Renewable Energy Laboratory; pp. 11-1 – 11-63. Colorado Energy. Our Clients. www.coloradoenergy.com. n.d. 2012. Web. 2 February. 2016. < http://www.coloradoenergy.com/clients.html>. Danko, Pete. ”New Mexico Joins The Geothermal Power Ranks.” The New York Times, 16 Jan., 2014. Web. 27 Aug. 2015. Diehl, Timothy H., and Melissa A. Harris. "Withdrawal and consumption of water by thermoelectric power plants in the United States, 2010." US Geological Survey Scientific Investigations Report 5184 (2014). Electric Power Research Institute (EPRI). Air-cooled condenser design, specification, and operation guidelines; Report No. 1007688; Electric Power Research Institite: Palo Alto, CA, December 2005. Energy Information Administration. Analysis of the Impacts of the Clean Power Plan. www.eia.gov. n.d. 22 May. 2015. Web. 18 October. 2015. <http://www.eia.gov/analysis/requests/powerplants/cleanplan/>. Energy Information Administration. Annual Energy Outlook 2014; April 2014. Energy Information Administration. Assumptions to the Annual Energy Outlook 2014; June 2014. Energy Information Administration. Annual Energy Outlook 2015; April 2015. Energy Information Administration. Electrical Power Annual; www.eia.gov. n.d. 12 December. 2013. Web. 24 December. 2014. <http://www.eia.gov/electricity/annual/>.

http://www.coloradoenergy.com/

http://www.coloradoenergy.com/clients.html

http://www.eia.gov/

http://www.eia.gov/analysis/requests/powerplants/cleanplan/

http://www.eia.gov/

http://www.eia.gov/electricity/annual/


Energyy Information Administration. Form EIA-860 detailed data. www.eia.gov. n.d. n.p. Web. 20 August. 2015. <http://www.eia.gov/electricity/data/eia860/>. Energyy Information Administration. Form EIA-923 detailed data. www.eia.gov. n.d. n.p. Web. 20 August. 2015. <http://www.eia.gov/electricity/data/eia923/>. Environmental Protection Agency. “Carbon Pollution Standards.” http://www2.epa.gov. 6 November. 2014. n.p. Web. 20 November. 2014. <http://www2.epa.gov/carbon-pollution-standards/clean-power-plan-proposed-rule-technical-documents#NODA>. Environmental Protection Agency. Analysis of the Proposed Clean Power Plan. www2.epa.gov. 15 October. 2015. n.p. Web. 24 November. 2015. <http://www.epa.gov/airmarkets/powersectormodeling/cleanpowerplan.html>. Environmental Protection Agency. Clean Power Plan Final Rule Technical Documents. http://www2.epa.gov. 13 August. 2015. n.p. Web. 27 August. 2014. < http://www2.epa.gov/cleanpowerplan/clean-power-plan-final-rule-technical-documents>. Environmental Protection Agency. Clean Power Plan Proposed Rule: GHG Abatement Measures. http://www2.epa.gov. 21 June. 2014. 10 June. 2014. Web. 17 November. 2014. <http://www2.epa.gov/sites/production/files/2014-06/documents/20140602tsd-ghg-abatement-measures.pdf >. Environmental Protection Agency. Clean Power Plan State Goal Visualizer. http://www2.epa.gov. 26 August. 2015. n.p. Web. 26 August. 2015. <http://www2.epa.gov/cleanpowerplantoolbox>. Environmental Protection Agency. eGRID. http://www.epa.gov. 5 August. 2014. n.p. Web. 20 October. 2014. <http://www.epa.gov/cleanenergy/energy-resources/egrid/>. Environmental Protection Agency. EPA Power Sector Modeling Platform v.5.13. www.epa.gov. 25 June. 2014. n.p. Web. 20 October. 2014. <http://www.epa.gov/airmarket/progsregs/epa-ipm/BaseCasev513.html>. Environmental Protection Agency. EPA’s Power Sector Modeling Platform v.5.15. http://www.epa.gov. 3 August. 2015. n.p. Web. 27 August. 2015. <http://www.epa.gov/powersectormodeling/psmodel515.html>. Farmington, New Mexico (City of). Capital Improvement Plan Supplemental Book: July 1, 2015 to July 30, 2016. http://fmtn.org. 25 Nov. 2015. City of Farmington. Web. 14 March. 2016. <http://fmtn.org/DocumentCenter/Home/View/6780>. Federal Energy Regulatory Commission. Form 1- Electric Annual Utility Report. http://www.ferc.gov. 25 April. 2012. n.p. Web. 20 October. 2014. <http://www.ferc.gov/docs-filing/forms/form-1/viewer-instruct.asp>. “Federal Plan Requirements for Greenhouse Gas Emissions From Electric Utility Generating Units Constructed on or Before January 8, 2014; Model Trading Rules; Amendments to Framework Regulations; Proposed Rule,” Title 40 Code of Federal Regulations, Parts 60, 62, and 78. 2015 ed.

http://www.eia.gov/


http://www.eia.gov/



http://www2.epa.gov/carbon-pollution-standards/clean-power-plan-proposed-rule-technical-documents#NODA

http://www2.epa.gov/carbon-pollution-standards/clean-power-plan-proposed-rule-technical-documents#NODA

http://www.eia.gov/

http://www.epa.gov/airmarkets/powersectormodeling/cleanpowerplan.html





http://www2.epa.gov/sites/production/files/2014-06/documents/20140602tsd-ghg-abatement-measures.pdf

http://www2.epa.gov/sites/production/files/2014-06/documents/20140602tsd-ghg-abatement-measures.pdf



http://www.epa.gov/

http://www.epa.gov/cleanenergy/energy-resources/egrid/


http://www.epa.gov/airmarket/progsregs/epa-ipm/BaseCasev513.html

http://www.epa.gov/

http://fmtn.org/DocumentCenter/Home/View/6780

http://www.ferc.gov/

http://www.ferc.gov/docs-filing/forms/form-1/viewer-instruct.asp


Feldman, David, et al. "Photovoltaic System Pricing Trends: Historical, Recent, and Near-Term Projections–2014 Edition." Golden, CO: National Renewable Energy Laboratory, PR-6A20-62558. Accessed August 20 (2015): 2014. Groves, Jack, Todd Krankkala, and Greg Nugent. "Afton Generating Station, Including Unique Hybrid Cooling." ASME 2009 Power Conference. American Society of Mechanical Engineers, 2009. L. Carroll and R. Weston Hudkins, Advanced Pipeline Design, [Online]. Available: http://www.ou.edu/class/che-design/a-design/projects-2009/Pipeline%20Design.pdf. Macknick, J., et al. "Operational water consumption and withdrawal factors for electricity generating technologies: a review of existing literature." Environmental Research Letters 7.4 (2012): 045802. National Energy Technology Laboratory. Coal-Fired Power Plants in the United States: Examination of the Costs of Retrofitting with CO2 Capture Technology and the Potential for Improvements in Efficiency; DOE/NETL- 402/102309. NETL; 2010 http://www.netl.doe.gov/energy-analyses/refshelf/PubDetails.aspx?Action=View&Source=Main&PubId=289 (accessed Dec. 31, 2013). National Energy Technology Laboratory. Estimating Plant Costs Using Retrofit Difficulty Factors; DOE/NETL- 341/082313, Pittsburgh, PA, August 2013. National Energy Technology Laboratory. The Western Wind and Solar Integration Study Phase 2; www.nrel.gov. 19 September. 2014. n.p. Web. 20 November. 2014. <http://www.nrel.gov/docs/fy13osti/55588.pdf>. National Energy Technology Laboratory. “NATCARB/ATLAS” www.netl.doe.gov. n.d. n.p. Web. 1 December. 2014. <http://www.netl.doe.gov/research/coal/carbon-storage/natcarb-atlas>. National Oceanic and Atmospheric Administration. NOAA National Centers for Environmental Information. http://gis.ncdc.noaa.gov. n.d. n.p. Web. 27 August. 2015. <http://gis.ncdc.noaa.gov/map/viewer/#app=cdo>. National Renewable Energy Laboratory. Renewable Electricity Futures Scenario Viewer. http://www.nrel.gov. n.d. n.p. Web. 20 August. 2015. <http://www.nrel.gov/analysis/re_futures/data_viewer/#>. Newcomer, Adam, and Jay Apt. "Storing syngas lowers the carbon price for profitable coal gasification." Environmental science & technology 41.23 (2007): 7974-7979. Parker, Nathan. "Using natural gas transmission pipeline costs to estimate hydrogen pipeline costs." (2004). PCL Construction. “LunaEnergy Facility—600 MW Combined Cycle Power Plant.” http://www.pcl.com. n.d. 2016. Web. 2 February. 2016. < http://www.pcl.com/Projects-that-Inspire/Pages/Luna-Energy-Facility-600-MW-Combined-Cycle-Power-Plant.aspx>.



http://www.sciencedirect.com/science?_ob=RedirectURL&_method=externObjLink&_locator=url&_cdi=277910&_issn=18766102&_origin=article&_zone=art_page&_plusSign=%2B&_targetURL=http%253A%252F%252Fwww.netl.doe.gov%252Fenergy-analyses%252Frefshelf%252FPubDetails.aspx%253FAction%25253DView%252526Source%25253DMain%252526PubId%25253D289

http://www.sciencedirect.com/science?_ob=RedirectURL&_method=externObjLink&_locator=url&_cdi=277910&_issn=18766102&_origin=article&_zone=art_page&_plusSign=%2B&_targetURL=http%253A%252F%252Fwww.netl.doe.gov%252Fenergy-analyses%252Frefshelf%252FPubDetails.aspx%253FAction%25253DView%252526Source%25253DMain%252526PubId%25253D289


http://www.nrel.gov/docs/fy13osti/55588.pdf

http://www.netl.doe.gov/



http://gis.ncdc.noaa.gov/

http://gis.ncdc.noaa.gov/map/viewer/#app=cdo


http://www.pcl.com/


Public Service Company of New Mexico. “PNM Integrated Resource Plan: 2014 - 2033.” https://www.pnm.com. n.d. n.p. Web. 20 March. 2016. <https://www.pnm.com/irp>. Rao, Anand B., and Edward S. Rubin. "Identifying cost-effective CO2 control levels for amine-based CO2 capture systems." Industrial & engineering chemistry research 45.8 (2006): 2421-2429. Rode, David C., and Paul S. Fischbeck. "The Value of Using Coal Gasification as a Long-Term Natural Gas Hedge for Ratepayers." (2006). Rubin, Edward S., Anand B. Rao, and Michael B. Berkenpas. Development and application of optimal design capability for coal gasification systems. Carnegie-Mellon University, 2007. Rubin, Edward S., et al. “Technical Documentation: The Economics of CO2 Transport by Pipeline and Storage in Saline Aquifers and Oil Reservoirs.” Carnegie Mellon University. Accessible at: http://www.cmu.edu/epp/iecm/IECM_Publications/2008ra%20McCoy%20et%20al,%20IECM%20Trans%20&%20Storage%20Tech.pdf (2008). Rubin, Edward S., et al. “Modeling of Integrated Environmental Control Systems for Coal-Fired Power Plants.” www.cmu.edu. n.d. n.p. Web. 20 October. 2014. <http://www.cmu.edu/epp/iecm/iecm_doc.html>. Sanyal, Subir K. "Cost of geothermal power and factors that affect it." Proceedings World Geothermal Congress. 2005. Sargent & Lundy 2009, “Coal-Fired Power Plant Heat Rate Reductions.” SL-009597, Final Report, January 2009, available at http://www.epa.gov/airmarkets/resource/docs/coalfired.pdf. Solar Energy Industries Association. SEIA Major Solar Projects List Update. http://www.seia.org. n.d. n.p. Web. 20 August. 2015. <http://www.seia.org/research-resources/major-solar-projects-list/seia-major-solar-projects-list-update>. Tidball, Rick, et al. "Cost and performance assumptions for modeling electricity generation technologies." Contract 303 (2010): 275-3000. Tidwell, Vincent C., et al. "Mapping water availability, projected use and cost in the western United States." Environmental Research Letters 9.6 (2014): 064009. Tri-State Generation and Transmission Association Inc. “Integrated Resource Plan/Electric Resource Plan.” http://www.tristate.coop. 12 November. 2015. Tri-State Generation and Transmission Association, Inc. Web. 20 March. 2016. <http://www.tristategt.org/resourceplanning/documents/Tri-State_IRP-ERP_Final.pdf>. Webster, Mort, Pearl Donohoo, and Bryan Palmintier. "Water-CO2 trade-offs in electricity generation planning." Nature Climate Change 3.12 (2013): 1029-1032. Wiser, Ryan, and Mark Bolinger. "2013 Wind Technologies Market Report." (2014).

https://www.pnm.com/



http://www.cmu.edu/


http://www.seia.org/


Zhai, Haibo, Edward S. Rubin, and Peter L. Versteeg. "Water use at pulverized coal power plants with postcombustion carbon capture and storage." Environmental science & technology 45.6 (2011): 2479-2485. Zhai, H., K. Kietzke, and E. S. Rubin. "IECM Technical Documentation: Probabilistic Comparative Assessment Using the IECM." (2012).


Appendix A: Tables

Table A1 List of New Mexico fossil fuel power plants modeled in the IECM.

Table A2 Annual, site-specific temperature (°F) used in the IECM.

Table A3 Annual, site-specific percent relative humidity used in the IECM.

Table A4 Distributions for probabilistic 2030, site-specific temperature and relative humidity used in the IECM.

Table A5 USGS estimation of 2010 site-specific water source and annual water consumption.

Table A6 Estimation of alternative water availability in New Mexico (Tidwell et al., 2014).

Table A7 Estimation of alternative water cost in New Mexico (Tidwell et al., 2014).

Table A8 Historical CO2 emission rate intensity (lbs/MWh) and total emissions (millions of tons of CO2) for New Mexico fossil fuel fleet.

Table A9 Historical power generation for New Mexico, inclusive of hydropower.

Table A10 Coefficients of installed utility photovoltaic capital cost scaling factor (2010$/W) regression model as a function of summertime peak capacity (MW): 𝒂 × (𝑴𝑾)𝒃.

Table A11 Coefficients of future installed utility photovoltaic capital cost (2010$/W) regression model as a function of duration from 2009 (MW): 𝒂 × (𝒅𝒖𝒓𝒂𝒕𝒊𝒐𝒏)𝒃.

Table A12 Coefficients of utility photovoltaic operating and maintenance costs (2010$/kW) regression model as a function of duration from 2009 (MW): 𝒂 × (𝒅𝒖𝒓𝒂𝒕𝒊𝒐𝒏)𝒃.

Table A13 Normalized values for wind power O&M costs for years subsequent to installation. The O&M cost for any given year is determined by multiplying the appropriate normalized value by the initial O&M cost for the year of installation, as given by Wiser and Bolinger (2014) and the AEO 2014. The normalized values for service after year twelve are equivalent to those for year twelve.

Table A14 Coefficients of geothermal capital cost scaling factor (2010$/MW) regression model as a function of capacity (MW): 𝒂 × 𝒆−𝒃∗(𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚−𝟒).

Table A15 Coefficients of geothermal operating and maintenance costs (2010$/MW) regression model as a function of capacity (MW): 𝒂 × 𝒆−𝒃∗𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚.

Table A16 Peak summertime atmospheric conditions for dry-cooling and hybrid-cooling NGCC plants.

Table A17 As received properties of proxy coals used in the IECM.

Table A18 As received properties of natural gas used in the IECM.

Table A19 Annual, delivered site-specific prices ($/MMBtu) for coal used in the IECM.


Table A20 Annual, delivered site-specific prices ($/MMBtu) for natural gas used in the IECM.

Table A21 Parameters for stochastic model.

Table A22 Scenario 1 net generation, LCOE, and makeup water for New Mexico generation sources for state goal, rate-based approach in 2030.





Table A27 Scenario 1 net generation, LCOE, and makeup water for New Mexico generation sources for technology standard, rate-based approach in 2030.



Table A30 Scenario 1 net generation, LCOE, and makeup water for New Mexico generation sources for mass-based approach in 2030.



Table A33 Replacement availability of alternative water sources as supplementary makeup water for the wet cooling system at each site for the state option of the rate-based approach for New Mexico fossil fuel generation sources. The resulting cost is the total cost for the plant makeup water inclusive of cooling water and the associated cost for the water derived from the alternative, if any. The makeup water cost for the current water source is taken as $0/kgal.

Table A34 Replacement availability of alternative water sources as supplementary makeup water for the wet cooling system at each site for the technology option of the rate-based approach for New Mexico fossil fuel generation sources. The resulting cost is the total cost for the plant makeup water inclusive of cooling water and the associated cost for the water derived from the alternative, if any. The makeup water cost for the current water source is taken as $0/kgal.

Table A35 Replacement availability of alternative water sources as supplementary makeup water for the wet cooling system at each site for the mass-based approach for New Mexico fossil fuel generation sources. The resulting cost is the total cost for the plant makeup water inclusive of cooling water and the associated cost for the water derived from the alternative, if any. The makeup water cost for the current water source is taken as $0/kgal.


Table A36 Impact of the availability and cost of alternative water sources on the levelized cost of water and wet cooling system VOM at each site for the state option of the rate-based approach for New Mexico fossil fuel generation sources. The resulting water cost is the total cost for the plant makeup water inclusive of cooling water and the associated cost for the water derived from the alternative, if any. The makeup water cost for the current water source is taken as $0/kgal. If an alternative water source is not available, the conditions are designated as “NA.”

Table A37 Impact of the availability and cost of alternative water sources on the levelized cost of water and wet cooling system VOM at each site for the mass-based approach for New Mexico fossil fuel generation sources. The resulting water cost is the total cost for the plant makeup water inclusive of cooling water and the associated cost for the water derived from the alternative, if any. The makeup water cost for the current water source is taken as $0/kgal. If an alternative water source is not available, the conditions are designated as “NA.”


Table A1 List of New Mexico fossil fuel power plants modeled in the IECM.

Plant Name ORSIS Plant Code

Unit ID Type Fuel Type Cooling Type

San Juan 2451 1 Pulverized Coal-steam

Bituminous Recirculating tower







Escalante 87 1 Pulverized Coal-steam

Sub-bituminous

Recirculating tower

Afton 55210 NGCC Natural Gas Hybrid cooling tower

Bluffview 55977 NGCC Natural Gas Recirculating tower

Hobbs 56458 NGCC Natural Gas Dry cooling

Luna 55343 NGCC Natural Gas Recirculating tower

Table A2 Annual, site-specific temperature (°F) used in the IECM. Plant name WBAN86 ID 2010 2012 2030

San Juan 23090 52.8 54.7 53.5 Escalante 93057 49.4 53.6 49.2

Afton 23078 60.5 60.1 61.4 Bluffview 23090 52.8 54.7 53.1

Hobbs 93033 62.8 67 63.9 Luna 23078 60.7 62.4 61.4

86 WBAN is the acronym for the Weather Bureau Army Navy identification number.


Table A3 Annual, site-specific percent relative humidity used in the IECM. Plant name WBAN87 ID 2010 2012 2030

San Juan 23090 41.5 33.5 37.5 Escalante 93057 43.9 36.2 41.5

Afton 23078 35.6 29.9 32.7 Bluffview 23090 41.5 33.5 38.3

Hobbs 93033 44.2 32.9 37.4 Luna 23078 35.6 29.9 32.7

Table A4 Distributions for probabilistic 2030, site-specific temperature and relative humidity used in the IECM.

Plant name WBAN88 ID Temperature Distribution (°F)

Relative Humidity Distribution (%)

San Juan 23090 Uniform (22.99, 79.92) Uniform (31.47, 1.161) Escalante 93057 Uniform (47.87, 50.43) Uniform (37.98, 45.11)

Afton 23078 Uniform (59.74, 62.95) Uniform (28.46, 36.97) Bluffview 23090 Uniform (22.99, 79.92) Uniform (31.47, 1.161)

Hobbs 93033 Uniform (62.29, 65.51) Uniform (27.05, 47.68) Luna 23078 Uniform (59.74, 62.95) Uniform (28.46, 36.97)

Table A5 USGS estimation of 2010 site-specific water source and annual water consumption. Plant name Water source Consumption (Mgal/day)

San Juan Surface water 11.2 Escalante Ground water 2.17

Afton Ground water 0.18 Bluffview Municipality 0.27

Hobbs NA NA Luna Ground water 1.18

87 Ibid. 88 Ibid.


Table A6 Estimation of alternative water availability in New Mexico (Tidwell et al., 2014). Plant name Water

basin (8-digit

HUC)

Unappropriated ground water

(gal/day)

Appropriated water

(gal/day)

Brackish ground water

(gal/day)

Wastewater (gal/day)

San Juan 14080105 33,411,421 1,694,101 1,748,617 0 Escalante 13020207 0 92,925 1,373,739 0

Afton 13030102 0 2,013,988 24,980,246 5,332,885 Bluffview NA NA NA NA NA

Hobbs NA NA NA NA NA Luna 13030202 0 1,743,840 999,210 2,357,135

Table A7 Estimation of alternative water cost in New Mexico (Tidwell et al., 2014).

Plant name Water basin (8-

digit HUC)

Unappropriated ground water

($/kgal)

Appropriated water ($/kgal)

Brackish ground water

($/kgal)

Wastewater ($/kgal)

San Juan 14080105 0.60 0.73 5.11 NA Escalante 13020207 NA 0.73 3.65 NA

Afton 13030102 NA 0.73 3.44 0.66 Bluffview NA NA NA NA NA

Hobbs N/A NA NA NA NA Luna 13030202 NA 0.73 3.13 1.91

Table A8 Historical CO2 emission rate intensity (lbs/MWh) and total emissions (millions of tons of CO2) for New Mexico fossil fuel fleet.89

2010 2012 Plant type Unit

number Intensity

(lbs/MWh) Mass

(Mton) Intensity

(lbs/MWh) Mass

(Mton) OGST NA 1,313 0.97 1,313 0.97 San Juan 1 2,332 2.73 2,332 2.47 San Juan 2 2,314 2.24 2,356 2.32 San Juan 3 2,261 2.92 2,350 2.23 San Juan 4 2,354 3.85 2,354 3.89 Escalante NA 2,436 2.04 2,436 1.38 Afton NGCC NA 973 0.14 927 0.22 Bluffview NGCC

NA 937 0.20 936 0.22

Hobbs NGCC

NA 897 1.34 897 1.34

Luna NGCC NA 905 0.86 913 0.83 Total NA NA 17.29 NA 16.79

89 OGST emissions and generation held constant at 2012 levels.


Table A9 Historical power generation for New Mexico, inclusive of hydropower. Plant Type/Year 2010 2012

Pulverized Coal (TWh) 11.8 11.4 NGCC (TWh) 5.6 5.7 OGST (TWh) 1.5 1.5 Wind (TWh) 1.8 2.2 Solar (TWh) 0.0 0.3

Hydropower (TWh) 0.2 0.2 Total Net Generation (TWh) 20.9 21.3

Percent Renewable Energy (%) 9.8 12.9 Table A10 Coefficients of installed utility photovoltaic capital cost scaling factor (2010$/W) regression model as

a function of summertime peak capacity (MW): 𝒂 × (𝑴𝑾)𝒃. Model Coefficient a b R2

Values 2.1027 -0.03157 0.908

Table A11 Coefficients of future installed utility photovoltaic capital cost (2010$/W) regression model as a function of duration from 2009 (years): 𝒂 × (𝒅𝒖𝒓𝒂𝒕𝒊𝒐𝒏)𝒃.

Model Coefficient a b R2

Values 4.7994 -0.3341 0.982 Table A12 Coefficients of utility photovoltaic operating and maintenance costs (2010$/kW) regression model as

a function of duration from 2009 (years): 𝒂 × (𝒅𝒖𝒓𝒂𝒕𝒊𝒐𝒏)𝒃. Model Coefficient a b R2

Values 39.738 -0.28146 0.843


Table A13 Normalized values for wind power O&M costs for years subsequent to installation. The O&M cost for any given year is determined by multiplying the appropriate normalized value by the initial O&M cost for the year of installation, as given by Wiser and Bolinger (2014) and the AEO 2014. The normalized values for

service after year twelve are equivalent to those for year twelve. Service year after installation

Installation year

0 1 2 3 4 5 6 7 8 9 10 11 12

1998 1 1 0.990 0.923 1.130 1.381 1.607 1.763 1.336 1.609 1.622 2.995 2.490 1999 1 1 0.990 0.923 1.130 1.381 1.607 1.763 1.336 1.609 1.622 2.995 2.490 2000 1 1 0.990 0.923 1.130 1.381 1.607 1.763 1.336 1.609 1.622 2.995 2.490 2001 1 1 0.990 0.923 1.130 1.381 1.607 1.763 1.336 1.609 1.622 2.995 2.490 2002 1 1 0.990 0.923 1.130 1.381 1.607 1.763 1.336 1.609 1.622 2.995 2.490 2003 1 1 0.990 0.923 1.130 1.381 1.607 1.763 1.336 1.609 1.622 2.995 2.490 2004 1 1 0.990 0.923 1.130 1.381 1.607 1.763 1.336 1.609 1.622 2.995 2.490 2005 1 1 0.930 0.942 1.015 0.984 0.892 2.024 1.231 1.609 1.622 2.995 2.490 2006 1 1 0.930 0.942 1.015 0.984 0.892 2.024 1.231 1.609 1.622 2.995 2.490 2007 1 1 0.930 0.942 1.015 0.984 0.892 2.024 1.231 1.609 1.622 2.995 2.490 2008 1 1 0.930 0.942 1.015 0.984 0.892 2.024 1.231 1.609 1.622 2.995 2.490 2009 1 1 1.013 1.127 1.340 0.984 0.892 2.024 1.231 1.609 1.622 2.995 2.490 2010 1 1 1.013 1.127 1.340 0.984 0.892 2.024 1.231 1.609 1.622 2.995 2.490 2011 1 1 1.013 1.127 1.340 0.984 0.892 2.024 1.231 1.609 1.622 2.995 2.490 2012 1 1 1.013 1.127 1.340 0.984 0.892 2.024 1.231 1.609 1.622 2.995 2.490 2013 1 1 1.013 1.127 1.340 0.984 0.892 2.024 1.231 1.609 1.622 2.995 2.490 2014 1 1 1.013 1.127 1.340 0.984 0.892 2.024 1.231 1.609 1.622 2.995 2.490 2015 1 1 1.013 1.127 1.340 0.984 0.892 2.024 1.231 1.609 1.622 2.995 2.490 2016 1 1 1.013 1.127 1.340 0.984 0.892 2.024 1.231 1.609 1.622 2.995 2.490


Table A14 Coefficients of geothermal capital cost scaling factor (2010$/MW) regression model as a function of

capacity (MW): 𝒂 × 𝒆−𝒃∗(𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚−𝟒). Model Coefficient a b R2

Values 10,437 -0.002498 1

Table A15 Coefficients of geothermal operating and maintenance costs (2010$/MW) regression model as a function of capacity (MW): 𝒂 × 𝒆−𝒃∗𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚.

Model Coefficient a b R2 Values 22.36 -0.002514 1

Table A16 Peak summertime atmospheric conditions for dry-cooling and hybrid-cooling NGCC plants.

Plant name Temperature (°F) Relative Humidity (%) Afton 80.25 33.46

Hobbs 82.49 39.29

Table A17 As received properties of proxy coals used in the IECM. Variable Illinois #6 bituminous Wyoming Powder River

Basin sub-bituminous

Heating value (Btu/lb) 11,670 8,340 Carbon (% wt.) 63.75 48.18

Hydrogen (% wt.) 4.5 3.31 Oxygen (% wt.) 6.88 11.87

Chlorine (% wt.) 0.29 0.01 Sulfur (% wt.) 2.51 0.37

Nitrogen (% wt.) 1.25 0.7 Ash (% wt.) 9.7 5.32

Moisture (% wt.) 11.12 30.24


Table A18 As received properties of natural gas used in the IECM. Variable Units Natural gas

Heating value Btu/lbs 22,480 Methane (CH4) vol. % 93.1 Ethane (C2H6) vol. % 3.2

Propane (C3H8) vol. % 1.1 Carbon Dioxide (C2O) vol. % 1.0

Oxygen (O2) vol. % 0.0 Nitrogen (N2) vol. % 1.6

Hydrogen Sulfide (H2S) vol. % 0.0 Density lbs/cu ft 0.0456

Table A19 Annual, delivered site-specific prices ($/MMBtu) for coal used in the IECM. Plant name 2010 2012 2030

San Juan 2.496 2.452 2.371 Escalante 2.275 2.257 2.182

Table A20 Annual, delivered site-specific prices ($/MMBtu) for natural gas used in the IECM. Plant name 2010 2012 2030

Afton 5.475 3.612 6.547 Bluffview 4.6 3.403 6.168

Hobbs 4.911 3.403 6.168 Luna 4.892 3.352 6.076


Table A21 Parameters for stochastic model. Variable Units Nominal Value Distribution

Delivered coal price % -3.32 Normal (2.5, 5) Delivered natural gas

price % 81.3 Normal (-0.7, 34.2)

ERC 2010$/MWh 25 Uniform (nominal ± 25) Solar capital cost

(2007-2009) $/kW Year dependent Uniform (nominal ± 3.24)

Solar capital cost (2007-2009)

$/kW Year dependent Uniform (nominal ± 1050.9)

Solar capital cost (2010)


Solar capital cost (2011)


Solar capital cost (2012-2016)


Solar O&M cost $/kW Year dependent Uniform (nominal ± 3.24) Wind capital cost

(<5MW) $/kW Year dependent Uniform

(nominal ± 1335.0) Wind capital cost

(5-20MW) $/kW Year dependent Uniform





(nominal ± 525.9) Wind O&M cost $/kW Year dependent Uniform (nominal ± 4.21)




LCOE ($/MWh)

Plant Makeup Water

(Ggal/year) OGST NA 1,476,861 NA 1.22

San Juan 1 2,366,901 45.2 2.01 San Juan 2 0 NA 0.0 San Juan 3 0 NA 0.0 San Juan 4 3,719,416 40.8 3.05 Escalante NA 1,892,950 41.3 1.54

Afton NGCC NA 1,552,000 54 0.26 Bluffview NGCC NA 427,300 103.7 0.14

Hobbs NGCC NA 3,458,000 60.7 0.0 Luna NGCC NA 3,669,000 57.9 1.11

Wind NA 6,388,011 72.5 0.0 Solar NA 3,463,368 120.4 0.0

Geothermal NA 62,677 204.7 0.0





LCOE ($/MWh)

Plant Makeup Water





Wind NA 4,974,809 69.7 0.0 Solar NA 2,697,177 106.8 0.0


Table A24 Scenario 3 net generation, LCOE, and makeup water for New Mexico generation sources for state

goal, rate-based approach in 2030. Plant name Unit

number Net Generation

(MWh) LCOE

($/MWh) Plant Makeup

Water (Ggal/year)

OGST NA 1,476,861 NA 1.22 San Juan 1 2,366,901 45.2 2.01 San Juan 2 0 NA 0.0 San Juan 3 0 NA 0.0 San Juan 4 3,719,416 62 3.55 Escalante NA 1,892,950 41.3 1.54



Wind NA 5,139,799 70.1 0.0 Solar NA 2,786,629 102.4 0.0






LCOE ($/MWh)

Plant Makeup Water





Wind NA 4,200,122 67.5 0.0 Solar NA 2,277,167 122.3 0.0





LCOE ($/MWh)

Plant Makeup Water


San Juan 1 0 NA 0.0 San Juan 2 0 NA 0.0 San Juan 3 0 NA 0.0 San Juan 4 3,108,269 43.6 2.58 Escalante NA 1,581,915 44 1.22



Wind NA 4,217,821 67.5 0.0 Solar NA 2,286,763 122.2 0.0






LCOE ($/MWh)

Plant Makeup Water





Wind NA 6,807,296 73.1 0.0 Solar NA 3,690,690 120.1 0.0


Table A28 Scenario 4 net generation, LCOE, and makeup water for New Mexico generation sources for

technology standard, rate-based approach in 2030. Plant name Unit


(MWh) LCOE


Water (Ggal/year)

OGST NA 1,476,861 NA 1.22 San Juan 1 1,157,751 60 0.94 San Juan 2 0 NA 0.0 San Juan 3 0 NA 0.0 San Juan 4 1,819,323 52.1 1.42 Escalante NA 925,922 55.3 0.72



Wind NA 4,728,263 69.1 0.0 Solar NA 2,563,508 121.7 0.0






LCOE ($/MWh)

Plant Makeup Water


San Juan 1 0 NA 0.0 San Juan 2 0 NA 0.0 San Juan 3 0 NA 0.0 San Juan 4 2,571,620 45.1 1.97 Escalante NA 1,308,794 46.8 1.01



Wind NA 4,742,905 69.1 0.0 Solar NA 2,571,446 121.6 0.0


Table A30 Scenario 1 net generation, LCOE, and makeup water for New Mexico generation sources for mass-

based approach in 2030. Plant name Unit


(MWh) LCOE


Water (Ggal/year)

OGST NA 1,476,861 NA 1.22 San Juan 1 1,851,212 49.6 1.56 San Juan 2 0 NA 0.0 San Juan 3 0 NA 0.0 San Juan 4 2,909,048 44.2 2.37 Escalante NA 1,480,524 45.5 1.20



Wind NA 3,677,671 65.4 0.0 Solar NA 1,993,911 123.1 0.0






LCOE ($/MWh)

Plant Makeup Water


San Juan 1 2,138,288 47.9 2.09 San Juan 2 0 NA 0.0 San Juan 3 0 NA 0.0 San Juan 4 3,360,167 70 2.74 Escalante NA 1,710,115 43.8 1.60

Afton NA 1,552,000 54 0.26 Bluffview NA 427,300 103.7 0.14

Hobbs NA 3,458,000 60.7 0.0 Luna NA 3,669,000 57.9 1.11 Wind NA 3,050,122 62 0.0 Solar NA 1,653,675 124.4 0.0


Table A32 Scenario 3 net generation, LCOE, and makeup water for New Mexico generation sources for mass-

based approach in 2030. Plant name Unit


(MWh) LCOE


Water (Ggal/year)

OGST NA 1,476,861 NA 1.22 San Juan 1 2,100,263 48.1 2.01 San Juan 2 0 NA 0.0 San Juan 3 0 NA 0.0 San Juan 4 3,300414 62 3.15 Escalante NA 1,679,705 42.5 1.21

Afton NA 1,552,000 54 0.26 Bluffview NA 427,300 103.7 0.14

Hobbs NA 3,458,000 60.7 0.0 Luna NA 3,669,000 57.9 1.11 Wind NA 3,133,244 62.5 0.0 Solar NA 1,698,741 124.2 0.0



Table A33 Replacement availability of alternative water sources as supplementary makeup water for the wet cooling system at each site for the state option of the rate-based approach for New Mexico fossil fuel generation sources. The resulting cost is the total cost for the plant makeup water inclusive of cooling water and the

associated cost for the water derived from the alternative, if any. The makeup water cost for the current water source is taken as $0/kgal. Unappropriated

Groundwater Unappropriated

Water Brackish

Groundwater Wastewater

Plant Percent

available Cost

($/kgal) Percent

available Cost

($/kgal) Percent

available Cost

($/kgal) Percent

available Cost

($/kgal) San Juan 100 0.49 14.8 0.09 15.3 0.64 0 NA Escalante 0 NA 2.8 0.02 39.8 1.19 0 NA

Afton 0 NA 100 0.59 100 2.78 100 0.53 Bluffview NA NA NA NA NA NA NA NA

Hobbs NA NA NA NA NA NA NA NA Luna 0 0 70.9 0.42 40.6 1.02 95.8 1.47

Table A34 Replacement availability of alternative water sources as supplementary makeup water for the wet cooling system at each site for the technology option of the

rate-based approach for New Mexico fossil fuel generation sources. The resulting cost is the total cost for the plant makeup water inclusive of cooling water and the associated cost for the water derived from the alternative, if any. The makeup water cost for the current water source is taken as $0/kgal. Unappropriated


Water Brackish


Plant Percent

available Cost

($/kgal) Percent

available Cost

($/kgal) Percent

available Cost

($/kgal) Percent

available Cost

($/kgal) San Juan 100 0.49 14.8 0.09 15.3 0.64 0 0 Escalante 0 0 2.8 0.02 39.8 1.19 0 0

Afton 0 0 100 0.59 100 2.78 100 0.53 Bluffview NA NA NA NA NA NA NA NA



Table A35 Replacement availability of alternative water sources as supplementary makeup water for the wet cooling system at each site for the mass-based approach for New Mexico fossil fuel generation sources. The resulting cost is the total cost for the plant makeup water inclusive of cooling water and the associated cost for the

water derived from the alternative, if any. The makeup water cost for the current water source is taken as $0/kgal. Unappropriated


Water Brackish


Plant Percent

available Cost

($/kgal) Percent

available Cost

($/kgal) Percent

available Cost

($/kgal) Percent

available Cost

($/kgal) San Juan 100 0.49 19.1 0.12 20 0.83 0 0 Escalante 0 0 3.6 0.02 51.2 1.53 0 0

Afton 0 0 100 0.59 100 2.78 100 0.53 Bluffview NA NA NA NA NA NA NA NA


Table A36 Impact of the availability and cost of alternative water sources on the levelized cost of water and wet cooling system VOM at each site for

the state option of the rate-based approach for New Mexico fossil fuel generation sources. The resulting water cost is the total cost for the plant makeup water inclusive of cooling water and the associated cost for the water derived from the alternative, if any. The makeup water cost for the current water

source is taken as $0/kgal. If an alternative water source is not available, the conditions are designated as “NA.” Unappropriated


Water Brackish


Plant Cost

($/MWh) Water VOM

($/MWh)

Cost ($/MWh)

Water VOM

($/MWh)

Cost ($/MWh)

Water VOM

($/MWh)

Cost ($/MWh)

Water VOM

($/MWh) San Juan 1 0.34 0.86 0.063 0.58 0.45 0.96 NA NA San Juan 4 1.23 3 0.23 2 1.60 3.38 NA NA

Escalante NA NA 0.025 0.93 1.5 2.40 NA NA Afton NA NA 0.1 0.82 0.46 1.19 0.09 0.81 Luna NA NA 0.15 0.54 0.37 0.72 0.53 0.86


Table A37 Impact of the availability and cost of alternative water sources on the levelized cost of water and wet cooling system VOM at each site for the mass-based approach for New Mexico fossil fuel generation sources. The resulting water cost is the total cost for the plant makeup water inclusive of cooling water and the

associated cost for the water derived from the alternative, if any. The makeup water cost for the current water source is taken as $0/kgal. If an alternative water source is not available, the conditions are designated as “NA.”

Unappropriated Groundwater

Unappropriated Water

Brackish Groundwater

Wastewater

Plant Cost ($/MWh)

Water VOM

($/MWh)

Cost ($/MWh)

Water VOM

($/MWh)

Cost ($/MWh)

Water VOM

($/MWh)

Cost ($/MWh)

Water VOM

($/MWh) San Juan 1 0.63 1.64 0.16 1.16 1.075 2.082 NA NA San Juan 4 0.96 2.41 0.24 1.69 1.63 3.080 NA NA

Escalante NA NA 0.02 0.78 1.17 1.93 NA NA Afton NA NA 0.1 0.82 0.46 1.19 0.09 0.81 Luna NA NA 0.15 0.54 0.37 0.72 0.53 0.86


Appendix B: Figures

Figure B1 EIA projection of the impact of the CPP on the Henry Hub natural gas price (EIA, 2015).

Figure B2 Historical forecast error in the natural gas wellhead price by the EIA Annual Energy Outlook as determined by Rode and Fischbeck (2006).

Figure B3 Historical utility photovoltaic capital costs as documented by Barbose, et al. (2014). Figure B4 Projected range in installed photovoltaic capital costs based upon expert elicitations, as

documented by Feldman, et al. (2014).

Figure B5 Historical and projected utility photovoltaic operation and maintenance costs, as documented by Bolinger,, M., Weaver, S., & Zuboy, J. (2015).

Figure B6 Historical wind turbine capital costs, as documented by Wiser, Ryan, & Bolinger. (2014). Figure B7 Historical wind turbine operation and maintenance costs, as documented by Wiser, Ryan, &

Bolinger. (2014).

Figure B8 Cumulative distribution function estimations of the plant makeup water rate in 2030 for the New Mexico fossil fuel fleet under the rate-based approach for the CPP. Estimations are based upon an uncertainty simulation in the IECM that only accounts for the uncertainty in the average annual temperature and humidity.

Figure B9 Cumulative distribution function estimations of the plant makeup water rate in 2030 for the New Mexico fossil fuel fleet under the mass-based approach for the CPP. Estimations are based upon an uncertainty simulation in the IECM that only accounts for the uncertainty in the average annual temperature and humidity.

Figure B10 Cumulative distribution function estimations of the cooling makeup water rate in 2030 for the New Mexico fossil fuel fleet under the rate-based approach for the CPP. Estimations are based upon an uncertainty simulation in the IECM that only accounts for the uncertainty in the average annual temperature and humidity.

Figure B11 Cumulative distribution function estimations of the cooling makeup water rate in 2030 for the New Mexico fossil fuel fleet under the mass-based approach for the CPP. Estimations are based upon an uncertainty simulation in the IECM that only accounts for the uncertainty in the average annual temperature and humidity.

Figure B12 Sensitivity of avoidance cost to ERC price for San Juan EGU #4 with and without 40% capture CCS system. When fitted with CCS, increasing the capacity factor of the main boiler provides the steam and electricity for the system. The two lines intersect when the ERC price is $33.3/MWh and result in an avoidance cost of $53/ton.

Figure B13 Sensitivity of avoidance cost to MA price for San Juan EGU #4 with and without 40% capture CCS system. When fitted with CCS, an auxiliary natural gas boiler provides the steam and electricity for the system. The two lines intersect when the MA price is $221.9/ton of CO2 and result in an avoidance cost of $89.7/ton of CO2.


Figure B14 Sensitivity of avoidance cost to MA price for San Juan EGU #4 with and without 40% capture CCS system. When fitted with CCS, increasing the capacity factor of the main boiler provides the steam and electricity for the system. The two lines intersect when the MA price is $96.8/ton of CO2 and result in an avoidance cost of $65.4/ton of CO2.

Figure B15 Sensitivity of avoidance cost to ERC price and percent increase in natural gas price relative to 2012 for San Juan EGU #4 with a 40% capture CCS system. Increasing the capacity factor of the main boiler provides the steam and electricity for the CCS system. Areas shaded red represent the combinations of ERC price and natural gas price increases that have an avoidance cost of $53/ton of CO2 or greater. In these regions, CCS without an auxiliary natural has a cost effective avoidance cost.

Figure B16 Sensitivity of avoidance cost to MA price and percent increase in natural gas price relative to 2012 for San Juan EGU #4 with a 40% capture CCS system. An auxiliary natural gas boiler provides the steam and electricity for the CCS system. Areas shaded red represent the combinations of MA price and natural gas price increases that have an avoidance cost of $89.7/ton of CO2 or greater. In these regions, CCS with an auxiliary natural gas boiler has a cost effective avoidance cost.

Figure B17 Sensitivity of avoidance cost to MA price and percent increase in natural gas price relative to 2012 for San Juan EGU #4 with a 40% capture CCS system. Increasing the capacity factor of the main boiler provides the steam and electricity for the CCS system. Areas shaded red represent the combinations of MA price and natural gas price increases that have an avoidance cost of $65.4/ton of CO2 or greater. In these regions, CCS without an auxiliary natural has a cost effective avoidance cost.

Figure B18 Distribution of predicted 2030 price of natural gas for Hobbs NGCC plant. This variation is driven by the historical AEO forecast error documented by Rode and Fischbeck (2006). Approximately 1% of the simulations result in a natural gas price that is less expensive than that for 2012, which leads in part to the negative avoidance costs.

Figure B19 Distribution of 2030 NGCC plants generation-weighted average LCOE. The variation is due to fluctuation in natural gas price.

Figure B20 Relationship between the percent reduction in CO2 mass and CO2 avoidance cost for fossil fuel emission intensity under the different scenarios from the implementation of the CPP in New Mexico for 2030. The fossil fuel emission intensity is the summation of the pounds of CO2 emitted by all endogenous sources divided by all endogenous net generation, exclusive of renewable sources existing prior to 2030. The mass emission is the summation of the tons of CO2 emitted by all endogenous sources. The scenarios for the rate-based approach cluster according to scenario characteristics. Scenario T1 dominates all other scenarios for achieving CPP compliance at the lowest avoidance cost.

Figure B21 Relationship between the percent reduction in CO2 mass and the increase in LCOE under the different scenarios from the implementation of the CPP in New Mexico for 2030. The mass emission is the summation of the tons of CO2 emitted by all endogenous sources. The scenarios for the rate-based approach cluster according to scenario characteristics. Scenarios M1 and M3 dominate all other scenarios for achieving CPP compliance at the lowest increase in LCOE.

Figure B22 Impact of alternative makeup cooling water availability and cost on the levelized water cost and the water system VOM in the implementation of the state option for the rate-based approach on the New Mexico PC EGUs and NGCC plants.


Figure B1 EIA projection of the impact of the CPP on the Henry Hub natural gas price (EIA, 2015).


Figure B2 Historical forecast error in the natural gas wellhead price by the EIA Annual Energy Outlook as determined by

Rode and Fischbeck (2006).

Figure B3 Historical utility photovoltaic capital costs as documented by Barbose, et al. (2014).


Figure B4 Projected range in installed photovoltaic capital costs based upon expert elicitations, as documented by

Feldman, et al. (2014).

Figure B5 Historical and projected utility photovoltaic operation and maintenance costs, as documented by Bolinger,, M.,

Weaver, S., & Zuboy, J. (2015).


Figure B6 Historical wind turbine capital costs, as documented by Wiser, Ryan, & Bolinger. (2014).

Figure B7 Historical wind turbine operation and maintenance costs, as documented by Wiser, Ryan, & Bolinger. (2014).


Figure B8 Cumulative distribution function estimations of the plant makeup water rate in 2030 for the New Mexico fossil fuel fleet under the rate-based approach for the CPP. Estimations are based upon an uncertainty simulation in the IECM

that only accounts for the uncertainty in the average annual temperature and humidity.

Figure B9 Cumulative distribution function estimations of the plant makeup water rate in 2030 for the New Mexico fossil

fuel fleet under the mass-based approach for the CPP. Estimations are based upon an uncertainty simulation in the IECM that only accounts for the uncertainty in the average annual temperature and humidity.


Figure B10 Cumulative distribution function estimations of the cooling makeup water rate in 2030 for the New Mexico

fossil fuel fleet under the rate-based approach for the CPP. Estimations are based upon an uncertainty simulation in the IECM that only accounts for the uncertainty in the average annual temperature and humidity.

Figure B11 Cumulative distribution function estimations of the cooling makeup water rate in 2030 for the New Mexico

fossil fuel fleet under the mass-based approach for the CPP. Estimations are based upon an uncertainty simulation in the IECM that only accounts for the uncertainty in the average annual temperature and humidity.


Figure B12 Sensitivity of avoidance cost to ERC price for San Juan EGU #4 with and without 40% capture CCS system. When fitted with CCS, increasing the capacity factor of the main boiler provides the steam and electricity for the system.

The two lines intersect when the ERC price is $33.3/MWh and result in an avoidance cost of $53/ton of CO2.

Figure B13 Sensitivity of avoidance cost to MA price for San Juan EGU #4 with and without 40% capture CCS system. When fitted with CCS, an auxiliary natural gas boiler provides the steam and electricity for the system. The two lines

intersect when the MA price is $221.9/ton of CO2 and result in an avoidance cost of $89.7/ton of CO2.


Figure B14 Sensitivity of avoidance cost to MA price for San Juan EGU #4 with and without 40% capture CCS system. When fitted with CCS, increasing the capacity factor of the main boiler provides the steam and electricity for the system.

The two lines intersect when the MA price is $96.8/ton of CO2 and result in an avoidance cost of $65.4/ton of CO2.

Figure B15 Sensitivity of avoidance cost to ERC price and percent increase in natural gas price relative to 2012 for San Juan EGU #4 with a 40% capture CCS system. Increasing the capacity factor of the main boiler provides the steam and

electricity for the CCS system. Areas shaded red represent the combinations of ERC price and natural gas price increases that have an avoidance cost of $53/ton of CO2 or greater. In these regions, CCS without an auxiliary natural

has a cost effective avoidance cost.


Figure B16 Sensitivity of avoidance cost to MA price and percent increase in natural gas price relative to 2012 for San

Juan EGU #4 with a 40% capture CCS system. An auxiliary natural gas boiler provides the steam and electricity for the CCS system. Areas shaded red represent the combinations of MA price and natural gas price increases that have an avoidance cost of $89.7/ton of CO2 or greater. In these regions, CCS with an auxiliary natural gas boiler has a cost

effective avoidance cost.

Figure B17 Sensitivity of avoidance cost to MA price and percent increase in natural gas price relative to 2012 for San

Juan EGU #4 with a 40% capture CCS system. Increasing the capacity factor of the main boiler provides the steam and electricity for the CCS system. Areas shaded red represent the combinations of MA price and natural gas price increases that have an avoidance cost of $65.4/ton of CO2 or greater. In these regions, CCS without an auxiliary natural has a cost

effective avoidance cost.


Figure B18 Distribution of predicted 2030 price of natural gas for Hobbs NGCC plant. This variation is driven by the

historical AEO forecast error documented by Rode and Fischbeck (2006). Approximately 1% of the simulations result in a natural gas price that is less expensive than that for 2012, which leads in part to the negative avoidance costs.

Figure B19 Distribution of 2030 NGCC plants generation-weighted average LCOE. The variation is due to fluctuation in

natural gas price.


Figure B20 Relationship between the percent reduction in CO2 mass and CO2 avoidance cost for fossil fuel emission intensity under the different scenarios from the implementation of the CPP in New Mexico for 2030. The fossil fuel

emission intensity is the summation of the pounds of CO2 emitted by all endogenous sources divided by all endogenous net generation, exclusive of renewable sources existing prior to 2030. The mass emission is the summation of the tons of

CO2 emitted by all endogenous sources. The scenarios for the rate-based approach cluster according to scenario characteristics. Scenario T1 dominates all other scenarios for achieving CPP compliance at the lowest avoidance cost.

Figure B21 Relationship between the percent reduction in CO2 mass and the increase in LCOE under the different

scenarios from the implementation of the CPP in New Mexico for 2030. The mass emission is the summation of the tons of CO2 emitted by all endogenous sources. The scenarios for the rate-based approach cluster according to scenario

characteristics. Scenarios M1 and M3 dominate all other scenarios for achieving CPP compliance at the lowest increase in LCOE.


Figure B22 Impact of alternative makeup cooling water availability and cost on the levelized water cost and the water system VOM in the implementation of the state option for the rate-based approach on the New Mexico PC EGUs and

NGCC plants.


Appendix C: Scaling the IECM Simulation Model for NGCC Plant Profile from NEEDS and eGRID Database: LCOE, Fuel, Plant Makeup Water

Equation C1: Generic Power Regression Expression

Equation C2: LCOE Capacity Scaling Factor

Equation C3: Fuel and Water VOM Heat Rate Scaling Factor

Equation C4: LCOE Equation for Capacity and VOM Heat Rate Scaling Factors

Equation C5: Fuel and Water VOM Equation for Capacity and Heat Rate Scaling Factors

Equation C6: Plant Makeup Water use (tons/hr) Equation for Capacity and Heat Rate Scaling Factors

Table C1 Simulated output metrics for the IECM default NGCC plant configured with a wet cooling tower, cooling system as a function of net capacity. The IECM default settings are used for all model inputs, with the exceptions that no cost is associated with the water, all costs are reported in constant 2010 dollars, and GE-7FA model gas turbines are used.

Table C2 Simulated output metrics for the IECM default NGCC plant configured with an air-cooled condenser, cooling system as a function of net capacity. The IECM default settings are used for all model inputs, with the exceptions that no cost is associated with the water, all costs are reported in constant 2010 dollars, and GE-7FA model gas turbines are used.

Table C3 Coefficients for wet cooling tower, cooling system LCOE (2010$/MWh) regression model as a function of capacity (MW): 𝒂 × (𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚)𝒃.

Table C4 Coefficients for air condenser cooling system LCOE (2010$/MWh) regression model as a function of capacity (MW): 𝒂 × (𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚)𝒃.

Table C5 Coefficients for wet cooling tower, cooling system water VOM (2010$/MWh) regression model as a function of capacity (MW): 𝒂 × (𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚)𝒃.

Table C6 Coefficients for air condenser cooling system water VOM (2010$/MWh) regression model as a function of capacity (MW): 𝒂 × (𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚)𝒃.

Table C7 Coefficients for wet cooling tower, cooling system fuel VOM (2010$/MWh) regression model as a function of capacity (MW): 𝒂 × (𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚)𝒃. Coefficient b is independent of cooling system type and b may be taken as 0.

Table C8 NGCC plant characteristics, as reported in the NEEDS and eGRID 2010 databases, to determine the scaling factors for the IECM simulated NGCC plants. The net heat rate from the IECM simulation is also given for reference.

Table C9 Coefficients for wet cooling tower, cooling system water plant makeup water use (tons/hr) regression model as a function of capacity (MW): 𝒂 × (𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚)𝒃.


Figure C1 LCOE regression trendline for the IECM default NGCC plant configured with a wet cooling tower, cooling system. Data points for increasing capacity correlate to the incremental addition of one GE-7FA model gas turbines.

Figure C2 LCOE regression trendline for the IECM default NGCC plant configured with an air-cooled condenser, cooling system. Data points for increasing capacity correlate to the incremental addition of one GE-7FA model gas turbines.

Figure C3 Water VOM regression trendline for the IECM default NGCC plant configured with a wet cooling tower, cooling system. Data points for increasing capacity correlate to the incremental addition of one GE-7FA model gas turbines.

Figure C4 Cooling system VOM regression trendline for the IECM default NGCC plant configured with an air-cooled condenser, cooling system. Data points for increasing capacity correlate to the incremental addition of one GE-7FA model gas turbines.

Figure C5 Fuel VOM regression trendline for the IECM default NGCC plant configured with a wet cooling tower, cooling system. The trendline and regression coefficients for the air-cooled condenser, cooling system show similar insensitivity to capacity variations. Data points for increasing capacity correlate to the incremental addition of one GE-7FA model gas turbines.

Figure C6 Plant makeup water use (tons/hour) regression trendline for the IECM default NGCC plant configured with a wet cooling tower, cooling system. Data points for increasing capacity correlate to the incremental addition of one GE-7FA model gas turbines. No such relationship exists for the air-cool condenser case.


The capacity, steam cycle heat rate, and parasitic load for the simulated NGCC plant cannot be directly adjusted in the IECM to reflect those for the modeled plant to yield the reported net generation, capacity factor, and net heat rate. Therefore, this study adjusts the model and number of gas turbines and the capacity factor in the IECM to achieve the reported net generation. The gas turbine model is assumed to depend upon the age of the plant and is chosen in all cases to be the GE-7FA model, while the number of turbines is taken as that described in the eGRID database. Selecting the model and number of gas turbines then determines the simulated net capacity, which is larger than the peak summertime capacity reported in the NEEDS database. Given the larger capacity and a more efficient heat rate, the capacity factor is decreased from that given in the eGRID database, so that the reported net generation can be simulated. Some of the financial and performance output metrics for this larger and more efficient simulated plant must then be scaled to achieve the metrics for a plant simulated with the full reported operational characteristics. This scaling factor will depend upon the configuration of the NGCC plant cooling system and the output metric. The associated scaling factors related to the difference in capacity are determined by simulating multiple NGCC plants with a given cooling system configuration and multiple numbers of the same gas turbine model (Table C1 and Table C2).90 Here, the output for the desired metric is then used as the dependent variable in a power regression to determine how that metric changes as a function of the net capacity (Equation [C1]). The ratio of the simulated and the modeled capacities is raised to the resulting coefficient, b, to derive the scaling factor by which the simulated metric is multiple to determine the modeled metric. The scaling factors related to the difference in net heat rate are done in a similar manner with b taken as one.

Equation C1: 𝑌𝑖 = 𝑎 ∗ 𝑋𝑖𝑏

where 𝑌 = dependent metric, 𝑋 = independent metric, 𝑎 = coefficient 1, 𝑏 = coefficient 2, 𝑖 = number of turbines.

90 The IECM default settings are used for all model inputs, with the exceptions that no cost is associated with the water, and all costs are reported in constant 2010 dollars.


Table C1 Simulated output metrics for the IECM default NGCC plant configured with a wet cooling tower, cooling system as a function of net capacity. The IECM default settings are used for all model inputs, with the exceptions that no cost is associated with the water, all costs are reported in constant 2010 dollars, and GE-7FA model gas turbines are used.

Turbines (Number)

Capacity (MW)

LCOE ($/MWh)

Fuel VOM ($/MWh)

Water VOM ($/MWh)

Plant Makeup Water (tons/hr)

1 271.2 65.08 48 0.4593 349.9 2 542.4 63.8 48.01 0.4509 699.8 3 813.6 63.31 48.01 0.4479 1,050 4 1,085 63.04 48.02 0.4462 1,400 5 1,356 62.84 48.02 0.445 1,750

Table C2 Simulated output metrics for the IECM default NGCC plant configured with an air-cooled condenser, cooling system as a function of net capacity. The IECM default settings are used for all model inputs, with the exceptions that no cost is associated with the water, all costs are reported in constant 2010 dollars, and GE-7FA model gas turbines are used.

Turbines (Number)

Capacity (MW)

LCOE ($/MWh)

Fuel VOM ($/MWh)

Cooling VOM ($/MWh)

1 266.7 67.38 48.49 0.806 2 533.4 69.93 48.47 0.7913 3 800.1 65.39 48.5 0.7857 4 1,067 65.1 48.51 0.7827 5 1,333 64.91 48.5 0.7807

LCOE and Fuel The LCOE in Table C1 and Table C2 is composed of two levelized and annualized components—the fixed costs and the variable operating and maintenance costs—each of which can be functions of the capacity (MW) and net heat rate (Btu/kWh) of the NGCC plant. Therefore, regressing the LCOE on the capacity accounts for the variation of each of these components due to variations in capacity. This relationship to the modeled LCOE is then expressed as the product of the simulated LCOE and the ratio of the capacities raised to the power coefficient from the regression (Equation [C2]). The variation due to heat rate cannot be regressed since the heat rate cannot be varied in the IECM for the NGCC plants; however, we can assume that the relationship for the fuel and water VOM is linear with regard to heat rate91 so that the scaling factor can be defined as the ratio of the heat rates (Equation [C3]). The addition of the heat rate adjustment to the VOM for the fuel and water to the capacity adjustment to the LCOE results in the modeled LCOE (Equation [C4]). The modeled VOM for the fuel and water can also be determined as the sum of the capacity and net heat rate adjusted components in a similar manner to that for the LCOE (Equation [C5]). This expression is required to compare the marginal cost for the different power sources that is used to determine the dispatch order.

91 The plants with higher hear rates will consume more fuel and water to produce the same amount of electricity as plants with lower heat rates. This increased consumption is seen to be a proportional increase.


Equation C2:

𝐿𝐶𝑂𝐸𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦 𝑎𝑑𝑗𝑢𝑠𝑡𝑒𝑑 = 𝐿𝐶𝑂𝐸𝐼𝐸𝐶𝑀 ∗ �𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝑚𝑜𝑑𝑒𝑙𝑒𝑑𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝐼𝐸𝐶𝑀

�𝑏

, [𝐶2] Equation C3: 𝑉𝑂𝑀ℎ𝑟 𝑎𝑑𝑗𝑢𝑠𝑡𝑒𝑑 = 𝑉𝑂𝑀𝐼𝐸𝐶𝑀 ∗ �ℎ𝑒𝑎𝑡 𝑟𝑎𝑡𝑒𝑚𝑜𝑑𝑒𝑙𝑒𝑑

ℎ𝑒𝑎𝑡 𝑟𝑎𝑡𝑒 𝐼𝐸𝐶𝑀� , [C3]

Equation C4:

𝐿𝐶𝑂𝐸𝑚𝑜𝑑𝑒𝑙𝑒𝑑 = 𝐿𝐶𝑂𝐸𝐼𝐸𝐶𝑀 ∗ �𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝑚𝑜𝑑𝑒𝑙𝑒𝑑𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝐼𝐸𝐶𝑀

�𝑏

+ 𝑉𝑂𝑀ℎ𝑟 𝑎𝑑𝑗𝑢𝑠𝑡𝑒𝑑𝑤𝑎𝑡𝑒𝑟 + 𝑉𝑂𝑀ℎ𝑟 𝑎𝑑𝑗𝑢𝑠𝑡𝑒𝑑

𝑓𝑢𝑒𝑙 , [𝐶4] Equation C5:

𝑉𝑂𝑀𝑚𝑜𝑑𝑒𝑙𝑒𝑑 = �𝑉𝑂𝑀𝐼𝐸𝐶𝑀 ∗ �𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝑚𝑜𝑑𝑒𝑙𝑒𝑑

𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝐼𝐸𝐶𝑀�𝑏

� ∗ �ℎ𝑒𝑎𝑡 𝑟𝑎𝑡𝑒𝑚𝑜𝑑𝑒𝑙𝑒𝑑

ℎ𝑒𝑎𝑡 𝑟𝑎𝑡𝑒 𝐼𝐸𝐶𝑀� , [𝐶5]

where hr is the net heat rate, 𝑏 = coefficient 2. To use these relationships, one must determine the appropriate coefficients and the heat rate and capacity ratios. The power coefficients for the above equations are dependent upon the cooling system type and the modeled metric. Graphing each capacity dependent relationship indicates that the coefficient for each metric (with the exception of the fuel VOM) is negative and less than one, with a variation over the data range of less than 10% of the nominal value, (Figures C1-C4). The fuel VOM is insensitive to the plant capacity for both types of cooling systems (Figure C5), however. Tables C3-C7 give the corresponding model regression coefficients. The net plant heat rates and summertime peak capacities for these plants are given in Table C8.


Figure C1 LCOE regression trendline for the IECM default NGCC plant configured with a wet cooling tower, cooling system. Data points for increasing capacity correlate to the incremental addition of one GE-7FA model gas turbines.

Figure C2 LCOE regression trendline for the IECM default NGCC plant configured with an air-cooled condenser, cooling system. Data points for increasing capacity correlate to the incremental addition of one GE-7FA model gas

turbines.


Figure C3 Water VOM regression trendline for the IECM default NGCC plant configured with a wet cooling tower, cooling system. Data points for increasing capacity correlate to the incremental addition of one GE-7FA model gas

turbines.

Figure C4 Cooling system VOM regression trendline for the IECM default NGCC plant configured with an air-cooled

condenser, cooling system. Data points for increasing capacity correlate to the incremental addition of one GE-7FA model gas turbines.


Figure C5 Fuel VOM regression trendline for the IECM default NGCC plant configured with a wet cooling tower, cooling system. The trendline and regression coefficients for the air-cooled condenser, cooling system show similar

insensitivity to capacity variations. Data points for increasing capacity correlate to the incremental addition of one GE-7FA model gas turbines.

Table C3 Coefficients for wet cooling tower, cooling system LCOE (2010$/MWh) regression model as a function of capacity (MW): 𝒂 × (𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚)𝒃.

Model Coefficient a b R2 Values 73.4 -0.02181 0.976

Table C4 Coefficients for air condenser cooling system LCOE (2010$/MWh) regression model as a function of capacity (MW): 𝒂 × (𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚)𝒃.


Table C5 Coefficients for wet cooling tower, cooling system water VOM (2010$/MWh) regression model as a function of

capacity (MW): 𝒂 × (𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚)𝒃. Model Coefficient a b R2

Values 0.5118 -0.01968 0.971

Table C6 Coefficients for air condenser cooling system water VOM (2010$/MWh) regression model as a function of

capacity (MW): 𝒂 × (𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚)𝒃. Model Coefficient a b R2

Values 0.8992 -0.01994 0.973


Table C7 Coefficients for wet cooling tower, cooling system fuel VOM (2010$/MWh) regression model as a function of capacity (MW): 𝒂 × (𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚)𝒃. Coefficient b is independent of cooling system type and b may be taken as 0.


Table C8 NGCC plant characteristics, as reported in the NEEDS and eGRID 2010 databases, to determine the scaling factors for the IECM simulated NGCC plants. The net heat rate from the IECM simulation is also given for reference.

NGCC Plant

Turbines (Number)

Summertime Peak Capacity

(MW)

Net Plant Heat Rate (Btu/kWh)

IECM Simulation Net Plant Heat Rate

(Btu/kWh) Afton 1 236 8,631 7,082

Bluffview 1 65 8,295 7,087 Hobbs 2 526 7,449 7,170

Luna 2 558 7,418 7,082


Plant Makeup Water While the levelized cost of the water used in the cooling system should decrease with increasing capacity due to economies of scale, the rate at which this water is used should not decrease. Plotting the simulated plant makeup water use for the as a function of capacity indicates that this is an increasing, linear relationship (Figure C6). Regressing on these values yields a power coefficient, b, of one (Table C9). Therefore, the scaling factor for the simulated rate of water use will take the form of the water VOM in the wet cooling system (Equation C6).

Figure C6 Plant makeup water use (tons/hour) regression trendline for the IECM default NGCC plant configured with a

wet cooling tower, cooling system. Data points for increasing capacity correlate to the incremental addition of one GE-7FA model gas turbines. No such relationship exists for the air-cool condenser case.

Table C9 Coefficients for wet cooling tower, cooling system water plant makeup water use (tons/hr) regression model as a

function of capacity (MW): 𝒂 × (𝒄𝒂𝒑𝒂𝒄𝒊𝒕𝒚)𝒃. Model Coefficient a b R2

Values 1.2888 1.00189 1.000 Equation C6: 𝑀𝑎𝑘𝑒𝑢𝑝 𝑊𝑎𝑡𝑒𝑟𝑚𝑜𝑑𝑒𝑙𝑒𝑑

= �𝑀𝑎𝑘𝑒𝑢𝑝 𝑊𝑎𝑡𝑒𝑟𝐼𝐸𝐶𝑀 ∗ �𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝑚𝑜𝑑𝑒𝑙𝑒𝑑

𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦𝐼𝐸𝐶𝑀�𝑏

� ∗ �ℎ𝑒𝑎𝑡 𝑟𝑎𝑡𝑒𝑚𝑜𝑑𝑒𝑙𝑒𝑑

ℎ𝑒𝑎𝑡 𝑟𝑎𝑡𝑒 𝐼𝐸𝐶𝑀� , [𝐶6]


Appendix D: Panhandle A and B Equations and Constants Equation D1: Panhandle A

Equation D2: Panhandle B

Table D Constants used in Panhandle A and B equations.


Panhandle A The Panhandle A equation calculates the flow of gas from the physical properties of the gas and the characteristics of the pipe. Here, the flow is affected indirectly by the roughness of the pipe in two ways. Firstly, the roughness increases the effective length of the pipe through turbulent losses. Secondly, the efficiency of the pipe is also decreased for similar reasons related to friction. Each of these conditions requires the diameter of the pipe to increase to maintain the required flow of NG. Equation D1:

𝑄𝑔 = 435.87 × E(𝑇𝑏𝑃𝑏

)1.0788 �𝑃12 − 𝑒𝑠𝑃22

𝐿𝑒𝑍𝑇𝑓𝐺0.8538�0.5394

× 𝑑2.6182

𝐿𝑒 =𝐿(𝜀𝑠 − 1)

𝑠

𝑠 = 0.0375𝐺 �∆𝐻𝑇𝑓𝑍

�

where Qg = gas-flow rate, MMscfD, E = efficiency factor of pipe, d = pipe inside diameter, in., ε = pipe roughness, in., Pb = base pressure, psia, P1 = upstream pressure, psia, P2 = downstream pressure, psia, L = length, miles, ΔH = elevation difference, ft, Tb = base temperature, °F, Tf = temperature of flow, °F, G = gas gravity, and Z = compressibility factor of gas.


Panhandle B The Panhandle B equation calculates the flow of gas from the physical properties of the gas and the characteristics

of the pipe. Here, the flow is affected indirectly by the roughness of the pipe in two ways. Firstly, the roughness

increases the effective length of the pipe through turbulent losses. Secondly, the efficiency of the pipe is also

decreased for similar reasons related to friction. Each of these conditions requires the diameter of the pipe to

increase to maintain the required flow of NG. This equation differs from the Panhandle A equation in the constants,

which result from adaption for calculating flows with higher Reynolds numbers.

Equation D2:

𝑄𝑔 = 737 × E(𝑇𝑏𝑃𝑏

)1.02 �𝑃12 − 𝑒𝑠𝑃22

𝐿𝑒𝑍𝑇𝑓𝐺0.961�0.51

× 𝑑2.53

𝐿𝑒 =𝐿(𝜀𝑠 − 1)

𝑠

𝑠 = 0.0375𝐺 � ∆𝐻

𝑇𝑓𝑍�

where Qg = gas-flow rate, MMscfD, E = efficiency factor of pipe, d = pipe inside diameter, in., ε = pipe roughness, in., Pb = base pressure, psia, P1 = upstream pressure, psia, P2 = downstream pressure, psia, L = length, miles, ΔH = elevation difference, ft, Tb = base temperature, °F, Tf = temperature of flow, °F, G = gas gravity, and Z = compressibility factor of gas.


Table D Constants used in Panhandle A and B equations. Parameter Symbol Units Value Pipe efficiency factor E none 0.95

Pipe roughness ε inch 0.000700 Base Pressure Pb pound per

square inch absolute (psia)

14.70

Upstream pressure P1 psia 1,500 Downstream pressure P2 psia 800

Elevation difference ΔH feet 100 Base temperature Tb °F 60 Flow temperature Tf °F 70

Gas gravity G none 0.6 Compressibility factor Z none 0.8793

Viscosity μ centipoise 0.01