Novel planning techniques for the optimal allocation of DSOs owned energy storage

Transcription

1 The Norwegian Smart Grid Conference September 2017 Clarion Hotel Congress Trondheim Novel planning techniques for the optimal allocation of DSOs owned energy storage Prof. Fabrizio Pilo, Ph.D. Department of Electrical and Electronic Engineering University of Cagliari - Italy

2 Index Introduction Overview of regulatory framework on storage in EU Italian regulation for storage Private storage (distribution system) TSO DSO CBA and MO for storage allocation Preliminary results Conclusions Trondheim, 19/09/ Norwegian Smart Grid Conference 2

3 Storage and DSO DSOs may be allowed to own storage for network operation Transient condition Market should not be stopped CBA must be positive Storage is remunerated (WACC) Economic CBA not affordable for small DSOs Standardization issues Is the storage application connected to MV network? no (A simplified CBA methodology is envisaged for LV applications) AEEGSI decision 646/2015; compare with : CEER «New role of DSOs» conclusions C15-DSO Are the rules enabling Distributed Resources to take part to the ancillary service market defined? no yes Is the DSO able to demonstrate, through a CBA case (ex-ante approved methodology), the cost-effectiveness of this storage application? yes CBA positive no (Storage cannot participate to the market) yes (MV) Smart Functionalities for Observability and Voltage Regulation no are already active on the given MV network? yes (Smart grid functions are present) ALLOWED no CBA negative Is the storage unit qualifiable for the ancillary services market (included its size is above the minimum power threshold)? NOT ALLOWED yes (Storage treated exactly as DG) Trondheim, 19/09/ Norwegian Smart Grid Conference 4

4 Novel planning techniques for DES allocation Multi-objective optimization Storage optimal position, size (power and energy) and the daily control law for any representative network Set of Pareto optimal with no a priori choice of benefits CBA applied to the solutions of the Pareto front Clustering applied to the results Simple to use look-up tables as final output r = r + 1 Input data for project definition N max_des ; [P min, P max ] ; [d min, d max ] Selection of Objectives in MO r = 1 Input data from r th network (topology, generation and consumption MO optimization. The Pareto set of optimal solutions is found CBA applied to the Pareto set Positive CBA are grouped into clusters A B1 B2 C1 C2 C3 C4 T1 NO NO YES NO NO YES YES T2 NO NO NO NO NO NO NO T3 NO NO NO NO NO NO NO T4 NO NO NO NO NO NO NO T5 NO NO NO NO NO NO NO T6 NO NO NO NO NO NO NO T7 NO NO NO NO NO NO NO T8 NO NO NO NO NO NO NO T9 NO NO NO NO NO NO NO T10 NO NO NO NO NO NO NO T11 NO NO NO NO NO NO NO T12 NO NO NO NO NO NO NO r = N networks? Look up table for final decision Trondheim, 19/09/ Norwegian Smart Grid Conference 5 STO P

5 Benefits from DES all monetary? Benefits (not all monetized) Investment deferral Reduction of energy losses Power Congestions Reactive power compensation Voltage regulation Service continuity and resiliency RES integration (less curtailment) Black start Unbalances (in Italy only LV) Multi-objective optimization NSGA-II used for finding the Pareto front Real coding (NEW!) to include also daily energy scheduling Daily pattern (24 hours) for storage Trondheim, 19/09/ Norwegian Smart Grid Conference 6

6 Clustering of results Input data for project definition N max_des ; [P min, P max ] ; [d min, d max ] Selection of Objectives in MO r = 1 Input data from r th network (topology, generation and consumption r = r + 1 MO optimization. The Pareto set of optimal solutions is found CBA applied to the Pareto set Positive CBA are grouped into clusters r = N networks? Look up table for final decision STOP Trondheim, 19/09/ Norwegian Smart Grid Conference 7

7 Example - Representative rural networks HV/MV substation MV/LV trunk node MV/LV lateral node DG (existing PV) DG ( new PV ) Trunk branch Emergency connection Lateral branch HV/MV substation MV/LV trunk node MV/LV lateral node DG (existing PV) DG ( new PV ) Trunk branch Emergency connection Lateral branch Trondheim, 19/09/ Norwegian Smart Grid Conference 8

8 Results Acceptation 80% Small Sizes (T1, P n < 500 kw e d n < 5 h ) Overhead Rural with high shares of renewables (PV) A B1 B2 C1 C2 C3 C4 T1 NO NO YES NO NO YES YES T2 NO NO NO NO NO NO NO T3 NO NO NO NO NO NO NO T4 NO NO NO NO NO NO NO T5 NO NO NO NO NO NO NO T6 NO NO NO NO NO NO NO T7 NO NO NO NO NO NO NO T8 NO NO NO NO NO NO NO T9 NO NO NO NO NO NO NO T10 NO NO NO NO NO NO NO T11 NO NO NO NO NO NO NO T12 NO NO NO NO NO NO NO Trondheim, 19/09/ Norwegian Smart Grid Conference 9

9 Testing the methodology with real networks Real network examined The calculations that a DSO can perform are simulated The methodology is good if all positive cases for DSO are judged in the same way in the look up table Number of classes is crucial Trondheim, 19/09/ Norwegian Smart Grid Conference 10

10 Results Pareto front distribution of solutions after MO optimization Distribution of Pareto optimal solutions with positive CBA Trondheim, 19/09/ Norwegian Smart Grid Conference 11

11 Cross check and validation More classes are necessary to better capture the behaviour of underground (urban) contexts Tuning of the algorithms is under completion Very promising results Sensitivity analysis A B1 B2 C1 C2 C3 C4 T1 YES NO YES NO NO YES YES T2 NO NO NO NO NO NO YES T3 YES NO NO NO NO NO NO T4 NO NO NO NO NO NO NO T5 NO NO NO NO NO NO NO T6 NO NO NO NO NO NO NO T7 NO NO NO NO NO NO NO T8 NO NO NO NO NO NO NO T9 NO NO NO NO NO NO NO T10 NO NO NO NO NO NO NO T11 NO NO NO NO NO NO NO T12 NO NO NO NO NO NO NO A B1 B2 C1 C2 C3 C4 T1 NO NO YES NO NO YES YES T2 NO NO NO NO NO NO NO T3 NO NO NO NO NO NO NO T4 NO NO NO NO NO NO NO T5 NO NO NO NO NO NO NO T6 NO NO NO NO NO NO NO T7 NO NO NO NO NO NO NO T8 NO NO NO NO NO NO NO T9 NO NO NO NO NO NO NO T10 NO NO NO NO NO NO NO T11 NO NO NO NO NO NO NO T12 NO NO NO NO NO NO NO Trondheim, 19/09/ Norwegian Smart Grid Conference 12

12 Conclusions The regulation for enabling DSO to own storage is still on the way EU directives (winter package 2016) are not in favor to allow DSO owning storage Derogations are allowed under strict conditions (Italy) DES can be remunerated only if CBA is positive (for the society). Italian Regulator financed a research project for finding the conditions that can entitle DSO to own storage as regulated bodies and obtain remuneration of investments A methodology based on MO genetic algorithms, CBA, and clustering techniques presented The more complex the methodology the simpler the application. It will be as simple as using a look up table. Designed for small DSO and LV applications also. Results showed that only few cases exist where DES is more convenient than other investments and only for very small scale (<500 kw) Trondheim, 19/09/ Norwegian Smart Grid Conference 13

13 Thank you! Questions? Trondheim, 19/09/ Norwegian Smart Grid Conference 14