Linear Regression-Based Relative Efficiency Analysis Using Empirical Charging Data

Transportation electrification is a major topic, with a constantly increasing share of electrical vehicles (EVs). So far, one of the factors that has received little attention is the relative charging efficiency, which describes the efficiency related to the vehicle's benchmark. It directly influences the charging energy demand and should therefore be studied. It is impacted by various dependent factors like the used charger, battery, auxiliary systems, and external conditions. These factors are rarely analyzed as a whole due to the complexity of models and the lack of empirical datasets. This paper presents a methodology using a linear regression-based relative efficiency analysis, examining empirical charging data with the primary objective of creating usable generic models. It analyzes the impact of crucial factors such as temperature, charging power, state of charge, and vehicle age. During the analysis, multiple linear regression is performed on the individual vehicles, vehicle groups, and the whole dataset. It is shown that the vehicles have highly individual behavior and influences, so that an analysis of the whole dataset does not lead to suitable results. However, multiple vehicle groups with superior fitting can be created. The model results can be used subsequently to generate an understanding of the influence of the factors and include charging efficiency in EV energy demand models.