Towards Sustainable Intra-Dispatch Power Balancing

1. Task 2111.002 Towards Sustainable Intra-Dispatch Power Balancing Nipun Popli npopli@andrew.cmu.edu Marija Ilić milic@ece.cmu.edu

2. System Balancing A Two-Step Approach Scheduling Unit Commitment Economic Dispatch Error Balancing Stabilization Tracking

3. Intra-Dispatch Real Power Balancing • Need for Tracking • Prediction Error in Load • Intra-Dispatch Variationsin Load • Conventional Approach • Hour Ahead Load Curve projected- Data-History • Generators are Ramped Up • Average Ramp-Rates- MW/5-minute • Remaining Error picked-up by AGC

4. Renewable Portfolio Std. : Expected Scenario • Tehachapi Wind Generation in April ‘05 • NYISO-17% Projected Increase in Load Following Requirements[1] (2009-2020) Picture: D. Hawkins & C. Loutan, Integration of Renewable Resources, PSERC Tele-Seminar-Webinar Presentation,, October 2007 [1] Growing Wind- Final Report of the New York ISO 2010 Wind Generation Study, September 2010 ISO New York

5. How Wind affects Intra-Dispatch Tracking? • Tracking System Imbalances under Wind Penetration • Increased Variability in 5-minute time-frame • Wind Speed: Lack of Rich Data History • Short-Term Forecasting Error Prone • Inconsistent Correlation with Load Swings • Consequences • Increase in Load/Wind following Requirements • Hard-to-define Ramp Rates for Wind • Law of Averages MW/5-minute may not hold good • Large Error left-over for AGC

6. Typical Intra-Dispatch Balancing by Utilities SOURCE: Y. Makarov, C.Loutan, J. Ma, P. Mello, S. Lu ‘Impacts of Wind Generation on Regulation and Load Following Requirements in the California System’ PES General Meeting-Conversion & Delivery of Electrical Energy in 21st Century 2008 IEEE

7. Renewable Portfolio Std. : Violation of CPS • NERC’s Control Performance Standards not met Non-Zero Mean ACE Frequency Offsets Out of Bound M. Ili ́c , N. Popli, J. Y. Joo, and Y. Hou, A Possible Engineering and Economic Framework for Implementing Demand Side Participation in Frequency Regulation at Value, IEEE Power Engineering Society General Meeting, July 2011

8. Intra-Dispatch Real Power Balancing Managing System’s Heterogeneity

9. Sources & Sinks of Energy • Sources • Sinks • Controllable • Dispatchable • Time-Scale of Technology (Combustion/Steam/Hydro/Nuclear/Storage) • Controllable • Responsive • Size & Time-Scale Type (Inductive or Resistive) • Residential/Commercial/Industrial

10. Intra-Dispatch Power Balancing • Proposed Approach • Automated Feedback Control • Based on Quasi Stationary Concept • Formal Mathematical Model forPower Balancing • Power Imbalance: Feedback Signal • AGC responds to Frequency Offsets

11. Generation Resources Sustainable Intra-Dispatch Real Power Balancing Case Study for Flores, Azores Island N. Popli, M. Ili ́c, Modeling & Control Framework to Ensure Intra-Dispatch Regulation Reserves, Book-Chapter, The Case of Low-Cost Green Azores Islands

12. Flores Island Network Network • 45-Bus Radial • 2 MW Demand • Mix of Generation: 4-Hydro (1.7 MW), 4-Diesel (2.5 MW), 2-Wind (0.66 MW) Wind • Wind ~ 14% Installed Capacity • Scheduled Wind ~ 21% Total Demand

13. Flores: Wind Forecast & Measurements Challenges & Issues • Prediction Error over 10-minutes • Lack of Intra-10 minutes Wind Speed Data • 10% decrease in Load (i.e. Increase in Wind) within an hour • Hard-to-define Wind Ramps • Not enough Primary Reserves for Stability Noha A Karim, M. Ili ́c , Multi-temporal Wind Power Forecasting Models for Operations and Planning, , Book-Chapter, The Case of Low-Cost Green Azores Islands

14. Mechanical Dynamics of Energy Resources A Three-Way Relation Generators Steady State Droop • Hydro 0 Challenges & Issues • Prediction Error over 10-minutes • Lack of Intra-0 minutes Wind Speed Data • 10% decrease in Load (i.e. Increase in Wind) within an hour • Hard-to-define Wind Ramps • Not enough Primary Reserves for Stability • Diesel 0 N. Popli, M. Ili ́c, Modeling & Control Framework to Ensure Intra-Dispatch Regulation Reserves, Book-Chapter, The Case of Low-Cost Green Azores Islands

15. Generator’s Steady State Droop Steady State Droop/Steady State Gain Gain Droop • Hydro • Diesel N. Popli, M. Ili ́c, Modeling & Control Framework to Ensure Intra-Dispatch Regulation Reserves, Book-Chapter, The Case of Low-Cost Green Azores Islands

16. Wind Loss: Case of Higher Intra-Dispatch Variability Rate of Energy Produced=Rate of Energy Consumed + Rate of Wind Energy Loss Align Time-Scales of Generation Resources with Time-Scale of Power Imbalances

17. Tracking Real Power Imbalances • Power Model Measured Imbalance Can’t be Predicted • Delayed Response: Feedback Approach • Correction Error: Picked by AGC

18. Flores: Tracking Quasi-Stationary Wind Variations Balancing Flores Island • Steady-State Droop subjected to Network Constraints • Generator respond to Quasi-Stationary Wind Variations • Feedback Response to Sustained Deviations • Set Points Updated 30-seconds • Error/Residual Imbalance picked-up by AGC/Slack N. Popli, M. Ili ́c, Modeling & Control Framework to Ensure Intra-Dispatch Regulation Reserves, Book-Chapter, The Case of Low-Cost Green Azores Islands

19. Conventional Modeling Approach A Better Modeling Approach ? Mechanical Sub-System Electrical Sub-System Energy Dynamics ?

20. Mapping Energy Dynamics Typical Wind Ramps Source: Caiso.com Renewable Profile for 20th February 2012

21. Mapping Energy Dynamics Stand-Alone Steam Power Plant Mechanical ? Steam Power Plant: Ion Boldea, Synchronous Generators, CRC Press, 2006

22. Mapping Energy Dynamics Stand-Alone Hydro Power Plant Head ? Surge Tank? Mechanical & Penstock

23. Intra-Dispatch Demand Response • Sources(Dispatch able/Controllable) & Sinks of Energy • What we need to balance ?Load (history known), wind (predicted-based on wind speed information) • High Degree of Heterogeneity in System? • HOW TO BALANCE IT SUSTAINABLY? (Ostrom, simplistic approach for balancing) • Secondary Control (ACE based, CPS criteria, no well defined control protocols for load following at system level ) • Cradle to Grave analysis of how the energy is produced and consumed ? • Align time scales of generation with that of imbalances or change system condition • Real Life System (Azores Island) Hydro vs Diesel • Conventional thinking in terms of Mechanical/Electrical sub-systems (Mechanical for SSSA and Electrical mainly for Transient). These involves dynamics. On a longer time scale the operator worries about ED and UC. Since the load pattern is known in advance due to past operating experience. Ramp rate is just being used as a steady state constraint so far in optimal scheduling of system resources , which we call as economic dispatch. • Fluctuations vs Sustained Deviations (later becoming more of a problem as wind percentage increase (CAISO). Problem will move from present stabilization to tracking in the future (just think of it as an hard-to-predict negative load). Need more stand-by reserves or ramp-up online energy resources • What do we mean time scale alignment ??? Rate of Consumption of Energy = Rate of Production of Energy – Rate of Production of Wind Energy (at nominal frequency i.e. 60 Hz) • LOOK AT MONOGRAPH CHAPTER FOR SUBSECTIONS (wind info, feed fwd/back, savings on regulation reserves, violation of CPS criteria, Detailed Governor Model (primary source))

24. Questions ???