Integrated Modeling and Forecasting of Electric Vehicles Charging Profiles Based on Real Data

TL;DR

This paper develops a data-driven model for residential EV charging profiles based on real usage data, enabling impact assessment on power distribution networks amid increasing EV adoption.

Contribution

It introduces a novel multi-charging algorithm and probability-based modeling approach derived from real data to simulate EV charging behavior.

Findings

Charging patterns vary significantly among users.

The model accurately predicts daily charging profiles.

Impact analysis shows potential stress on distribution networks.

Abstract

In the context of energy transition and decarbonization of the economy, several governments will ban the sale of new combustion vehicles by 2050. Thus, growing penetration of electric vehicles (EVs) in distribution networks (DN) is predicted. Impact analyzes must be performed to determine if mitigation means are needed to accommodate a large quantity of EVs in the DN. Furthermore, the habits of the local population resulting in different EVs charging patterns needs to be realistically considered. This article proposed an individual residential EVs multi-charging algorithm based on the observed charging behavior of 500 measured residential EVs located in a large North American utility. Probability functions are derived from the analysis of these charging patterns. These can be used to model daily charging profiles of individual EV to assess, in a quasi-static time-series perspective, their impact either on a single costumer or a whole DN. An impact evaluation study is also presented.