Charging Behavior Analysis Based on Operation Data of Private BEV Customers in Beijing

Abstract

Charging behavior is essential to understanding the real performance and evaluating the sustainability of battery electric vehicle (BEV) development and providing the basis for optimal infrastructure deployment. However, it is very hard to obtain the rules, due to lack of the data support, etc. In this research, analyzing the charging behavior of users with private charging piles (PCPs) is carried out based on the real vehicle data of 168 BEV users in Beijing, covering 8825 charging events for a one-year duration. In this study, the charging behaviors are defined by five indexes: the starting state of charge (SOC) of batteries, charging location selection, charging start time, driving distance, and duration between two charging events. To further find the influencing rules of the PCPs owning state, we setup a method to divide the data into two categories to process further analysis and comparison. Meanwhile, in order to better observe the impact of electric vehicle charging on the power grid, we use a Monte Carlo (MC) simulation to predict the charging load of different users based on the analysis. In addition, an agent-based trip chain model (ABTCM), a multinomial logistic regression (MLR), and a machine learning algorithm (MLA) approach are proposed to analyze the charging behavior. The results show that with 40% or lower charging start SOC, the proportion of users without PCPs (weekday: 55.9%; weekend: 59.9%) is larger than users with PCPs (weekday: 45.5%; weekend: 42.6%). Meanwhile, users without PCPs have a certain decrease in the range of 60–80% charging start SOC. The median charging time duration is 51.44 h for users with PCPs and is 17.25 h for users without PCPs. The charging peak effect is evident, and the two types of users have different power consumption distributions. Due to the existence of PCPs, users have lower mileage anxiety and more diverse charging time choices. The analysis results and method can provide a basis for optimal deployment and allocation of charging infrastructure, and to make suitable incentive policies for changing the charging behavior, targeting the carbon neutral objectives.