Fluctuation-sensitive model-predictive communication for distributed energy resources

Abstract

In the course of increasing numbers of electric vehicles (EVs) and renewable energy sources such as photovoltaic (PV) systems coordinated operation of these units becomes a promising option. In particular a concept currently in development is the charging of EVs at arbitrary locations with own produced renewable energy e.g. from a home PV system. However this requires extensive communication to the EV due to the fluctuating power feed-in of the PV system. In order to reduce the amount of transmissions per distributed energy resources (DERs) model-predictive communication can be applied. This approach has already been proved valid for PV systems in Virtual Power Plants (VPPs) where the power production data has to be transmitted continuously to central entities which balance the fluctuating feed-in from PV, wind-power and other systems. In this paper we show that the existing PV model can be enhanced with sensitivity to clouds and other shadowing effects. The proposed changes can be easily transferred to other DERs and further reduce the amount of communication — in case of a PV system especially under cloudy weather conditions. For the validation of the model publicly available real-life data from a PV system over a period of one year has been used. The proposed model-predictive communication scheme saves up to 68.6% of traffic while only inducing a mean normalized error rate of 1.09 %.