Study On Wavelet Denoising Method For Fault Rate Prediction Model Of Transmission Lines In Inner Mongolia

Transmission line is an indispensable infrastructure in the power system.If the transmission line fails,it will cause incalculable loss to the whole power system.How to study and predict the failure rate of transmission line reasonably plays an important role in the security of the whole power system.The data of transmission line fault rate contains a lot of redundant interference information,which has the characteristics of nonlinearity and randomness.In this paper,the method of wavelet de-noising is used to de-noising the data of transmission line fault rate in Inner Mongolia,to remove the redundant interference in the data,and to model and predict the de-noising data of fault rate,so as to improve the prediction accuracy.Firstly,this paper compares the traditional four kinds of wavelet de-noising methods qualitatively,constructs an improved wavelet threshold de-noising method,and proves that compared with the traditional wavelet de-noising method,the information loss is reduced,and the de-noising effect is better.Secondly,two traditional wavelet de-noising methods and the improved wavelet threshold de-noising method are used to denoise the transmission line fault rate data in Inner Mongolia,and the denoising effect is compared.Finally,we use the time series autoregressive moving average(ARMA)model,BP neural network model and the combination of the two prediction(ARMA-BP neural network)model to predict the original data and the denoised data respectively,and compare the prediction results and prediction errors.Theoretical analysis and simulation test show that the transmission line fault rate can effectively reduce the redundant information interference in the data sequence after wavelet denoising pretreatment,and retain the characteristics of the signal.The signal-to-noise ratio(SNR)of transmission line fault rate data after wavelet improved threshold denoising is higher,the root mean square error(RMSE)is lower,and the denoising effect is better.Compared with using the original data to build the model directly,the prediction accuracy is obviously improved and the prediction error can be effectively reduced.The prediction accuracy based on the improved wavelet threshold denoising prediction model is better than that of other traditional wavelet denoising methods,which shows that choosing the wavelet denoising method with better denoising effect can improve the prediction accuracy more effectively,which proves that the transmission line fault rate prediction model based on wavelet denoising method is effective.