integration of renewables in swiss energy system schaffen wissen – heute fÜr morgen integration...
TRANSCRIPT
WIR SCHAFFEN WISSEN – HEUTE FÜR MORGEN
Integration of Renewables in Swiss Energy System
Evangelos Panos :: Energy Economics Group:: Paul Scherrer Institute
Kannan Ramachandran :: Energy Economics Group:: Paul Scherrer Institute
72th Semi-annual ETSAP meeting, 11-12th December, Zurich
Overview of the Swiss Energy System, 2015
Page 2
0
20
40
60
80 Solar
Wind
Other
Nuclear
Hydro
ELECTRICITY GENERATION (TWh)
0
5
10
15
20
25Solar
Wind
Other
Nuclear
Hydro
ELECTRICITY CAPACITY (GW)
0
200
400
600
800
1000 Heat
RES
Wood
Electricity
Other
Gas
Oil
Coal
FINAL ENERGY CONSUMPTION (PJ)
0
10
20
30
40Power generation
Services and agriculture
Industry
Residential
Transport
ENERGY RELATED CO2 EMISSIONS (Mt)
Swiss energy strategy 2050:
aims at gradually phasing out nuclear and promoting renewables and demand side efficiency
56%
38%
13.7
3.3
Solar
• We study integration measures for variable renewable generation (VRES) from wind
and solar PV in Switzerland for the horizon 2015 – 2050, such as:
Reinforcing and expanding the grid network
Deploying local storage, complementary to pump hydro, like batteries and ACAES
Deploying dispatchable loads such as P2G, water heaters and heat pumps
• The study was performed in the context of the ISCHESS project, which is a
collaboration between the Paul Scherrer Institute and the Swiss Federal Institute of
Technology (ETH Zurich), funded by the Swiss Competence Center Energy and
Mobility (CCEM) http://www.ccem.ch/ischess
Objectives of the research
Page 3
• High intra-annual resolution with 288 typical hours (3 typical days, 4 seasons, 24h/day)
• In this research, the model includes:
Higher detail in the electricity sector, variability in the RES generation, ancillary
services and power plant dispatching constraints
But, it excludes oil transport and has aggregate representation of industry
Methodology – The Swiss TIMES Energy Systems Model (STEM)
Page 4
Swiss TIMES Energy system Model (STEM)
Fuel supply
module
Fuel
distribution
module
Demand modulesElectricity supply
module
Resource module
Electricity import
Uranium
Natural gas
Hydrogen
Electricity export
Electricity
Gasoline
Diesel
Renewable· Solar
· Wind
· Biomass
· Waste
Electricity
storage
Hydro resource· Run-of rivers
· Reservoirs
CO2
Demand
technologies
Residential- Boiler
- Heat pump
- Air conditioner
- Appliances
Services
Industires
Hydro plants
Nuclear plants
Natural gas
GTCC
Solar PV
Wind
Geothermal
Other
Taxes &
Subsidies
Fuel cell
Energy
service
demands
Person
transportat
ion
Lighting
Motors
Space
heating
Hot water
Oil
Transport
Car fleetICE
Hybrid vehicles
PHEV
BEV
Fuel cell
Bus
Rail
Mac
roec
onom
ic d
river
s (e
.g.,
popu
latio
n, G
DP
, flo
or a
rea,
vkm
)
Inte
rnat
iona
l ene
rgy
pric
es (o
il, n
atur
al g
as, e
lect
ricity
, ...)
Tec
hnol
ogy
char
acte
rizat
ion
(Effi
cien
cy, l
ifetim
e, c
osts
,…)
Res
ourc
e po
tent
ial (
win
d, s
olar
, bio
mas
s, …
.)
Biofuels
Biogas
vkm-Vehicle kilometre
tkm-tonne kilometre
LGV-Light goods vehicles
HGV-Heavy good vehicles
SMR-steam methane reformer
GTCC-gas turbine combined cycle plant
Oil refinery
Process
heat
FreightsTrucks
HGV
Rail
Natural gas
Heating oilOther
electric
• Different grid levels, characterised by transmission costs and losses
• A set of power plants can be connected to each level:
Power plants are differentiated by costs, efficiency, technical constraints & availability
• A linearised approximation of the Unit Commitment problem is formulated
Representation of the electricity sector in STEM
Page 5
Very High Voltage Grid Level 1
NuclearHydro DamsImports Exports
Distributed Power Generation
Run-of-river hydroGas Turbines CCGas Turbines OCGeothermal
High Voltage Grid Level 3
Large Scale Power Generation
Medium Voltage Grid Level 5
Wind FarmsSolar ParksOil ICEWaste Incineration
Large scale CHP district heating
Wastes, BiomassOilGasBiogasH2
Low Voltage Grid Level 7
Large Industries & Commercial
CHP oilCHP biomassCHP gasCHP wastesCHP H2Solar PVWind turbines
Commercial/Residential Generation
CHP gasCHP biomassCHP H2Solar PVWind turbines
Lead-acid batteriesNaS batteriesVRF batteries
PEM electrolysis
Pump hydro
CAES
Lead-acid batteriesNaS batteriesVRF batteries
Li-Ion batteriesNiMH batteries
PEM electrolysis
Lead-acid batteriesLi-Ion batteriesNiMH batteries
+ 4 nodes for nuclear
power plants
• With a reduction algorithm from FEN/ETHZ we map the detailed transmission grid to an
aggregated grid with 𝑁 = 15 nodes and 𝐸 = 319 lines, based on a fixed disaggregation
of the reduced network injections to the detailed network injections
Representation of electricity transmission grid
Page 6
MAPPING
−𝐛𝐝 ≤ 𝐇𝐝 × 𝐠𝐝 − 𝐥𝐝 ≤ 𝐛𝐝
𝐇𝐝 is the PTDF matrix of the detailed network, 𝐃 is the fixed dissagregation matrix,𝐠𝐝 , 𝐠𝐚 are vectors with elelectricity injections,
𝐥𝐝 , 𝐥𝐚 are vectors of electricity withdrawals, and 𝐛𝐝 , 𝐛𝐚 are vector of line capacities; superscripts d=detailed, a=aggregated network
The matrix 𝐃 is not unique, since there are infinite ways in which an aggregate injection can be distributed between multiple nodes;
here, it allocates power injections according to the original distribution of generation capacity in the detailed power flow model
−𝐛𝐚 ≤ 𝐇𝐝 × 𝐃 × 𝐠𝐚 − 𝐥𝐚 ≤ 𝐛𝐚
DETAILED NETWORK AGGREGATED NETWORK
• To capture variability, we need the variance of the mean availability of wind and solar
per typical hour in STEM
• By bootstrapping a 20-year meteorological sample, we calculate the distribution of
mean availability per typical hour in STEM
• We move ± 3 st.dev in this distribution to introduce variability of RES per typical hour
Representation of stochastic RES variability
Page 7
Bootstrapped Distribution of Mean Photovoltaic
Capacity Factors (availability): Summer (left), Winter (right)
• Storage capacity must accommodate downward variation of the Residual Load Curve
(RLDC) and upward variation of non-dispatchable generation
• The dispatchable peak generation capacity (incl. storage) must accommodate upward
variation of the RLDC and downward variation of non-dispatchable generation
Source: Singh, A. PSI, 2016
• Power plants commit capacity to the reserve market based on their operational
constraints and shadow price of electricity generation and reserve provision
• In each of the typical hours the demand for reserve is endogenously calculated from
the joint probability distribution function (p.d.f.) of the individual p.d.f. of forecast
errors of supply and demand
The sizing is based on both probabilistic and deterministic assessment
We move ± 3 s.d. on the joint p.d.f to estimate the reserve requirements
Ancillary services markets – provision of reserve
Page 8
The standard deviation of the joint probability
distribution of forecast errors is :
𝜎2𝑠𝑜𝑙𝑎𝑟
+ 𝜎2𝑤𝑖𝑛𝑑
+ 𝜎2𝑙𝑜𝑎𝑑
And the reserve requirement in hour 𝑡 is given as:
By assuming uncorrelated random variables:
𝑅𝑡 = 3 ∗ 𝜎2𝑠𝑜𝑙𝑎𝑟
∙ 𝐺𝑡2
𝑠𝑜𝑙𝑎𝑟+ 𝜎2
𝑤𝑖𝑛𝑑∙ 𝐺𝑡
2
𝑤𝑖𝑛𝑑+ 𝜎2
𝑙𝑜𝑎𝑑∙ 𝐿𝑡
2 + 𝑎 ∙ 𝑃𝑚𝑎𝑥
𝐺𝑖 generation from option 𝑖 , 𝐿 load, 𝑃𝑚𝑎𝑥 largest grid element
A range of “what-if” scenarios was assessed along three main dimensions:
1. Future energy policy and energy service demands, assuming nuclear phase out by 2034:
Energy service demand growth: low growth vs high growth
Electricity imports: allow net imports vs autarky
Climate change mitigation policy: mild policy vs 70% reduction in CO2 emissions
in 2050 from 2010 levels
2. Location of new gas power plants and installed capacity as % of the total national capacity
3. Grid expansion: allowing grid reinforcement beyond the “Grid 2025” plans of Swissgrid or not
in total about 100 scenarios were assessed with the STEM model based on the
Cartesian product of the above combinations across the three dimensions
Long term scenarios analysed with STEM
Page 9
Corneux (NE) Chavalon (VS) Utzenstorf (BE) Perlen (LU)
Schweizerhalle
(BL)
Case 1 20.0 20.0 20.0 20.0 20.0
… … … … … …
Case 11 0.0 33.3 33.3 33.3 0.0
… … … … … …
Case 26 33.3 33.3 0.0 0.0 33.3
• Electricity consumption increases 4 – 30% from 2015 ( 0.1 – 0.8% p.a)
• New gas power plants replace existing nuclear capacity
• Under climate policy VRES could provide up to 28% of the domestic supply
Electricity consumption continues to increase and gas, VRES & imports replace nuclear by 2050
Page 10
-100
1020304050607080
2015
2050 (max)
2050 (min)
ELECTRICITY GENERATION & CONSUMPTION IN 2050 (TWh)
The results correspond to ranges among the 100 scenarios assessed
• The requirements for secondary reserve almost double in 2050 from today’s level and
peak demand shifts from winter to summer
• Hydrostorage is the main contributor, followed by gas turbines (after 2040)
• Flexible CHPs and batteries could participate in the reserve provision by forming
“virtual plants”, especially in those scenarios with high reserve requirements
Electricity consumption continues to increase and gas, VRES & imports replace nuclear by 2050
Page 11
SECONDARY RESERVE MAXIMUM DEMAND AND PROVISION IN 2050 (MW)
0100200300400500600700800
2015
2050 (max)
2050 (min)
Maximum contribution per technology
The results correspond to ranges among the 100 scenarios assessed
• High shares of VRES require electricity storage peak capacity of ca. 30 – 50% of the
installed capacity of wind and solar PV (together)
• Above 14 TWh of VRES generation there is a significant storage deployment
• About 13% of the excess summer VRES production is seasonally stored in P2G (~ 1 TWh)
Storage needs increase with VRES deployment
Page 12
0.00.51.01.52.02.53.03.54.04.55.05.5
8 10 12 14 16 18 20 22 24
Inst
alle
d s
tora
ge c
apac
ity
(GW
)
TWh of wind and solar PV electricity production
Pump hydro(pumpingcapacity)
Batteries (4hmaxdischarge)
P2G(only as
seasonal
storage)
• Small scale batteries are driven by
distributed solar PV installations
• Medium scale batteries are driven by
wind and large scale PV and CHP
• Large scale batteries complement pump
hydro storage when it is unavailable
Max. Total battery capacity 5.5 GW
Large scale batteries 0.5 GW
Small scale
batteries 3 GW
Medium scale
batteries 2 GW
WIND AND SOLAR PV ELETRICITY VS STORAGE CAPACITY
Each data point in the graph corresponds to a different long term scenario and year
Contribution of each sector to the total electricity stored in water
heaters and heat pumps
Residential 85-97%
Industry <2%
Services 3-23%
• Electricity storage in water heaters and heat pumps could account for 8 – 24% of the
total electricity consumption for heating
• Accelerated deployment of electric load shifting when electricity consumption for heat
exceeds 13 TWh
• Large potential for load shifting is in water heating (resistance heating) followed by
space heating in buildings
Dispatchable loads help in easing electricity load peaks in the stationary end-use sectors
Page 13
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
8 13 18
Ele
ctri
city
sto
red
in w
ate
r h
eat
ers
an
d h
eat
pu
mp
s (T
Wh
)
Electricity consumption for heating (TWh)
ELECTRICITY STORED IN WATER HEATERS AND HEAT PUMPS
VS ELECTRICITY CONSUMPTION IN HEATING IN 2050 (TWh)
Each data point in the graph corresponds to a different long term scenario
• Restrictions in grid expansion results, due to congestion, in:
Higher system costs, up to 90 BCHF (+10% when grid expansion is allowed) in 2020-50
Non-cost optimal deployment of electricity supply options
Less electrification of demand and reliance on fossil-based heating
The system-wide benefits from the electricity grid expansion outweigh the costs
Page 14
1100
1200
1300
1400
1500
1600
1700
No gridexpansion
Gridexpansion
No gridexpansion
Gridexpansion
Reference Stingent climate policy
Max
Min
CUMULATIVE UNDISCOUNTED ELECTRICITY AND HEAT SYSTEM COST, BHCF/yr, 2020 – 2050
The results correspond to ranges among the 100 scenarios assessed
• Grid expansion results in cost savings, mainly in heating sectors, directly (e.g. technology
change via heat pumps) and indirectly (e.g. less costs for imported fuels)
Under a stringent climate change mitigation policy and high demand, these costs could
be up to 3 BCHF/yr
About 1/3 of the savings occur in the electricity supply and 2/3 in the stationary sectors
The system-wide benefits from the electricity grid expansion outweigh the costs
Page 15
-3500-3000-2500-2000-1500-1000
-5000
5001000
Electricitypowerplants
ElectricityT&D
Netimports ofelectricityand fuels
Heatsupply
Total
Max
Min
ELECTRICITY AND HEAT SYSTEM COSTS SAVINGS DUE TO GRID EXPANSION, MCHF/yr, 2020-2050
The results correspond to ranges among the 100 scenarios assessed
• Electricity consumption continues to increase by 0.1 – 0.8% p.a., and especially
after 2030 due to new electricity uses; it could reach over 70 TWh/yr by 2050
• VRES can contribute up to 24 TWhe (or 28% of the domestic supply) after the
nuclear phase-out
• Storage peak capacity investments about 30 – 50% of the installed wind and solar
PV capacity; beyond 14 TWhe accelerated deployment of storage is inevitable
• About 13% of the excess electricity production from VRES in summer is seasonally
stored in P2G pathways, driven by the high seasonal electricity price differences
under stringent climate change mitigation scenarios
• Water heaters and heat pumps could contribute in easing electricity peaks and
they could shift 8 – 24% of the electricity consumption for heat
Conclusions
Page 16
• Grid reinforcement results in net economic benefits for the whole electricity and
heat supply system of Switzerland on the order of 0.5 – 3.0 BCHF/yr. over the
period 2030 - 2050
• Rebound effects prevail in grid reinforcement, and a social planner shall prioritize
the expansion of those lines with high benefit/cost ratio for the energy system
(insights are provided form the dual values of the grid constraints in the model)
• Further work is needed to overcome some important limitations, such as:
Regional representation also for the heat supply and not only for electricity
Consideration of N-2 grid security constraints
More detail in technical representation of storage technologies (e.g. depth of
discharge, ramping rates, max discharging times, etc. )
Conclusions ctd.
Page 17
Page 18
Wir schaffen Wissen – heute für morgen
Thank you for your attention. Evangelos Panos
Energy Economics Group
Laboratory for Energy Systems Analysis
Paul Scherrer Institute
• Without batteries and grid, there is 30 – 50% less deployment of wind and solar electricity
compared to the case when both options are available
Batteries are important for the integration of VRES to cope with their variability
Grid expansion is important to integrate large amounts of VRES production ( >16 TWh)
• Total system costs can be 10 – 14% higher if both batteries and grid expansion are unavailable
In particular climate policy costs could increase by more than 50% (from 103 to 160 BCHF)
Storage and grid expansion are required to realise the VRES potential and lower climate policy costs
Page 19
0
5
10
15
20
25
Reference Climate policy
TWh
of
win
d a
nd
so
lar
PV
NoBatteries,No gridexpansion
WithBatteries,Gridexpansion
IMPACT OF BATTERIES AND GRID EXPANSION IN
THE DEPLOYMENT OF WIND AND SOLAR PV POWER
1100
1150
1200
1250
1300
1350
1400
1450
Reference Climate policy
Un
dis
cou
nte
d c
um
ula
tive
co
st B
CH
F
NoBatteries,No gridexpansion
WithBatteries,Gridexpansion
IMPACT OF BATTERIES AND GRID EXPANSION IN
THE TOTAL SYSTEM COSTS
Results on the left graph corresponds to ranges; results on the right graph is for a scenario with high demand and electricit y imports