基于PSO-SVM的采煤机摇臂齿轮箱故障诊断研究

Fault Diagnosis of Gear Box for Shearer Arm Based on PSO-SVM

针对采煤机摇臂齿轮箱故障诊断的难题,提出了一种基于时频特征和PSO-SVM的故障诊断方法。考虑到SVM模型参数和故障特征对诊断结果有着重要的影响,提出了利用PSO对SVM参数进行优化,并同时选择最佳的特征子集,以获得性能最优的SVM分类器,最后将故障特征向量输入到优化的SVM分类器中进行故障诊断。轴承和齿轮的故障诊断实验结果表明,PSOSVM获得了比常规SVM更好的故障诊断性能。

Aiming at the problem of fault diagnosis on gear box for shearer arm, a novel fault diagnosis approach based on time-frequency characteristics and PSO-SVM is proposed in this paper. Because the SVM parameters and fault features have a great impact on the diagnosis results, PSO is simultaneously used for SVM parameter optimization while discovering the optimal subset of features, and consequently the optimal SVM classifier is obtained. Finally, the optimized feature vectors are input to the optimal SVM classifier for fault diagnosis. Experimental results of fault diagnosis for the bearing and gear failure datasets indicate that the PSO-SVM classifier achieves higher performance than normal SVM.