Short-Term Load Forecasting Based On RBF Neural Network And Fuzzy Control

With the increasing demand for power and the implementation of sustainable development strategy,the improvement of power system becomes very important.At present,the development of smart grid technology is fast,and an important module of smart grid is short-term load forecasting,so it has important application value to study different forecasting models and improve the accuracy of forecasting.First of all,this paper mainly introduces three popular methods in short-term load forecasting: short-term load forecasting method based on time series,short-term load forecasting method based on fuzzy control and short-term load forecasting method based on neural network.It simply summarizes several short-term load forecasting methods and analyzes their advantages and disadvantages.The comparison provides the idea for the short-term load forecasting method proposed in this paper,and also provides the reference for the further research and practical application of the researchers in the field of short-term load forecasting.Secondly,this paper proposes a short-term load forecasting method based on time series.The modified model divides time series data into four parts: trend item,period item,holiday item and error item;in the experimental part,this chapter provides a set of preprocessing method flow to preprocess the data,among which,aiming at the problem that the sampling rate of data collected by the current smart grid is not constant,a data smoothing algorithm is proposed,which has strong practicability It can process the data with different sampling rates into data with approximately the same time interval,so as to adapt to various prediction models flexibly.For the load data after cleaning,we use the time series forecasting method proposed in this paper to predict the rest day and work day load.The prediction results show that the model has good prediction effect and meets the actual application requirements.Finally,this paper proposes a forecasting model based on the combination of RBF neural network and fuzzy control,which further improves the accuracy of short-term load forecasting.Radial basis function neural network adopts shallow structue.Each unit of radiel basis function neural network is usually affected by a small part of the input mode,so the interaction between each unit is very small,which makes radial basis function neural network have better generalization performance.In addition,the fuzzy control technology is introduced into the radial basis function neural network.The error and the rate of error change generated by the radial basis function neural network are fed back to the fuzzy controller,and an adjustment is output to reduce the prediction error.Compared with the time series prediction model and the single radial basic function neural network prediction model,the accuracy of the prediction model proposed in this paper is obviously improved.The experimental results show that the model based on RBF neural network and fuzzy controller can effectively reduce the average relative error of the existing model.This study is meaningful and has practical application value.