Modeling EV fleet Load in Distribution Grids: A Data-Driven Approach

This paper proposes a modeling method for electric vehicle (EV) charging loads in the distribution grids. Different from previous work that modeling under general car travel distance based statistics, we advocate using real world power consumption data, which collected from the charging port meters. The essential charging behavior characteristics were retrieved from such high-resolution data. The vehicle behavior indicates that the distribution of the initial SOC when EV put into charge is not necessarily Iognormal type. Possible causes for such non-ideality are discussed. The new model proposed here can incorporate the stochastic nature of EV charging to improve researchers' analysis. An explicit description of the model along with its operating dynamics and a practice to analyze the total power load of a mid-sized EV fleet is provided. We demonstrate that the proposed model can more correctly reflect the total power need of a fleet and can be adopted as a load-forecasting tool of EV fleet charging load.