【正文】
003. 1823. ?。ň?輯 曹大剛)Data fusion in cable fault prediction and location system based on wavelet transformAbstract: A data fusion method based on wavelet transform for the plex nonlinear system modeling of cable fault prediction and location is brought forward. First, the careless mistake data are eliminated, and then the error trend of the measured validation data is detected by using wavelet transform. Consequently, the reliability of the measured validation data is measured according to the error trend, and the weights of the measured validation data are determined. Simulation shows that the mean value of prediction error of the neural network model constructed by using the fused data is quite less than that by using the data before fusion, and the data fusion method based on wavelet transform ensures the reliability of the training samples used for modeling and the rationality of the data rights.Key words: cable;fault location;wavelet;neural network;data fusion作 者 簡 介 汪 梅, 女,安徽金寨人,生于1965年,1990年畢業(yè)于西安礦業(yè)學(xué)院自動化系,現(xiàn)任西安科技大學(xué)副教授,博士生,主要從事控制理論與控制工程領(lǐng)域的科研與教學(xué)工作。近年來,先后在國際會議、核心期刊上發(fā)表中、英文學(xué)術(shù)論文及教學(xué)論文20多篇,其中多篇被EI、ISTP等檢索;在科學(xué)出版社合作出版教材《系統(tǒng)辨識及其MATLAB仿真》;主持陜西省教委專項科學(xué)研究基金項目“煤礦井下電纜故障定位理論與技術(shù)的研究”(JK00267)及陜西省科技攻關(guān)項目“智能電纜故障預(yù)測與定位裝置的研制”(2003K06G19);并參加了陜西21世紀(jì)初高等教育教學(xué)改革工程項目“雙語教學(xué)中教學(xué)規(guī)律的研究及教學(xué)方法的改革”。8 /