基于IGWO-SVM的轴承故障分类预测

Research on IGWO-SVM in Bearing Fault Classification and Prediction

随着旋转机械设备的集成化程度不断提高,轴承发生故障的概率以及故障诊断的难度都在增加。为了解决常规故障诊断出现误报和漏报等问题,课题组在采集已清洗的轴承数据基础上提出了一种新型分类预测算法。课题组通过改进的灰狼算法来收敛支持向量机的参数,并对数据集进行训练优化,以精准地对轴承故障进行判断和预测。研究结果表明判断精度可高达99.4%;通过与其他现有主流分类预测模型进行比较,进一步验证了该优化方法的优异性。该优化方案可以很好地应用于SCADA等实时状态监测系统并进行精准故障分类预测。

With the continuous improvement of the degree of integration of rotating machinery equipment,the probability of bearing failures and the difficulty of fault diagnosis are increasing.In order to solve the problem such as misrepresentation and omission in conventional fault diagnosis,a new classification prediction algorithm based on the collecting the cleaned bearing data was put forward.The improved grey wolf algorithm was used to converge the parameters of support vector machine and train the dataset for optimization to achieve the precise judgment and projections for bearing fault.The results show that the accuracy can be as high as 99.4%,and the superiority of the optimization method is further validated by comparing it with other existing mainstream classification prediction model.The optimization scheme can be well applied to SCADA system and other real-time status monitoring systems for accurate fault classification and prediction.