CAJ | 학술논문

针对滚动轴承故障信号的非平稳特性，提出了一种基于 EEMD 和 SVM 的滚动轴承故障诊断方法。首先，利用 EEMD 将原始信号进行分解，将其分解为多个固有模态函数（IMF)之和，选取若干含有主要故障信息的 IMF 分量做进一步分析。然后利用这些 IMF 中的特征信息构建故障特征向量，并使用 PCA 对特征向量进行降维。最后，运用GA 优化的 SVM 对滚动轴承故障进行诊断，实验结果表明其诊断准确率较高，比 BP 神经网络更有优势。
침대곤동축승고장신호적비평은특성，제출료일충기우 EEMD 화 SVM 적곤동축승고장진단방법。수선，이용 EEMD 장원시신호진행분해，장기분해위다개고유모태함수（IMF)지화，선취약간함유주요고장신식적 IMF 분량주진일보분석。연후이용저사 IMF 중적특정신식구건고장특정향량，병사용 PCA 대특정향량진행강유。최후，운용GA 우화적 SVM 대곤동축승고장진행진단，실험결과표명기진단준학솔교고，비 BP 신경망락경유우세。
According to the non-stationary characteristics of rolling bearing fault signal, a fault diagnosis method of rolling bearings based on ensemble empirical mode decomposition (EEMD) and support machine vector (SVM) is proposed. EEMD was used to decompose the fault signal into sum of several IMFs, and some IMFs containing main fault information was selected for further study. Then, the features of these IMFs were used to build fault feature vector information and PCA was used to reduce the dimension of feature vectors. Finally, the GA-optimized SVM was used for rolling bearing fault diagnosis, and test results showed that the SVM perform much better than the BP neural net-works.