Coordinated EV Charging Based on Charging Profile Clustering and Rule-Based Energy Management

Abstract

<div class="section abstract"><div class="htmlview paragraph">In this work, a novel approach is introduced comprising a combination of unsupervised machine learning (ML) scheme and charging energy management of electric vehicles (EV). The main goal of this implementation is to reduce the load peak of charging EV’s, which are regular users of electric vehicle supply equipment (EVSE) of a certain building and, at the same time, to meet their electric and behavioral demands. The unsupervised ML considers certain features within the charging profiles in addition to the behavioral characteristics of the EV based on its intended use. Moreover, these features are extracted from large sets of history measurement data of the EVSE, which are stored in the data bank. The ML categorizes the EVs within certain clusters having defined specifications. After that and based on these clusters, several rules are extracted in order to manage the charging energy while meeting both the electric and behavioral demands of the EVs and thus enabling a smart and coordinated charging process. The energy management is carried out in terms of both the current situation of the charging points and the participating vehicles. As a proof of concept, the FKFS research EVSE are employed to collect the data from a regular group of service, fleet and guest EVs. Then a rule based energy management system is implemented to reduce the load peak of EV for a given real scenario. The approach is successful of reducing the charging load peak substantially while meeting electric energy demands of the participating vehicles and increasing the availability of the charging points significantly. The combined action of the EV clustering scheme and energy management provides a smart and coordinated charging, which guarantees reduced load peaks of the building in addition, to ensuring proper charging of the EVs.</div></div>