基于模糊熵和分形维数的滚动轴承早期故障检测

Incipient Fault Detection of Rolling Bearing Based on Fuzzy Entropy and Fractal Dimension

滚动轴承是风机、变速箱等旋转机械中最常发生故障的元件,一旦出现故障,将会使得整个系统停机甚至导致灾难性后果。因此,对滚动轴承进行早期故障检测有着重要的意义。本文提出利用模糊熵和分形维数结合的方法对滚动轴承振动信号进行故障特征的提取,利用樽海鞘群算法优化的支持向量机(SSA-SVM)来构建早期故障检测的模型。通过滚动轴承正常运行和早期故障时采样的历史数据,训练和测试早期故障检测模型,可以实现对故障未知的振动信号进行早期故障检测。通过实验,证实了该方法在滚动轴承早期故障检测上的有效性。

Rolling bearing failures are the most frequent faults in rotating machinery,such as turbomachinery and gearbox,which can cause the system downtime or even lead to catastrophic consequences.Therefore,incipient fault detection is of great significance for rolling bearing stable operating.In this work,a feature extraction method combining fuzzy entropy and fractal dimension is proposed to extract faulty features in early stage for vibration signals.Using the Salp Swarm Algorithm optimized Support Vector Machine(SSA-SVM)to construct the classification model.With historical database of the rolling bearing vibration signal,the model can be trained and tested,then the incipient faults could be detected.The experiment results showed that this method can effectively detect the failures in early stage of rolling bearings.