EPRI â Con Edison - Electric Power Research Institute
EPRI – Con Edison Day-ahead Active and Reactive Power Controls Scheduling to Improve Transmission System
Alberto Del Rosso: EPRI Project Manager May 19, 2011
T&D Efficiency Initiative Timeline
© 2011 Electric Power Research Institute, Inc. All rights reserved.
2
Strong Executive Leadership Team Formed Commissioner Jon Wellinghoff, Chairman, FERC Arshad Mansoor, VP, EPRI US Executive Steering Committee Members: Nick Brown, President, & CEO Southwest Power Pool Terry Boston, President & CEO, PJM Interconnection Steve DeCarlo, Sr. V.P. Transmission, New York Power Authority Mike Hervey, V.P. T&D Operations, Long Island Power Authority Mike Heyeck, Sr. V.P. Transmission, American Electric Power Rob Manning, Executive V.P. Power Systems, Tennessee Valley Authority Yakout Mansour, President & CEO, California ISO Pedro Pizarro, Exec V.P. Power Operations, Southern California Edison John McAvoy, Sr. VP Central Operations, Consolidated Edison Rich Mandes, V.P. Transmission, Alabama Power Steve Whitley, President & CEO, New York Independent System Operator International Steering Committee Members: Barry MacColl, Technology Strategy & Planning, ESKOM Magdalena Wasiluk-Hassa, Director, Innovation, PSE Operator Ian Welch, R&D Strategy Manager, National Grid © 2011 Electric Power Research Institute, Inc. All rights reserved.
3
Overview Industry Wide Demonstrations 17 Companies with 33 Proposed Projects Technologies Identified to Improve Efficiency and Utilization
Transmission Efficiency Improving Opportunity
1. Reduce System Losses
Technologies to Improve Transmission Efficiency 1A. Coordination Voltage Var Control 1B. Voltage Upgrade/EHV AC/HVDC
2. Reduce Line/Equipment Losses
1C. Loss Minimization Optimization
3. Increase Line/System Utilization
2B. Low Loss/LEED Substation
2A. Advanced Conductors/ Low Loss Design Equipment
3A. Dynamic Rating 3B. Smart Transmission Control
© 2011 Electric Power Research Institute, Inc. All rights reserved.
4
Collaboration Across Demonstrations
© 2011 Electric Power Research Institute, Inc. All rights reserved.
5
Project: Integrated Active and Reactive Power Control for Con Edison Transmission System
© 2011 Electric Power Research Institute, Inc. All rights reserved.
6
Background • Con Edison Transmission System: – In-City densely loaded transmission system consisting of both 138 KV and 345 KV areas that are tied together. Mostly underground cables. – Generation capacity: 14000 MW; Peak Load: 12,000 MW, 5000 MVAr – Several Phase Angle Regulators (PARs): • regulate power flows on ties from adjacent utilities • balance real power flows between the 345kV, 138kV and 69kV portions of the internal transmission system. – Transformers that bridge two active transmission areas (i.e. 138kV and 345kV) move reactive power between these areas while adjusting voltage
© 2011 Electric Power Research Institute, Inc. All rights reserved.
7
Background • Con Edison Current Practice: – On summer peak day the system experience about 3 hours of ramping load on the order of 1000MW per hour. – Operators place almost all of the capacitors (3300 MVARs) in service during the morning ramp-up period, and to take most of them out of service during the nightly ramp-down period • “get ahead of the load curve” approach. – Generators, phase angle regulators, and shunt capacitors are adjusted as needed – High ramping periods require considerable time and attention from the operators for voltage concerns. – Currently no plan for the coordinated dispatch of reactive power resources or reserves on a system-wide basis. – Current approach successful in practice yet not optimal More efficient switching strategy could be implanted © 2011 Electric Power Research Institute, Inc. All rights reserved.
8
Projects • Real and Reactive Power Optimization: – Investigate opportunities for a more effective and utilization of active and reactive power flow controls to improve transmission efficiency
• Day-ahead Active and Reactive Power Controls Scheduling: • More optimal placed of reactive resources allocation over time • 24 hour schedule for system operators to support their decision making capacity over the course of day
• Reactive Power Forecasting to Assist VAr Planning: – Investigate the key variables in determining the reactive power demand for Con Edison system – Develop reactive power forecasting models over a range of timehorizons, at difference system levels © 2011 Electric Power Research Institute, Inc. All rights reserved.
9
End Stage – EPRI Forecast and OPF Applications - PI Data - Historic MV & MVAR load data - Weather data - Manual input
Active and Reactive Load Forecast for all Substations
- 24 Hr unit commitment and generator dispatch - Interface flows
Pre-processor to incorporate data
Reactive Power Optimization for 24 Hr
Post-processor & Report Builder
OUTPUT --Equipment
out of
Service -Topological changes
© 2011 Electric Power Research Institute, Inc. All rights reserved.
10
24 Hr scheduling of reactive power resources: - Capacitors and Reactors -Selected transformer taps -Phase shifters, FATCS - Voltage setting at selected buses Reactive Load & Reactive power reserve
OPF Application Module • Objective Function: – Min: 1 Σ MW losses + 2 Σ MVAr losses
• Control Variables – Shunt Capacitors and Reactors – Selected Transformer taps – Transformer phase shift angles: 17 Phase Shifters – Generators within Con Edison area are allowed to adjust their scheduled voltages to meet global objective: 50 Generators – Generators are not subject to direct regulation by Con Edison, but are expected to adhere to their voltage schedules and respond on request to provide additional support up to their stated limitations
© 2011 Electric Power Research Institute, Inc. All rights reserved.
11
OPF Application Module • Constraints: – Generator MVAr limitss: QGEN min ≤ QGEN ≤ QGEN max – Transformer tap ranges: TAP min ≤ TAP ≤ TAP max – Bus Voltage Constraint: Maximum (1.05 p.u.) and minimum “normal” (1.0 p.u.) voltage limits – Branch Flow limits
• Operational Constraints: • No more than 1 daily switching cycle (on/off) on tap and shunt capacitors/reactors Different control strategies: – Optimize every hour – Optimized at selected number of hours per day • Interface flows within specified limits
© 2011 Electric Power Research Institute, Inc. All rights reserved.
12
OPF Application Module • Optimizer: PSSTME OPF V32 • A Python script is been developed to optimize OPF process, input data and post-processing • Solution is based on a scheduled MW dispatch from day ahead schedule which is secure in terms of thermal constraints • An accurate AC model of the network is used • The OPF produces reactive schedules to support voltage security, without compromising MW security • Possible to migrate to other OPF tool in future implementation
© 2011 Electric Power Research Institute, Inc. All rights reserved.
13
Active and Reactive Load Forecast Module • Provide 24 hr active and reactive power load forecast for 62 area stations • Independent Reactive Power Forecast: based on the fact that reactive load patterns do not necessary follow their associated real power load patterns in a purely proportional manner • The computational algorithm utilizes historical information in combination with anticipated forecast conditions (temperature, weather) to predict a load curves at area station level
© 2011 Electric Power Research Institute, Inc. All rights reserved.
14
Industrial Customers
Residential Customers
Commercial Consumers
Short-Term Load Forecasting • Use of artificial neural networks and data mining provides accurate load forecasts • Correlations among variables might provides more insight • Help system engineer gain insights of power profile and characteristics • Enable utilization of existing transmission capacity.
© 2011 Electric Power Research Institute, Inc. All rights reserved.
15
Short-Term Load Forecasting • Expected MAPE for forecasting at each substation level (Not aggregated)
Real load (Hour ahead)
2%
Real load (24 hour ahead)
5-7%
Reactive load (Hour ahead)
5%
Reactive load (24 hour ahead)
~10%
© 2011 Electric Power Research Institute, Inc. All rights reserved.
16
Estimation of Savings and Benefits
© 2011 Electric Power Research Institute, Inc. All rights reserved.
17
X2R Losses [MAVr]
6000 5000 4000 3000 2000 1000 0 1
3
5
7
9
11
13 Hour
Optimized
I2R Losses [MW]
• Several 24 hr periods have been analyzed • Baseline operating conditions set up based on PI historic data: – load, generation, shunt capacitor/reactor, selected, transformer taps, voltage level at selected buses). • Potential significant savings in losses • Results indicate significant differences in reactive resource allocation as load changed. • To be used as metric by system operators or engineers to support their ongoing control efforts throughout daily operation
15
17
19
21
23
19
21
23
Base Case
180 160 140 120 100 80 60 40 20 0 1
3
5
7
9
11
13 Hour
Optimized
15
17
Base Case
Next steps • Finalize development of input/output interfaces and load forecast • Implementation and testing in the control room • Improve capability for more flexible operator intervention: – “what if” analysis – System performance metrics – Input data manipulation • Develop and implement visualization dashboard • Evaluate need/convenience to use other OPF software
© 2011 Electric Power Research Institute, Inc. All rights reserved.
18
Together…Shaping the Future of Electricity
© 2011 Electric Power Research Institute, Inc. All rights reserved.
19
Background slides
© 2011 Electric Power Research Institute, Inc. All rights reserved.
20
Optimization strategies Example: optimization performs 4 times a days Optimized hours
14000 12000 10000 8000 6000 4000 2000
MW
0 0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Non fully optimized hours
© 2011 Electric Power Research Institute, Inc. All rights reserved.
21
Alberto Del Rosso: EPRI Project Manager May 19, 2011
T&D Efficiency Initiative Timeline
© 2011 Electric Power Research Institute, Inc. All rights reserved.
2
Strong Executive Leadership Team Formed Commissioner Jon Wellinghoff, Chairman, FERC Arshad Mansoor, VP, EPRI US Executive Steering Committee Members: Nick Brown, President, & CEO Southwest Power Pool Terry Boston, President & CEO, PJM Interconnection Steve DeCarlo, Sr. V.P. Transmission, New York Power Authority Mike Hervey, V.P. T&D Operations, Long Island Power Authority Mike Heyeck, Sr. V.P. Transmission, American Electric Power Rob Manning, Executive V.P. Power Systems, Tennessee Valley Authority Yakout Mansour, President & CEO, California ISO Pedro Pizarro, Exec V.P. Power Operations, Southern California Edison John McAvoy, Sr. VP Central Operations, Consolidated Edison Rich Mandes, V.P. Transmission, Alabama Power Steve Whitley, President & CEO, New York Independent System Operator International Steering Committee Members: Barry MacColl, Technology Strategy & Planning, ESKOM Magdalena Wasiluk-Hassa, Director, Innovation, PSE Operator Ian Welch, R&D Strategy Manager, National Grid © 2011 Electric Power Research Institute, Inc. All rights reserved.
3
Overview Industry Wide Demonstrations 17 Companies with 33 Proposed Projects Technologies Identified to Improve Efficiency and Utilization
Transmission Efficiency Improving Opportunity
1. Reduce System Losses
Technologies to Improve Transmission Efficiency 1A. Coordination Voltage Var Control 1B. Voltage Upgrade/EHV AC/HVDC
2. Reduce Line/Equipment Losses
1C. Loss Minimization Optimization
3. Increase Line/System Utilization
2B. Low Loss/LEED Substation
2A. Advanced Conductors/ Low Loss Design Equipment
3A. Dynamic Rating 3B. Smart Transmission Control
© 2011 Electric Power Research Institute, Inc. All rights reserved.
4
Collaboration Across Demonstrations
© 2011 Electric Power Research Institute, Inc. All rights reserved.
5
Project: Integrated Active and Reactive Power Control for Con Edison Transmission System
© 2011 Electric Power Research Institute, Inc. All rights reserved.
6
Background • Con Edison Transmission System: – In-City densely loaded transmission system consisting of both 138 KV and 345 KV areas that are tied together. Mostly underground cables. – Generation capacity: 14000 MW; Peak Load: 12,000 MW, 5000 MVAr – Several Phase Angle Regulators (PARs): • regulate power flows on ties from adjacent utilities • balance real power flows between the 345kV, 138kV and 69kV portions of the internal transmission system. – Transformers that bridge two active transmission areas (i.e. 138kV and 345kV) move reactive power between these areas while adjusting voltage
© 2011 Electric Power Research Institute, Inc. All rights reserved.
7
Background • Con Edison Current Practice: – On summer peak day the system experience about 3 hours of ramping load on the order of 1000MW per hour. – Operators place almost all of the capacitors (3300 MVARs) in service during the morning ramp-up period, and to take most of them out of service during the nightly ramp-down period • “get ahead of the load curve” approach. – Generators, phase angle regulators, and shunt capacitors are adjusted as needed – High ramping periods require considerable time and attention from the operators for voltage concerns. – Currently no plan for the coordinated dispatch of reactive power resources or reserves on a system-wide basis. – Current approach successful in practice yet not optimal More efficient switching strategy could be implanted © 2011 Electric Power Research Institute, Inc. All rights reserved.
8
Projects • Real and Reactive Power Optimization: – Investigate opportunities for a more effective and utilization of active and reactive power flow controls to improve transmission efficiency
• Day-ahead Active and Reactive Power Controls Scheduling: • More optimal placed of reactive resources allocation over time • 24 hour schedule for system operators to support their decision making capacity over the course of day
• Reactive Power Forecasting to Assist VAr Planning: – Investigate the key variables in determining the reactive power demand for Con Edison system – Develop reactive power forecasting models over a range of timehorizons, at difference system levels © 2011 Electric Power Research Institute, Inc. All rights reserved.
9
End Stage – EPRI Forecast and OPF Applications - PI Data - Historic MV & MVAR load data - Weather data - Manual input
Active and Reactive Load Forecast for all Substations
- 24 Hr unit commitment and generator dispatch - Interface flows
Pre-processor to incorporate data
Reactive Power Optimization for 24 Hr
Post-processor & Report Builder
OUTPUT --Equipment
out of
Service -Topological changes
© 2011 Electric Power Research Institute, Inc. All rights reserved.
10
24 Hr scheduling of reactive power resources: - Capacitors and Reactors -Selected transformer taps -Phase shifters, FATCS - Voltage setting at selected buses Reactive Load & Reactive power reserve
OPF Application Module • Objective Function: – Min: 1 Σ MW losses + 2 Σ MVAr losses
• Control Variables – Shunt Capacitors and Reactors – Selected Transformer taps – Transformer phase shift angles: 17 Phase Shifters – Generators within Con Edison area are allowed to adjust their scheduled voltages to meet global objective: 50 Generators – Generators are not subject to direct regulation by Con Edison, but are expected to adhere to their voltage schedules and respond on request to provide additional support up to their stated limitations
© 2011 Electric Power Research Institute, Inc. All rights reserved.
11
OPF Application Module • Constraints: – Generator MVAr limitss: QGEN min ≤ QGEN ≤ QGEN max – Transformer tap ranges: TAP min ≤ TAP ≤ TAP max – Bus Voltage Constraint: Maximum (1.05 p.u.) and minimum “normal” (1.0 p.u.) voltage limits – Branch Flow limits
• Operational Constraints: • No more than 1 daily switching cycle (on/off) on tap and shunt capacitors/reactors Different control strategies: – Optimize every hour – Optimized at selected number of hours per day • Interface flows within specified limits
© 2011 Electric Power Research Institute, Inc. All rights reserved.
12
OPF Application Module • Optimizer: PSSTME OPF V32 • A Python script is been developed to optimize OPF process, input data and post-processing • Solution is based on a scheduled MW dispatch from day ahead schedule which is secure in terms of thermal constraints • An accurate AC model of the network is used • The OPF produces reactive schedules to support voltage security, without compromising MW security • Possible to migrate to other OPF tool in future implementation
© 2011 Electric Power Research Institute, Inc. All rights reserved.
13
Active and Reactive Load Forecast Module • Provide 24 hr active and reactive power load forecast for 62 area stations • Independent Reactive Power Forecast: based on the fact that reactive load patterns do not necessary follow their associated real power load patterns in a purely proportional manner • The computational algorithm utilizes historical information in combination with anticipated forecast conditions (temperature, weather) to predict a load curves at area station level
© 2011 Electric Power Research Institute, Inc. All rights reserved.
14
Industrial Customers
Residential Customers
Commercial Consumers
Short-Term Load Forecasting • Use of artificial neural networks and data mining provides accurate load forecasts • Correlations among variables might provides more insight • Help system engineer gain insights of power profile and characteristics • Enable utilization of existing transmission capacity.
© 2011 Electric Power Research Institute, Inc. All rights reserved.
15
Short-Term Load Forecasting • Expected MAPE for forecasting at each substation level (Not aggregated)
Real load (Hour ahead)
2%
Real load (24 hour ahead)
5-7%
Reactive load (Hour ahead)
5%
Reactive load (24 hour ahead)
~10%
© 2011 Electric Power Research Institute, Inc. All rights reserved.
16
Estimation of Savings and Benefits
© 2011 Electric Power Research Institute, Inc. All rights reserved.
17
X2R Losses [MAVr]
6000 5000 4000 3000 2000 1000 0 1
3
5
7
9
11
13 Hour
Optimized
I2R Losses [MW]
• Several 24 hr periods have been analyzed • Baseline operating conditions set up based on PI historic data: – load, generation, shunt capacitor/reactor, selected, transformer taps, voltage level at selected buses). • Potential significant savings in losses • Results indicate significant differences in reactive resource allocation as load changed. • To be used as metric by system operators or engineers to support their ongoing control efforts throughout daily operation
15
17
19
21
23
19
21
23
Base Case
180 160 140 120 100 80 60 40 20 0 1
3
5
7
9
11
13 Hour
Optimized
15
17
Base Case
Next steps • Finalize development of input/output interfaces and load forecast • Implementation and testing in the control room • Improve capability for more flexible operator intervention: – “what if” analysis – System performance metrics – Input data manipulation • Develop and implement visualization dashboard • Evaluate need/convenience to use other OPF software
© 2011 Electric Power Research Institute, Inc. All rights reserved.
18
Together…Shaping the Future of Electricity
© 2011 Electric Power Research Institute, Inc. All rights reserved.
19
Background slides
© 2011 Electric Power Research Institute, Inc. All rights reserved.
20
Optimization strategies Example: optimization performs 4 times a days Optimized hours
14000 12000 10000 8000 6000 4000 2000
MW
0 0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Non fully optimized hours
© 2011 Electric Power Research Institute, Inc. All rights reserved.
21
