Decentralized demand side management of plug-in hybrid vehicles in a Smart Grid

Abstract

Research predicts that in 2030, around 30% of all vehicles in Belgium will be plug-in hybrid electric vehicles (PHEVs). Because most PHEVs are charged after working hours, the existing peak load in the evening will increase significantly. Large peak loads cause more expensive production and can even damage the electricity infrastructure. In a Smart Grid, the charging of PHEVs can be controlled to reduce peak load, denoted as demand side management (DSM). The goal of our research is to compare several solutions for DSM of PHEVs. This paper takes a first step by benchmarking a multi-agent solution against an optimal quadratic programming (QP) scheduler solution. Simulations show that a QP scheduler is able to optimally flatten peak loads while sufficiently charging the PHEV batteries. However, this solution is unfeasible in practice because it scales poorly and requires complete information on when and how much PHEVs need to charge beforehand, which is not available. The MAS solution proves to be scalable and adaptable to incomplete and unpredictable information while peaks are still reduced with an efficiency up to 95% compared to the QP scheduler.