Research And Application Of Short-term Load Forecasting Method For 10kV Feeder

The power system load is a dynamic process,which is affected by regional economic development,political factors,natural conditions and other factors.Therefore,how to make use of historical load data and collect other historical influencing factors to establish a suitable short-term load forecasting model for rapid and high Predicting the load value for a certain period of time in the future with precision has become a key point for maintaining the safe operation of the power system and balancing the supply and demand in the power market.Short-term load forecasting refers to predicting the load value of one day to one week in the future,which is of great guiding significance for the stable improvement of power supply quality and reliability.It is also a prerequisite for narrowing the difference in load forecasting,reducing opera ting costs and profit loss.Therefore,continuous exploration It is becoming increasingly important to study advanced,fast and effective short-term load forecasting methods.Firstly,the article discusses load forecasting algorithm models such as trend extrapolation,time series method,gray model,expert system,neural network,wavelet analysis,etc.The above load forecasting methods are common at home and abroad,and then the article selects neural network algorithm for research and analysis.and introdu ces network structure and basic principles of the BP network and Elman network algorithm in details.The article makes use of the historical load data of a 10 k V feeder in a substation in Panyu District,Guangzhou City.First of all,pre-process the historical data to avoid affecting the prediction effect due to abnormal data.At the same time,through the analysis of the load characteristics and correlation analysis of historical data,the influence of temperature and day type on the load of 10 k V feeder is analyzed.In this paper,MATLAB software is used for algorithmic programming.By calling the neural network toolbox function,the BP network and Elman network models are established and trained respectively,and the comparative analysis of the verification samples and the simulation prediction results is used.According to the analysis of the simulation results,we can see that the rationality of the neural network load prediction model used in this article.In addition,according to the short-term load forecasting results using different neural network algorithms,comparative analysis is made in terms of error and convergence speed,which shows that the Elman neural network algorithm has better prediction performance than the BP neural network algorithm in the application of short-term power load forecasting.In order to further improve the prediction accuracy,this paper adopts improvement of the excitation function and network structure of the Elman network algorithm,so as to obtain better performance in short-term load forecasting.Finally,based on the application of the short term load forecasting model in this paper,some problems and some personal suggestions for improvement are put forward.