基于参数优化VMD的轴承故障诊断方法研究

Research on Fault Diagnosis Method of Rolling Bearing Based on Parameter-optimized VMD

针对早期滚动轴承故障诊断准确率低、信号特性不平稳且难以获取大量样本等问题,提出基于最大相关峭度解卷积(MCKD)、乌燕鸥算法优化变分模态分解(STOA-VMD)和粒子群算法优化支持向量机(PSO-SVM)的滚动轴承故障诊断模型。首先使用MCKD处理信号提高信噪比,再通过STOA-VMD对信号进行分解,特征参量选用均方根熵值,输入PSO-SVM实现故障分类,并由实验和仿真验证了该方法可使故障诊断准确率明显提高。

Aiming at the problems of low accuracy of early rolling bearing fault diagnosis, non-stationary signal characteristics and being difficult to obtain a large number of samples, a rolling bearing fault diagnosis model based on maximum correlated kurtosis deconvolution(MCKD), sooty tern optimization algorithm optimizing variational modal decomposition(STOA-VMD) and particle swarm optimization optimizing support vector machine(PSO-SVM) was proposed. Firstly the signal is processed by MCKD to improve the signal-to-noise ratio, and then the signal is decomposed by STOAVMD. Takes root mean square entropy as the fault characteristic parameter and import into PSO-SVM, the fault classification is realized, and the fault diagnosis rate of this method is significantly improved by experiment and simulation verification.