基于快速谱相关和PSO-SVM的变工况滚动轴承状态识别

State identification of rolling bearing under variable working condition based on fast spectral correlation and PSO-SVM

滚动轴承故障诊断进入"大数据"时代需要不断发展和完善故障智能识别技术,而已有方法在变工况下的故障识别准确率较低。针对此问题,提出了一种基于快速谱相关和粒子群优化支持向量机(PSO-SVM)的变工况滚动轴承状态识别方法。对滚动轴承的振动信号进行快速谱相关,得到快速谱相关谱;在快速谱相关谱中选取4个循环频率,并求其能量均值,得到信号的特征能量矩阵;将它作为特征向量输入PSO-SVM进行训练和测试。试验结果表明:在运用PSO-SVM进行变工况滚动轴承状态识别的过程中,由快速谱相关谱得到的特征能量矩阵能更好地体现滚动轴承在不同工况下的多状态特征,且PSO-SVM的自适应能力强,不需要人为设置参数,具有更高的识别率。

The fault diagnosis of rolling bearing entering the era of "dig data" requires continuous development and improvement of intelligent fault identification technology. However,the fault identification accuracy of existing me-thods is low under variable working conditions. To solve this problem,a state identification method based on fast spectral correlation and PSO-SVM(Particle Swarm Optimization Support Vector Machine) is proposed for the rolling bearing under variable working conditions. The vibration signal of rolling bearing is fast spectral correlated to obtain the fast spectral correlation spectrum. Four cyclic frequencies are selected from the fast spectral correlation spectrum and the mean value of their energy is calculated to obtain the characteristic energy matrix of the signals,which is taken as the feature vector and input into PSO-SVM for training and testing. The testing results show that,the cha-racteristic energy matrix obtained from the fast spectral correlation spectrum can better reflect the multi-state characteristics of rolling bearing under different working conditions in the process of using PSO-SVM to identify the states of rolling bearing under variable working conditions. PSO-SVM has strong adaptability,does not need to set artificial parameters,and has higher identification rate.