基于SCA-VMD和排列熵的轴承故障诊断研究

Bearing Fault Diagnosis Based on SCA-VMD and PE

轴承在早期故障时信号微弱、振动数据的获取或者缺失,导致轴承故障诊断的准确率降低,为此提出了一种同一故障类型不同损伤尺寸的数据集构建方法,构建正余弦优化算法(Sine Cosine Algorithm,SCA)优化变分模态分解(Variational Mode Decomposition,VMD)参数的轴承故障诊断模型。首先,对比经验模态分解以及变分模态算法的时域频域波形,再采用SCA算法对模态分解个数k和惩罚系数α寻找最优组合;然后,计算模态分量的排列熵值,选取峭度值最大的四个模态分量构建特征数据集,支持向量积进行参数优化,构建最优故障诊断模型,对不同故障诊断模型结果进行对比;最后,基于SCA-VMD和排列熵的轴承故障诊断研究准确度率为99.3%,数据构建的方法更符合实际的工况。

The accuracy of bearing fault diagnosis is reduced due to weak signal and acquisition or loss of vibration data in early bearing faults.Therefore,a data set construction method with different fault scales for the same damage fault type was proposed to construct a bearing fault diagnosis model with sine-cosine optimization algorithm(SCA)optimizing variational mode decomposition(VMD)parameters.First,comparing the time-domain and frequency-domain waveforms of empirical mode decomposition(EMD)and VMD algorithm,SCA algorithm is used to find the optimal combination of the number of mode decomposition k and penalty coefficient a.Then,the permutation entropy of the modal components is calculated,and the four modal components with the largest kurtosis value are selected to construct the feature data set.The optimal fault diagnosis model is constructed by using vector machine(SVM)for parameter optimization,and the results of different fault diagnosis models are compared.Finally,the accuracy rate of bearing fault diagnosis research based on SCA-VMD and alignment entropy is 99.3%,and the data construction is more in line with the actual working conditions and has high application.