基于GOA优化支持向量机滚动轴承故障诊断

针对滚动轴承早期故障难以辨别，提出了一种采用变分模态分解法(visual molecular dynamics, VMD)提取特征，基于蚱蜢算法(grasshopper optimization algorithm, GOA)和优化支持向量机(support vector machines, SVM)的故障诊断方法。首先采用贪心策略预处理滚动轴承的振动信号数据，然后基于变分模态分解处理振动信号数据得到多个本征模态分量(intrinsic mode function, IMF),其次计算各IMF分量的能量和相关时频特征构成多模态特征矩阵，最后利用蚱蜢算法优化的支持向量机进行故障的诊断和识别。通过实验测试大量数据得出的滚动轴承故障诊断结果表明VMD-GOA-SVM不仅可以识别滚动轴承不同的故障类型，同时相比传统方法亦有较高的准确度和运行效率。

Fault Diagnosis of rolling Bearing based on GOA optimized Support Vector Machine

In the early stage of rolling bearing failure, the failure signal is weak, so that the failure is difficult to distinguish. This paper proposes a fault diagnosis method based on Grasshopper Optimization Algorithm (GOA) optimization Support Vector Machines (SVM) using variational modal decomposition method (Visual Molecular Dynamics, VMD). First, the greedy strategy is used to preprocess the vibration signal data of the rolling bearing, and then the vibration signal data is processed based on the variational modal decomposition to obtain multiple intrinsic mode components (Intrinsic Mode Function, IMF), and then the energy and related time-frequency of each IMF component are calculated. The features constitute a multi-modal feature matrix, and finally the support vector machine optimized by the grasshopper algorithm is used to diagnose and identify the fault. The diagnostic results of rolling bearing faults obtained through experimental testing of a large amount of data show that VMD-GOA-SVM can not only identify different types of rolling bearing faults, but also has higher accuracy and operating efficiency than traditional methods.