Plug-in Electric Vehicle’s State-of Charge Prediction by Random Forest based Bayesian Optimization for Coordination Strategy in Distribution System Congestion Management

Abstract

The transportation electrification brings several challenges in transport as well as electrical sectors. The global fulfillment of plug-in electric vehicle (PEVs) demand is a big challenge for giant car manufacturers. Consequently, the sharp rise in PEVs popularity forces the electrical sector to handle the high electrical load during grid-to-vehicle (G2V) mode. The congestion level achieves as soon as electrical load increases due to the PEV charging in G2V mode in uncoordinated way. This paper adopted coordinated way of PEVs charging in distribution system considering G2V and vehicle-to-grid (V2G) mode. In this work, the workplace car park consists of solar powered charging-cum-parking lot (SPCPL). The forecasting of initial state-of-charge (SOC) of PEVs has been done by considering the combination of random forest method (RFM) and Bayesian optimization (BO). The particle swarm optimization (PSO) based coordination strategy has been established for the effective congestion management in 38 bus distribution system integrated with SPCPL.