chapter new technologies for generating and storing

Chapter 4

New Technologies for Generatingand Storing Electric Power

Advanced Batteries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Summary occurrent Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . .

List of TablesTable No.

4-1. Developing Technologies Considered in OTA’s Analysis . .4-2. Selected Alternative Generating and Storage Technologies:

Typical Sizes and Applications . . . . . . . . . . . . . . . . . . . . . . .

P’l<qt’

. . . . . . . . . . . . . . 7.5

. . . . . . . . . . . . . . . . . . . 764-3. Advantages and Disadvantages of Solar Thermal Electric Systems. . . . . . . . . . . . 884-4. Developing Technologies: Major Electric Plants Installed

or Under Construction by May 1, 1985 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .129

Chapter 4

New Technologies for Generatingand Storing Electric Power

INTRODUCTION

A number of new technologies for generatingand storing electricity are being developed asalternatives to large-scale, long lead-time conven-tional powerplants. Of increasing interest aretechnologies which are small scale and highly ef-ficient, which are capable of using alternativefuels, and/or which impose substantially lowerenvironmental impacts than conventional gen-erating options.

This chapter focuses on new technologieswhich, while generally not fully mature today,could f igure importantly in the electric-supplytechnologies insta//ed in the 7990s. Not exam-ined in detail are those technologies which al-ready are considered technically mature or thosewhich are very unlikely to achieve wide deploy-ment during that time. Only grid-connected ap-plications are considered. The chapter also doesnot examine closely technologies which recentlyhave been covered in other OTA reports.1

The new technologies covered in this reportare summarized in table 4-1. In table 4-2, thetechnologies are grouped according to size andapplication, along with their primary competitors.They range in size from units less than 1 MWeto units greater than 250 MWe. The technologiescan be divided between those in which the elec-trical power production is available upon utilitydemand (dispatchable) and those where it is not.In the table, dispatchable applications are furtherbroken down according to base, intermediate,and peak load applications. Among applicationswhere the utility cannot summon electrical poweron command are intermittent technologies (e.g.,wind turbines and direct solar equipment), when

1 These include nuclear power, convent io na I technologies usedIn cogeneratlon, and conventional equ Ipment which uses biomass;

see p. IV of th is repor t .

Table 4-1 .—Developing Technologies Considered inOTA’s Analysisa

Photovoltaics:Flat plate systems (tracking and nontracking)Concentrators

Solar thermal electric:Solar pondsCentral receiversParabolic troughsParabolic dishes

Wind turbinesGeothermal:

Dual flashBinary (large and small)

Atmospheric fluid ized-bed combustorsIntegrated gasification combined-cycleBatteries

Lead acidZinc chloride

Compressed-air energy storage (large and small)Phosphoric-acid fuel cells (large and small)

75

76 . New Electric Power Technologies: Problems and Prospects for the 1990s

Table 4-2.—Selected Alternative Generating and Storage Technologies: Typical Sizes and Applications

Typical configurations in the 1990s

GENERATING TECHNOLOGIESSolar Technologies

Introduction

In seeking ways to directly exploit the Sun’senergy to produce electric power, two alterna-tives are being pursued. Solar thermal-electrictechnologies rely on the initial conversion of lightenergy to thermal energy; the heat typically isconverted to mechanical energy and then to elec-tric power. Alternatively, photovoltaic cells maybe used to directly convert the light energy intoelectrical energy. Between the two technologies,many variations are being developed, each withits own combination of cost, performance, andrisk, and each with its own developmentalhurdles.

Most solar electric technologies promise note-worthy advantages over conventional technol-ogies. 2 These include:

1. Free, secure, and renewable energy source:These are especially important attributeswhen contrasted with price and availabil-ity uncertainties of oil and natural gas.

2. Widely available energy source: Figure 7-11 in chapter 7 illustrates the distributionof the solar resource in the United States.

~Note that some solar technologies may use a su pplernental fuelsuch as oil, gas, or biomass. In such instances, the hybrid systemwill not have some of the advantages and disadvantages listed; andthe system will possess some advantages and disadvantages notlisted.

Ch. 4—New Technologies for Generating and Storing Electric Power . 77

3.

4.

5.6.7.

8.

9.

10,

No off-site, fuel-related impacts: The deliv-ery of solar energy imposes no environ-mental impacts off-site, unlike the deliveryof conventional fuels which frequently re-quire a series of steps (exploration, extrac-tion, refining, transportation, etc.) whicheach impose environmental impacts quitedistinct from those at the power plant siteitself.No fuel supply infrastructure required: T h edelivery of solar energy does not requirethe development of an ancillary fuel-supplyinfrastructure as is the case with conven-tional fuels. A solar plant can operate re-motely at any site; a coal plant could notdo so without the prior development of aninfrastructure which extracts, refines, andtransports the coal to the plant.Short lead-times.Wide range of installation sizes.Dec/ining costs: Many of t he solar thermalsystems are experiencing declining costs,a fact which reduces risk in any plans toinvest in the technologies.Relative/y small water needs: Some of thesolar technologies require little water be-yond that used for cleaning.Litt/e or no routine emissions: Other thanthermal discharges and other than run-offfrom washing operations, most solar tech-nologies do not routinely emit large quan-tities of wastes into the air, water, or soil. jSiting flexibility.

Though graced by many advantages, solar elec-tric systems–like any other generating technol-ogies—also have disadvantages. Among them are:

1. /intermittent supp/y of energy: Solar energyis subject to uncontrollable and sometimesunforeseeable variations. It is not alwaysthere when needed. Most obviously, it com-pletely ceases to be available every day forextended periods (night) or its power is con-siderably diminished anytime clouds passbetween the Sun and the surface of theEarth. in addition, seasonal and annual fluc-tuations in average solar radiation can be sig-

‘Dl~posa I of the tec h nology at the end of Its usefu I I Ifet I me may,

however, create serious ~~aste problems.

2.

3

4.

5.

6.

nificant. 4 And these fluctuations put stresson hardware and can cause control problems.Capita/ intensive: Current solar electric tech-nologies are characterized by very high cap-ital costs per kWe.Land extensive: Solar systems use a lot ofland per unit of power. where land is ex-pensive, land acquisition can greatly in-crease installation costs. where solar con-centrators are spread over a large surfacearea, soils and microclimates and local eco-systems can be affected.Water usage: Some solar thermal systemsroutinely require large quantities of water;and all likely will require periodic cleaning.Where units are used in arid areas, this maybe a problem.Exposure to the elements and to malevo-lence: Many of the system components arefully exposed. They therefore suffer fromerosion, corrosion, and other damage fromthe wind, from moisture (including hail) andcontaminants in the air, and from tempera-ture extremes. The systems also may be easytargets for vandals or saboteurs. For thesereasons most solar electric installations areenclosed by fences, often with some kindof barrier for wind protection. And wherereflective mirrors are used, they frequentlyare designed not to shatter and to withstandthe elements. A permanent security forcealso may be required where the systems aredeployed, but their land extensivenessmakes security difficult.Cost and difficuky of access: The likelihoodthat the systems will be built in remote loca-tions raises problems relating to site accessduring construction and for maintenance.Transmission access may also be difficult orexpensive to obtain.

Photovoltaics

introduction.–A photovoltaic (PV) cc// is athin wafer of semiconductors material which con-

qlt \vas reported, for example, that the solar flux at Solar Onea solar-thermal central receiver plant in California, has been 25 per-cent lower than in the base year (1 976) used for planning purposesfor the plant, This may be due to the increased atmospheric par-ticulate load imposed by recent volcanic eruptions.

‘A semiconductor IS a material characterized by a conducti~ltylying between that of an Insulator and that of a metal.

78 • New Electric Power Technologies: Problems and Prospects for the 1990s

verts sunlight directly into direct current (DC)electricity by way of the photoelectric effect, Cellsare grouped into modules, which are encapsu-lated in a protective coating. Modules may beconnected to each other into panels, which thenare affixed to a support structure, forming an ar-ray (see figures 4-1 and 4-2). The array may befixed or movable, and is oriented towards theSun. Any number of arrays may be installed toproduce electric power which, after conversionto alternating current (AC), may be fed into theelectric grid. I n a PV installation, all the compo-nents other than the modules themselves are col-lectively termed the balance-of-system.

At present, PV systems are being pursued inmany different forms. Each seeks some particu-lar combination of cost and performance for themodule and for the balance-of-system. In a con-centrator moduIe, lenses are used to focus sun-Iight received at the module’s surface onto amuch smaller surface area of cells (see figure 4-2); all available concentrator systems follow theSun with two-axis tracking systems. A flat-platemodule is one in which the total area of the cells

used is close to the total area of sunlight hittingthe exposed surface of the module. Variousmechanisms such as mirrors can be used to di-vert light from adjacent spaces onto the exposedsurface of the modules. Flat-plate systems maybe fixed in position or may track the Sun witheither single or two-axis tracking systems.

The parallel development of these two typesof PV modules and systems reflects a basic tech-nological problem: it is difficult to produce PVcells which are both cheap and highly efficient.Cheap cells tend to be inefficient; and highly ef-ficient cells tend to be expensive. Some PV sys-tems which are being developed for deploymentin the 1990s are emphasizing cells which are rela-tively cheap and inefficient; such cells are usedin flat-plate modules. Others are using a smallernumber of high-cost, high-efficiency cells in con-centrator systems.

In either case, if PV systems are to competewith other grid-connected generating technol-ogies in the 1990s, their cost and overall risks willhave to come down and their performance will

Figure 4.1 .—Features of a Flat-Plate Photovoltaic System

Ch. 4—New Technologies for Generating and Storing Electric Power • 79

Figure 4-2.—Schematic of a Conceptual Design for a High Concentration Photovoltaic Array

Institute, 1984) EPRI AP-3263

have to improve. There is considerable disagree-ment over which particuIar PV system is thestrongest contender for this market. Marketpenetration will depend on the current state ofthe particular variety of PV system; the potentialfor cost reductions and performance improve-ments; and on the reduction in risk perceptionamong prospective investors in grid-connectedinstalIations.

At present, two types of modules appear to bethe leading contenders. One is the flat-plate mod-ule based on tandem cells made from amorphoussilicon; the other is the concentrator module,probably using crystalline silicon.

There are a number of reasons why the con-centrator module could be the photovoltaic tech-nology of choice in central station applications

in the near term. The crystalline silicon cell is rela-tively well understood, as are the techniques formaking such cells. The cells have been manufac-tured for many years, and information on cell per-formance after extended exposure to the ele-ments is rapidly accumulating. Concentratormodules may offer a favorable combination ofcost and performance in the Southwest, whereearly central station deployment probably will begreatest. And the prospects are good that cost andperformance improvements can be made duringthe balance of the century. Many of the improve-ments do not appear to require basic technicaladvances but rather incremental improvementsand mass production.

Flat-plate modules using amorphous siliconmeanwhile are expected to continue to develop.But basic technical improvements must be made

80 . New E/ectric Power Technologies: Problems and Prospects for the 1990s

before extensive grid-connected deployment willoccur. The current technology is too inefficient.Efficiencies must be improved, and the perform-ance of the improved modules must be estab-lished over time and under actual conditions. Thiscombination of technical improvements and theneed to establish a clear, long-term performancerecord will take a substantial period of time. Be-cause of this, amorphous-silicon flat-plates maynot offer a superior choice for central station ap-plications until the latter part of the 1990s or later.There is a small possibility, however, that rapidimprovement in the cost and performance ofamorphous modules is likely and that they willcompete successfully with concentrators duringmost of the 1990s.

Regardless of technology, the commercial pros-pects for PV systems will depend heavily on con-tinued technical development and the volumeof production. Factors influencing either techni-cal development or production levels thereforewill strongly affect cost and performance in the1990s.

The Typical Grid-Connected PhotovoltaicPlant in the 1990s.—ln the 1990s, central stationapplications probably will be favored over dis-persed applications. Indeed, by May 1985, ap-proximately 19 MWe of PV power in multimega-watt central station installations were connectedto the grid in the United States—this was mostof the grid-connected PV capacity in the coun-try. This capacity was divided roughly equally be-tween concentrator and flat-plate modules. By1995, as much as 4,730 MWe could be locatedin such installations nationwide.b Capacity prob-ably will be concentrated in California, Florida,Hawaii, Arizona, and New Mexico.7

In this analysis, it is assumed that the typicalgrid-connected photovoltaic system in the 1990sis a centralized photovoltaic system (see figure4-3). Unless otherwise stated, the numbers re-

bThis IS the range pro~’lded by Pieter Bos (Polydyne Inc.), as esti-mated in a submission at the OTA Workshop on Solar Photovol-talc Power (Washington, DC, June 12, 1984) and discussed by May-cock and Sherlekar (Paul D. Maycock and Vic S. Sherlekar,Photo\ wltaic Technology, Performance, Cost and Market Forecastto 1995. A Strategic Technology& Market Analysis (Alexandria, VA:Photovoltaic Energy Systems, Inc., 1984), pp. 130-1 36.),

‘Ibid.

. .


Figure 4-3.—A View of a Recently Installed Photovoltaic Central Station

SOURCE ARCO Solar, Inc

such as dust abatement measures, vegetationcontrol efforts, and periodic cleaning of themoduIes.

The lead-time required to deploy a PV plantis potentially quite short, perhaps 2 years—in-cluding planning, licensing, permitting, construc-tion, and other elements. Construction itselfshould be quick and simple. Licensing and per-mitting should proceed very rapidly becausemany of the environmental impacts are low rela-tive to those associated with conventional tech-nologies. However, large PV plants will be newto most areas in the 1990s; and the land-extensivecharacter of the technology raises problemswhich could engender controversy, leading toregulatory delays.

System Cost and Performance.–Operat ingavailabilities of 90 to 100 percent are anticipatedfor the multimegawatt PV installation of the1990S .10 This will be affected primarily by thenumber of PCSS required and their quality. Mostoperating large PV systems today are charac-terized by operating availabilities below this—between 80 and 90 percent–usually as a resultof problems with the PCSS. I n order to reach theexpected range of operating availabilities, PCSSmust be developed which can operate reliably.

I (Ioperat I ng ay a I la bl [ Ity of I nd Ivld ua I arrays k~ I I I be between q 5

and 100 percent, depending mostly on the performance of trw ker~,if they are used. Recent experwrwe with tra( kers su~ges[~ that theiroperatl ng availahl I Itles shou Id not ta I I below that ra nm.


Such systems now are being developed and maybe available during the 1990s.

Equipment lifetimes of up to 30 years are an-ticipated, but they will depend mostly on the life-time of the modules, the trackers (if they are used)and the PCS, and the lifetimes for all three typesof equipment are still uncertain. By far the mostimportant component in this regard is the mod-ule; degradation and failures may seriouslyshorten its life.

Capacity factors will differ noticeably from sys-tem to system, depending on the general designfeatures as well as on the location of the systemand atmospheric conditions. In this analysis, ca-pacity factors ll for fixed flat-plate systems varylittle–they range from 20 to 25 percent in Bos-ton to 25 to 30 percent in Albuquerque. The ca-pacity factors for tracking flat-plate systems areassumed to range from 30 to 40 percent, ’ 2

though, this has yet to be verified nationwide.The capacity factor for concentrator systemsvaries by a larger margin by location—from 20to 25 percent in Boston and Miami, to 30 to 35percent in Albuquerque.

The modules and the balance of system (BOS)jointly determine capital costs and efficiency.Module cost and efficiency, as discussed above,depends on whether the system utilizes flat-platemodules or concentrator modules. Regardless ofmoduIe or whether the array is fixed or tracking,BOS efficiencies are likely to fall within the samerough range. The costs of the BOS, however, willvary greatly, depending on whether or not atracking system is used.

The typical multimegawatt flat-plate module PVstation in the 1990s probably will produce elec-tric power with an efficiency between 8 to 14 per-cent. Capital costs are expected to range between$1,000/kWe and $8,000/kWe in Albuquerque,and higher elsewhere in the country. Instal lationsusing concentrator modules should be moreefficient-with a 12 to 20 percent efficiency. Cap-ital costs for concentrator moduIes i n Albuquer-que should be between $1,000/kWe and $5,000/

I I Capacity factor 15 the ratio of the annual energy output ( kwe

(AC)) of a plant to the energy output (kWe (AC)) It would have had

if it operated continuously at Its nominal peak operating cond It ions.I ZOTA ~tatt’ I ntervlew with D,G. Schueler, Manager , So lar Energy

Depar tment , Sanc~ia Laboratories, Albuquerque, NM, Aug. 7, 1984.

kWe; costs will be higher in areas with lowerlevels of direct sunlight.

Cost reductions and performance improve-ments in PV systems will require the deploymentof highly automated processes capable of massproducing cells as well as efficiently producingother components, such as tracking equipmentand lenses for concentrators.

Another important element of cell costs will bethe cost of silicon. If cell costs are to be drivendown, either the quantity of silicon consumedper kilowatt-electrical of cell produced must bereduced; or silicon costs must be lowered eitherthrough new production techniques or by an ex-pansion of silicon production capacity. Morematerial-efficient cells are being developed whichrequire less silicon per kilowatt-electrical pro-duced.13 Efforts are also underway to develop sili-con production processes which can producelow-cost silicon. There is a fair chance that thesesilicon production processes will be successfullydeveloped and available in the 1990s.14 And evi-dence indicates that the additional silicon pro-duction capacity will be built when needed .15

PV plants should have low operating and main-tenance costs—probably ranging from 4 to 28mills/kWh in the 1990s. These estimates arehighly uncertain, though, and will only becomemore definite as more systems are placed in thefield. Questions about module lifetimes, trackerproblems, and difficulties with the PCS makeoperating and maintenance (O&M) cost projec-tions uncertain.

Two other areas of uncertainty may increaseO&M costs. The first is that dirt accumulating onthe modules may reduce their efficiency. ’b Rain

13A god discussion of silicon and its importance as a driving force

behind the deve lopment o f a l ternat ive PV technolog ies can be

found in: Paul D. Maycock and V IC S. Sherlekar, H-rotovo/faIc Tech-nology, Performance, Cost and Market Forecast to 199.5. A Strate-gic Technology & Market Analysis, op. cit., 1984.

IALeonard J. Reiter, A probabj/istjc Ana/ysls o?’ .$ I/ICOn C05t ( Pasa-

dena, CA: Jet Propulsion Laboratory, 1983), DOE/JPL/1012.I SRObert V. Steele, ‘ ‘strategies on Poly, Photo ~’o/taics ]ntern.?-

tiona/, vol. 11, No. 4, August/September 1984, pp. 6-8.lbFor example, in a module performance Wah JatiOn program con-

ducted by the MIT Lincoln Laboratory and the Jet Propulsion Lab-oratory, “the greatest single cause of power loss has been soil ac-cumu Iation. ’ (Edward C. Kern, Jr., and Marvin D. Pope,Development and Evaluation o(Solar Photoi oltaic System~: FinalReport (Lexington, MA: ,MIT, 1983), DOE/ET/20279-240.


has been found to be an effective cleaner, butoften the best areas for photovoltaics have littlerain. With flat-plate systems, dirt does not seemto be as much of a problem as with concentra-tors. The second problem is wind damage. Mostexisting photovoltaics and solar thermal plantshave suffered damage from wind blown sand,though methods to prevent this are being de-veloped.

Solar Thermal= Electric PowerplantsTechnology Descriptions. -Solar thermal-elec-

tric plants convert radiant energy from the Suninto thermal energy, a portion of which subse-quently is transformed into electrical energy.Among the systems, there are four which, withsome feasible combination of reduced costs andrisks and improved performance, could be de-ployed within the 1990s in competition with

other technologies and without special and ex-clusive Government subsidies. They are centralreceivers, parabolic troughs, parabolic dishes,and solar ponds. Brief descriptions of these tech-nologies are provided below.

Centra/ Rece;ver.–A central receiver is charac-terized by a fixed receiver mounted on a tower(see figure 4-6). Solar energy is reflected from alarge array of mirrors, known as heliostats, ontothe receiver. Each heliostat tracks the Sun on twoaxes. The receiver absorbs the reflected sunlight,and is heated to a high temperature. Within thereceiver is a medium (typically water, air, liquidmetal, or molten salt) which absorbs the re-ceiver’s thermal energy and transports it awayfrom the receiver, where it is used to drive a tur-bine and generator, though it first may be stored.

Parabo/ic Dishes. –parabolic dishes consist ofmany dish-shaped concentrators, each with a re-

Ch. 4—New Technologies for Generating and Storing Electric Power ● 8 5

ceiver mounted at the focal point. The concen-trated heat may be utilized directly by a heat en-gine placed at the focal point (mounted-engineparabolic dish); or a fluid may be heated at thefocal point and transmitted for remote use (re-mote-engine parabolic dish). Each dish/receiverapparatus includes a two-axis tracking device,support structures, and other equipment (seefigures 4-7, 4-8, and 4-9).

Solar Troughs.–With a parabolic trough, theconcentrators are curved in only one dimension,forming long troughs. The trough tracks the Sunon one axis, causing the trough to shift from eastto west as the Sun moves across the sky. A heattransfer medium, usually an oil at high tempera-ture (typically 200 to 400° C), is enclosed in atube located at the focal line. The typical instal-lation consists of many troughs (see figure 4-1 O).

The oil-carrying tubes located at their focal linesare connected on each end to a network of larger

pipes. The oil is circulated through the tubesalong the focal lines, flows into the larger pipes,and is pumped to a central area where it can bestored in tanks or used immediately. In eithercase, it ultimately passes through a heat ex-changer where it transfers energy to a workingfluid such as water or steam which in turn isrouted to a turbine generator. At the Solar EnergyGenerating units in southern California, the onlylarge trough installations in the United States, theoil’s heat is supplemented with a natural gas-firedcombustion system to obtain adequate steamtemperatures to drive the turbine. After passingthrough the steam generator. The oil is used topreheat water destined for the steam generator;the oil may be returned to the trough field.

So/ar Pond.–In an ordinary body of openwater, an important mechanism which influencesthe thermal characteristics of the reservoir is nat-ural convection. Warmer water tends to rise to

the surface; and if the water is warmer than theambient air, it tends to lose its heat to the atmos-phere. A solar pond (see figure 4-11) is designedto inhibit this natural process. The creation ofthree layers of water, with an extremely denselayer at the bottom and the least dense layer atthe top, interferes with the movement of warmerbottom waters toward the surface. Salt is usedto increase the density of the bottom layer, form-ing a brine, to the point where its temperaturecan go as high as 227° F. The heat in this bot-tom layer can then be drawn off through a heatexchanger, where the brine transfers its heat toan organic working fluid which in turn can drivean engine to produce electric power.

General Overview. –Within most of the abovementioned technologies, many variations nowexist or could exist. The discussion here is con-fined to those variations which appear to affectprospects for solar thermal-electric systems in the1990s. The discussion is intended to be a briefsurvey rather than exhaustive examination of thetechnologies.

Each technology is characterized by a particu-lar set of advantages and disadvantages (see ta-ble 4-3) which together define its prospects thiscentury. All of the technologies share a principaldisadvantage in that costs and performance arecurrently uncertain. The level of uncertainty canonly be reduced sufficiently as commercial-sizedunits are deployed and operated. In some cases,research and development hurdles must still besolved before commitments are likely to be madeto early commercial units, Until this occurs, thechances for widespread commercial applicationfor any one technology during the 1990s are quitesmall, regardless of the technology’s ultimatepromise. At least one operating system is requiredto reduce cost and performance uncertainty toa level where it no longer is a primary impedi-ment to extensive investment; and perhaps sev-eral units—including early commercial units—would be necessary. The time and expense asso-ciated with these early demonstration and com-mercial units are critical elements in determin-ing the commercial prospects of the solar thermaltechnologies in the 1990s.

While the need to reduce uncertainty is ofprime importance, efforts to improve cost andperformance through continued research and de-velopment also couId enhance the prospects forthe technologies in the 1990s. The primary R&Dneeds are different for each technology thoughgenerally efforts directed towards the develop-ment of low-cost, durable, and efficient concen-trators, receivers, and heat engines are most im-portant. ’ 7 Also very important will be the need

‘7More cktdtted Intormatmn on R&D needs for solar thermal tech-nologies can be tou nd In Sandla Natlondl Laboratories, U.S. De-partment of Energy, ~li e Year Research and Dmdoprnent P/.?n1985-1989, draft (Lltermore, CA: Sandia Natmnal Laboratories, De-cem her 1984); and Edw a rci 1, t+. LI n, A Re\ Iew of the Sa/t GradientSo/ar Pond Techno/og} (Pasadena, CA Jet Propulsion Laboratory,1982I, DOE/SF-2 J52-l

for adequate data on the solar resource acrossthe country.18

Solar Ponds and Centra/ Receivers.– Neithersolar ponds nor central receivers appear to re-quire major technical breakthroughs before theycan be commercially applied. But because nocommercial-scale solar ponds or central receiversare now operating in the United States, becausenone are now under construction in this coun-try, and because of the long lead-times expectedfor the installations, it will be very difficult to de-ploy enough demonstration and early commer-

18 B p G u ~>tcl, Soi.jr Thernl,]l Research Program, Ann ua I Resea r[ h.,Plan, Fiscal Yedr 1985 (Golden, C() Solar Energy Rew~arch I nstl -tute, 1984).

38-743 0 - 85 - 4

88 • New Electric power Technologies: Problems and Prospects for the 1990s

Figure 4-il.—Solar Pond Powerplant Concept

Condenser1 4

Y

Generator

Pump

ry

1982).

Table 4-3.—Advantages and Disadvantages of Solar Thermal Electric Systems

Technology

Parabolic dishes Parabolic dishes CentralCharacteristic Solar ponds Solar troughs (mounted-engine) (remote engine) receiver

Yes (+)Yes (+)

Yes (+)low ( – )low (+)low (–)no (–)

yes (+)yes (–)med (0)

Yes (+)No (–)

No (–)high (+)high (–)high (+)no (–)

no (–)no (+)low (+)

Total, all categories:+ (major advantages) 5 5 5– (major disadvantages) 5 4 5● (moderate advan/disadvan) o 1 0

Total, categories 3-10:+ (major advantages) 5 3 4– (major disadvantages)● (moderate advan/disadvan)

3 4 40 1 0

Yes (+)Yes (+)

Yes (+)low ( – )low (+)low (–)no (–)

yes (–)no (+)med (0)—

541

341

No (–)No (–)

Yes (+)med (0)low (+)low (–)no (–)

yes (+)no (+)med (Ž)

542

422

SOURCE: Office of Technology Assessment


cial units quickly enough to sufficiently reduceuncertainty about cost and performance. Further-more, the expense of such early units is highenough that it is very unlikely that any entity orgroup of entities outside of government presentlywould invest in such units without special gov-ernment incentives. In short, the time and ex-pense associated just with reducing risks for thesetwo technologies strongly mitigate against theprovision of sizable amounts of electric powerby the technologies within the 1990s. It is likelythat only a sizable and immediate governmentintervention to encourage rapid deployment ofdemonstration units and subsequent units couldreduce uncertainty to the point where it no longeris a major impediment to commercial investmentin the 1990s.

Should such intervention occur and if the po-tential advantages of the solar ponds and centralreceivers are realized, both technologies offerfavorable balances of advantages and disadvan-tages which could stimulate considerable privateinvestment in the 1990s. Between the two tech-nologies, the central receiver probably would bemost widely deployed. Solar ponds must be lo-cated in areas where land, water, and salt areplentiful. Such sites are far less common than thesites available to central receivers, which requireconsiderably less land, less water, and do not re-quire such large quantities of salt. Siting optionstherefore are greater with the central receiver. 19

Furthermore, the central receiver is a more ma-ture technology. A 10 MWe (net) pilot facility,Solar One, has operated successfully in southernCalifornia since 1982; and a small experimentalfacility, rated at 0.75 MWe (gross) has been oper-ated in New Mexico .*” Small central receiversalso have been built and operated overseas, butno solar pond has ever produced electric powerin the United States. However, a 5 MWe unit isin operation in Israel and several ponds have——

lgFOr a ~lscussion Of the so[ar pond’s prospects in California, andof the II m Itat Ions regard I ng sites, see Marshal F. Merriam, E/ectri-clty Generation from Ncm-Corrvectlve Solar Ponds in Ca/lt’ornia (Ber-keley, CA: Universitywlde Energy Research Group, December1983), UER-109,

20john T. Holmes, “The Solar Molten Salt Electric Experiment, ”Advanced Energy Systems–Their Role in Our Future: Proceedingsof the 19th /ntersociety Energy Conversion Engineenng Conference,Aug. 19-24, 1984, (San Francisco, CA: American Nuclear Society,1984), Paper 849521,

been built and operated in the United States andelsewhere for applications other than the produc-tion of electricity. The solar pond concept how-ever is considered to be well established and thesuccessful commercial deployment of the tech-nology is not expected to require any major tech-nical breakthroughs. *1

Parabolic Troughs and Dishes.– Unlike theponds and central receivers, parabolic dishes andparabolic troughs already have been deployedin commercial-scale units. Indeed, commercialinstallations financed by private investors assistedby the Renewable Energy Tax Credits now areoperating. Further demonstration and early com-mercial units are being planned over the next 5years, though the extent to which the plans arerealized depends heavily on Government tax pol-icies or funding, As current units continue tooperate, and as new units are added, the levelof uncertainty and risk associated with the tech-nologies will continue to drop.

At present, the parabolic trough is the mostmature of the solar thermal electric technologies,with commercial units operating, under construc-tion, and planned. Nearly 14 MWe (net) of pri-vately financed capacity already is operating insouthern California at the Solar Electric Gener-ating System-1 (S EGS-I); an additional 30 MWe(net), the Solar Electric Generating System-n(SEGS-11), now is being built. Additional capac-ity— 150 MWe or more—may be added by early1989. if the energy tax credits are extended insome form. Whatever the case, by 1990 morecommercial experience will have been loggedwith this technology than any other solar thermal-electric alternative. The resultant low level of riskwill constitute an important advantage for thistechnology. Other important advantages will bethe technology’s inherent storage capacity andthe relatively wide variety of markets to whichit could be applied—including industrial process-heat applications.

21 See: I ) N4assach usetts I nstltute of Technology, A state-o~-fhe-

Art Study of Noncon\wcti\e So/ar Ponds for Power Generation (PaloAlto, CA: Electric Power Research Institute, January 198S), EPRI AP-3842. 2) Edward I.H. Lln, A Retlew ot’the Sa/t Gradient SoL~r PondTechnology, op. cit., 1982,

90 Ž New Electric Power Technologies: Problems and Prospects for the 1990s

But the troughs are saddled with several seri-ous disadvantages which to some extent will con-strain deployment. The technology’s low effi-ciency is its most serious disadvantage. Anotheris the need by the SEGS units for a supplemen-tary fuel such as oil or gas, and for considerablevolumes of water. Finally, the system of conduitsthrough which the heat-absorbing oil flows maydevelop problems or the oil itself may degradeat an excessive rate; these potential problemshave not yet materialized at the SEGS-I installa-tion, but further operating experience is requiredbefore long-term performance can be proven.

Two types of parabolic dishes may be de-ployed in the 1990s: the mounted-engine para-bolic dish and the remote-engine parabolic dish.Each offers many design options. The primary ad-vantages of the mounted-engine units are theirhigh efficiencies, low water consumption, andability to operate without supplemental fuel. Thesmall size of the basic electricity-producing mod-ule also carries with it advantages. The systemmay be installed in many sizes, and multi moduleinstallations may produce electric power long be-fore the full installation is completed; individualmoduIes or groups of moduIes may begin oper-ating while others are being installed. Togetherthese advantages provide the technology withconsiderable siting flexibility and potentially veryshort lead-times.

The largest disadvantage of the mounted-engineunit is the relatively high level of uncertaintyabout its performance, and the possibility that theengines may require an excessive amount ofmaintenance. Only three commercial-scale dem-onstration units—at about 25 kWe (net) each—had been deployed by May 1985, and few arescheduled to be deployed by 199o; no commer-cial installation yet exists, or is under construc-tion. Other disadvantages include the lack of stor-age capacity and the inability to readily adapt thetechnology to cogeneration or nonelectric appli-cations.

The remote-engine dishes, like the troughs, en-joy the advantage of being used at present in acommercial installation. A 3.6 MWe system, builtby LaJet, Inc., now is operating in southern Cali-fornia. Also, like the troughs, the remote-enginetechnology may use as few as one or two engines;

engine-related O&M costs therefore could bemuch lower than those of the mounted-engineparabolic dishes. The remote-engine technologyin addition may be easily used for nonelectric ap-plications. The LaJet design at present does notrequire a supplemental fuel.

But the remote engine technology is inefficient;much heat is lost as the heat transfer fluid ispumped from the collector field to the turbines.Also, the system has little storage capacity; elec-tricity production therefore cannot be deferredfor very long. And unlike the mounted-engineunits, the remote-engine technology consumessizable volumes of water.

Both dishes and troughs suffer from the sameserious problem—they lack the cost and perform-ance certainty which can only be gained throughmore commercial-size operations. This mitigatesagainst private sector investment which is not insome manner accompanied by government sup-port. At the current pace, it is uncertain whetherthe situation will change over the next 5 to 10years.

Generally, capital costs will have to be reducedand performance improved if the technologiesare to be deployed. To some extent this can befostered by research oriented towards incre-mental improvements of the commercial-scalesystems now operating. The most useful researchwould concentrate on low-cost, durable, andhighly reflective reflector materials and inexpen-sive, long-lasting receivers and engines. But if thetechnologies are to be extensively deployed inthe 199os, the greatest overall need is to reduceuncertainty and thereby increase demand to thepoint where economies of scale can drive costsdown.

By virtue of the fact that commercial-scale sys-tems now are operating for troughs and dishes,the level of cost and performance uncertaintyamong the troughs and dishes will be consider-ably lower than the uncertainty associated withthe central receiver and ponds in the 1990s. Be-tween troughs and dishes, uncertainty will belowest for troughs, highest for the mounted-engine dishes, and somewhere in between forremote-engine dish systems.


The mounted-engine dishes, in particular couldbenefit from greater deployment of commercial-scale units. A considerable reduction in uncer-tainty and greatly improved commercial pros-pects might result. Under such conditions, themounted-engine parabolic dishes could eliminatethe current lead enjoyed by parabolic troughsamong the solar thermal technologies. If the en-gines perform well, the parabolic dish technol-ogy could well become the prevalent choice forsolar thermal electricity production in the 1990s.

Typical Solar Thermal-Electric Installation forthe 1990s.—The precise cost and performanceof the solar thermal-electric systems in the 1990swill vary widely according to system design, loca-tion, overall market size, risk, and many otherfactors. No attempt here is made to fully discussthe cost, performance, and uncertainty of all themany solar thermal technologies. Rather, a sin-gle technology–the mounted-engine parabolicdish–is examined in fuller detail and used forreference purposes. The cost and performancenumbers shown i n appendix A, table A-2 for themounted-engine parabolic dish installation in the1990s represent reasonable estimates, but obvi-ously should be viewed with caution.

By 1995, mounted-engine parabolic-dish plantsmight account for up to 200 MWe of installed ca-pacity, The deployment level depends mostly onthe extent of Government support over the next5 to 10 years–primarily the Renewable EnergyTax Credit—and avoided cost rates.

The reference plant used in this analysis con-sists of 400 electricity producing modules, eachindependently tracking the Sun and producingelectric power. The plant would have a gross ca-pacity of 10.8 MWe and a net capacity of 10.2MWe–the 0.6 MWe difference goes primarily todriving the tracking motors which keep the dishproperly oriented toward the Sun during the day,and to cooling the engine. Other equipment re-quired on the site include a central control unit,electric power subsystems, buildings, mainte-nance facilities, and other equip ment.22

z~whc re Stlrllng Pnglnes are used, the other equipment Includes

system~ which pressu rlze hydrogen for use i n the Stirling engines,

The amount of time required to build the plantshould be very short, perhaps 2 years. The great-est uncertainty in this estimate lies with permit-ting and licensing. A large area of land—approx-imately 67 acres—would be required for theinstallation; the impacts of the developmentwould be extensive. The most obvious impactwould be visual, arising from the modules, roads,and transmission lines (see figure 4-7). Serious im-pacts on the soil and vegetation of the area couldalso occur. Installations in the 1990s probablywould be concentrated in arid areas which havefragile soil and plant communities. Regulatory de-lays could result from concerns over all these im-pacts. Indeed, such problems reportedly have de-layed the planned expansion of LaJet’s Solarplant1 facility in southern California (see box 7B inchapter 7).

The overall operating availability23 of the instal-lation could be quite high for several reasons.Routine maintenance could be conducted atnight. Should a module not be working duringthe day, its incapacity would not impede theoperation of other modules. As long as large num-bers of unpredictable failures do not occur (asfor example might happen after a severe anddamaging storm), then high operating availabil-ities for the system as a whole can be maintained.The reference system used in this analysis ischaracterized by operating availabilities of 95percent.

The expected plant lifetime is 30 years. Manyof the components are relatively simple and dura-ble. The power conversion unit (PCU) located atthe focal point, which uses relatively unproventechnology, is the component which creates thegreatest uncertainty about plant lifetime, It is an-ticipated, however, that with a regular and per-haps expensive maintenance pregram, this uncer-tainty can be greatly reduced, although furtherdevelopment is needed to assure this.

z }Operatlng ava(labl Ilty here refers to the average percentage Ofmodules capable of operating between sunrise and sunset. A 95percent availability indicates that during the average day, s per-cent of the modules are not operating.


Wind Turbines

Introduction

A wind turbine converts wind into useful me-chanical or electrical energy. Wind turbines maybe classified according to the amount of electri-city they generate under specified wind condi-tions. A small turbine generates up to 200 kWe,an intermediate-sized turbine can deliver from100 to 1,000 kWe, and a large system may pro-duce more than 1 MWe.

Since the early 197os, the development of windtechnologies for electric power production hasfollowed two relatively distinct paths–one di-rected towards small turbines and the othertowards the large machines. As the efforts relat-ing to the large turbines bogged down with tech-nical and economic problems, the small turbines—aided by State and Federal tax incentives—progressed very rapidly. In the early 1980s, windturbines were extensively deployed, mostly inCalifornia, where 8,469 turbines were operatingby the end of 1984. The total capacity of theseunits was approximately 550 MWe. Almost allwere erected at windy locations, in clusters called“wind farms. ” By the end of 1984, many thou-sands of wind machines, with a total installed ca-pacity of over 650 MWe, were producing elec-tric power in the United States, and almost allwere small turbines (see figure 4-I 2).

As operating experience accumulated with thesmall machines, both manufacturers and inves-tors began to gravitate towards intermediate-sizedmachines. By the end of 1984, intermediate-sizedmachines were being deployed in small numbers.It is widely believed that if large numbers of windturbines are to be manufactured and deployedin the 1990s, in free competition with othergenerating technologies, intermediate-sized ma-chines probably will be favored over both smalland large machines. Only the intermediate-sizedmachines promise sufficiently cheap power with-out imposing unacceptable risks (figure 4-13 il-lustrates a intermediate-sized vertical-axis windturbine).

While it appears that the total installed capac-ity of wind turbines in the United States may ex-ceed 1,000 MWe by 1985, the rate of subsequent

Figure 4-12.—Maintenance Crews Performinga Routine Inspection of a Small Wind Turbine

SOURCE: U.S. Windpower, Ed Linton, Photographer

deployment is a matter of speculation. Given theshort time within which a wind farm can be de-ployed and operated–from 1 to 2 years, exclud-ing wind data gathering—growth u rider favora-ble circumstances could be extremely rapid. itis possible that the market potential for wind tur-bines could be as high as 21,000 MWe for the1990-2000 period.24

The areas most favored for wind farms are thosewith good wind resources, heavy reliance on oilor gas, and with an expected need for additionalgenerating capacity. They are located mostly inCalifornia, the Northeast, Texas, and Oklahoma.There are, however, less extensive but neverthe-less promising opportunities elsewhere in thecountry, especially in parts of the Northwest,Michigan, and Kansas.25 Most–though not all–

Zqscience Applications International Corp., Ear/y Market Potef?-

tial for Utility Applications of Wind Turbines, Preliminary Dratl (PaloAlto, CA: Electric Power Research Institute, December 1984), EPRIResearch Project 1976-1.

ZSlbld,

Ch. 4—New Technologies for Generating and Storing Eleclric Power ● 9 3

Figure 4-13. —A 500 kW Vertical”Axis Wind Turbine

SOURCE Southern California Edison Co.

turbines probably will continue to be installed inrelatively large wind farms rather than individu-ally or in small clusters. Figure 7-10 in chapter7 indicates the distribution of the wind resourceand the areas of the United States where winddevelopment is most favored.

A Typical Wind Farm in the 1990s

The reference wind farm used in this analysisis summarized in appendix A, table A-4. A typi-cal wind farm in the 1990s may consist of up toseveral hundred, 200 to 600 kWe wind turbines;the reference wind farm consists of 50 turbineswith ratings of 400 kWe each. Installations couldvary widely in the number of turbines deployedor in their exact ratings. But based on current pro-jections, the important cost and performancecharacteristics would be common to the average

facil ity considered by investors during thatdecade.

In addition to the turbines themselves, relatedequipment will be necessary at the site, includ-ing power conditioning equipment, system pro-tection devices, security fencing, metering devicesfor measuring turbine output, wind measuringequipment for monitoring site conditions andequipment performance, control buildings, anda fabrication yard where equipment is stored andassembled .26

The turbines of the reference wind farm wouldbe distributed over an area of anywhere from 300to 2,OOO acres, depending on the topography,prevailing wind direction, the shape and orien-tation of the property on which the farm is lo-cated, and the size of turbines being used. Theturbines are spaced to avoid excessive interfer-ence with each other. Because installation andmaintenance of the turbines requires vehicularaccess, at least one road leads to a wind farm andto each individual wind turbine (see figure 4-14)unless topography, surface characteristics, andregulations allow access without roads. Since theperformance of the turbines and the cost of theirpower depends directly on wind exposure, allmajor obstructions such as trees would be re-moved. 27

It is evident that the major environmental im-pacts of wind farms will result from their initialconstruction as well as from their high visibility,their extensive road networks, and from the activ-ities of maintenance crews on the roads andaround the turbines.28 Among the other impacts,the severity of which may be assessed less read-ily but which nevertheless are considered poten-tially serious, are those associated with the noisecreated by turbine operation.

Concern over environmental impacts couldseriously delay the deployment of wind turbines.

zbsanl sa~l~r, @ al,, t~fl~~} Lanci O\~mer\ ~uldf? [%km, OR: ~~K-

gon Department of Energy, 1984), p. 17,ZTlt shou Id be noted that many prl me MI nd sites, bel ng expoied

to frequent h Igh velocity winds, are I n hospitable en~)ronrnents for:rees and therefore treq uently are devoid of large, upright treeswh Ich COU Id be con fdered serious obstructions.

“’’Wind Farms, Timber Logging May Have Similar En\ ironmentalImpacts, Har\ard’s Turner Says, ” So/,?r /nte//igence Report, Not.26, 1984, p. 375,

94 ● New E/ectric Power Technologies: Problems and Prospects for the 1990s

Figure 4-14.—Aerial View of a Wind Farm in the Altamont Pass in California


sited in windy areas. Under these favorable con-ditions, typical capacity factors between 20 and35 percent are expected in the 1990s as theintermediate-sized turbine technology matures.

As the performance of wind turbines improvesand is better understood, one especially large un-certainty still remains in estimating the annualoutputs of turbines in the 199os: the quality ofthe wind resource to which the turbines will beapplied. Today’s turbines are exploiting some ofthe best wind resources available. But new siteswill be required, and the average quality of newsites probably will decline. High capacity factorswiII be progressively more difficuIt to maintain.At present, it is difficult to predict what wind re-gimes will characterize new sites exploited in the1990s, because sufficiently detailed, site-specificwind data are not yet available in most instances.Information is accumulating, however, and it sug-gests that there remain considerable areas of landavailable with high-quality wind resources.

The lifetime of a wind farm is somewhat diffi-cult to determine because individual turbines andeven components of turbines can be replaced asneeded; in a sense, the wind farm itself can out-live any of its individual components. Generally,the components of a wind farm in the 1990s willbe designed to last 20 to 30 years, though somekey components–such as the rotor–may fail andbe replaced before that time.

Wind Farm Costs.–The average capital costfor wind turbines installed on California windfarms in 1984 was $1,860/kWe.2g This capital costhowever is heavily inflated as a result of thefinancing arrangements associated with currentprojects; one observer has estimated that in factactual costs would be closer to $1,330/kWe if thefinancing mechanisms typical of utilities wereused .30

The capital costs of the typical wind farm in the1990s may range from $900 to $1,200/kWe. Thereduced capital cost will resuIt both from design

zgconver5ation between MI ke Bat ham, Californ id Energy COrn-

mlsston, and OTA staff, Feb. 5, 1985. See also ‘ ‘California Adds 366

MWe of Wind Capacity; Size, Capacity Factor Up, ” So la r EnergyIntelligence Report, Jan. 28, 1985, p. 30.

JODonalcf A. Baln, wind Energy Specialist, Renewable Resources,

Oregon Department of Energy, conversation with OTA staff, June11, 1985.

improvements and from the more competitivemarket expected when the current favorable taxtreatment is phased out. Termination or phase-out of the Federal and California State tax credits,for example, would very likely contribute to de-creases in the capital costs.

Operating and maintenance costs for the windfarms of the 1990s could range between 6 and14 mills/kWh. Available evidence indicates thatcosts for small turbines i n 1984 ranged between15 and 25 mills/kWh.sl The high O&M costswhich thus far have been incurred can be at-tributed to the fact that the first generations ofmachines, those deployed in the early 1980s,were plagued with mechanical problems. Changesin two areas will stimulate the reduced O&Mcosts: smaller numbers of turbines per kilowatt-hour generated and improved turbine design, Ofcentral importance will be the maintenance ofhigh operating availability.

An important cost associated with wind-gen-erated electric power is the cost of access to thewind itself—if indeed access can be gained at anycost. The fee charged by the landowner typicallyis either in the form of a minimum rent, royaltypayments, or some combination of the two.32

Costs of access have increased substantially; land-owners have already begun to appreciate thevalue of prime sites, particuIarly i n California .33There, in 1984, annual land charges commonlyamounted to 6 to 13 percent of gross revenuesfrom the sale of the electricity over the lifetimeof the contract negotiated between the developerand the landowner. 34

The prospects for wind turbines in the 1990swould be enhanced by research and develop-ment. Among the most important R&D items arethe need to better understand turbulence andpredict its effects; to more readily and accurately

g’ “Wind Turbine Operating Experience and Trends, ” EPRI Jour-na/, November 1984, pp. 44-46.

32 For a discussion of the determ i nat ton ot’ w I nd resource va I ue

and contractual arrangements see: Sam Sadler, et al., Windy LandOwners’ Guide, op. cit., 1984.

gJTeknekron Research, cost Estimates and Cost-Forecasting Meth-odologies for Selected Non-Conventional Electrical GenerationTechrro/ogies (Sacramento, CA: California Energy Commission, May

1982).gdconversation between Mike Batham (Ca l i fo rn ia Energy Com-

mission) and OTA sta f f , Nov. 30, 1984.


model structural dynamics; to better predict noiseproblems; to accurately model wind farm costand performance; and to develop, test, andcharacterize materials and components. The de-velopment of cheap, reliable, and accurate windmeasurement instruments as well as better un-derstanding and prediction of the wind’s char-acteristics also are needed. Detailed and accurateassessments of the wind resource nationwide arenecessary too.35

Geothermal Power

moderate temperature range (1 50 to 250o F).37

Geothermal development has in the past focusedon the high-quality, vapor-dominated reservoirs,which are confined to limited areas of the UnitedStates,

The equipment required to exploit these high-quality resources is commercially available, andno major changes in the basic characteristic ofthe technology is likely during this century. Thereare available improved technologies, however,which not only could more economically exploitthe high-quality resource, but also may economi-cally tap the much more plentiful resources oflesser quality. Among these are single-flash, dual-flash, binary, and total flow systems.

The single-flash technology has been commer-cially deployed in the United States. Because thisanalysis focuses on technologies which are notalready technologically mature, the single-flashtechnology will not be examined here. The totalflow systems also will not be discussed, since theyeither require considerable further technical de-velopment, or will be applied only to a smallnumber of high-quality sites i n the United States.The total flow systems therefore are unlikely toconstitute more than a small fraction of geother-mal capacity additions in the 1990s. The dual-flash and binary systems will constitute the mostimportant new technologies applied to the liquid-dominated geothermal resource in the 1990s,and, therefore, are the subject of this analysis.

Geothermal Power Technology

Before the resource is exploited to producedelectric power, it must be located and assessed.This itself is a time-consuming, expensive proc-ess involving its own particu tar set of technologiesand problems. Ultimately, resource assessmentrequires building roads, transporting dril l ingequipment to the site, constructing the rigs anddrilling. Once the resource has been satisfactorilymeasured, the thermal energy next must bebrought to the surface where it can be used. Thistoo involves particular technologies and difficul-

~7M. Nathenson, “High-Temperature Geothermal Resources inHydrothermal Convection Systems In the United States, ” Proceed-ings otthe Se~enth Annu,il Geotherm,jl Conference and workshop,Altas Corp. (cd. ) (Palo Alto, CA: Electrlc Power Research Institute,1983), EPRI AP-3271 , pp. 7-1 to 7-2.


ties. An especially important technological hur-dle which remains to be satisfactorily overcomeis the development of cheap, reliable "down-hole” pumps capable of moving the brine fromthe underground reservoir, and subsequentlyreinfecting it. While the brine is at the surface,a portion of its thermal energy is drawn off andused to produce electric power.

Dual-Flash Systems.–Figure 4-15 illustrates thetypical dual-flash unit. When liquid-dominated,high temperature brine–300 to several thousandpounds per square inch (psia) and 410 to 6000 F—reaches the surface, a portion of the brine“flashes” into steam. First, a high pressure flash-tank processes the geothermal brine into satu-rated steam and spent brine, The steam entersone inlet of a dual-inlet turbine, while the un-flashed brine goes on to a second, lower pres-sure flash-tank. The second-stage flash-tankproduces further steam which is routed to theother inlets of the turbine. The remaining un-flashed brine then is reinfected underground.

After exiting the turbines, the steam passesthrough a condenser, where it transfers its heatto a stream of cooling water. The cooling wateris then routed to a cooling apparatus. Current de-signs use “wet cooling” devices in which the hotwater is sprayed into the air and discharges itsheat mostly through evaporation, The remainingwater is recirculated to the condenser to repeatthe cycle, along with “make-up” water requiredto compensate for evaporative losses. The con-densed steam from the turbine is reinfected intothe geothermal reservoir to help maintain reser-voir pressure.

The make-up water requirements may be ex-tremely large. The 50 MWe reference plant usedin this analysis would require about 3 million gal-lons of make-up water daily, roughly six times theamount of water required by an atmosphericfluid ized-bed combustor of comparable net gen-erating capacity. The water requirements couldbe reduced with “dry-cooling” systems; but theseare very expensive and reduce the plant’s over-all efficiency.

Figure 4-15.—Schematic of Dual-Flash Geothermal Powerplant

Turbine-generator

First stage

Io2

flash vessel(high pressure)

Concentrated

1 I

07

98 ● New Electr ic Power Technologies: Problems and Prospects for the 1990s

Alternatively, the water requirement could begreatly reduced by meeting it in part with the con-densed steam from the condenser (instead of re-infecting it into the reservoir). This can only bedone, however, to the degree allowed by con-tractual agreements and regulations. The field de-veloper may require that all or part of the con-densed steam be reinfected into the geothermalreservoir to maintain the quality of the resource.Or regulators may require some degree of rein-fection in order to reduce subsidence problems.

The basic turbine, condenser, and coolingtower subsystem are similar to traditional steampowerplant designs, although there are significantdifferences. An important factor in the use of flashtechnology is the existence of noncondensablegases and/or entrained solids in the brine. Thesecontaminants can cause scaling, corrosion, anderosion within the flash equipment, surface pip-ing, and reinfection well casing. Development ofhighly saline resources has been slowed by theseproblems. Considerable research has been con-ducted to develop and demonstrate reliableremoval technologies for these resources.

Although, operational dual-flash units abroadtotal 396 MW,38 there is little commercial experi-ence with these systems in the United States.None is now operating, and only one 47 MWe(net) dual-flash unit is under construction (see fig-ure 4-1 6). Nevertl~eless, the dual-flash system willbe used increasingly to exploit moderate to hightemperature hydrothermal resources because itis more efficient than the single-flash system .39

Appendix A, table A-5 contains cost and per-formance estimates for dual-flash units in the1990s. By the reference year 1995, dual-flash geo-thermal units will most likely range in size from40 to 50 MWe. The expense of smaller units

‘8R. Dipippo, “Worldwide Geothermal Power Development:1984 Overview and Update,” Altas Corp. (cd.), Proceedings of theEighth Annual Geothermal Conference and Workshop (Palo Alto,CA: Electric Power Research Institute, 1984), EPRI AP-3686, pp. 6-1 through 6-15.

jgThe Electric power Research Institute (persona! communication

between E. Hughes (EPRI) and OTA staff, Oct. 4, 1984) predictsthat most of the planned flash plants at the Salton Sea and Brawleyresources will use dual-flash technology.

Figure 4-17.–Simplified Process Flow Diagram of Binary Cycle Technology

I

surface condensers (instead of direct contact con-densers).

The major advantages of binary cycles relateto efficiency, modularity, and environmental con-siderations. First, working fluids in binary cyclescan have thermodynamic characteristics superiorto steam, resulting in a more efficient cycle overthe same temperature difference. QB Second, bi-nary cycles operate efficiently at a wide range ofplant sizes. Especially attractive are small plantswhich, in addition to encouraging short lead-times, have many other important advantages aswell. Third, since the brine is kept under pres-sure and reinfected after leaving the heat ex-changer, air pollution, e.g., hydrogen sulfide,from binary plants can be tightly controlled.There are also several other cost and efficiencyadvantages of binary technology over the clual-flash systems. Nevertheless, the dual-loop designof binary cycles is more complex and costly thana flash design.

~Ep B[ai r, et al,, Geothermal Energy. Investment Decisions and

Comrnercl~/ Development (New York: John Wiley & Sons, 1982).

Binary cycle technology is in developmentalstages with few large operational generating units.By the end of 1985, one large 45 MWe (net) bi-nary plant will have been installed near Heber,California (see figure 4-1 8); in addition, small bi-nary plants, with a total capacity of about 30MWe, will be operating. 49 This will account formost of the binary capacity installed worldwide.Development is expected to proceed, and exten-sive commercial deployment is feasible in the1990s.

The expected cost and performance of binarygeothermal plants in the 1990s are summarizedin appendix A, table A-5. Data is provided for tworeference plants, a large plant of about 50 MWe(net) and a small plant of about 7 MWe (net). Thelarge plant could require up to 20 acres of landfor the powerplant and for the maze of pipingrequired for both the brine and the working fluid.The small unit might occupy 3 acres or Iess.so Aswith the the dual-flash technology, very large

‘9 Ronald DiPippo, “Worldwide Geothermal Power Development:1984 Overview and Update, ” op. cit., 1984.

~Opersonal communication between H. Ram (Ormat, Inc. ) andOTA staff, Oct. 6, 1984.

Ch. 4—New Technologies for Generating and Storing E/ectric Power ● 101

Figure 4-18.—Artist’s Conception of the Heber, CA, Binary Geothermal Installation

SOURCE San Diego Gas & Electrlc Co

volumes of cooling water are required; indeed,the water requirements are even larger than thedual-flash units. The large plant would requireover 4 million gallons each day. The smaller plantwould need about 0.6 million gallons per day.

Small plants of about 5 to 10 MWe (net) canbe erected and operating on a site in only 100days. But prior licensing, permitting, and otherpreconstruction activities could extend the lead-time to 1 year.5’ Construction of larger binaryunits shouId take only 1 ½ to 2 years.52 But heretoo, overall lead-times will be longer because ofpreconstruction activities, including licensing and

“Wood & Associates, a geothermal energy developer, has hadpermitting problems at the county, State, and Federal level at Itssite near Mammoth Lakes, CA.

‘zSan Diego Gas & Ele[ trlc also had problems getting their largebinary plant through the permitting process, Its problems, howeier,were encountered during the California Publ K Uti Iities Comml+slon’s plant dpprok al process.

permitting. About 5 years total might be requiredfor the first plant at a resource, and 3 years mightbe necessary for subsequent additions. With bothsmall and large plants, problems about water re-quirements and environmental impacts couldseriously extend the licensing and permittingprocess.

Binary cycle plants are designed to operatecontinually in base load operation. Availabilityis expected to be between 85 and 90 percent,and capacity factors are likely to be in the 75 to80 percent range. Binary pIants should last at least30 years.

The net brine effectiveness of binary cycleplants may vary between 7 and 12 Wh/lb of steamat a 4000 F resource. Advanced binary technol-ogy in the larger sizes should increase present ef-fectiveness values at Heber from 9.5 to 12 Wh/

102 ● New Electric Power Technologies: Problems and Prospects for the 1990s

●

●

●

●

●

●

responsiveness to changes in desired output,short lead-times,easily recovered waste heat,fuel flexibility,off-site manufacturing, andability to operate unattended.

Most of the commercial demonstration plantscrucial to the future of the fuel cell will not be-gin operations until the late 1980s, though by May1985, a 4.5 MWe demonstration plant was oper-ating successfully in Japan and thirty-eight 40-kWe demonstration units were operating in theUnited States. ss Should the performance of thedemonstration units be very good, limited quan-tities of commercial fuel cells may be producedat the earliest at the end of this decade or thebeginning of the next.5G 57

The level of deployment depends heavily onthe success of the demonstration units; the periodof time deemed necessary to generate investorconfidence; and the willingness of the vendorsto share the risk and cost of the early units, Theperceptions and decisions of investors, vendorsand buyers cannot be accurately and confidentlypredicted, but current evidence suggests that theearly 1990s may see the beginnings of fuel-cellmass production and the first commercial appli-cations. As much as 1,200 MWe of fuel cell pow-erplant capacity may be operating by 1995.

The low production levels will drive installedcapital costs down somewhat, but they will re-main far above possible costs in a mature mar-ket. High-volume mass production is unlikely tooccur until a sizable market is anticipated—in themid-l990s at the earliest. Such a market may de-velop as investors observe the continued opera-tion of the demonstration units and the initialoperation of the early commercial installations.

Most important to the prospective investors willbe operating and maintenance costs, economic

ssjw staniunas, et al,, United Technologies Corp., Eo//ow’-On40-kWtn Field Test Support, Annual Report (July 1983 -Jurre 1984)(Chicago, IL: Gas Research Institute, 1984), FCR-6494, GR1-84/0131.

JGpeter H Unt, Ana/ysjs of Equipment Manufacturers and Vendors

in the Electric Power Industry for the 1990s as Related to Fuel Cells(Alexandria, VA: Peter Hunt Associates, 1984), OTA contractor re-pOrt OTA US-84-1 1.

SzBattelle, Columbus Division, Fina/ Report on Alternative Gen-

eration Technologies (Columbus, OH: Battelle, 1983), VOIS. I and11, pp. 13-11.

Ch. 4—New Technologies for Generating and Storing E/ectric Power ● 103

life, and reliability. In particular, investors arelikely to be sensitive to the rate at which fuel cellperformance degrades over time under variousoperating conditions as well as the cost and dif-ficulty of replacing cells when their performancebecomes unacceptable. There is uncertaintyamong investors over these two points, both ofwhich are crucial to the fuel cell’s uItimate com-mercial prospects. ShouId problems be encoun-tered in either area in early commercial proto-types, commercial deployment will be delayed.If powerplant operation is favorable, subsequentmarket growth in the latter half of the I990S couldbe very rapid.

Basic Description

The typical fuel cell powerplant will consist ofthree highly integrated major components: thefuel processor, the fuel cell power section, andthe power conditioner. The fuel processor ex-tracts hydrogen from the fuel which can be anyhydrogen-bearing fuel, though most installationsin the 1990s are expected to employ natural gas.

The hydrogen is then fed into the fuel-cellpower section, the heart of which are “stacks”of individual fuel cells. The operation of a singlefuel cell is schematically illustrated in figure 4-19, The cells are joined in series (the stacks),which, in turn, are combined to form a power-plant. There are several types of fuel cells beingdeveloped. These are categorized according tothe type of electrolyte-the medium in which theelectrochemical reaction occurs—they use. Thefirst-generation fuel cells use phosphoric-acid asthe electrolyte. These cells are the most devel-oped and are likely to account for most of thefuel cells deployed in the 199os.

Other less mature, fuel cell designs which em-ploy alternative electrolytes promise superior per-formance; molten carbonate cells are the closestto commercial application, but are not expectedto be commercially deployed until the late 1990sat the earliest. They therefore are not likely to ac-count for an important share of fuel cell power-plants installed in the 1990s. M w

S8Peter H u nt, AflJ/YSIS Ot Equipment Manufacturers and Vendors

In the Electrlc Power Industry for the 1990s as Related to Fuel Cells,OP. C i t . , 1984.

5gu ,s, Congress, office of Technology Assessment Workshop onFuel Cells, Washington, DC, June 5, 1984.

The electrical power which flows from the fuelcell stacks is direct current (DC). With some volt-age regulation, this DC power can be used if theload is capable of operating with direct current.Otherwise, a power conditioner is required totransform the direct current into alternating cur-rent. This allows it to be fed into the electricalgrid and to be used by alternating-current elec-trical motors.

The components of the fuel cell plant are tightlyintegrated to reduce energy losses through theproper management of fuel, water, and heat (seefigure 4-20). Various parts of the plant benefitfrom the byproducts of other parts of the instal-lation. Further efficiency gains result when by-product heat from different parts of the plant aretapped for external use. The fullest exploitationof the fuel cell’s heat may yield total energy effi-ciencies of up to 85 percent for the entire plant.The heat can be used for domestic hot water, forspace heating, or to provide low-level processheat for industrial uses.

Typical Fuel Cell Powerplantsfor the 1990s

The expected cost and performance of typicalfuel-cell powerplants for the reference year 1995are summarized in appendix A, table A-7. Be-cause no complete powerplants identical to thosewhich might be deployed at that time exist today,these values remain estimates.

The units deployed in the 1990s probably willbe built around two sizes of fuel cell stacks. Thelarger stacks are likely to be capable of generat-ing approximately 250 to 700 kWe (gross, DC)each and the smaller stacks about 200 to 250kWe (gross, DC) each. The plants built aroundthe small stacks will be installed mostly in largemultifamily dwellings, commercial buildings, andin light industries; most will probably be used tocogenerate both electricity and useful heat. Thetypical system would consist of at least two com-plete self-contained modules (see figure 4-21 ),each of which might produce about 200 kWe(net, AC).

Plants using the larger fuel cell stacks most likelywill be deployed primarily by electric utilities, andby industries which would use them in cogener-ation applications. Installation capacities probably


Figure 4-19.—Schematic Representation of How a Fuel Cell Works

1. Hydrogen gas flows over negativeelectrode (anode).

111

Electrons

11

Ieotrone, and o

will range from several megawatts on up. Thereference installation used in this analysis is 11MWe (net). Many of the major componentswouId be fabricated in factories and shipped tothe site on pallets.

The lead-time of a small fuel cell powerplantsshould be about 2 years. These units are relativelysmall, unobtrusive, and quickly and easily erected.Modules subsequently added at the same sitecould require as little as a few months. Regula-tory delays are unlikely because of relatively mi-nor siting and environmental considerations.

Installations utilizing the larger stacks, however,may encounter more serious regulatory prob-lems, Unlike the approximately 480 to 600 square

feet required by an installation of two, 200 kWeunits, an 11 MWe installation would occupyabout 0.5 to 1.2 acres of land (see figure 4-22).Because the plants frequently may be located inthe midst of populated areas, the opportunity forregulatory conflicts with these larger plants is con-siderably greater. Partly offsetting these factors,though, are the environmental advantages asso-ciated with fuel cell powerplants. Hence, a lead-time of 3 to 5 years is anticipated with the largerunits, considerably longer than the small plant’slead-time, but also much shorter than that of mostconventional powerplants. As with the smallerfuel cell installations, capacity subsequently ad-ded to an already existing fuel cell plant shouldrequire considerably shorter lead-times.

L

Mechanical systems

Water

— — — — — — -

Watertreatment

1

Powerplantcontroller

106 ● New E/ectric Power Technologies: Problems and Prospects for the 1990s

Figure 40-21.— Design for a 200 kWe Fuel Cell Module

While basically similar to units which might be deployed In the 1990s, this design differs in several important details,

The operating availability of large fuel cell pow-erplants may range between 80 and 90 percent.Availability is heavily dependent on the qualityof design—its simplicity, the extent to which ithas redundant components, the number of parts,and their reliability—and on the availability ofspare parts and repair people when they areneeded.

The fuel cell can be applied to any duty cycle.The fuel cell has excellent load following capa-bilities and high efficiency over a wide variety ofoperating levels.

The fuel cell powerplant’s lifetime is assumedto be approximately 20 to 30 years with periodicoverhaul of the fuel cell stacks and other com-ponents. over time, the powerplant’s efficiencydrops. The timing of overhauls will vary; sched-u Ies will be a function of the performance reduc-

tion over time and of other factors such as thecost of fuel .60 I n some cases the stacks must peri-odically be removed and replaced with newones. The old unit then is shipped back to a man-ufacturing plant where its catalyst (in the fuelprocessing section) and perhaps other compo-nents are removed, processed, and recycled.While the overhaul schedule and costs are un-certain, it is assumed here that all stacks arereplaced after the equivalent of 40,000 hours ofoperation at full capacity.

60J. R, Lance, et al., Westinghouse Electric Corp., “Economics andPerformance of Utility Fuel Cell Power Plants, ” Acf~anced Energy

Systems– The/r Role In Our Future: Proceedings of the 19th /rrter-soclety Energy Conversion Engineering Conference, August 19-24,/984 (San Francisco, CA: American Nuclear Society, 1984), paper849133.

Ch. 4—New Technologies for Generating and Storing Electric Power Ž 107

co

Figure 4-22.—Typical Arrangement of 11 MWe Fuel Cell Powerplant

Auxiliarypower system \

Powernditioner \ 1

Liquidfuel system

\Natural

gas system

supply

—

/-.

Efficienciesbl for large fuel cell plants are ex-pected to be between 40 and 44 percent. Smallplants may have efficiencies of approximately 36to 40 percent. The estimated efficiencies are thosewhich might characterize a plant over its lifetime;efficiencies of new stacks could be higher whilethose of older stacks might be below that level.

The installed capital costs of fuel cell powerplants are expected to range from $700 to $3,000/kWe for large units to $950 to $3,000/kWe forsmall plants; expected values for 1995 are $1 ,430/kWe and $2,240/kWe respectively. The low num-bers can be expected where units are commer-cially produced in large numbers; the high num-bers are representative of prototype units andinclude nonrecurring costs. By far the largest ex-pense is the fuel-cell power section itself; it is ex-pected to account for about 40 percent of the

‘]’ Based on higher heatln~-~ alue

costs of a mature 11 MWe plant .62 The largestdecrease in capital costs over the next decadewill come from increases in the levels of fuel-cellproduction. However, technical improvementsin the fuel-cell plant itself may substantially re-duce costs as well. Already, over the past sev-eral years, design changes have reduced costs byan appreciable amount.

Operating and maintenance costs may rangebetween 4.3 and 13.9 mills/kWh. The biggest ele-ment in O&M costs is the cost of periodicallyreplacing cells stacks. For specific applications,the actual O&M costs will depend on the over-haul period for the cell stacks and the materialand labor costs for each overhau 1.

b~u nlted TeCh rlol~~les Power Systems, sfu~y OfJ ~~05@Or/C ACI~

Fuel Cells Using Coal-Deri~ed Fuels (South Windsor, CT: UnitedTechnologies Power Systems, Apr. 27, 1981 ), prepared for Tennes-see Valley Authority, contract No. TV-52900A, FCR-2948.


Fuel costs are expected to be approximately27 to 33 mills/kWh, accounting for the major por-tion of electricity costs from fuel cells. Fuel costsalso constitute the fuel cell’s greatest advantageover some of its competitors such as the gas tur-bine, due to the fuel cell’s high efficiency whichyields substantially lower per kilowatt-hour fuelcosts. Major variations in fuel costs per kilowatt-hour will result primarily from fluctuations in fuelprices. Fuel cell efficiency variations, due to tech-nical improvements or maintenance practices(especially stack reloading schedules), also wouldbe reflected in different fuel costs. In the longerterm, post-2000, it is expected that natural gaswill have to be replaced with more abundantfuels. Primary candidates are synthetic fuels–especially methanol—from coal and biomass.

Stacks are of central importance in determin-ing capital, O&M, and fuel costs. The develop-ment and extended demonstration of cheap (perkWe) and reliable stacks which can operate athigh efficiencies for extended periods are criti-cal to the success of the technology. Technologi-cal improvements which could be especially im-portant in this regard are the development ofinexpensive, corrosion-resistant cell structuralmaterials and less expensive and more effectivecatalysts to operate at higher pressures and tem-peratures, and improved automated fabricationand handling processes for large area cells. Alsoimportant is the development of cheap, reliableand efficient small-scale reformers (fuel process-ing units) and the improvement of various otherstandard components.

Combustion Technologies

Integrated Gasification/Combined-CyclePowerplants

introduction.-A coal gasification/combined-cycle powerplant centers around two elements.First is a gasification pIant which converts a fuelinto a combustible gas; other equipment purifiesthe gas, Second is a combined-cycle powerplantin which the gas fuels a combustion turbinewhose hot exhaust gases are used to generatesteam which drives a steam turbine. While thegasification system can be quite separate and dis-tinct from the combined-cycle system, they can

be integrated so that some of the heat dischargedin the gasification sequence is exploited in thecombined-cycle system, and a portion of the heatdischarged by the combined-cycle unit may berouted back for use in the gasification plant (seefigure 4-23). This section focuses on such inte-grated units, commonly referred to as IGCCS.

The primary attractions of the IGCC are its fuelefficiency and its low sulfur dioxide, carbonmonoxide, nitrogen oxide, and particulate emis-sions. The high efficiency allows for fuel savingsand hence reduced operating costs. The poten-tial for very low emissions makes the technologyparticularly attractive for using coal to generateelectric power. Another advantage allowed bythe IGCC is “phased construction. ” Some partsof the plant may be installed and operated be-fore the rest of the plant is completed;63 this canbe financially advantageous and is considered amajor selling point for the technology. The IGCCalso may be very reliable. In addition, the tech-nology requires less land and water than a con-ventional scrubber-equipped, pulverized coalboiler powerplant. Furthermore, its solid wastesare less voluminous and less difficuIt to handlethan those of its scrubber-equipped competitorand of the atmospheric fluidized-bed combustor(AFBC). Current estimates are that solid wastesfrom an IGCC will be 40 percent of a pulverizedcoal boiler and 25 percent of an AFBC of com-parable size.

The evidence suggests that the IGCC offers afavorable combination of cost and performancewhen compared to its competitors (see also chap-ter 7). Nevertheless, a combination of two fac-tors—lead-times and risk—may mitigate againstits extensive deployment within the 1990s. Be-cause of its modular nature and positive environ-mental features, potentially the IGCC has lead-times of no more than 5 to 6 years. It is likely,however, that the first plants, at least, will requirelonger times–up to 10 years–because of regula-tory delays, construction problems and opera-tional difficulties associated with any new,complex technology. It may take a number of

GJFOr example, the gas-turbine/generator sets may be installedbefore the gaslfiers and operated off of natural gas. When the gasl -fiers are completed, the synthetic gas then may be used instead.

Coal

Nitrogento atmosphere

IProcess I . Flash I i

ClausIncinerated

steam Effluent gas sulfur tall gas

treatment - recovery anincineration1

I4

Saturated steam

Po-we/ Generation (Argonne, IL Argonne National Laboratory 1983)

I Plantauxiliary

– Jpower


commercial plants before the short lead-time po-tential of the IGCC is met, unless strong steps aretaken to work closely with regulators and to as-sure quality construction for these initial plants.Such steps may be facilitated if the early plantsare in the 200 to 300 MWe range rather than thecurrent design target of 500 MWe. Should thelonger lead-times be the case, projects must beinitiated no later than the end of 1991 and per-haps as early as the end of 1989 if they are tobe completed within the 1990s.

In addition to these possible longer lead-times,limited experience with IGCC demonstrationplants may also constrain extensive deploymentin the 1990s. Although there has been extensiveexperience with the gasification phase, currently,there is only one IGCC plant in operation, the100 MWe Cool Water plant in California (see fig-ure 4-24). In addition, there is another demon-stration plant using a different gasifier, under con-struction in a large petrochemical plant (discussedmore in chapter 9). The Cool Water plant hasbeen very successful in meeting its constructionscheduled and budget, and in its early operation.As a result it has given confidence to utilities intheir consideration of whether to commit to an

b~An interesting d iscuss ion of the plannlng and const ruc t ion of

an IGCC can be iound in: Cool Water Coal Gasification Program& Bechtel Power Corp., Cool Water Coal Gasification Program–Second ,4nnua/ Progress Report, interim report (Palo Alto, CA: Elec-tric Power Research Institute, October 1983), EPRI AP-3232.

IGCC. Despite this success, however, more oper-ating experience is likely to be required beforethere will be major commitment to the IGCC bya very cautious electric utility industry. The Elec-tric Power Research Institute, a major sponsor ofthe IGCC Cool Water project, anticipates threeto four commitments by the end of 1986. If theseprojects go forward and the shorter lead-time po-tential of the IGCC is proven, then significantdeployment in the mid to late 1990s is quitepossible.

Description of a Typical IGCC in the 1990s.–Plausible cost and performance features of a rep-resentative IGCC are described in appendix A,table A-6. The reference year considered in thereport is 1990, at which time two plants, gener-ating altogether approximately 200 MWe prob-ably will be operating in the United States. Thereference plant capacity is 500 MWe, consistingperhaps of five gasifiers,65 though installations assmall as 250 MWe might be preferred. While ca-pable of being built with capacities even smallerthan 250 MWe, such smaller installations wouldbe more costly per unit of capacity.66 The plantwould consist of three types of equipment: thegas production, cooling and purification facilities;the combined-cycle system (including gas tur-bines and steam turbines); and the balance of theplant. Included in the constituents of the latterare fuel receiving and preparation facilities, watertreatment systems, ash and process-waste dis-posal equipment, and in most cases an oxygenplant.

Gszaininger Engi r-reefing CO., capacity factors and costs of Elec-tricity for Conventional Coal and Gasification-Combined CyclePower Plants (Palo Alto, CA: Electric Power Research Institute, 1984),EPRI AP-3551 .

66According t. one source (Electric bA@’ Research Institute, Eco-

nomic Assessment of the Impact of Plant Size on Coal Gasification—Combined-Cyc/e P/ants (Palo Alto, CA: Electrlc Power Research in-stitute, 1983), AP-3084), the Ievellzed cost of electricity for a vari-ety of IGCCS using Texaco gasifiers increased as plant size de-creased. The economies of scale were relatively small among plantsof capacities greater than 250 MWe. But as plant size diminishesbelow 250 MWe, Ievellzed costs increase very significantly. Selec-tion of a 500 MWe module also was favored by participants in aJune 1984 OTA-sponsored workshop on IGCCS, though it was sug-gested that installations as small as 250 MWe might seriously beconsidered.

Ch. 4—New Technologies for Generating and Storing Electric Power ● 111

The IGCC can be built in phases. The designpermits the operation of portions of the pIant be-fore other segments are completed. The gas tur-bines could be installed first and operated withnatural gas. The steam turbines then could beadded, allowing the production of still more elec-tric power. Finally, the gas facilities could beadded to complete the plant and to allow its oper-ation based on synthetic gas. Hence, some elec-trical power could be produced before the en-tire plant is completed.

The typical plant would require a rather largearea of land and considerable quantities of waterduring its lifetime of approximately 30 years. Anestimated 300 to 600 acres would be needed forthe facilities, and for disposal of solid wastes. And3 to 5 million gallons of water, on average, wouldbe required to run the plant daily. These quanti-ties are large, but as noted above, they are smallerthan those which characterize conventional pul-verized coal plants equipped with scrubbers.

The operating availability of the reference IGCCplant is 85 percent. There is uncertainty associ-ated with availability estimates, as these plantswould be the first commercial units and couldexperience problems which would result in loweravailability rates. Of particuIar concern is thereliability of the combined-cycle system; com-bined-cycle system design as well as operatingand maintenance practices will largely determinecombined-cycle reliability.

The IGCC facility commonly would be used toprovide base load power at efficiencies rangingfrom 3S to 40 percent. This corresponds to a heatrate of between 8,533 and 9,751 Btu/kWe-hour.67

It is worth noting that the Cool Water demon-stration plant, which had a design heat rate of11,400 Btu/kWh, has consistently met that tar-get in operation to date. While the gasifier de-sign certainly has an important effect on effi-ciency, the most important factor in efficiencywithin the anticipated range probably would be

FIT I t I ~ I m ~)() rtcl nt tO note that I Gcc heat rates are part IC u IJ rlv wn -~ltl~ e to J mblent temperatures, Heat rates go down it Ith a mblenttemperatu rej. See, (or example, table 3-1 In: Zaln Inger Engl neer-Infi Co , Cap,]t It) F<]( tor~ .;nd Ccxti oi E/ectrKlt\ for Con\ entlon,]/Coa/ ,]nd G,]sltlc,]rlon-C~~n?blned C\c/e Po\~ w P/.]nt\, op [ it,, 1984,It IS a~surned here that amhlent temperatu rej are held constantthroughout the year at 88 F,

the gas turbines. To reach the high efficienciesprojected for the IGCC will require high-tem-perature, advanced combustion turbines. Theprojected efficiency range is somewhat higherthan the 35 to 36 percent efficiency expected forconventional plants with scrub bers. 68

Conventional turbines would yield efficienciesat the low end of the efficiency range, while ad-vanced turbines might yield higher efficiencies. 69

The choice of turbine type could significantly af-fect O&M costs in addition to efficiency. ’o For ex-ample, an advanced turbine design, while prom-ising higher efficiencies couId also entaiI greatertechnical problems and therefore higher O&Mcosts. The choice of turbine also would affect cap-ital costs. Higher efficiency turbines wouId resuItin a higher electrical output for a given gasitierand feed system; and the steam plant would berelatively smaller. Both changes would reducecapital costs per kilowatt- hour.71

Capital costs probably will range from $1,200to $1,350/kWe (net). For units in the 250 MWerange, costs are expected to be somewhat higher,about $1,600/kWe, By far the largest expensewould be the gas production and purification fa-cilities. These might account for approximately40 percent of total costs. The cost would varyespecially with gasifier design; there are indica-tions that substantial capital cost differences mayexist among leading gasifier designs,72 though themagnitude of these differences not clearly estab-lished. Costs also will vary significantly accord-ing to the degree to which redundancy is de-signed into the system. Another 40 percent of thecost would include buildings, coal receiving andpreparation equipment, an oxygen plant, wastehandling equipment, water equipment, and the

68B. M, Banda, et al,, ‘‘Comparlw)n of I ntegrat~d Coat Ga\ltl~.vtlon Combined Cycle Power Plants L\’ith Current and ,f+di an( edGas Turbines, ” Aci~,?nceci Energy SjS$ten?\- Their Rolt’ In Our F(I-ture’ Prm eedlng$ of the 19th Interwx let~ [nerg} Con~ er~lon Eng/-rwering Conrerem e, ,Auguht 19-24, 1 %94 (San Franc IV o, CA: 4mer-

Ican Nuclear %clety, 1984), paper 849507.b91 bld‘°For a dlscuislon ot relei ~nt turbine de~ elopmenri, w’e Erl( jett~

“Tokyo Congress Highlights Ettlclenc\ and N[)I Control, ” Ga5 Tur-bine \l’or/d, J,~nu,]r\-Ft>bru,ir) 1 WA, pp. 26-N).

“‘Genera I E Iw t rl[ Co, , Re~ Ie\t ,I17d Corrtment,]r\ on Design orAd\ arrceci Fo~\I/ Fue/ S}\tem< ( Falrtleld CT: General Elec trlc Co.,1982).

‘IOTA stati telephone c on~ erwtlon it Ith Bert Lou L\, El(>ctricPou er Research I n~tltute, June 6 1984


combined-cycle system .73 Finally there are accessroads, site preparation, and various civil engineer-ing tasks; together these might represent roughly20 percent of capital costs.

Operating and maintenance costs could rangefrom 6 to 12 mills/kWh, a figure roughly equiva-lent to the costs characteristic of a conventionalpulverized coal plant. The greatest source of un-certainty in the estimate concerns the perform-ance of the gasifiers, of the syngas coolers (if theyare used) and of the gas turbines.

Fuel costs for the reference plant are projectedto range from 15 to 17 mills/kWh based on 1990coal costs of $1 .78/MMBtu (see “Definitions” inappendix A for discussion of fuel costs). It is herewhere the possible cost advantage of the IGCCover the conventional scrubber equipped plantis greatest. Because of its higher efficiency, theIGCC’s fuel costs would be less than those of itsconventional counterparts. As discussed above,an important determinant of overall efficiency isthe gas turbine’s efficiency. Advanced turbineswhich are expected to be available by the early199os would be much more efficient than presentturbines. Their use could allow fuel costs to fallto the low end of the estimated range. Since fuelcosts account for a large portion of the cost ofgenerating electricity from the IGCC, the antici-pated improvement in turbine efficiency will af-fect the competitive position IGCC significantly .74

Atmospheric Fluidized=Bed Combustion

Introduction.–The AFBC is a combustiontechnology which will provide an economic alter-native to conventional pulverized coal plants inthe 1990s. Its relatively low volumes of sulfur di-oxide and nitrogen oxide emissions, great fuelflexibility, small commercially available size ( <100 MWe), easily handled solid wastes, respon-siveness to demand changes, and other features

TIH G Hernphill and M, B, jennings (Raymond Ka iser Engineers,

Inc . ) , “Offsites, Utilitles, and Genera l Fac i l i t ies fo r Coal Conver-sion Plants, o‘ Advanced Energy Systems— Their Role in Our Future:

Proceedings of the 19th Intersociety Energy Conversion Engineer-

ing conference, August 19-24, 1984 (San Francisco, CA: AmericanNuclear Society, 1984), paper 849195.

7qB. M. Bancia, et al., “Comparison of Integrated Coal Gasifica-tion Combined Cycle Power Plants With Current and AdvancedGas Turbines, ” op. cit., 1984.

offer advantages which may allow it to competesuccessfully with conventional plants, particularlyin areas where high sulfur coals are used. Invest-ment outside the utility industry in AFBC cogen-eration units already is growing rapidly, Greaterinvestment by utilities is likely in the 1990s,though various factors may keep the number oflarge utility-owned AFBCs operating by the endof the century below that which cost alone wouldset (see chapter 9).

There are two basic types of fluid ized-bed com-bustors: the atmospheric fluid ized-bed combus-tor (AFBC) and the pressurized fluidized-bedcombustor (PFBC). The PFBC operates at highpressures, and therefore can be much more com-pact than the AFBC. The PFBC also may producemore electricity for a given amount of fuel. De-spite these potential advantages, the PFBC hasmore serious technical obstacles to overcomeand is less well developed than the AFBC. It hasnot yet been successfuIly demonstrated on acommercial scale, nor are any commercial-scaledemonstrations now under construction in theUnited States. It is unlikely that more than a fewcommercial units could be completed and oper-ating before the end of the century, though thePFBCs longer term potential is quite promising.

The AFBC, the focus of this analysis, operatesat atmospheric pressures. Small-scale AFBCs al-ready are used commercially around the worldfor process heat, space heat, and in various otherindustrial applications; and are producing elec-trical power abroad as well as in very smallamounts in the United States. Three types ofAFBC installations may be important over thenext 15 years: large electric-only plants (100 to200 MWe), cogeneration installations, and non-electric systems. The electric-only units are likelyto be deployed by utilities, whereas the cogen-eration and nonelectric units probably would bebuilt and operated by others.

The cogeneration unit is an installation oper-ated to provide both electricity and usable ther-mal energy, while the nonelectric systems areused to supply usable heat only. Electric-onlyAFBCs may be new “grass-roots” plants builtfrom the ground up; or they may be “retrofits”to existing plants which have been modified to


77A Cyc lone ,s a mechan Ica I device wh IC h sepa rateS Particles from

gases by using centrifugal force.78A ~y~tem of fabric fl[ters (bags) for dust removal from stack gases.


Fluidized-bed combustors are commonly cate-gorized by the degree to which solids are en-trained in the gas-flow through the bed and towhich solids are recycled to the bed after pass-ing through the cyclone. The primary types offluidized-beds are il lustrated in figure 4-25;among these, bubbling beds and the circulatingbeds are the most important.79

The bubb/ing bed AFBC is characterized by lowgas velocities through the bed. The result is a bedfrom which only the smaller particles are en-trained with the gas; after being entrained, thesolids on the average are recycled through thebed less than once. Conversely, the gas flow ve-locities through the circulating bed are rapid. Thebed itself becomes less distinct with greater en-

79 These I n turn can be further subdivided. Among the bubbling

beds are the conventional bubbling bed, multibed and in-bed cir-culating models. Circulating systems include conventional and multi-solids bed (or hybrid} systems. (Bruce St. John, NUS Corp., ArM/y -

sis and Comparison ot FiL’e Generic FBC Systems, paper presentedat Fluldized Bed Combustion Conference, sponsored by the Gov-ernment Institutes, May 1984. )


trainment levels, as larger portions of the fuel andsorbent repeatedly are cycled through the com-buster. The fuel and limestone are thoroughlymixed as combustion of the fuel takes place.

Each of the two types of AFBC possesses cer-tain operating characteristics and peculiarities. Animportant shortcoming shared by both technol-ogies is the fact that neither have been built toproduce electric power on a scale–100 to 200MWe—attractive to utilities. And both face seri-ous technical challenges in moving from the verysmall nonelectric industrial boilers which havetypified AFBC applications so far to these largersizes.

The bubbling bed has an important advantagein that it is the older of the two types and thereis greater operating experience in the UnitedStates (in small, nonelectric, industrial applica-tions). The bubbling-bed combustor can morereadily be retrofitted to some preexisting conven-tional boilers, But the design also has its draw-backs. perhaps the most serious are the fuel-feedproblems encountered as the unit is scaled-up.it is difficult to design a reliable feed mechanismthat adequately distributes fuel to the bed; theproblem becomes progressively more difficult asthe bed is enlarged. An elaborate feed design isrequired; and the size and moisture content ofthe fuel must be carefully controlled.

By the end of this decade several large bub-bling-bed AFBCS will be operating in the UnitedStates. One is a grass-roots, 160 MWe demon-stration plant in Paducah, Kentucky (see figure4-26). Two others are retrofit units. Among theunits, two different feed systems will be used.Should serious problems be encountered in thefeed systems of the units, the deployment of thebubbling beds with capacities between 100 and200 MWe in the 199os may be seriously delayed.Favorable operation would encourage commer-cial orders of large units. Other problems asso-ciated with some bubbling bed designs, whichmay impede commercial deployment, are ero-sion and corrosion of materials which are in con-tact with the bed itself or particulate laden gases.These difficulties, should they persist, could re-sult in unacceptably high O&M costs.

With the circulating bed, by virtue of its greatergas’ velocities and higher levels of particulate re-cycling, the fuel-feed problem may be far less ofa problem, at least with smaller units. A simplerfeed mechanism can be used, and larger varia-tions in fuel size and moisture are tolerated. Effi-cient combustion and sorbent utilization is morereadily achieved. Nitrogen oxide and carbonmonoxide emissions also tend to be lower.

Being a newer “second-generation’ technol-ogy, there is less experience operating even smallcirculating-bed AFBCs. But this disadvantage israpidly disappearing. Many vendors now areoffering circu Iating beds; and almost all the ma-jor cogeneration units and many of the nonelec-tric AFBC projects now being built employ cir-culating beds.

While it is not clear whether plants using bub-bling beds, circulating beds, or some hybrid ofthe two will be favored for large grass-root plantsin the 1990s, the recent commercial trends indi-cates that the circulating beds are becoming thetechnology of preference for small cogenerationuses and a sizable share of nonelectric applica-tions. Favorable experience with these units, aswell as the single large retrofit unit using the cir-culating bed, couId decisively favor the competi-tive position of large circulating bed AFBCs in the199os. As with the large bubbling bed demon-stration units, however, difficulties with the dem-onstration retrofit unit could seriously retard thecommercial deployment of large units.

Typical AFBC Plant for the 1990s.–A largeAFBC plant typical of the kind which might mightbe deployed for electricity production in the1990s is described in appendix A, table A-5. Thetable and the following discussion focus on all-electric, grass-roots plants. By the early 1990s,U.S. utilities will have only one such plant onwhich to base evaluations of the technology. Thisis the 160 MWe demonstration unit which cur-rently is being constructed at TVA’s ShawneeSteam Plant in Kentucky; startup is scheduled for1989. Investors, however, also by the early 1990swill benefit from the technical progress and in-formation resulting from two large demonstrationretrofit units, one of 100 MWe and the other of

116 Ž New Electric Power Technologies: Problems and Prospects for the 1990s

Figure 4-26.– 160 MW AFBC Demonstration Plant in Paducah, KY

Authority

125 MWe, which also will have operated for sev-eral year-s by the early 1990s. Also important willbe experience gained from the operation of a fullycom mercia1, 125 MWe cogeneration u nit—alsoa~ retrotit—being installed by a private firm inFlorida. And many hundreds of megawatts of\mall AFBs will have been installed by 1990.

The reference AFBC plant considered in theanalysis has a generating capacity of approxi-mately 150 MWe (net). The gross electrical powerproduction of the plant actually would exceednet capacity, because power is required to oper-.~te the equipment which circulates the solids andforces air into the bed. Any commercial units con-sidered in the early 1990s are not likely to ex-ceed by very much the size of the demonstra-tion units; AFBCs are subject to scale-up problems

which probably will inhibit during the 1990s de-ployment of any commercial units much largerthan the demonstration plants.

Many features of the AFBC installations de-ployed in the 1990s, regardless of type, are likelyto be much the same. They will require accessto coal and limestone supplies; this usually meansrailroad access. A rather sizable piece of land willbe required, not only for the AFBC itself but forcoal and limestone handling and processing fa-cilities, storage areas for the limestone and coal,disposal areas for the solid waste generated bythe plant, and ponds of various sorts. Disposalof spent limestone may be one of the most seri-ous problems for the AFBC. Current estimates arethat about 1,200 tons per MWe year need to bedisposed of for 3.5 percent sulfur, Illinois coal.


For a 150 MWe plant, about 90 to 218 acrescould be required; the exact amount depends onseveral conditions. Access to water also will berequired; the 150 MWe reference plant is ex-pected to require about 1.5 million gallons eachday.

Like any large powerplant, the AFBC is ex-pected to require a considerable amount of timeto deploy. An AFBC in the 100 to 200 MWe rangepotentially has a lead-time of no more than 5years because of its smaller size and environ-mental benefits. As with the IGCC, however,lead-times of the first plants are likely to begreater, and could be as long as 10 years. Thisincludes up to 5 years for design, preconstruc-tion, and licensing activities; and 2 to 5 years forconstruction. Favorable regulatory treatment, andrapid and quality construction could result inlead-times close to the potential.

If in fact large, grass-root AFBC plants take upto 10 years to build from initial commitment,orders for them must be made by 1990 for theAFBCs to contribute appreciably to generating ca-pacity before the close of the century. Given thefact that the three large demonstration plants andnumerous small cogeneration units will be oper-ating by then, there is a possibility that consider-able numbers of large plants indeed will be i niti-ated by that time.

The operating availability of an AFBC power-plant may be around 85 to 87 percent. But con-siderable uncertainty surrounds this figure. Dif-ficulties with the fuel feed system in bubbling-bedAFBCs could severely reduce operating availabil-ity. Or erosion or corrosion associated with bothbubbling-bed and circulating-bed AFBCS couldhave similar effects.

AFBCs are expected to be used primarily asbase load plants, though their demand-followingcapabilities will allow their use in intermediateapplications. An AFBC plant is expected to lastfor approximately 30 years, and to operate withan efficiency of approximately 35 percent—some-what higher than a conventional pu Iverizecl coalplant equipped with scrubbers.

The capital cost of a large AFBC probably willbe pegged at a level roughly comparable to thatof its main competitors, the conventional scrubber-equipped plants and the IGCC. The estimate inthis analysis is $1,260 to $1,580/kWe. Fuel costsare expected to be approximately 17 mills/kWhassuming coal costs of $1.78/MMBtu. O&M costsare expected to fall between 7 and 8 mills/kWh,but high uncertainty is associated with this esti-mate. Should technical problems be experiencedwith the fuel feed system, or shouId serious ero-sion or corrosion problems arise, power produc-tion could suffer and expensive repairs and modi-fications could be required. Consequently, O&Mcosts could escalate.

The major opportunities for research whichcou Id yield technical improvements in the AFBCor reduce uncertainty about performance lie i nthe three large demonstration projects which cur-rently are underway. These projects offer thechance to experiment with basically different de-signs and to compare technologies. Of particu-lar importance will be research relating to the fuelfeed systems and to designs and materials whichcan reduce erosion and corrosion of system com-ponents.

ENERGY STORAGE TECHNOLOGIES

Introductioning periods when the marginal cost of electricity

There are several tasks that electric energy stor- is high. I n addition, storage equipment can beage equipment, employed by utilities, can per- used as spinning reserve, the backup for gener-form. The most common is load-/eve/ing, i n ating systems which fai1, or as system regulation,which inexpensive base load electricity is stored the moment-by-moment balancing of the utility’sduring periods of low demand and released dur- generation and load.


Energy storage technologies also may be usedby utilities’ customers in either remote or grid-connected applications. The latter typically in-volve the use of storage devices by utility custom-ers wishing to avoid the high price of electricityduring peak periods. Cheaper power is pur-chased during base periods and stored for useduring higher cost, peak periods.

Modular storage technologies, such as batteriesand flywheels, can be deployed in either a utility-owned or in a nonutility-owned dispersed fash-ion. However, economic considerations currentlyseem to favor large utility- or third-party-ownedinstallations. While storage technologies may atsome point be installed in conjunction with largedeployments of intermittent generating plants,such as photovoltaics or wind, storage facilitiesin the 1990s will most likely be used to storethe inexpensive output of large, conventionalplants .80

There are two storage technologies whichcould, under some circumstances, see significantdeployment in the 199os: advanced batteries andcompressed air energy storage (CAES). Batteriesare a well-established technology, familiar mostlyin mobile applications, but only recently have ad-vances in chemistry and materials made it possi-ble to construct large-scale systems with suffi-ciently long lifetimes and low capital costs toattract utility interest.

A CAES plant is a central station storage tech-nology i n which off-peak power is used to pres-su rize an underground storage cavern with air,which is later released to drive a gas turbine. Thetechnology has been demonstrated in Europe,but not in the United States.

Compared to batteries, CAES plants have sev-eral advantages. They are in a more advancedstage of technical development and are likely tobe less expensive than batteries on a dollar perkilowatt-hour basis when long discharge times(roughly 5 hours or more) are required. However,compared to batteries, CAES plants are less mod-ular, and thus carry more financial risk perproject.

~{~As Cu rr~nt IY IS the c ~rnrn~rl practice with pu roped hyciroelec-

trlc tacl Iltles; there may be some exceptions, howei er, in certainisolated areas with large potential tor renewable, such as Hawal i,

Among the storage technologies not likely tomake a significant additional contribution in the1990s are pumped hydro, flywheels, and super-conducting magnet energy storage. While thereare numerous pumped hydro plants in existencein the United States, it has become difficult to sitethese plants if they involve a large, above-groundreservoir. If all the water is stored underground,the plants are economic only in very large units. BlFlywheels, while possibly competitive in small in-stallations, e.g., cars or homes, cannot competeeconomically with batteries or CAES in larger in-stallations. B2 Finally, superconducting magneticenergy storage is not likely to be commercial be-fore the next century.

Compressed Air Energy Storage

Introduction

A CAES plant uses a modified gas turbine cy-cle in which off-peak electricity—stored in theform of compressed air–substitutes for roughlytwo-thirds of the natural gas or oil fuel necessaryto run an equivalent conventional plant (see fig-ure 4-27). in a conventional plant, the turbinemust power its own compressor to supply thecompressed air necessary for operation. Thismakes only a third of the turbine’s power avail-able to produce electricity. I n a CAES plant, how-ever, off-peak electricity is used to drive the com-pressor (through the generator running in reverseas a motor) which charges an underground stor-age cavern with compressed air. Later the air isreleased and passes through a burner where ahydrocarbon fuel such as natural gas is burned.BsThe resulting hot gases then pass through a tur-bine which, freed from its compressor, can drivethe electric generator with up to three times itsnormal fuel efficiency. The gases discharged from

8 I peter E, ‘jc ha u b, Potomac E Iect ric Power CO., comments o n

OTA electric power technologies November 1984 draft report, Jan.29, 1985.

8~ja me5 H. Sw15her and Robert R. Reeves, ‘‘Energy Storage Tech-nology, ’’Energy Systems Handbook (New York: John Wiley & Sons,February 1983).

B3 I n ,Icjd it ion to the CA ES tech no logy descrl bed here, there J resevera 1, more advanced CAES systems which reduce or ellm I natethe need for natural gas or hydrocarbon fuel. These systems wouldbe more expensive than the more conventional CAES systems, .]ndWhi Ie none hake yet been demonstrated, they c ou Id k)e developedfor the 1990s with sufficient utility Interest.


Figure 4-27.– First Generation CAES Plant

Exhaust

A f t e r

coo le r I(Cools alrbefore It

enters cavern

~—.

t

I I Turbines I I

LI

H

— Iair

the turbine pass through a “recuperator” wherethey discharge some of their heat to the incom-ing air from the cavern; this, too, increases theoverall efficiency of the plant.

Three types of caverns may be used to storethe air: salt reservoirs, hard rock reservoirs, oraquifers (see figure 4-28). Each has its advantagesand disadvantages. About three-fourths of theUnited States rests on geology more or less suit-able for such reservoirs (see figure 7- I 2 in chap-ter 7). The salt domes are concentrated mostly

in Louisiana and eastern Texas. Salt caverns are“solution-mined” by pumping water into the de-posit and having it “dissolve” a cavern. The re-suIting reservoir is virtually air-tight. These saltcaverns are pressurized to up to 80 atmospheres,have a depth of 200 to 1,000 meters, and a vol-ume of 1,000 cubic meters/MWe.

Rock caverns are located throughout theUnited States. They must be excavated with un-derground mining equipment. A typical CAESplant using a rock cavern would be coupled to

38-743 0 - 85 - 5


Figure 4-28.—Geological Formations for CAES Caverns

Wells

Imperv ious

cap rock Air shaft

Solutlon-minedcavity

I

Salt dome

Compensation

Salt caverns are mined by a techniquecalled solution-mining. A narrow well isdrilled into a salt dome and fresh water iscontinuously pumped in to dissolve thesalt while the resultant brine is pumpedout. The process is continued until thedesired storage volume is reached, The

Hard rock caverns are mined with standardexcavation techniques. A compensationreservoir on the surface maintains a con-stant pressure in the cavern as the com-pressed air is injected and withdrawn. Thisminimizes the volume of rock it is necessaryto excavate.

necessary volume is larger than that need-ed in a hard rock or aquifer systembecause without water present, thepressure of the compressed alr drops as itis withdrawn from the cavern,

SOURCE “Eighty Atmospheres In Reserve” .E/W Journa/, April 1979

an above-ground compensating reservoir whichwould maintain a constant pressure in the cav-ern as it discharges. The maintenance of constantpressure offers several important operational ad-vantages. In addition to maintaining the desiredpressure, the reservoir also allows for a muchsmaller cavern than is the case with salt reservoirs.Thus, only about 600 cubic meters/MWe are

needed underground, though a pool of about 700cubic meters/MWe of water is required on thesurface.84

M u~e of ,1 c Onl pen 5at I ng reservoir I 5 less 5u Ita hle for sa It re5er-

voi r~ because the salt dissolves i n the water. This can on Iy be pre-vented by the u5e ot’ water saturated Wtth salt, an approa( h whlc hcou Id resu It i n major en~ Iron mental ~]roblems,

Ch. 4—New Technologies for Generating and Storjng Electrjc Power . 121

The aquifer reservoirs are naturally occurringgeological formations found in much of the Mid-west, the Four Corners region, eastern Pennsyl-vania and New York. An advantage of this kindof reservoir is that it does not require any exca-vation. It consists of a porous, permeable rockwith a dome-shaped, nonporous, impermeablecap rock overlying it. Compressed air is pumpedinto the reservoir, forcing the water downwardfrom the top of the dome. Later, as air is drawnfrom the reservoir, the water returns to its origi-nal place beneath the dome. An important advan-tage of this kind of reservoir lies in the fact thatthe volume of the reservoir is quite flexible, al-lowing for a variety of plant capacities and oper-ating schedules.

With the exception of the recuperator, thetechnologies required for CAES plants—the tur-bomachinery and the reservoir-related technol-ogies such as mining equipment—are well-estab-lished technologies. The turbomachinery is onlya slight modification of currently used equipmentand there are several manufacturers, Americanand foreign, that offer CAES machinery with fuIIcommercial guarantees, While there are somequestions as to the dynamic properties of the airas it enters and leaves a cavern, there is littledoubt that the technology exists to build andmaintain underground storage facilities. Thesecaverns have been used for years to store natu-ral gas and other hydrocarbons, and the samefirms that supply the oil and gas industry haveoffered to provide utilities with CAES caverns thatcan be warranted and insured. 85

Despite the relative maturity most of the com-ponents which make up the CAES plant, therehas been no experience in the United States withCAES itself–though a CAES plant using a salt cav-ern has been in operation since 1978 at Huntorf,West Germany (see figure 4-29) and has per-formed well. This lack of domestic experiencewith the technology constitutes the largest hur-dle facing CAES. There is a general reluctanceamong utilities in this country to be the first tomake a commitment to build a plant. While sev-eral utilities have made preliminary planning— .

‘ r’ Perw~n<]l c (JrrtJ\\)fjn(j(’n[ c het~~ ~~[~n Arnold FIC k(>tt f EltI[ t rl(Pow t>r R(+e,]rc h Initltut[’) ,lnd OTA \t,ltt, Ju ly 2, 1984

Figure 4-29.—The Huntorf Compressed Air EnergyStorage Plant in West Germany

[n the foreground is the wellhead, where compressed air IS Injectedand released The rest of the plant is in the background

SOURCE BBC Brown Boven, Inc

studies, the Soyland Electric Cooperative in 11-Iinois is the only American utility that has ordereda plant. This plant, however, was for various rea-sons canceled and no project has been initiatedsince then.

Typical CAES Plant for the 1990s

CAES plants in the 1990s are likely to be avail-able in two modular unit sizes, 220 MWe, com-monly called maxi-CAES, and 50 MWe, mini -CAES. These sizes are determined by the sizesof existing models of turbomachinery—the tur-bines, compressors, generator/motor, and a gear-box which connects them.

A CAES plant must be sited in an area with ac-cess to water and fuel. The turbomachinery re-quires about 2,000 ga[lons/MWe of water perday, and a plant with a rock cavern needs addi-tional water for the compensation reservoir. Bothmini- and maxi-CAES plants burn about 4,000Btu/kWh of fuel and emit the standard combus-tion byproducts, such as nitrogen oxide, but atonly a third of the level of a similar size conven-tional gas turbine. CAES plants also have noiselevels similar to those of more conventionalplants. There are several waste disposal problemsinvolved with building the caverns. If a rock cav-ern is used, it is necessary to dispose of a large


volume of waste rock,8b and when a salt cavernis used, the brine pumped out of the cavern mustbe disposed of. Land requirements would rangefrom around 15 acres for a maxi-CAES plant to3 acres for a mini-CAES plant.

The lead-times expected for CAES plants willprobably range from 4 to 8 years. The large plantswould occupy the higher end of the range, whilethe smaller units would fall at the lower end. Theprimary source of uncertainty in lead-time esti-mates concerns licensing and permitting, whichis expected to take 2 to 4 years. Regulatory hur-dles will vary depending on the type of reservoirused. Among the regulatory impediments arethose relating to the disposal of the hard rock orbrine from the mining operation, and relating towater usage and impacts. Also problematic maybe the requirements of the Powerplant and in-dustrial Fuel Use Act of 1978. Even though a CAESplant is an oil and gas saving device, the fact thatit uses these fuels means a utility must receivean exemption from the act to operate one. Aprecedent was established when such an exemp-tion was granted to the Soyland Electric Coop-erative, but under the current regulations, exemp-tions would be required for every CAES plant .87

The properties of the two sizes are similar (seetable A-8, appendix A); the mini-CAES turbo-machinery costs somewhat less—$392/kWe vs.$51 5/kWe for the maxi-CAES. The storage cav-erns can be formed out of three types of geologi-cal formations: aquifers, salt deposits, and hardrock. In general, aquifers are the least expensive,followed closely by salt. Rock caverns, whichmust be excavated, are by far the most expen-sive. on a total dollars per kilowatt basis, cavernsfor maxi-CAES plants are less expensive thanthose for mini-CAES.

BGThiS problem is greatly alleviated by the fact that the excavated

material can be used in constructing the compensating reservoiror other facilities (Peter E. Schaub, comments on OTA electric powertechnologies November 1984 draft report, op. cit., 1985).

HTp, L, Hendrickson, Lega/ and Regulatory Issues ~ffWf/ng COnl-

pressed Air Energy .Storage (Rlchland, WA: Pacific Northwest Lal]-oratory, July 1981 ), PNL-3862, UC-94b.

Advanced Batteries

Introduction

Batteries are more efficient than mechanicalenergy storage systems, but their principal advan-tage is flexibility. Batteries are modular so thatplant construction lead-times can be very shortand capacity can be added as needed. Batterieshave almost no emissions, produce little noise(though because of pumps and ventilation sys-tems, they are not silent), and they can be sitednear an intended load, even in urban areas. Abattery’s ability to rapidly begin charging or dis-charging (reaching full power in a matter ofseconds, as opposed to minutes for a CAES sys-tem) makes it valuable for optimizing utility oper-ations. However, battery-storage installations donot benefit very much from economies of scaleeither in capital costs or in maintenance costs,so that if large blocks of storage are required,CAES may be less expensive. Also, though costeffective and reliable in numerous remote appli-cations, battery technology has not yet achievedthe combination of low cost, good performance,and low risk necessary to stimulate investmentin grid-connected applications.

There are two types of utility-scale batterieswhich under some circumstances could be par-ticularly important in the 1990s: advanced lead-acid batteries, and zinc-chloride batteries. Lead-acid batteries are in wide use today mostly inautomobiles and other mobile applications; ad-vanced lead batteries constitute an incrementalimprovement over the existing technology. Zinc-chloride batteries are a newer technology, andconstitute a fundamental departure from the con-ventional lead-acid battery. I n both cases, indi-vidual modules similar to commercial moduleswhich might be deployed in the 1990s, have beentested at the Battery Energy Test Facility in NewJersey. 88 Though neither type of battery has beendeployed yet in a multi megawatt commercial in-stallation, plans to do so during the late 1980sare being developed and implemented.

Other battery technologies meanwhile are be-ing pursued. Among these, the most promisingappear to be zinc-bromide batteries and sodium-

~Bsee ch, g for further detai Is on this f~c i Iity.

—

Ch. 4—New Technologies for Generating and Storing Electric Power . 123

Typical Battery Installation for the 1990s

If battery technology is deployed in the 1990sby utilities, the general requirements of a typicalplant, regardless of the battery technology em-ployed, are expected to be a peak power out-put of 20 MWe and a storage capacity of about100 MWh. Such a plant could consist of about10 to 50 factory built modules, along with con-trol and power conditioning equipment, housedin a protective building (see figures 4-30 and q-31). Battery installations outside the util ity-industry might be considerably smaller.

The total land necessary will depend on boththe so-called “energy footprint” (energy densityi n kilowatt-hour per square meter) of the particu-lar battery technology as well as the amount ofspace necessary for easy maintenance. Each ofthe reference battery installations discussed herewill require about 0.02 to 0.03 acres. There areno fuel and only minimal water requirements.

The lead-time required to deploy battery instal-lations is expected to be very short. Because ofthe comparatively low environmental impacts ofthe installation, licensing could proceed quiterapidly. And since the battery moduIes are fac-tory built, construction can be very rapid too. Thelead-time of the plant should be less than 2 years.There is, however, uncertainty regarding the timerequired for Iicensing and permitting. Concernover possible accidents and disposal of hazard-ous materials, discussed in greater detail below,couId be a source of reguIatory delays particu -


Figure 4-30.—Generic Battery System

Enclosure

Enclosure

Ch. 4—New Technologies for Generating and Storing E/ectric Power • 125

Figure 4-31 .—A Commercial Load-Leveling Zinc-Chloride Battery System

Englneerlng Conference San Francisco CA, Aug 19-24, 1984


Figure 4-32.— Lead-Acid Batteries

The il lustrations below show how a lead-acid battery stores electrlc energy. Advanced lead-acid batteries differ (n the construction of the elec-

trodes, etc., but the basic operation is the same as the more traditional designs.

In its fully charged state, the negative electrode consists of spongylead with a small mixture of antimony (around 10 percent), while thepositive electrode is lead dioxide. The electrolyte is sulfuric acid.

Electron flow

.

Fully charged

When fully discharged, the battery’s electrodes are almost entirelylead sulfate and the electrolyte IS largely water.

electrode electrodePbS0 4 PbS0 4

Negative Positiveelectrode electrode

\ ADischarging

To charge the battery, electrons are driven back into the negative

– Discharged

4 4 .2

Electron flow

l . -

electrode electrode

Charging

4


The safety hazards of advanced lead-acid bat-teries occur primarily if the battery is over-charged. In this instance, it will generate poten-tially explosive mixtures of hydrogen and oxygenwhich must be ventilated. Stibine and arsine canalso be formed from the materials used i n theelectrodes. Finally, there are also dangers fromacid spills and fire. When the battery is decom-missioned, the lead must be recycled, and theacid disposed of. However, there is much experi-ence with lead-acid batteries, and few safetyproblems are anticipated if well-established main-tenance and safety procedures are followed.

Zinc-Chloride Batteries.–The zinc-chloridebattery has been under development since theearly 1970s. It is a flowing electrolyte battery (seefigure 4-33). During charging, zinc is removedfrom the zinc-chloride electrolyte and depositedonto the negative graphite electrode in the bat-tery stack, while chlorine gas is formed at thepositive electrode. The gas is pumped into thebattery sump, where it reacts with water at 10°C to form chlorine hydrate, an easily managableslush. During discharge, the chlorine hydrate isheated to extract the chlorine gas, which ispumped back into the stack, where it absorbs thezinc and releases the stored electrical energy.

A principal advantage of the zinc-chloride bat-tery is that it promises to be ultimately less ex-pensive than the lead-acid battery, due primar-ily to the inexpensive materials that go into itsconstruction. However, the technology, whichrequires pumps and refrigeration equipment, ismore complex—it is sometimes described as be-ing more like a chemical plant than a battery. 94

Since no commercial design zinc-chloride bat-tery has yet been operated, any cost projectionsmust be taken with some caution. (See appen-dix A, table A-9.)

Estimates indicate that at a production level ofabout 700 MWe/year, zinc-chloride batteriescould be sold at a price less that-t $500/kWe, Be-cause zinc-chloride batteries will most likelymake their first appearance in grid-connected use(unlike lead-acid batteries which are already soldin other markets), this price is likely to depend

‘W)TA statt;nter~ Iew~ with ( 1 ) Arnold Fickett, op. cit., 1984 andi2) J.J, KelleY, EXIDE Corp., Aug. 29, 1984,

strongly on the volume produced. If only 50MWe/year were made, the price could be about$860/kWe; and early commercial units could costas much as $3,000 /kWe.

The zinc-chlorine battery may have a longerlifetime than the lead-acid battery. The best cellshave run for 2,500 cycles, and while there havebeen numerous problems with pumps and plumb-ing, no basic mechanisms have been identifiedwhich would limit the lifetime to less than 5,000cycles. 95 However, the 500 kWh test module atthe BEST facility has only run for less than 60 cy-cles, and several tough engineering problemshave yet to be overcome before the battery canhave a guaranteed lifetime long enough for com-mercialization. In addition, the AC to AC round-trip efficiency, which is currently in the low 60to 65 percent range for the large battery systems,must be increased to 67 to 70 percent; values inthis range have been attained by smaller prototypes.

The O&M requirements of zinc-chloride sys-tems are even more uncertain than for lead-acidsystems. However, the expected longer lifetimes,and the less expensive replacement costs for thestacks and sumps (estimated to be about one-third the initial capital cost of the battery) shouldlead to Ievelized replacement O&M costs in the3 to 9 mills/kWh range. For lack of better dataon operating experience, the annual O&M costsare estimated to be the same as for lead-acid, I

to 4 mills/kWh, though because of the increasedcomplexity of the system, they probably will behigher.

Another major advantage of the zinc-chloridebattery over the lead-acid battery is that their re-action rates are controllable. This is due to thefact that, in a charged zinc-chloride battery, thezinc and the chlorine are separated in the stacksand sumps. The rate at which the battery dis-charges is controlled by the speed at which thepumps allow the reactants to recombine. This notonly makes the battery more flexible in its oper-ation, but provides a major safety advantage inthat if a zinc-chloride cell malfunctions, its dis-charge can be stopped by shutting off the chlo-rine pumps. In contrast, the reactants in a

~JOTA stdf~ I nter~lew with Arnold Fickett, op. cit. ~ J 984.

-—


Figure 4-33.—Zinc-Chloride Flowing Electrolyte Batteries

H

H

Fully charged

When the battery IS fully charged, the negative electrode IS platedwith zinc and the electrolyte has only a weak concentration of zincchloride. The battery sump IS filled with chilled chlorine hydrate.

Elefl

Zinc [ /ZnCll(aq)

C I1 = H 20’ H20

Charging

Heat

As the battery is charged, chlorine ions from the electrolyte combineat the positive electrode to form chlorine gas and release twoelectrons. These electrons are driven to the negative electrode bythe charging generator There they combine with zinc being platedonto the negative electrode, The chlorine gas is pumped to the sumpwhich has been chilled to below 10 “C, The gas reacts with the coldwater and forms an easily storable solid, chlorine hydrate.

flow

Discharge

The battery IS discharged by heatingsump, which then releases the chlorine gas. This gas is pumped tothe stack, where it combines with electrons from the positiveelectrodes and breaks into chlorine ions. At the negative electrodethe zinc atoms release electrons and enter the electrolyte as zincIons.

Negative electrode: Zn - Zn + + e -

Ch. 4—New Technologies for Generating and Storing Electric Power Ž 129

charged lead-acid battery cell remain in the samebattery case, so that short-circuited terminalscould lead to a sudden release of the storedenergy.

A major concern of regulatory ofticials consid-ering zinc-chloride plants is likely to be the safetyproblems associated with the accidental releaseof chlorine. Since the chlorine is stored in a solidform, there is no danger of a sudden release of

large quantities of the gas. However, the sameprocedures used in industrial plants manufactur-ing or using this gas must be followed. In addi-tion, the sum ps must be sufficiently insulated sothat in the event of a malfunction of the refriger-ation system, the chlorine will stay frozen in thechIoride hydrate phase long enough for repairsto be made.

SUMMARY OF CURRENT ACTIVITY

The tables in appendix A at the end of this re- i n the U n ited States. The extent to which capac-port summarize the cost and performance char- ity already has been deployed or is being con-acteristics discussed in this chapter. Table 4-2 structed provides an additional indication of thesummarizes the information detailed in the ap- cost, performance and risk associated with thependix. Table 4-4 provides an overview of the technologies.plants which currently are installed or operating

Table 4.4.—Developing Technologies: Major Electric Plants Installed or Under Construction by May 1, 1985

Technology Capacity Location Primary sources of funds Status

Wind turbinesa . . . . . . . . . 550+ MWe (gross)b

Solar thermal electric:Centrai receiver . . . . . . . .

Paraboiic trough . . . . . . . .

Parabolic dish . .

Solar pond. . . . . . . . . . . . .Photovoltaics:

Fiat piate ... . . . . . . . .

Concentrator. . . . . . .

Geothermal:Dual flash ., . . . . . . . . . . .

Binary:Smail ., . . . . . . . . . . . . .

Large . . . . . . . . . . . . . . .

100+ MWe (gross)c

? MWed

10 MWe (net)e

0.75 MWe

14 MWe (net)30 MWe (net)

0.025 MWe (net)f

2 x 0,025 MWe (net)f

2 x 0.025 MWe (net)f

3.6 MWeNone

1 MWe (de, gross)1 MWe (de, gross)1 MWe (de, gross)

6.5 MWe (de, gross)0.75 MWe (de, gross)

4.5 MWe (de, gross)1.5 MWe (de, gross)3.5 MWe (de, gross)

10 M’We10 MWe47 MWe (net)32 MWe (net)

2 x 3.5 MWe3 x 0.3 MWe3 x 0.4 MWe

10 MWe1 x 0.75 MWe (gross)3 x 0.35 MWe (gross)3 x 0.45 MWe (gross)4 x 1.25 MWe (gross)3 x 0.85 MWe (gross)

45 MWe (net)

California wind farmsU.S. wind farms outside

of CaliforniaAll U.S. wind farms

Daggett, CA

Albuquerque, NM

Daggett, CADaggett, CAPalm Springs, CAVarious iocationsVarious locations,Warner Springs, CA

SacramentoSacramento, CAHesperia, CACarrisa Plains, CACarrisa Plains, CABorrego Springs, CADavis, CABarstow, CA

Brawley, CASalton Sea, CAHeber, CASalton Sea, CA

Mammoth, CAHammersly Canyon, ORHammersly Canyon, OREast Mesa, CAWabuska, NVLakeview, ORLakeview, ORSuifurviiie, UTSulfurville, UTHeber, CA

NonutilityNonutility

Nonutility

Utility, nonutility, andGovernment

Utility, nonutiiity, andGovernment

NonutilityNonutiiityGovernmentNonutilityNon utilityNonutility

Utility and GovernmentUtility and GovernmentNonutilityNonutilityNonutilityNonutilityNonutilityNonutiiity

Utility lnonutiiityUtilitylnonutilityNonutilityNonutility

NonutilityNonutilityNonutilityNonutilityNonutiiityNonutiiityNonutiiityNonutiiityNonutilityUtility, nonutility, and

Government

InstailedInstalled

Under construction (1986)

installed

Installed

InstalledUnder construction (1986)InstalledInstailedUnder constructionInstalled

InstalledUnder construction (1985)instailedInstalledUnder constructionInstaiiedgInstal ledInstaiiedg

InstalledInstalledUnder construction (1985)Under construction (1985)

InstalledInstalledInstaliedh

installedInstaliedInstaiiedh

Instailedh

Under construction (1985)!Under construction (1985)1

Installed

—— —— . . —— — — — — — —


Table 4-4.—Developing Technologies: Major Electric Plants Installed or Under Constructionby May 1, 1985-Continued

Technology Capacity Location Primary sources of funds Status— —Fuel cells:

Large, . . . . . . . . . . . . . . . . .SmallJ . . . . . . . . . . . . . . . . .

Small cogeneration . . . . .

IGCCn . . . . . . . . . . . . . . . . . . .

Batteries:Lead acid”. . . . . . . . . . . . .Zinc chloride. . . . . . . . . .

CAES:Mini . . . . . . . . . . . . . . . . . .Maxi . . . . . . . . . . . . . . . . . .

28 MWe2.8 MWe

24 MWe20 MWe

100 MWe

Pontiac, MlWashington, DC

Enfield, MEChinese Station, CADaggett, CA

0.5 MWe Newark, NJ Utility and GovernmentNoneP

NoneNone

Installed

Under construction

Under construction (1989)Under construction (1987)Under construction (1986)Under construction (1986)Under construction (1985)Under construction (1986)Under construction (1987)Under construction (1986)Under construction (1986)Under construction (1987)Installed

Under construction (1986)Installed

Under construction (1986)Under construction (1986)Installed

Installed

alncludes small. and medium-sized wind turbines.bApproximately 550 Mwe were operating in California at the end of 1984, It is not known how much additional capacity was Installed by May 1985cApproximately 100 Mwe were operating outside of California at the end of 1984, It is not known how much additional capacity had been installed outside Cal if Ornla

by May 1985,dlt is not known how much capacity was under construction on May 1, 1985.eThis facility, the ~lar One pilot plant, is not a Commerc ial. scale plant and differs in other impo~ant ways from the type of system which might be deployed commer-

cially in the 1990s.fThis installation consists of only one electricity producing module; a commercial installation probably would COnSiSt of hundreds of modules.90nly 10 percent of the modules were operating at the time because of problems with the Power conversion sYStems.hlnstatled but not operating, pending contractual negotiations with utilitiOS.iThe equipment ~dules have been delivered to the site; site preparation, however, has not started.iThGs.e un~ts are not cornrnerc~a~.sc~e units.klncluding the Electric Power Research lnStitUtO.IThis ,s the total capacity which may be generated from the four AFBC boilers which Will be installed,mThis is the tot~ capacity which may be generated from the two AF8C boilers which will be installed.nwhile this inst~lation, the Cool Water unit, uses commercial. scale components, the installation itself IS nOt a Commercial-SCa!.S installation,owhile this installation at the Batte~ Energy stora~e Test Facility uses a commercia~.scale batte~ module, the installation itseif iS not a COITrrl)OrCial-SCa[O inStaiiatiOn

PA 0.5.Mwe zinc chloride commercial-scale battery module was, however, operating at the Battery Energy storage Test facility until early 1985.

SOURCE Office of Technology Assessment.

chapter new technologies for generating and storing

Documents