基于弹性核凸包张量机的机械设备热成像故障诊断方法

A Fault Diagnosis Method of Mechanical Equipment Based on FKCH-TMand Thermal Images

针对机械设备热成像故障诊断中图像样本向量化维数灾难和结构信息破坏的问题,提出了基于弹性核凸包张量机的机械设备热成像故障诊断方法。首先定义样本特征张量空间的弹性核凸包,然后求解弹性核凸包间的最优超平面来构建弹性核凸包张量机模型,最后以机械设备不同故障状态下的热成像图像为输入,实现机械设备的故障诊断。实验结果表明:该方法可以准确、稳定地对机械设备进行故障诊断。

Aiming at problems of dimensionality disaster and structure information destruction in thermal imaging fault diagnosis of mechanical equipment,a fault diagnosis method of mechanical equipment thermal images was proposed based on FKCH-TM.Firstly,flexible kernel convex hull(FKCH)of each class in feature tensor space was defined.Then,the optimal hyperplane between the FKCH was found,and FKCH-TM model was built.Finally,thermal images of mechanical equipment under different fault conditions were directly used as the input of FKCH-TM for fault diagnosis.Experimental results show that the proposed method may identify and diagnose the faults of mechanical equipment accurately and stably.