• CSCD核心库收录期刊
  • 中文核心期刊
  • 中国科技核心期刊

Electric Power Construction ›› 2020, Vol. 41 ›› Issue (2): 47-57.doi: 10.3969/j.issn.1000-7229.2020.02.006

Previous Articles     Next Articles

Load Probability Modeling and Scenario Generation for Electric Vehicle Charging Station Considering Time Correlation

JIANG Hao1, LIN Shunjiang1, LU Yi2, HE Sen1   

  1. 1.School of Electric Power Engineering, South China University of Technology, Guangzhou 510640, China; 2. Power Dispatch Control Center of Shenzhen Power Supply Bureau Co., Ltd., Shenzhen 518001, Guangdong Province, China
  • Online:2020-02-01
  • Supported by:
    This work is supported by National Natural Science Foundation of China(No. 51977080), Fundamental Research Funds for Central Universities(No. 2019MS015) and Research Program of Shenzhen Power Supply Bureau Co., Ltd.(No. SZKJXM20160174).

Abstract: Due to the influence of random factors such as electric vehicle (EV) owners travel behaviors and traffic condition, the load of EV charging station has strong randomness. Establishing an appropriate probability model to describe the stochastic load of EV charging station is of great significance to the secure operation analysis of distribution network. Taking the load of each period in daily load curve of EV charging station as a random variable, according to historical data, the probability distribution of each single-period load is established with versatile distribution which has the highest fitting accuracy. Because of the continuity of EVs charging behaviors, the charging stations loads in adjacent periods are correlated. Firstly, 96 periods of a day are divided into several continual period sets according to the correlation analysis between loads of different periods. In each set of continual periods, the correlation between the loads of several adjacent periods is relatively large. Then, the joint probability distribution of the loads of several adjacent periods in each set of continual periods is established by using Pair-copula method with D-vine structure. Furthermore, the Pair-copula joint probability distribution model is used to sample and generate load scenarios considering time correlation for all continual period sets in a day. The trend and size of fluctuation are used to represent the fluctuation indices of the generated load scenarios to measure the rationality. Finally, taking the actual historical data of an EV charging station in Shenzhen  as an example, the effectiveness of the proposed method is verified.

Key words: electric vehicle load, probability modeling, time correlation, Pair-copula, scenario generation

CLC Number: