Heuristic Procedure For The Centralized Control Of EV Charging In LV Networks

Abstract

High penetration levels of electric vehicles (EVs) in distribution networks will create significant operational challenges to distribution system operators and demand robust methods to optimize the charging problem. This work addresses the problem of controlling the charging of electric vehicles in low voltage distribution systems using an optimization framework formulated as a scheduling problem. A computationally efficient procedure with a detailed modeling of the network is sought. The objective is to obtain a charging schedule for all the EVs connected to the network satisfying a set of network operational constraints. The approach uses a complete AC three-phase modeling of the network and deploys a time-series power flow coupled with a greedy heuristic to obtain a smart charging procedure over a fixed planning period considering the variable behavior of loads and EV connection patterns. Distribution network constraints consider phase unbalance, transformer and line limitations, and node voltage limits. The approach is tested on an IEEE low voltage distribution test feeder.