Spatiotemporal distribution model of charging demand for electric vehicle charging stations based on markov chain

Abstract

Considering the diversity of travel habits of users, spatiotemporal distribution model of charging load for highly random electric vehicles(EV), based on Markov chain, is proposed. According to the residents' travel habits of the 2017 National Household Travel Survey (NHTS) in the United States, the destinations are divided into homes, work places, and other locations. Based on Markov Chain and Monte Carlo method, a highly random and complex process chain with unlimited processes is constructed. The distribution of start times and end times in one-day journey is fitted by the parameterless distribution with normal distribution as the kernel. The distribution of travel time, travel distance and dwell time is fitted by lognormal distribution. Then, a spatiotemporal distribution model of one-day vehicle travel is established. Considering the influence of EVs, the dual-input fuzzy algorithm of travel time and travel distance is used to calculate the power consumption of the trip. According to the crowd's travel anxiety and different travel needs, different charging fredquencies and different charging power levels are used for vehicle charging behavior. Finally, the Monte Carlo method is used to calculate and analyse the charging load of EVs in different classification scenarios, such as household income, home address, weekdays and weekends. The one-day electric vehicle load distribution model was successfully established in three locations. The results show that the load of electric vehicle charging stations is not only affected by factors such as holidays, but also by the composition of the urban population.