Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4945427 | International Journal of Electrical Power & Energy Systems | 2018 | 12 Pages |
Abstract
This paper compares traditional local voltage control strategy with coordinated, optimization-based ones in low voltage (LV) distribution systems with photovoltaics and battery energy storage systems. Optimization-based strategies are formulated within a model predictive control (MPC) framework. Three strategies based on MPC are proposed and implemented, namely, centralized, decentralized and distributed MPC. The formulated strategies for voltage control are compared in a case study using a modified CIGRÃ European 3-area low-voltage network. Results indicate that decentralized MPC gives a better voltage profile in the network when compared to local voltage control strategy, since the latter inherently fails to maintain voltages of buses in the network not connected to photovoltaics or battery storage system within limits. Centralized MPC strategy is able to provide the optimal voltage profile across the network but utilizes 13% higher reactive power from the control devices to achieve this when compared to decentralized MPC. The latter performs well as long as the reactive power reserves within an area is sufficient but faces drawbacks similar to that of local voltage control strategy when the reactive reserves are completely exhausted. Distributed MPC utilizes 1.3% higher amount of reactive power reserves compared to centralized MPC in order to provide a network voltage profile similar to that of the latter while also yielding architectural advantages of decentralized MPC.
Keywords
DSOMPPTMPCLVCPCCDistributed controlDistribution system operatorBESSMaximum power point trackingDistribution systemBattery energy storage systemPhotovoltaicproportional-integralPoint of common couplingMedium voltagelow voltageDistributed generatordecentralized controlcentralized controlModel predictive control
Related Topics
Physical Sciences and Engineering
Computer Science
Artificial Intelligence
Authors
Pavan Balram, Le Anh Tuan, Ola Carlson,