Robust Battery Scheduling in a Micro-Grid with PV Generation Xing Wang, Ph.D. GE Grid Software 2016 March 30, 2016

Transcription

1 1 Robust Battery Scheduling in a Micro-Grid with PV Generation Xing Wang, Ph.D. GE Grid Software 2016 March 30, 2016 Imagination at work

2 2 Outline Introduction Problem description Case studies Conclusion 2

3 3 Definition of Micro-Grid DOE: A group of interconnected loads and distributed energy resources (DER) with clearly defined electrical boundaries that acts as a single controllable entity with respect to the grid [and can] connect and disconnect from the grid to enable it to operate in both grid connected or island mode. Challenge: how to deal with intermittent nature of renewable DERs in a micro-grid? 3

4 NICE GRID Project: facts and figures AVIGNON CARROS Site description Carros: inhabitants, with residential and industrial areas 1,8 MWp of existing Solar PV 88 secondary substations 20 MW peak load at Primary substation 400 kv line MARSEILLE TOULON NICE Use cases Optimize massive PV integration in the distribution grid Test islanding within a low voltage area Test 3.5 MW load shedding Incentivize Prosumer behavior fg Seven solar districts 550 clients with LINKY smart meters (2500 Linky in total) 220 clients with electric water heaters 50 solar installations: 550 kwp 3 grid batteries: 1 x 250 kw/600 kwh and 2 x 33 kw / 106 kwh 20 residential batteries: 4.6 kw / 4 kwh CARROS

5 Nice Grid demonstrator objectives Reduction of power demand by shifting up to 5MW of load. This scenario could be triggered following a DSO, or a TSO request to relieve electrical constraints on the grid. Management of a massive distributed PV generation and its impacts on the distribution grid regarding voltage requirements. Islanding: its goal is to show that a specific network area can operate independently and be disconnected from the electrical network. The main goal of this operation mode is to improve the continuity of service provided to the customers. Islanding will be tested by the DSO in some prepared, anticipated or emergency situations. 5

6 Vision of the NiceGrid demonstrator 6

7 Problem Description -1 Reliably islanding a micro-grid for a certain period by optimally scheduling batteries to manage PV uncertainty Decision Variables Interchange power I(t) Battery charge schedule C(b,t) and discharge schedule D(b,t) Islanding duration IP(t), start and end time, IS(t) & IE(t) (if the islanding duration is not fixed) Major Uncertainties Being Managed Loads: predicted power consumption L(l,t) Intermittent resources: PV genertation forecast P(p,t) Operational Objectives: Trade off between reliability and economics Maintenance Scheduling (Preset Islanding Schedule with Optimal Battery Charge/Discharge Schedule) Peak Hour Islanding Scheduling (Optimal Islanding Schedule) Predicted Power Consumption L(l, t) Loads P 7 Power Interchange I(t) Main Grid Max Interchange Power Min Interchange Power P(p,t) Predicted Power Generation L(l, t) Charge Schedule C(t) Discharge Schedule D(t) DERs Batteries 7

8 Problem Description -2 Major Constraints: Battery Constraints - Charging & Discharging power constraints SoC Constraints - Fixed SoC Constraint & Technical SoC Constraint Battery Ramp Constraints - Technical Ramp Constraint Interchange Constraints - Interchange Power Constraints Operating Constraints - Islanding Period Constraints Operating Penalty Constraints - Charge/Discharge State Change Constraints P 8

9 Two Stage Robust Optimization Master Problem: determine islanding scheduling as well as battery charging and discharging schedule Sub problem: solve the dispatch problem under the worst-case PV generation scenario with the fixed islanding and charge/discharge decisions. The solution of the subproblem discovers the worst-case scenarios which are used to generate the cuts. P 9

10 Case Studies Test system BATTERY PV 2 Power Factor =0.9 BATTERY 1 PV 1 P 10

11 Case Studies Deterministic Parameters 2 Batteries 2 PV panels 2 Load Zones Operation Parameters Value Study Period Start T08:00:00Z Study Period End T18:00:00Z Objective Robust Longest Islanding Duration Maximum Interchange 200 kw Islanding Compensate 20 $/hour Interchange Price 1 $/kw or 4 $/kw Interchange Constraints 200 kw and 40 kw Peak Hour Interchange Penalty 50 $/kw Peak Hour T13:00 to T15:00 Battery Parameter Battery 1 Battery 2 Initial SoC 100 kwh 50 kwh Capacity 600 kwh 300 kwh Minimum Capacity 25 kwh 25 kwh Max Charge/Discharge Power 500 kw 250 kw Min Charge/Discharge Power 0 kw 0 kw Max Charge/Discharge Ramp 100 kwh 100 kwh Ramp Penalty 10 $/kwh 10 $/kwh Charge Price 1.25 $/kw 1.25 $/kw Discharge Price 0.75 $/kw 0.75 $/kw P 11

12 Case Studies Uncertain Parameters Set the positive/negative deviation bounds PV Generation to be 5%, 15%, and 25% for different commitment intervals. (1h-3h, 4h-7h, and 8h-10h) 3 deterministic scenarios: Nominal High Bound Low Bound

13 Case Studies (a) Nominal Scenario Battery Charge/Discharge Schedule (c) Lower Bound Scenario Battery Charge/Discharge Schedule (b) Robust Optimization Battery Charge/Discharge Schedule (d) Upper Bound Scenario Battery Charge/Discharge Schedule

14 Case Studies Model Type Optimal Obj. ($) Deterministic (Norm) 243 Deterministic (High) Deterministic (Low) Robust Model Nominal High Low Robust Islanding Start T8:00 T8:00 T13:00 T13:00 Point Islanding End T15:00 T18:00 T15:00 T14:30 Point Duration 7h 10h 2h 1.5h

15 Conclusion Battery and islanding scheduling in a micro grid with PV generation is a good use case for Robust Optimization Ensure reliability even with PV uncertainty Tractable performance due to the size of the problem

16 Thank you!