Adaptive & Proactive Planning

Transcription

1 The Johns Hopkins University Environment, Energy, Sustainability & Health Institute The Value of Better Transmission Planning Models: Adaptive & Proactive Planning The Evolution of System Planning, PJM General Session Wilmington, DE December 6, 2016 Benjamin F. Hobbs Schad Professor of Environmental Management Chair, CAISO Market Surveillance Committee Qingyu Xu (JHU) Pearl Donohoo-Vallett (The Brattle Group) Jonathan Ho (NREL) Saamrat Kasina (JHU) With thanks to: WECC and USDOE/LBNL Consortium for Electricity Reliability Technology Solutions (CERTS) and BPA for funding Vijay Satyal, Gary Simonson, Mike Bailey of WECC, the project Technical Advisory Committee, and Joe Eto (LBNL) for advice Yujia Zhu and Dr. Daniel Tylavsky (ASU) for their network reduction code and advice Jim McCalley (ISU) for collaboration on the BPA project Results are hypothetical; opinions expressed don t necessarily represent the position of the funding agencies or any of the above individuals; the authors are solely responsible for any opinions or errors.

2 Agenda Motivation Western Electricity Coordinating Council Methodology Results 2

3 Motivation Why long-term? Line construction is slow & costly potential for large regret Why proactive planning (co-optimize trans & gen)? [Sauma & Oren 06; Liu et al. 13] Transmission routing affects generation siting Gen investment savings important benefit of transmission Why adaptive planning in face of uncertainty? Long run: Uncertain fuel prices, load growth, policy (renewable & carbon) Short run: Load & renewable variability Adaptation: Learn & adjust Why stochastic programming? [van der Weijde & H 12; Munoz et al. 14] To find a solution that adapts well to several possible futures in a single run 3

4 Questions For WECC system using TEPPC 2013 database: 1. Is proactive & adaptive planning practical? 2,3. Are adaptive plans different & better? 4. What is the economic value of other model features? # hours unit commitment network representation (KCL; KCL/KVL; hybrid) pro-active planning 4

5 Contents Motivation Methodology JHSMINE Scenario development WECC network model Results 5

6 JHU Stochastic Multistage Integrated Network Expansion (JHSMINE): A Stochastic Program Stage 2014: Today s Choices Stage 2024: Tomorrow s Choices Stage 2014: Today s Choices Uncertainty (Multiple Scenarios) Stage 2024: Tomorrow s Choices Choose Yr 10 Investments in: Transmission Generation Choose Yr 20 Investments in trans / gen Operations Choose Yr 10 investments in: Transmission Generation Scenarios of $ Fuels Load growth Technology Policies Choose Yr 20 investments in trans / gen Operations Deterministic Approach: One model for each study case JHSMINE: Solve all cases at once in one model 6

7 Simple example of stochastic planning Simple example of stochastic planning Stage 1 Choice: Goalie Position Uncertainty Ball left (p = 0.5) Stage 2 Choice: Goalie Reaction For transmission: Does uncertainty mean we should: Delay investment wait & see? Build more as insurance diversify? Stand left Stand middle Where should Goalie stand? Ball right (p = 0.5) Scenarios of where Ball kicked 7

8 JHSMINE formulation: Stochastic MILP Optimize the objective: Minimize (probability-weighted, present worth) of cost over 40 yrs By choosing values of decision variables: Transmission investment (0-1) 10 yr portal lines (in addition to Common Case lines) 20 yr lines Gen investment & dispatch (co-optimized) Respecting constraints: Kirchhoff s laws (linear OPF) by hour Generator operating constraints Variable renewable availability by hour Unit commitment linearization RPS Siting restrictions Accounting for uncertainties: load/renewable conditions (hourly variability) IN STOCHASTIC MODEL: long-run scenarios 8

9 Long Run (30 year) Scenarios 5 Example Scenarios predefined by WECC (2013 TEPPC) Variable: Scenario Gas Price Carbon Price Load Growth State RPS Federal RPS DG Wind Cap. Cost Geo Cap. Cost Solar Cap. Cost DR Storage Peak Growth Instate RPS Coal Price IGCC w/ CCS Cap. Cost Base Case WECC 1: Econ. Recovery WECC 2: Clean Energy WECC 3: Short-Term Consumer Costs WECC 4: Long-Term Societal Costs Probabilities: Equiprobable or expert opinion Low Value Base Case Value High Value 9

10 WECC 300 Bus Network Preserve WECC paths between regions 244 preserved monitored lines 282 equivalenced unmonitored lines Isolated dots are WREZs for renewable development 300-bus network (developed by JHU, with help of ASU): Pipes & Bubbles or Linearized DC OPF (KCL/KVL) 10

11 Contents Motivation Methodology Results: 4 Sets of Questions: 1. Is stochastic proactive planning practical? 2. Do stochastic plans differ? 3. Are stochastic plans better? 4. What is the economic value of other model features? # hours, unit commitment, network These are hypothetical runs based on JHU databases and don t represent official WECC assumptions, policy, or results 11

12 1. Is stochastic programming practical for proactive transmission planning? Answer: Yes* Deterministic (15 candidate backbone lines x 2 stages) Stochastic (Same, but with 5 WECC scenarios) Network Pipes & Bubble Hybrid DC OPF P&B Hybrid DC OPF P&B Hybrid DC OPF P&B Hybrid DC OPF Hours per year # Constraints 0.23 M 0.26 M 0.26 M 0.46 M 0.51 M 0.52 M 1.15 M 1.25 M 1.26 M 2.25 M 2.49 M 2.51 M # Variables 0.18 M 0.19 M 0.19 M 0.36 M 0.36 M 0.36 M 0.90 M 0.93 M 0.93 M 1.74 M 1.86 M 1.86 M Solution Time 15 secs 4.25 min 5.78 min 0.78 mins 21 mins 35 mins 3 mins 3 Hour 4 Hour 0.25 Hour 16.8 Hour 24 Hour *Tradeoff: For a given amount of computation time, if you simplify the network (P&B) and operations (dispatch, no UC) then you can have more scenarios & hrs/yr 12

13 2. Do stochastic and deterministic plans differ? BPA Case Study Stage 1: 2015 Choices Stage 2: Choices Stage 1: 2015 Choices Uncertainty (Multiple Scenarios) Stage 2: Choices Choose immediate investments in: Transmission Generation Choose followon investments trans / gen; Operations Scenarios of $ Fuels, CO 2 Technology Policies Also: Significant wind & CC investment Effect of Uncertainty: WAIT AND SEE 1 st Stage Transmission, Deterministic (Favorable Wind Economics) 1st Stage Transmission, Stochastic Five Scenarios 13

14 2. Do stochastic and deterministic plans differ? Sept WECC-Wide Results (First Stage Lines, KVL Model, 48 hrs) Answer: Yes; stochastic model identifies lines that enhance robustness but that the deterministic approach misses Effect of Uncertainty: BUILD MORE & DIVERSIFY PW $ Deterministic Stochastic Backbone Line Cost 0.55 $B 0.72 $B WREZ Line Cost 3.23 $B 4.05 $B Transmission Benefit $B $B Base Case (Deterministic) Model 5 Scenario (Stochastic) Model 14

15 3. Are stochastic solutions better? If you build Base Case 1 st stage transmission lines instead of the optimal stochastic lines in 300 bus model, then E(cost penalty) = ~$6.5B = Value of Stochastic Solution Depends on other model features (KVL? # hours/year? Probabilities of scenarios?) Cf. ~$5B of variable 1 st stage transmission investment ~$26B net value (PW) of adding transmission in WECC (Sept WECC results) 15

16 3. Are stochastic solutions better? Are they more robust against scenarios not considered? Answer: Yes, the 5 Scenario 1 st stage lines perform better against the withheld 15 scenarios than the base case (deterministic) 1 st stage lines Base Case minus 5 Scenario Stochastic (with Equal Probabilities, 300 bus model) PW: $ billion Base Case 1 st stage lines perform better 5 scenario stochastic 1 st stage lines perform better (July WECC results) Included in 5 Scenario Model; other 15 scenarios not in model 16

17 4. What s the value of other model features? E.g., Dispatch vs. Unit Commitment Modeling Answer for UC: Yes, it can change decisions in some high coal cases 6000 Without UC MW With UC MW

18 Conclusions: which model features matter Planning model enhancements improve performance of near-term transmission investments Most value obtained from adaptation to uncertainty (stochastic programming) (WECC) [Hobbs et al., 2016]: $6.7B (cf. $25.7B value of new transmission) Some value for other features (WECC) [Hobbs et al., 2016]: $0.3B for 48 load/wind slices/yr (vs. 24) $0.36B for DC load flow (vs. pipes & bubbles ) $0.51B for unit commitment (vs. dispatch alone) Value of proactive planning/co-optimization (Eastern Interconnection) [Spyrou et al., 2016]: $3.5B (vs. Generation-first planning in which transmission is planned against a fixed generation siting scenario) Much more transmission built, changed generation siting 18

19 Bibliography Bibliography on Stochastic & Proactive Planning M. Awad, K.E. Casey, A.S. Geevarghese, J.C. Miller, A.F. Rahimi, A.Y. Sheffrin, M. Zhang, E. Toolson, G. Drayton, B.F. Hobbs, and F.A. Wolak, "Economic Assessment of Transmission Upgrades: Application of the California ISO Approach", Ch. 7, in X.-P. Zhang, Restructured Electric Power Systems: Analysis of Electricity Markets with Equilibrium Models, Power Engineering Series, J. Wiley & Sons/IEEE Press, July 2010, R. Johnson, A. Baechert, S. Koppolu, E. Spyrou, J. Ho, B.F. Hobbs, J. McCalley, A. Figueroa, and S. Lemos-Cano, Cooptimization of Transmission and Other Resources Study, Final Report, Prepared by Energy Exemplar, LLC, The Johns Hopkins University, and Iowa State University, Submitted to the Eastern Interconnection States Planning Council and National Association of Regulatory Utility Commissioners, Washington, DC, January 26, 2015, A. Liu, B.F. Hobbs, J. Ho, J. McCalley, V. Krishnan, M. Shahidehpour, and Q. Zheng, Co-optimization of Transmission and Other Supply Resources, Prepared for the Eastern Interconnection States Planning Council, NARUC, Dec. 2013, naruc.org/grants/documents/co-optimization-white-paper_final_rv1.pdf F.D. Munoz, B.F. Hobbs, and S. Kasina, An Engineering-Economic Approach to Transmission Planning Under Market and Regulatory Uncertainties: WECC Case Study, IEEE Transactions on Power Systems, 29(1), January 2014, J.P. Pfeifenberger and D. Hou, Transmission s True Value: Adding up the Benefits of Infrastructure Investments, Public Utilities Fortnightly, February 2012, J. Pfeifenberger, J. Chang, A. Sheilendranath, Toward More Effective Transmission Planning: Addressing the Costs and Risks of an Insufficiently Flexible Electricity Grid, Brattle Group, E. Sauma and S. Oren, J. Regulatory Economics, 2006 E. Spyrous, B.F. Hobbs, J. Ho, R. Johnson, and J. McCalley, What are the Benefits of Co-optimizing Transmission and Generation Investment? Eastern Interconnection Case Study, IEEE Trans. Power Systems (in review) A.H. van der Weijde and B.F. Hobbs, The Economics of Planning Electricity Transmission to Accommodate Renewables: Using Two-Stage Optimisation to Evaluate Flexibility and the Cost of Disregarding Uncertainty, Energy Economics, 34(5), Sept. 2012, B.F. Hobbs, Q. Xu, J. Ho, P. Donohoo, S. Kasina, J. Ouyang, S.W. Park, J. Eto, and V. Satyal, Adaptive Transmission Planning: Implementing a New Paradigm for Managing Economic Risks in Grid Expansion, IEEE Power & Energy Magazine, 19 14(4), July-August 2016,

20 Questions? 20