Research On Medium And Long Term Load Forecasting Of Power System Based On Grey Theory

With the continuous development and progress of the social economy,power load forecasting plays an important role in people's daily life.The change and performance of power load are important contents discussed in power system planning and operation.Scientific forecasting is the basis for correct decision-making and Guarantee.According to the change law of China's load data,it is a classic gray system,so this paper conducts research on medium and long-term load forecasting based on the gray model.However,with the increasing requirements of forecasting accuracy,the traditional gray model has been unable to meet the requirements of medium and long-term load forecasting,and many external factors that cannot be grasped affect the accuracy of load forecasting.Therefore,the gray model needs further exploration and improvement.To meet the requirements of prediction accuracy.This paper first analyzes and introduces the gray system theory.Through in-depth research on the traditional gray prediction model GM(1,1),it points out that the inherent shortcomings of the model itself lead to limitations in the prediction process,and improves the traditional gray prediction model.In order to weaken the problem of outliers in the original data,two methods of preprocessing the original data are proposed to enhance the overall trend of the original data sequence.For most cases where there are positive and negative fluctuations in the error sequence,the same-dimensional information replenishment technology is used to dynamically supplement the new information and eliminate the original old information;use Fourier error correction to superimpose the error prediction between the predicted value and the original data on the predicted value In the above,the error with the actual data is corrected,then the two are combined to establish an iterative gray prediction model based on Fourier error correction,which improves the prediction accuracy.In response to data fluctuations or outliers,a new parameter ? is introduced to dynamically adjust the gray model parameters a and u,and the weight of thetarget year data is appropriately increased.An equal-dimensional information supplemented gray dynamic parameter prediction model with error correction is proposed.Not only solves the error caused by too much data fluctuation,but also further improves the overall prediction accuracy of the model.In order to verify the performance of the two improved models,two cases were used to verify the prediction effect of the improved gray model,and the traditional gray model,time series method and artificial neural network method were used to compare the load prediction with the two improved gray models.The experimental results show that the two prediction models proposed in this paper have greatly improved the prediction accuracy compared with the traditional gray prediction model.