WEBINAR: From zonal to nodal architectures - Findings from first market simulations in the OSMOSE project 27 May 2021 - Electricity market ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
WEBINAR:
Electricity market designs for flexibility:
- From zonal to nodal architectures -
Findings from first market simulations
in the OSMOSE project
27 May 2021
Agenda
• Introduction to OSMOSE WP2
• PART 1: Zonal market architectures:
a) Study with RTE’s PROMETHEUS-ATLAS model
b) Study with Joint Market Model by UDE
• PART 2: Nodal market architectures:
a) Study with RTE’s PROMETHEUS-ATLAS model
b) Study with Joint Market Model by UDE
• Conclusions
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 2
INTRODUCTION TO OSMOSE WP2 Maxime Laasri, RTE
OSMOSE PROJECT: leveraging flexibilities
Flexibility is understood as a power system's ability to cope with variability and
uncertainty in demand, generation and grid, over different timescales.
HOLISTIC APPROACH OF FLEXIBILITY
NEW
FLEXIBILITY
SOURCES
?
NEW FLEXIBILITY NEEDS
4
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 4
OSMOSE PROJECT: key figures
✓ H2020 EU funded
✓ 28M€ budget
✓ 33 partners
✓ Leaders: RTE, REE,
TERNA, ELES, CEA, TUB
✓ 2018 – 2022
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 5
OSMOSE PROJECT: project structure
WP = Work Package
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 6
WP2: market designs & regulations
SIMULATIONS
OBJECTIVES
✓ Explore and propose
some market-based
solutions for the
development of an optimal
mix of flexibility sources in
Europe
✓ Create advanced tools
and methodologies for
*RT : Real Time ; DA : Day Ahead; ID : intra Day market design analysis
orange is for CEGrid-JMM model from UDE, and blue is for the PROMETHEUS-ATLAS model from RTE
PART 1: ZONAL market studies: a) Study with RTE’s PROMETHEUS-ATLAS model Maxime Laasri, RTE
Modeling a market environment in
LONG-TERM PLANNING SHORT-TERM DECISIONS
High marginal cost
scenario (80% net
load forecast
quantile of initial
study)
Medium marginal
cost scenario (direct Conversion Day-Ahead
Clearing …
output of initial to ATLAS Orders
study)
WP1 Current Goals Achieved Long-term vision
Scenario Low marginal cost (marginal costs, ATLAS
scenario (20% net hydraulic reservoir ATLAS is an agent-based model implemented in
load forecast management…) PROMETHEUS that can simulate complex market
quantile of initial environments. It can simulate market participants
study) formulating order books under uncertainty, perform
market coupling, and simulate each participant’s
portfolio optimization.
Antares is an open-source power system
simulator developed by RTE. It is NB: In this study, market participants are all price
embedded into PROMETHEUS, and takers and submit orders based on their costs only.
available as a standalone application: They are two agents per country: one generation
https://antares-simulator.org/ company, and one energy supplier
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 9Primary results of long-term planning simulations
• The study comprises
the same 33 countries System marginal cost over one week
as in WP1, and the (€/MWh)
same grid assumptions, 150
taken from e-Highway 100
50
• The number of thermal WP1 Current 0
Goals Achieved
clusters and the Scenario
stratification of their
variable costs has a be medium de medium es medium fr medium
direct impact on the
FR DE
system’s marginal cost 110,2 110,4
• NB: variable costs in 110,1
110
110,2
110
these first runs are 109,9 109,8
different from UDE’s 109,8 109,6
00:00:00
02:00:00
04:00:00
06:00:00
08:00:00
10:00:00
12:00:00
14:00:00
16:00:00
18:00:00
20:00:00
22:00:00
00:00:00
02:00:00
04:00:00
06:00:00
08:00:00
10:00:00
12:00:00
14:00:00
16:00:00
18:00:00
20:00:00
22:00:00
00:00:00
02:00:00
04:00:00
06:00:00
08:00:00
10:00:00
12:00:00
14:00:00
16:00:00
18:00:00
20:00:00
22:00:00
00:00:00
02:00:00
04:00:00
06:00:00
08:00:00
10:00:00
12:00:00
14:00:00
16:00:00
18:00:00
20:00:00
22:00:00
values. Final runs will
use the same values fr_mgcost de_mgcost
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 10Short-term water values
Hourly water values over one year in FRANCE (€/MWh)
• Water values are close
to the system’s overall
marginal cost for
average storage levels
• We observe the
expected seasonality
for high and low storage
levels: water gains
value around the winter
but looses value in the
summer
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 11Short-term water values
Hourly water values over one year in ITALY (€/MWh)
• Water values are close
to the system’s overall
marginal cost for
average storage levels
• We observe the
expected seasonality
for high and low storage
levels: water gains
value around the winter
but looses value in the
summer
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 12Cleared quantities on 11/03 and 12/03 DA market
DE FR
120000 90000
80000
100000
70000
80000
60000
60000 50000
40000 40000
30000
20000
20000
0
10000
-20000 0
-10000
total de_ccgt de_hard_coal de_lignite
de_ev de_w de_pv de_BioMass
total fr_hydro fr_ccgt fr_nuclear fr_ev fr_w fr_pv fr_ror
de_Waste de_ror
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 13Cleared quantities on 11/03 and 12/03 DA market
• High renewable penetration entails power stations being out-of-the-money, and opportunities for electric
vehicles charging
• CCGT provide flexibility to the market, with its own cost (see next slide)
ES
BE
50000
16000
14000 40000
12000
30000
10000
8000 20000
6000
10000
4000
0
2000
0 -10000
-2000
total es_hydro es_ccgt es_hard_coal
es_nuclear es_ev es_z_psp_gen es_w
total be_ccgt be_nuclear be_ev
es_pv es_BioMass es_Waste es_ror
be_w be_pv be_BioMass be_ror
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 14Clearing prices on 11/03 and 12/03 DA market
• As expected, day-ahead prices are more volatile ES
than the marginal cost due to the simulated market 160
environment (constrained market orders add 140
complexity and rigidity compared to a pure and 120
perfect competition-based system optimization) 100
• The depth of price drops induced by solar power 80
peaks can be radically different from one day to 60
the next, as illustrated for Spain 40
• Prices are also impacted by the level of 20
stratification in thermal units’ variable cost 0
es_price es_mgcost
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 15Clearing prices on 11/03 and 12/03 DA market
200
150
100
• Day-ahead prices are sometimes continuously FR
50
higher than the marginal cost anticipated during 0
long-term planning (see France, Germany, and
Belgium, almost always in the same price group as fr_price fr_mgcost
the grid is unconstrained on these days) 200
150
100
DE
50
0
de_price de_mgcost
200
150
100
BE 50
0
be_price be_mgcost
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 16PART 1: ZONAL market studies: b) Study with the Joint Market Model Prof. Dr. Cristoph Weber, Florian Boehnke
Agenda
• Introduction of the Joint Market Model (JMM)
• Preliminary Results
• Zonal Market Studies
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 18Market Design Studies
Modelling Landscape – Joint Market Model (JMM)
• Modelling of power plant operation and
market outcomes (unit commitment)
• Starting point: with well-functioning
competition the market outcome corresponds
to the outcome of a central optimization
(fundamental model)
• Objective function minimizes variable
generation costs
• Two-stage optimization approach for
modelling of forecast errors and redispatch Rolling Planning Approach
Geoscope
• Modelling of different markets:
day-ahead, intraday, balancing, heating
markets
• Detailed formulation of technical restrictions
• Geoscope: Europe
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 19Market Design Studies
Modelling Landscape – Joint Market Model (JMM)
Data Base Simulation Models
CHP Model
Market Model
• Underlying conventional
• Historical generation
Timeseries • Shadow prices • PTDFs
• Regional load & infeed hydro power • RAMs
• Power plant Data plants
• Timeseries REN • Price Zone
Vertical Load Model • Infeed run of river Configuration
• Timeseries Load • Average daily
• Grid data • Regional prices
• Price Zone Config. Timeseries (Vertical • Import/ Export at
• Fuel & CO2 Prices Load) • Vertical Load border nodes
• Availabilities
• … • Power Plants Grid Model
Cluster Algorithm
• Price Zone Configuration
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 20Zonal Market Design
Setup & Input Parameter
• Data input from WP1
• Scenario: “Current Goals”
• Scenario year: 2030
• Geoscope: 33 countries
• 2 case studies are part of this presentation:
• Reference Case (LP, no uncertainties, NTC)
• Uncertainties Case (LP, uncertainties for wind power generation, NTC)
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 21Zonal Market Design
Input Parameter I
Load and Renewable Infeed [MW] Load vs. Residual Load
600 600
500 500
400 400
Electricity [GW]
Load [GW]
300 300
200 200
100 100
0 0
245
489
733
977
1
1221
1465
2441
2685
2929
3905
4149
4393
4637
5613
5857
6101
7077
7321
7565
7809
8541
1709
1953
2197
3173
3417
3661
4881
5125
5369
6345
6589
6833
8053
8297
220
439
658
877
1
1096
1315
1534
1753
1972
2191
2410
2629
2848
3067
3286
3505
3724
3943
4162
4381
4600
4819
5038
5257
5476
5695
5914
6133
6352
6571
6790
7009
7228
7447
7666
7885
8104
8323
8542
Time [h] Time [h]
Load Solar RoR Wind Load Res_load
• Total Demand: 3.291 TWh
• Wind Power / PV / RoR / Total : 763 / 354 / 681 / 1798 TWh
• Max / Min Residual Load: 372 GW / 15 GW
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 22Zonal Market Design
Input Parameter II
• Conventional capacities Installed Capacity [GW]
see graph
250
• Carbon Price: 18 €/tCO2
200
• Fuel Prices (in €/MWh)
• Nat Gas: 24.0 150
• Hard Coal: 8.1 100
• Oil: 49.4
50
0
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 23Zonal Market Design
Results I Price
• Preliminary results in the chart Average Day2030Ahead
- ref_vs_uncertainty 2030 - uncertainty_eval
Market Prices (Current Goals 2030)
45
• Day Ahead price range between
30 to 40 €/MWh 40
• Considerably lower levels only in 35
ES and PT
30
• Insufficient transmission capacity
Netposition
[Euro/MWh]
25
• Uncertainty
100 does not affect DA
prices (on average!) 20
• Stochastic
80 uncertainty time series 15
• More volatile prices during the
year60 10
5
40
0
AL AT BA BE BG CH CZ DE DK EE ES FI FR GB GR HR HU IE IT LT LU LV ME MK NL NO PL PT RO RS SE SI SK
20
[TWh]
2030 - ref_vs_uncertainty
0 2030 - uncertainty_eval
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 24
AL AT BA BE BG CH CZ DE DK EE ES FI FR GB GR HR HU IE IT LT LU LV ME MK NL NO PL PT RO RS SE SI SKZonal Market Design
Results II Day Ahead Price Comparison for Germany
(reference - uncertainties)
Price Deviation in €/MWh // 8760 hours
• Volatile Day Ahead Prices when
20
Day Ahead Price Comparison for Germany (reference - uncertainties) in 2030
considering uncertainties
15
• Deviations more dense in Q1 &
Q4 due to overall wind power
10
Price Difference [Euro/MWh]
generation
• Multiple hours in Q2 & Q3 with 5
equal prices combined with strong
(arbitrary) peaks in price 0
difference
-5
-10
1195
1593
2190
2588
2986
3384
3782
4379
4777
5175
5573
5971
6568
6966
7364
7762
8160
1394
1792
1991
2389
2787
3185
3583
3981
4180
4578
4976
5374
5772
6170
6369
6767
7165
7563
7961
8359
8558
1
200
399
598
797
996
-15
25
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2]Zonal Market Design
Results III
• Initial update similar to DA-price vs. ID last price update for first week in February 2030 (12:00h), Geoscope GER
wind expectation of DA
UC 65
DA Price Last ID Price
• Temporal correlation of
FC Errors 60
• Subsequent updates lead
to price differences 55
between DA and ID
Price [Euro/MWh]
market prices, pending on 50
the FC information
45
40
35
1 13 25 37 49 61 73 85 97 109 121 133 145 157
Hours [h]
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 26Zonal Market Design
Results IV
• Positive deviation (dispatched quantity > planned Cumulated Deviation per Technology (1 year)
Diagrammtitel
quantity)
HYDRO PUMP NUCLEAR NAT GAS WASTE LIGNITE ELECTRIC COAL
• Depending on the lead time conv. power plants used for 15
ramping needs
• Most flexibility provided by Hydro Power
Cumulated Deviation [ 10^6 MWh]
10
• Negative Deviation (dispatched quantity < planned
quantity)
5
• Energy is stored (Pump / Electric (DSM))
• Downregulation of conv. Power plants
0
Generation
Generation per Technology
per Technology
900 -5
800
700
600
[TWh]
500
400 -10
300
200 Neg. Deviation Pos. Deviation
100
0
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 27PART 2: NODAL market studies: a) Study with RTE’s PROMETHEUS-ATLAS model Sandrine Bortolotti, RTE
Study perimeter
• Small region of central France simulated with a nodal
market
• The rest of France is modeled with consistent electric
regions
• The rest of Europe is modeled on a country scale
• RES forecast data come from historical measurements
(which preserve spatial correlations between forecasts),
and were scaled to match 2030 levels
• Selection of one interesting week to simulate at the
beginning of November
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 29National load forecasts
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 30Regional load forecasts (05_fr)
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 31Substation load forecasts (VICHYP3)
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 32Primary results on historical data analysis: RMSE on
load forecasts
2 takeaways:
• The geographical scope has a huge influence
on the magnitude of the forecast errors
• Regarding load forecast, aggregating some
substations reduce drastically the magnitude
of the errors.
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 33Primary results on historical data analysis
2 takeaways:
• The geographical scope has a huge influence
on the magnitude of the forecast errors
• The uncertainty only decrease significantly a
few hours before real time
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 34PART 2: NODAL market studies: b) Study with the Join Market Model Prof. Dr. Cristoph Weber, Florian Boehnke
Nodal Market Design
Methodology
• Nodal market design applies prices for electricity consumed or generated at a nodal level
• Market design to fully consider physical grid restrictions (congestion management)
(Zonal Markets NTC → Zonal Markets FBMC → Nodal Markets)
• Prices at adjacent nodes are equal in case of sufficient transfer capacities
• Implemented in many U.S. markets, e.g. California (CAISO), Texas (ERCOT)
• Main inputs
• Grid ENTSOE TYNDP (without coordinates)
• Zonal market design inputs (country level) → Nodalizing input parameters (nodal level)
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 36Nodal Market Design
Nodalizing Wind Power
• Nodalizing Wind on NUTS3 level (county)
• Estimating capacities from different data sources for
current regional assets
• Creation of infeed timeseries
• Weather information from Cosmo-EU Voronoi Areas with Nodes
• Multiplied by turbines’ power curves
• Scaling to 2030 infeed timeseries
• Wind turbines are assigned to nodes by Voronoi-
areas
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 37Nodal Market Design
Preliminary results
• Based on German case study
• Not based on TYNDP Grid
• Average day-ahead market prices for January 2030
• Published in MS8
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 38CONCLUSIONS Maxime Laasri, RTE
Key messages
• The investigation of different market designs in a
European context is necessary for ensuring the
viability of any potential prospective energy mix
• Forecast uncertainties are a key element to be
reflected in market design studies, as they
impact how market opportunities are leveraged by
market participants and how power system
operators respond subsequently
• Nodal market designs are challenging to model
and simulate in practice
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 40Upcoming activities
• A second webinar on WP2 will be organised in Fall, with focus on modeling
• Three webinars jointly organised with the EU-SYSFLEX project on 15-16-17 June
14h00 CET:
✓ High RES scenarios : from adequacy to stability challenges and new solutions
✓ IT challenges to activate and monitor flexibilities, Wednesday
✓ Value and demonstrations of flexibility provision by distributed sources
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 41Thanks for your attention
Maxime Laasri, RTE: maxime.laasri@rte-france.com
Sandrine Bortolotti, RTE: sandrine.bortolotti@rte-france.com
Prof. Dr. Cristoph Weber: christoph.weber@uni-due.de
Florian Boehnke: florian.boehnke@uni-due.de
The OSMOSE project has received funding from the European Union’s Horizon 2020
research and innovation programme under grant agreement No 773406
27/05/2021 ISGAN | ELECTRICITY MARKET DESIGNS FOR FLEXIBILITY [OSMOSE WP2] 42You can also read
