【摘要】：針對變壓器故障診斷神經(jīng)網(wǎng)絡模型存在網(wǎng)絡結構復雜、訓練時間長等問題,提出基于粗糙集及RBF神經(jīng)網(wǎng)絡的變壓器故障診斷方法。運用粗糙集理論中無決策分析,建立基于可分辨矩陣和信息熵的知識約簡算法,進行數(shù)據(jù)挖掘,尋找最小約簡;以處理后的數(shù)據(jù)集合作為訓練樣本,采用高斯函數(shù)作為徑向基函數(shù),分別求解方差及各層權值,建立變壓器故障診斷模型。通過測試對比,此算法雖然略微降低診斷正確率,但網(wǎng)絡結構簡單、訓練速度快、泛化能力強,對提高神經(jīng)網(wǎng)絡在變壓器故障診斷中的應用性能有較好的指導意義。
[Abstract]:Aiming at the problems of complex network structure and long training time in transformer fault diagnosis neural network model, a transformer fault diagnosis method based on rough set and RBF neural network is proposed. In this paper, a knowledge reduction algorithm based on discernible matrix and information entropy is established by using the no-decision analysis in rough set theory, data mining is carried out to find the minimum reduction, and the processed data set is used as the training sample. The Gao Si function is used as the radial basis function to solve the variance and the weights of each layer, and the transformer fault diagnosis model is established. The test results show that this algorithm has the advantages of simple network structure, fast training speed and strong generalization ability, although it slightly reduces the correct rate of diagnosis. It has a good guiding significance for improving the application performance of neural network in transformer fault diagnosis.
【作者單位】： 南京工程學院電力工程學院;江蘇省高校"配電網(wǎng)智能技術與裝備"協(xié)同創(chuàng)新中心;國網(wǎng)江蘇省電力公司;國網(wǎng)南通供電公司;
【基金】：江蘇省高校自然科學研究基金面上項目(13KJB470006) 江蘇省電力公司2014年科技項目(J2014090) 江蘇省電力公司2015年科技項目
【分類號】：TP183;TM407

【參考文獻】

相關期刊論文 前9條

1 趙文清;朱永利;;電力變壓器狀態(tài)評估綜述[J];變壓器;2007年11期

2 朱振玉;董春輝;;變壓器故障診斷的粗集決策新方法[J];電測與儀表;2010年07期

3 蘭華;鄭廣瑞;孫學林;侯彩霞;;基于Vague集的模糊聚類在變壓器故障診斷中的研究[J];電測與儀表;2011年01期

4 朱繼;喻瑛;王辰煒;饒s，

本文編號：2140847