广西大学学报(自然科学版)
廣西大學學報(自然科學版)
엄서대학학보(자연과학판)
JOURNAL OF GUANGXI UNIVERSITY (NATURAL SCIENCE EDITION)
2014年
2期
328-333
,共6页
机车走行部%滚动轴承%故障诊断%主成分分析方法(PCA)%最小二乘支持向量机(LSSVM)
機車走行部%滾動軸承%故障診斷%主成分分析方法(PCA)%最小二乘支持嚮量機(LSSVM)
궤차주행부%곤동축승%고장진단%주성분분석방법(PCA)%최소이승지지향량궤(LSSVM)
locomotive running gear%rolling bearing%fault diagnosis%principal component analysis ( PCA)%least squares support vector machine( LSSVM)
机车走行部的故障诊断是保障机车安全运行的重要手段之一。鉴于目前机车走行部滚动轴承故障诊断中存在的特征提取困难、故障诊断不准确等问题,提出一种基于最小二乘支持向量机( LSSVM)的故障诊断方法。通过主成分分析方法( PCA)对机车走行部滚动轴承的故障特征进行特征提取,将提取后的特征向量输入到故障诊断模型中,结合最小二乘支持向量机原理实现机车走行部滚动轴承的故障诊断。仿真结果表明,相比其他方法而言分类准确率达到了100%,模型构建时间为3.642 s,满足机车走行部滚动轴承的诊断要求。将最小二乘支持向量机引入并应用到机车走行部故障诊断领域中,为机车走行部故障诊断系统的研究与开发提供新思路和新方法。
機車走行部的故障診斷是保障機車安全運行的重要手段之一。鑒于目前機車走行部滾動軸承故障診斷中存在的特徵提取睏難、故障診斷不準確等問題,提齣一種基于最小二乘支持嚮量機( LSSVM)的故障診斷方法。通過主成分分析方法( PCA)對機車走行部滾動軸承的故障特徵進行特徵提取,將提取後的特徵嚮量輸入到故障診斷模型中,結閤最小二乘支持嚮量機原理實現機車走行部滾動軸承的故障診斷。倣真結果錶明,相比其他方法而言分類準確率達到瞭100%,模型構建時間為3.642 s,滿足機車走行部滾動軸承的診斷要求。將最小二乘支持嚮量機引入併應用到機車走行部故障診斷領域中,為機車走行部故障診斷繫統的研究與開髮提供新思路和新方法。
궤차주행부적고장진단시보장궤차안전운행적중요수단지일。감우목전궤차주행부곤동축승고장진단중존재적특정제취곤난、고장진단불준학등문제,제출일충기우최소이승지지향량궤( LSSVM)적고장진단방법。통과주성분분석방법( PCA)대궤차주행부곤동축승적고장특정진행특정제취,장제취후적특정향량수입도고장진단모형중,결합최소이승지지향량궤원리실현궤차주행부곤동축승적고장진단。방진결과표명,상비기타방법이언분류준학솔체도료100%,모형구건시간위3.642 s,만족궤차주행부곤동축승적진단요구。장최소이승지지향량궤인입병응용도궤차주행부고장진단영역중,위궤차주행부고장진단계통적연구여개발제공신사로화신방법。
Fault diagnosis of locomotive running gear is one of the important means to guarantee the safe operation of the locomotive. In view of existing fault diagnosis problems that the feature extrac-tion is difficult and fault detection is not accurate for the locomotive running gear rolling bearing, fault diagnosis method based on PCA-LSSVM is proposed in this paper. By using the method of prin-cipal component analysis ( PCA) , the fault characteristics of locomotive running gear rolling bearing are extracted and the extracted feature vector is input into the fault diagnosis model. The fault diag-nosis of locomotive running gear of rolling bearing is realized by combined with the principle of least squares support vector machine ( LSSVM ) . The simulation results show that compared with other methods the classification accuracy reached 100 percent and the model building time was 3 . 642 sec-onds, satisfying the demands of locomotive rolling bearing diagnosis. The application of the LSSVM to the field of fault diagnosis of locomotive running gear provides new ideas and methods for the re-search and development of locomotive running gear fault diagnosis system.