Probabilistic Wind Power Forecasting in Electricity Market Operations ...
Probabilistic Wind Power Forecasting in Electricity Market Operations: a Case Study of Illinois Zhi Zhou*, Audun Botterud*, Jianhui Wang Argonne National Laboratory, USA [email protected]; [email protected]
Ricardo Bessa, Hrvoje Keko, Jean Sumaili, Vladimiro Miranda INESC Porto, Portugal
Project website: http://www.dis.anl.gov/projects/windpowerforecasting.html FERC Technical Conference on Increasing Real-Time and Day-Ahead Market Efficiency Through Improved Software, June 28-30 2011
Outline Background and Motivation Wind power forecasting - Probabilistic density forecasting - Scenario generation reduction
System operation with wind power uncertainty - Two-settlement market - Stochastic unit commitment
Test Case - IL Power System - System operation analysis
Conclusion and future work
2
Outline Background and Motivation Wind power forecasting - Probabilistic density forecasting - Scenario generation reduction
System operation with wind power uncertainty - Two-settlement market - Stochastic unit commitment
Test Case - IL Power System - System operation analysis
Conclusion and future work
3
Project Overview: Wind Power Forecasting and Electricity Markets Goal: To contribute to efficient large-scale integration of wind power by developing improved wind forecasting methods and better integration of advanced wind power forecasts into system and plant operations. Collaborators:
Institute for Systems and Computer Engineering of Porto (INESC Porto), Portugal
Industry Partners:
Horizon Wind Energy and Midwest ISO (MISO)
Sponsor:
U.S. Dept. of Energy (Wind and Water Power Program)
The project consists of two main parts: Wind power forecasting – Review and assess existing methodologies – Develop and test new and improved algorithms
Integration of forecasts into operations (power system and wind power plants) – Review and assess current practices – Propose and test new and improved approaches, methods and criteria
http://www.dis.anl.gov/projects/windpowerforecasting.html 4
Background and Motivation – U.S. Wind Power Capacity Wind power has been rapidly integrated into the current power systems
5
Background and Motivation - Handling Uncertainties in System/Market Operation [MW] Source of uncertainty
Operating Reserve
Δ Load
Δ Generating capacity
Operating Reserves (spin and non-spin)
Δ Wind Power
?? ??
Wind power forecasting
Increase operating reserves? Change commitment strategy? - Stochastic UC
What are the impacts on the system? – Reliability (curtailment,..) – Efficiency (system cost, price..)
6
Outline Background and Motivation Wind power forecasting - Probabilistic density forecasting - Scenario generation and reduction
System operation with wind power uncertainty - Two-settlement market - Stochastic unit commitment
Test Case - IL Power System - System operation analysis
Conclusion and future work
7
Probabilistic forecasting with kernel density estimation Conditional wind power probabilistic forecasting
f P pt k | X xt k|t
f P , X pt k , xt k|t f X xt k|t
Joint or multivariate density function of p and x
Marginal density of x
Kernel density estimation (KDE)
8
Quantile-Copula Estimator for Conditional KDE Copula Definition
KDE ESTIMATOR
multivariate distribution function separated in: •marginal functions •dependency structure between the marginal, modeled by the copula copula density function
KDE ESTIMATOR
empirical cum. dist. Ui=FXe(Xi) and Vi=FYe(Yi)
R. Bessa, et. al. “Quantile-copula density forecast for wind power uncertainty 9 Trondheim Norway, 2011. modeling,” Proceedings IEEE Trondheim PowerTech 2011,
Illustration of Kernel Density Forecast Forecast the wind power pdf at time step t for each look-ahead time step
t+k of a given time-horizon knowing a set of explanatory variables (NWP forecasts, wind power measured values, hour of the day)
20
15
Wind Speed (m/s
10
) 5
1 0.8 0.6 0.4
Wind Po w er
(p.u.)
0.2
0 0
10
Scenario Generation and Reduction Kernel Density Forecast (KDF) methods (e.g. Quantile-copula in the IL case study) produce pdf forecasts of the wind power generation
Stochastic unit commitment model requires scenario representation of wind power forecast → account for the temporal correlation of forecast errors
A large number of scenarios generated with Monte-Carlo simulation based on quantile distribution (multivariate Gaussian error variable, covariance matrix) [Pinson et al. 09]
Three scenario reduction methods – Random selection – Scenario reduction method in GAMS [Gröwe-Kuska, Heitsch, et. al, 2003] (used in the IL case study) – Scenario clustering approach [Sumaili et al. 2011]
11
Scenario Generation and Reduction - Illustration A. Probabilistic forecast (KDF)
B. Large scenario set
C. Reduced scenario set (scenarios with different probabilities)
12
Scenario Reduction Reduces Variance of Scenario Set 10 reduced scenarios:
100 reduced scenarios 0.025
0.03 0.025 0.02
1000
0.015
10-SR1
0.01
10-SR2 10-SR3
0.005 0
Scenario Variance [p.u.]
Scenario Variance [p.u.]
0.035
0.02 0.015 1000 100-SR1
0.01
100-SR2
0.005
100-SR3
0 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46
Time [hour]
Time [hour]
SR1 - Random selection SR2 - ScenRed GAMS SR3 - Scenario clustering
13
Outline Background and Motivation Wind power forecasting - Probabilistic density forecasting - Scenario generation and reduction
System operation with wind power uncertainty - Two-settlement market - Stochastic unit commitment
Test Case - IL Power System - System operation analysis
Conclusion and future work
14
Steps in U.S. Electricity Market Operations (based on Midwest ISO)
Day ahead: Clear DA market using UC/ED
Post operating reserve requirements Submit DA bids
1100
ISO/RTO Forecast
Submit revised bids
1600
Post-DA Reliability UC
1700 DA – day ahead RT – real time UC – unit commitment ED – economic dispatch
Post results (DA energy and reserves)
Operating day: Clear RT market using ED (every 5 min)
Intraday Reliability UC Submit RT bids
-30min
Operating hour Post results (RT energy and reserves) 15
A Stochastic Unit Commitment (UC) Model w/Wind Power Uncertainty Formulation using wind power forecast scenarios (s) w/probabilities (probs): 𝑀𝑖𝑛
𝑠 𝐹𝐶𝑡,𝑖 + 𝐶(𝑅𝑁𝑆𝑡𝑠 ) + 𝐶(𝐸𝑁𝑆𝑡𝑠 ) +
𝑝𝑟𝑜𝑏𝑠 ∙ 𝑠
𝑡,𝑖
𝑠 𝑖 𝑔𝑒𝑛𝑡ℎ𝑒𝑟𝑚𝑎𝑙,𝑖,𝑡
𝑆𝐶𝑡,𝑖 𝑡,𝑖
𝑠 + 𝑔𝑒𝑛𝑤𝑖𝑛𝑑,𝑡 = 𝑙𝑜𝑎𝑑𝑡 − 𝐸𝑁𝑆𝑡𝑠 , ∀ 𝑡, 𝑠
𝑠 𝑠 𝑠𝑟𝑡ℎ𝑒𝑟𝑚𝑎𝑙,𝑖,𝑡 ≥ αsr (𝑂𝑅𝑟𝑒𝑔,𝑡 + 𝑂𝑅𝑤𝑖𝑛𝑑,𝑡 ) − 𝑆𝑅𝑁𝑆𝑡𝑠 , ∀ 𝑡, 𝑠 𝑖
𝑖
Objective function (min daily expected cost) Energy balance (hourly)
Spinning Reserve balance (hourly)
𝑠 𝑠 𝑛𝑠𝑟𝑡ℎ𝑒𝑟𝑚𝑎𝑙,𝑖,𝑡 ≥ (1 − 𝛼𝑠𝑟 ) (𝑂𝑅𝑟𝑒𝑔,𝑡 + 𝑂𝑅𝑤𝑖𝑛𝑑,𝑡 ) − 𝑁𝑆𝑅𝑁𝑆𝑡𝑠 , ∀ 𝑡, 𝑠 Non-spinning Reserve balance (hourly)
Commitment Constraints (i, t)
Unit commitment constraints (ramp, min. up/down)
A two-stage stochastic mixed integer linear programming (MILP) problem – First-stage: commitment – Second-stage: dispatch Wang J, Botterud A, Bessa R, Keko H, Carvalho L, Issicaba D, Sumaili J, and Miranda V, Wind power forecasting uncertainty and unit commitment, Applied Energy, in press, 2011. Z. Zhou, A. Botterud, J. Wang, R.J. Bessa, H. Keko, J. Sumaili, V. Miranda, “Application of 16 Probabilistic Wind Power Forecasting in Electricity Markets”, submitted
Operating Reserves
vs.
Stochastic UC
Forecast quantiles
Reduced scenario set
Hourly operating reserve requirement (spinning + non-spinning) + Deterministic UC
Commitment schedule Real-time dispatch
Realized generation
Stochastic UC + scenario set
Commitment schedule
Real-time dispatch 17
Outline Background and Motivation Wind power forecasting - Probabilistic density forecasting - Scenario generation and reduction
System operation with wind power uncertainty - Two-settlement market - Stochastic unit commitment
Test Case - IL Power System - System operation analysis
Conclusion and future work
18
Case Study Assumptions 210 thermal units: 41,380 MW
Generation Capacity
– Base, intermediate, peak units
4.78% Combine Cycle Turbine
Wind power: 14,000 MW
30.95%
20.60%
– 2006 wind series from 15 sites in Illinois
Gas Turbine Nuclear
19.71%
(NREL EWITS dataset) – 20% of load
Steam Turbine
Wind and Load in July-October 2006 40000
Peak load: 37,419 MW
Load/Wind Power (MW)
35000
– 2006 load series from Illinois
No transmission network
30000 25000 20000 Load
15000
Wind
10000 5000 0
to
October 31st, 2006) – Day-ahead unit commitment w/wind power point forecast – Real-time reliability assessment commitment (RAC) w/ wind power scenarios
1 111 221 331 441 551 661 771 881 991 1101 1211 1321 1431 1541 1651 1761 1871 1981 2091 2201 2311 2421 2531 2641 2751 2861
120 days simulation period (July
1st
Hour
Case study focus is to compare: -Operating reserves vs. stochastic UC -Probabilistic forecasting methods 19
Market Simulation Set-Up
Day-ahead point forecast
UC
ED
Day-ahead schedule and price
4hr-ahead probabilistic forecast
Actual wind power
RAC
ED
Real-time dispatch and price
20
UC Case Study: Deterministic and Stochastic Cases
Case
Add’l Reserve: 𝑂𝑅𝑤𝑖𝑛𝑑,𝑡 *
Forecast
UC strategy at RAC stage
P1
None
Perfect in both DA and RT
Deterministic
PF-F0
None
50% quantile
Deterministic
PF-F1
Fixed: avg. 50-10% quantile
50% quantile
PF-F2
Fixed: avg. 50-5% quantile
50% quantile
Deterministic
PF-F3
Fixed: avg. 50-1% quantile
50% quantile
Deterministic
PF-D1
Dynamic: 50-10% quantile
50% quantile
Deterministic
PF-D2
Dynamic: 50-5% quantile
50% quantile
Deterministic
PF-D3
Dynamic: 50-1% quantile
50% quantile
Deterministic
SF-S0
None
10 Scenarios
Stochastic
SF-S1
10% of wind scenario
10 Scenarios
Stochastic
SF-S2
20% of wind scenario
10 Scenarios
Stochastic
Deterministic
* This additional reserve is applied at the RAC stage only to handle wind power uncertainty. All cases use a regular reserve, 𝑂𝑅𝑟𝑒𝑔,𝑡 , equal to the largest contingency ( 1146 MW). 21
Overview of total cost (4-months period ) Costs 1600
Unserved load Unserved nonspinning reserve Unserved spinning reserve Start-up Fuel
1400
Cost (M$)
1200 1000 800
600 400 200 0 P1
PF-F0 PF-F1 PF-F2 PF-F3 PF-D1 PF-D2 PF-D3 SF-S0 SF-S1 SF-S2 Cases
Point forecast with no additional reserve too risky Stochastic unit commitment has the lowest total costs Dynamic reserves perform slightly better than fixed reserves Overall, more operating reserves lead to lower costs within the same categories 22
Overview of generation cost (4-months period) Total Generation Costs 660
Generation Costs (M$)
640 620 600
Start-up Fuel
580 560 540 520 P1
PF-F0 PF-F1 PF-F2 PF-F3 PF-D1 PF-D2 PF-D3 SF-S0 SF-S1 SF-S2 Cases
Stochastic UC model has slightly higher generation costs Additional generation costs are more than offset by the reduced curtailment costs 23
Total curtailment on load and reserve (4-months period)
Same trend on curtailment of load and spinning reserve. More load curtailments in cases with fixed reserve strategies More spinning reserve curtailment in cases with dynamic reserve strategies Least curtailment on both load and spinning reserve in cases with stochastic UC 24
Selected Over-forecasted Day (October 19th, 2006) Wind power (MW)
12000
S1
10000
S2 S3
8000
S4 S5
6000
S6
4000
S7 S8
2000
S9
0
S10
1 3 5 7 9 11 13 15 17 19 21 23 Time (Hour)
4 hour ahead Real
Efficiency (clearing prices) and reliability (load curtailment) 4000
1000
800 P1 PF-F0
600
PF-F2 400
PF-D2 SF-S2
200
0
Hourly Energy Price ($/MWHour)
Load Curtailment (MW)
1200
3500 3000 2500
P1 PF-F0
2000
PF-F2
1500
PF-D2 1000
SF-S2
500 0
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (Hour)
Time (Hour)
25
Selected Under-forecasted Day (September 22nd, 2006) 16000
Wind Power (MW)
14000
S1 S2
12000
S3
10000
S4 S5
8000
S6
6000
S7 S8
4000
S9
2000
S10 4 hour ahead
0 1
3
5
7
Real
9 11 13 15 17 19 21 23 Time (Hour)
Efficiency (clearing prices and wind curtailment) 1400 1200
25 PF-F0 20
PF-F1 PF-F2
15
PF-D1 PF-D2
10 SF-S0 SF-S1
5
SF-S2
Wind Power Curtailment (MW)
RT Hourly Energy Price ($/MWHour)
30
PF-F0
1000
PF-F1 800
PF-F2 PF-D1
600
PF-D2 SF-S0
400
SF-S1 200
SF-S2
0
0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324 Time (Hour)
1
3
5
7
9
11 13 15 17 19 21 23 Time (Hour)
26
Outline Background and Motivation Wind power forecasting - Probabilistic density forecasting - Scenario generation and reduction
System operation with wind power uncertainty - Two-settlement market - Stochastic unit commitment
Test Case - IL Power System - System operation analysis
Conclusion and future work
27
Conclusions Probabilistic wind power forecasts can contribute to efficiently schedule energy and operating reserves under uncertainty in wind power generation
Dynamic operating reserves (derived from forecast quantiles) + Well aligned with current operating procedures + Lower computational burden + Lower cost and increased reliability - Does not capture inter-temporal events - Uncertainty not represented in objective function
Stochastic unit commitment (with forecast scenarios) + Captures inter-temporal events through scenarios + Explicit representation of uncertainty in objective function + Lower cost and increased reliability - More radical departure from current operating procedures - High computational burden
28
Conclusions Others – Dynamic operating reserves and stochastic UC give similar results in the IL Test Case – Inaccurate forecasts can lead to large implications for system efficiency and reliability.
29
Comments and Questions.
Contacts: Zhi Zhou*, Audun Botterud*, Jianhui Wang Argonne National Laboratory, USA [email protected]; [email protected]
Ricardo Bessa, Hrvoje Keko, Jean Sumaili, Vladimiro Miranda INESC Porto, Portugal Project website: http://www.dis.anl.gov/projects/windpowerforecasting.html 30
Ricardo Bessa, Hrvoje Keko, Jean Sumaili, Vladimiro Miranda INESC Porto, Portugal
Project website: http://www.dis.anl.gov/projects/windpowerforecasting.html FERC Technical Conference on Increasing Real-Time and Day-Ahead Market Efficiency Through Improved Software, June 28-30 2011
Outline Background and Motivation Wind power forecasting - Probabilistic density forecasting - Scenario generation reduction
System operation with wind power uncertainty - Two-settlement market - Stochastic unit commitment
Test Case - IL Power System - System operation analysis
Conclusion and future work
2
Outline Background and Motivation Wind power forecasting - Probabilistic density forecasting - Scenario generation reduction
System operation with wind power uncertainty - Two-settlement market - Stochastic unit commitment
Test Case - IL Power System - System operation analysis
Conclusion and future work
3
Project Overview: Wind Power Forecasting and Electricity Markets Goal: To contribute to efficient large-scale integration of wind power by developing improved wind forecasting methods and better integration of advanced wind power forecasts into system and plant operations. Collaborators:
Institute for Systems and Computer Engineering of Porto (INESC Porto), Portugal
Industry Partners:
Horizon Wind Energy and Midwest ISO (MISO)
Sponsor:
U.S. Dept. of Energy (Wind and Water Power Program)
The project consists of two main parts: Wind power forecasting – Review and assess existing methodologies – Develop and test new and improved algorithms
Integration of forecasts into operations (power system and wind power plants) – Review and assess current practices – Propose and test new and improved approaches, methods and criteria
http://www.dis.anl.gov/projects/windpowerforecasting.html 4
Background and Motivation – U.S. Wind Power Capacity Wind power has been rapidly integrated into the current power systems
5
Background and Motivation - Handling Uncertainties in System/Market Operation [MW] Source of uncertainty
Operating Reserve
Δ Load
Δ Generating capacity
Operating Reserves (spin and non-spin)
Δ Wind Power
?? ??
Wind power forecasting
Increase operating reserves? Change commitment strategy? - Stochastic UC
What are the impacts on the system? – Reliability (curtailment,..) – Efficiency (system cost, price..)
6
Outline Background and Motivation Wind power forecasting - Probabilistic density forecasting - Scenario generation and reduction
System operation with wind power uncertainty - Two-settlement market - Stochastic unit commitment
Test Case - IL Power System - System operation analysis
Conclusion and future work
7
Probabilistic forecasting with kernel density estimation Conditional wind power probabilistic forecasting
f P pt k | X xt k|t
f P , X pt k , xt k|t f X xt k|t
Joint or multivariate density function of p and x
Marginal density of x
Kernel density estimation (KDE)
8
Quantile-Copula Estimator for Conditional KDE Copula Definition
KDE ESTIMATOR
multivariate distribution function separated in: •marginal functions •dependency structure between the marginal, modeled by the copula copula density function
KDE ESTIMATOR
empirical cum. dist. Ui=FXe(Xi) and Vi=FYe(Yi)
R. Bessa, et. al. “Quantile-copula density forecast for wind power uncertainty 9 Trondheim Norway, 2011. modeling,” Proceedings IEEE Trondheim PowerTech 2011,
Illustration of Kernel Density Forecast Forecast the wind power pdf at time step t for each look-ahead time step
t+k of a given time-horizon knowing a set of explanatory variables (NWP forecasts, wind power measured values, hour of the day)
20
15
Wind Speed (m/s
10
) 5
1 0.8 0.6 0.4
Wind Po w er
(p.u.)
0.2
0 0
10
Scenario Generation and Reduction Kernel Density Forecast (KDF) methods (e.g. Quantile-copula in the IL case study) produce pdf forecasts of the wind power generation
Stochastic unit commitment model requires scenario representation of wind power forecast → account for the temporal correlation of forecast errors
A large number of scenarios generated with Monte-Carlo simulation based on quantile distribution (multivariate Gaussian error variable, covariance matrix) [Pinson et al. 09]
Three scenario reduction methods – Random selection – Scenario reduction method in GAMS [Gröwe-Kuska, Heitsch, et. al, 2003] (used in the IL case study) – Scenario clustering approach [Sumaili et al. 2011]
11
Scenario Generation and Reduction - Illustration A. Probabilistic forecast (KDF)
B. Large scenario set
C. Reduced scenario set (scenarios with different probabilities)
12
Scenario Reduction Reduces Variance of Scenario Set 10 reduced scenarios:
100 reduced scenarios 0.025
0.03 0.025 0.02
1000
0.015
10-SR1
0.01
10-SR2 10-SR3
0.005 0
Scenario Variance [p.u.]
Scenario Variance [p.u.]
0.035
0.02 0.015 1000 100-SR1
0.01
100-SR2
0.005
100-SR3
0 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46
Time [hour]
Time [hour]
SR1 - Random selection SR2 - ScenRed GAMS SR3 - Scenario clustering
13
Outline Background and Motivation Wind power forecasting - Probabilistic density forecasting - Scenario generation and reduction
System operation with wind power uncertainty - Two-settlement market - Stochastic unit commitment
Test Case - IL Power System - System operation analysis
Conclusion and future work
14
Steps in U.S. Electricity Market Operations (based on Midwest ISO)
Day ahead: Clear DA market using UC/ED
Post operating reserve requirements Submit DA bids
1100
ISO/RTO Forecast
Submit revised bids
1600
Post-DA Reliability UC
1700 DA – day ahead RT – real time UC – unit commitment ED – economic dispatch
Post results (DA energy and reserves)
Operating day: Clear RT market using ED (every 5 min)
Intraday Reliability UC Submit RT bids
-30min
Operating hour Post results (RT energy and reserves) 15
A Stochastic Unit Commitment (UC) Model w/Wind Power Uncertainty Formulation using wind power forecast scenarios (s) w/probabilities (probs): 𝑀𝑖𝑛
𝑠 𝐹𝐶𝑡,𝑖 + 𝐶(𝑅𝑁𝑆𝑡𝑠 ) + 𝐶(𝐸𝑁𝑆𝑡𝑠 ) +
𝑝𝑟𝑜𝑏𝑠 ∙ 𝑠
𝑡,𝑖
𝑠 𝑖 𝑔𝑒𝑛𝑡ℎ𝑒𝑟𝑚𝑎𝑙,𝑖,𝑡
𝑆𝐶𝑡,𝑖 𝑡,𝑖
𝑠 + 𝑔𝑒𝑛𝑤𝑖𝑛𝑑,𝑡 = 𝑙𝑜𝑎𝑑𝑡 − 𝐸𝑁𝑆𝑡𝑠 , ∀ 𝑡, 𝑠
𝑠 𝑠 𝑠𝑟𝑡ℎ𝑒𝑟𝑚𝑎𝑙,𝑖,𝑡 ≥ αsr (𝑂𝑅𝑟𝑒𝑔,𝑡 + 𝑂𝑅𝑤𝑖𝑛𝑑,𝑡 ) − 𝑆𝑅𝑁𝑆𝑡𝑠 , ∀ 𝑡, 𝑠 𝑖
𝑖
Objective function (min daily expected cost) Energy balance (hourly)
Spinning Reserve balance (hourly)
𝑠 𝑠 𝑛𝑠𝑟𝑡ℎ𝑒𝑟𝑚𝑎𝑙,𝑖,𝑡 ≥ (1 − 𝛼𝑠𝑟 ) (𝑂𝑅𝑟𝑒𝑔,𝑡 + 𝑂𝑅𝑤𝑖𝑛𝑑,𝑡 ) − 𝑁𝑆𝑅𝑁𝑆𝑡𝑠 , ∀ 𝑡, 𝑠 Non-spinning Reserve balance (hourly)
Commitment Constraints (i, t)
Unit commitment constraints (ramp, min. up/down)
A two-stage stochastic mixed integer linear programming (MILP) problem – First-stage: commitment – Second-stage: dispatch Wang J, Botterud A, Bessa R, Keko H, Carvalho L, Issicaba D, Sumaili J, and Miranda V, Wind power forecasting uncertainty and unit commitment, Applied Energy, in press, 2011. Z. Zhou, A. Botterud, J. Wang, R.J. Bessa, H. Keko, J. Sumaili, V. Miranda, “Application of 16 Probabilistic Wind Power Forecasting in Electricity Markets”, submitted
Operating Reserves
vs.
Stochastic UC
Forecast quantiles
Reduced scenario set
Hourly operating reserve requirement (spinning + non-spinning) + Deterministic UC
Commitment schedule Real-time dispatch
Realized generation
Stochastic UC + scenario set
Commitment schedule
Real-time dispatch 17
Outline Background and Motivation Wind power forecasting - Probabilistic density forecasting - Scenario generation and reduction
System operation with wind power uncertainty - Two-settlement market - Stochastic unit commitment
Test Case - IL Power System - System operation analysis
Conclusion and future work
18
Case Study Assumptions 210 thermal units: 41,380 MW
Generation Capacity
– Base, intermediate, peak units
4.78% Combine Cycle Turbine
Wind power: 14,000 MW
30.95%
20.60%
– 2006 wind series from 15 sites in Illinois
Gas Turbine Nuclear
19.71%
(NREL EWITS dataset) – 20% of load
Steam Turbine
Wind and Load in July-October 2006 40000
Peak load: 37,419 MW
Load/Wind Power (MW)
35000
– 2006 load series from Illinois
No transmission network
30000 25000 20000 Load
15000
Wind
10000 5000 0
to
October 31st, 2006) – Day-ahead unit commitment w/wind power point forecast – Real-time reliability assessment commitment (RAC) w/ wind power scenarios
1 111 221 331 441 551 661 771 881 991 1101 1211 1321 1431 1541 1651 1761 1871 1981 2091 2201 2311 2421 2531 2641 2751 2861
120 days simulation period (July
1st
Hour
Case study focus is to compare: -Operating reserves vs. stochastic UC -Probabilistic forecasting methods 19
Market Simulation Set-Up
Day-ahead point forecast
UC
ED
Day-ahead schedule and price
4hr-ahead probabilistic forecast
Actual wind power
RAC
ED
Real-time dispatch and price
20
UC Case Study: Deterministic and Stochastic Cases
Case
Add’l Reserve: 𝑂𝑅𝑤𝑖𝑛𝑑,𝑡 *
Forecast
UC strategy at RAC stage
P1
None
Perfect in both DA and RT
Deterministic
PF-F0
None
50% quantile
Deterministic
PF-F1
Fixed: avg. 50-10% quantile
50% quantile
PF-F2
Fixed: avg. 50-5% quantile
50% quantile
Deterministic
PF-F3
Fixed: avg. 50-1% quantile
50% quantile
Deterministic
PF-D1
Dynamic: 50-10% quantile
50% quantile
Deterministic
PF-D2
Dynamic: 50-5% quantile
50% quantile
Deterministic
PF-D3
Dynamic: 50-1% quantile
50% quantile
Deterministic
SF-S0
None
10 Scenarios
Stochastic
SF-S1
10% of wind scenario
10 Scenarios
Stochastic
SF-S2
20% of wind scenario
10 Scenarios
Stochastic
Deterministic
* This additional reserve is applied at the RAC stage only to handle wind power uncertainty. All cases use a regular reserve, 𝑂𝑅𝑟𝑒𝑔,𝑡 , equal to the largest contingency ( 1146 MW). 21
Overview of total cost (4-months period ) Costs 1600
Unserved load Unserved nonspinning reserve Unserved spinning reserve Start-up Fuel
1400
Cost (M$)
1200 1000 800
600 400 200 0 P1
PF-F0 PF-F1 PF-F2 PF-F3 PF-D1 PF-D2 PF-D3 SF-S0 SF-S1 SF-S2 Cases
Point forecast with no additional reserve too risky Stochastic unit commitment has the lowest total costs Dynamic reserves perform slightly better than fixed reserves Overall, more operating reserves lead to lower costs within the same categories 22
Overview of generation cost (4-months period) Total Generation Costs 660
Generation Costs (M$)
640 620 600
Start-up Fuel
580 560 540 520 P1
PF-F0 PF-F1 PF-F2 PF-F3 PF-D1 PF-D2 PF-D3 SF-S0 SF-S1 SF-S2 Cases
Stochastic UC model has slightly higher generation costs Additional generation costs are more than offset by the reduced curtailment costs 23
Total curtailment on load and reserve (4-months period)
Same trend on curtailment of load and spinning reserve. More load curtailments in cases with fixed reserve strategies More spinning reserve curtailment in cases with dynamic reserve strategies Least curtailment on both load and spinning reserve in cases with stochastic UC 24
Selected Over-forecasted Day (October 19th, 2006) Wind power (MW)
12000
S1
10000
S2 S3
8000
S4 S5
6000
S6
4000
S7 S8
2000
S9
0
S10
1 3 5 7 9 11 13 15 17 19 21 23 Time (Hour)
4 hour ahead Real
Efficiency (clearing prices) and reliability (load curtailment) 4000
1000
800 P1 PF-F0
600
PF-F2 400
PF-D2 SF-S2
200
0
Hourly Energy Price ($/MWHour)
Load Curtailment (MW)
1200
3500 3000 2500
P1 PF-F0
2000
PF-F2
1500
PF-D2 1000
SF-S2
500 0
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (Hour)
Time (Hour)
25
Selected Under-forecasted Day (September 22nd, 2006) 16000
Wind Power (MW)
14000
S1 S2
12000
S3
10000
S4 S5
8000
S6
6000
S7 S8
4000
S9
2000
S10 4 hour ahead
0 1
3
5
7
Real
9 11 13 15 17 19 21 23 Time (Hour)
Efficiency (clearing prices and wind curtailment) 1400 1200
25 PF-F0 20
PF-F1 PF-F2
15
PF-D1 PF-D2
10 SF-S0 SF-S1
5
SF-S2
Wind Power Curtailment (MW)
RT Hourly Energy Price ($/MWHour)
30
PF-F0
1000
PF-F1 800
PF-F2 PF-D1
600
PF-D2 SF-S0
400
SF-S1 200
SF-S2
0
0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324 Time (Hour)
1
3
5
7
9
11 13 15 17 19 21 23 Time (Hour)
26
Outline Background and Motivation Wind power forecasting - Probabilistic density forecasting - Scenario generation and reduction
System operation with wind power uncertainty - Two-settlement market - Stochastic unit commitment
Test Case - IL Power System - System operation analysis
Conclusion and future work
27
Conclusions Probabilistic wind power forecasts can contribute to efficiently schedule energy and operating reserves under uncertainty in wind power generation
Dynamic operating reserves (derived from forecast quantiles) + Well aligned with current operating procedures + Lower computational burden + Lower cost and increased reliability - Does not capture inter-temporal events - Uncertainty not represented in objective function
Stochastic unit commitment (with forecast scenarios) + Captures inter-temporal events through scenarios + Explicit representation of uncertainty in objective function + Lower cost and increased reliability - More radical departure from current operating procedures - High computational burden
28
Conclusions Others – Dynamic operating reserves and stochastic UC give similar results in the IL Test Case – Inaccurate forecasts can lead to large implications for system efficiency and reliability.
29
Comments and Questions.
Contacts: Zhi Zhou*, Audun Botterud*, Jianhui Wang Argonne National Laboratory, USA [email protected]; [email protected]
Ricardo Bessa, Hrvoje Keko, Jean Sumaili, Vladimiro Miranda INESC Porto, Portugal Project website: http://www.dis.anl.gov/projects/windpowerforecasting.html 30
