Predictive control of eV charging for wide-area reduction in grid congestion

Abstract

We present a solution for unlocking the flexibility of electric vehicles (eVs) and ease their integration into the electrical grid. As opposed to most existing solutions that focus on controlling individual charging stations, the proposed solution controls the charging profiles of a group of vehicles to activate their demand-side flexibility. The solution aggregates the eVs into a virtual power plant to alleviate grid congestion, specifically to prevent overloading on a set of grid substations. More precisely, it operates a pool of vehicles that can be charged at any charging point served by grid substations within a given set. The proposed solution implements a model predictive control (MPC) algorithm under constraints to set the optimal charging profile for each eV. This allows the pool to respond to congestion management requests (or constraints) by limiting the peak power at the substations. The solution can also be used to respond to ancillary services requests. The proposed solution was tested on simulated data from 1000 eVs during two weeks in summer in the Greater Paris Area. State of charge (SoC) constraints were respected 99% of the time by the proposed solution compared to 92% by the benchmark approach.