Research On Active Load Forecasting Method Of Distribution Network Considering Electric Vehicles

With the progress and development of science and technology,the work of power system load forecasting has increasingly become an important work in the operation of power systems.Short-term load forecasting is an important part of power system load forecasting,and it is also a key research object of bus load forecasting.It plays a guiding role for the dispatching department.The dispatching department can carry out economic dispatch and optimal unit combination according to the results of load forecasting.The higher the prediction accuracy,the more beneficial it is to improve the utilization rate of power generation equipment and the effectiveness of economic dispatch.However,a high proportion of distributed power sources such as wind power,photovoltaics,and electric vehicles have been connected to the distribution network,which has caused load fluctuations to increase,making load changes more random and uncertain,and posing new challenges to short-term load forecasting.Predicting conventional loads can no longer meet the accuracy requirements of power systems for load forecasting.How to improve the accuracy of short-term load forecasting in the case of a high proportion of distributed power sources more effectively needs to be solved.It is necessary to consider distributed power sources.Analyze and predict the load characteristics of the load with strong volatility.Aiming at the above problems,this article uses artificial intelligence algorithms to solve the current problem of high-proportion distributed power access to the distribution network,and proposes an electric vehicle load forecast based on the combination of orthogonalized maximum information coefficients and long and short-term memory artificial neural networks.Method to quantify the impact of related factors on electric vehicle load forecasting,thereby reducing the dimensionality of the factors that affect the load,determining the most appropriate influencing factors and their quantities,and introducing weight trend variables on this basis,and improving them The previous methods are compared to verify the effectiveness of the proposed method.Based on the load forecasting of electric vehicles,combined with the load characteristics of electric vehicles and the distribution network,comprehensive load forecasting is carried out.The method of combining fuzzy C-means clustering and BP neural network is introduced.Since electric vehicle load is sensitive to seasonal characteristics,and regular load will also be affected by seasonal factors,the importance of seasons is more prominent in the comprehensive load,so consider The seasonal characteristics of the load are divided,and compared with the forecast results in the case of not dividing the seasons,based on this,the peak and valley characteristics of electric vehicles are considered in the comprehensive load,and the peak and valley characteristics of electric vehicles are not considered.FCM+BP model and FCM+SVR model are used to compare and predict the comprehensive load in the case that the peak-trough characteristics of electric vehicles are considered regardless of the four seasons.The proposed method of simulation verification can reduce the comprehensive load forecasting error and improve the accuracy.Based on the above research,the load error is analyzed and compared with the different proportions of electric vehicles in the comprehensive load.