基于多时间尺度融合的规模化电动汽车可行域边界区间预测OA

As a new type of flexible energy storage device,electric vehicles(EVs)can provide valuable flexible regulation capabilities for the power system by participating in the regulation of electric vehicle aggregators.With the continuous increase in the penetration rate of renewable energy,the regulation demands of the power system are becoming increasingly complex.At present,adjustable capacity prediction of EVs faces challenges in achieving higher accuracy and resilience against risks.To this end,this paper proposes a prediction method for feasible region boundary interval of large-scale EVs based on multi-timescale integration.Firstly,considering the influence of user credibility on the evaluation of the adjustable capability of EVs,access rules for participating in regulation and the feasible region correction methods are designed for different types of EVs,and then aggregated to obtain the feasible region boundary of large-scale EVs.Secondly,a trend-preserving data amplification method for time series is proposed to achieve the amplification of a single feasible region boundary dataset to multiple datasets with similar time series information but different specific values.Again,based on multiple datasets obtained through amplification,a multi-timescale integration prediction and probabilistic interval modeling method for the feasible region boundary of EVs combining Auto former and Gaussian mixture models is proposed.Finally,based on the large-scale historical charging records of Shenzhen,China,the effectiveness of the proposed method is verified.The results show that the proposed prediction method of feasible region boundary interval can provide more accurate and reliable information support for aggregators participating in the electricity market and power system regulation on the future adjustable capacity of EV resources.