| In the process of production and life in modern society,the accuracy of load forecasting is directly related to the normal production and life order of the society.Improving the level of load forecasting plays a great role in promoting the operation of the whole society.The main research contents of this paper are as follows:(1)Firstly,study the factors affecting load forecasting,master the processing methods of abnormal and missing values of collected sample data,use the load deviation rate method to smooth the data without distortion,improve the data persuasion,identify the key elements affecting the accuracy of load forecasting,quantify the elements that cannot be identified digitally,and build a complete load forecasting framework.(2)Then,in view of the defect that it is difficult to find the global optimal value in the standard BP neural network load forecasting model,which depends on the selection of the initial parameter weight and threshold of each neuron and the calculation results of the standard BP neural network,a gwo-bp model is proposed to continuously optimize the connection parameters between neurons in each layer of BP neural network by improving the gray wolf optimization algorithm until the load forecasting model meets the accuracy requirements.(3)Compare the standard test examples to verify the advantages of gwo-bp algorithm compared with the standard algorithm,and then apply it to the short-term load forecasting of Fushun Power Grid.The input data is historical load data,combined with relevant factors such as weather,meteorology,temperature and humidity,and compare the results with the actual value and the standard BP neural network model.Through the verification of actual examples,the average relative error of gwo-bp is 2.92%,and the average relative error of BP is 6.03%,Compared with gwo-bp algorithm,the minimum value of BP relative error increases by 0.483%,the maximum value increases by 8.231%,and the average value increases by 3.11%.Therefore,gwo-bp model has a good application prospect in guiding the load forecasting of Fushun Power Grid. |