基于EMD-AR与改进宽度学习系统的滚动轴承故障诊断方法

Fault Diagnosis Method of a Rolling Bearing on EMD-AR and Improved Broad Learning System

滚动轴承内、外圈和滚动体的故障振动信号通常伴随噪声,并且频率不尽相同,同时具有非线性、非平稳性等特点,传统的滚动轴承故障诊断方法效率较低,且大多未考虑变负载、噪声情况。针对上述问题,提出基于经验模态分解自回归模型和改进的宽度学习系统(broad learning system,BLS)的滚动轴承故障诊断方法。首先,对滚动轴承振动信号进行经验模态分解(empirical mode decomposition,EMD),得到一组固有模态函数(intrinsic mode function,IMF);对每个IMF建立自回归(autoregressive model,AR)模型,求得AR模型参数和残差余项,以此作为各类状态信号的特征矩阵;将该特征矩阵输入至改进的BLS,判断滚动轴承的故障位置及故障尺寸,实现滚动轴承的故障诊断与定位;并将该方法应用于变负载、有噪声工况下滚动轴承故障诊断与定位。同时设计2种对比方法,并将所提方法与两种对比方法的预测结果进行比较,实验结果表明所提方法诊断识别能力更强,为变负载、有噪声工况下的滚动轴承故障诊断与定位提供参考。

The vibration signal of rolling bearing inner ring,outer ring and rolling body failure is usually accompanied by noise,and the frequency is not the same,and has the characteristics of non-linearity,non-smoothness,etc.The traditional rolling bearing fault diagnosis method is less efficient,and most of them do not consider the variable load and noise situation.In response to the above problems,a rolling bearing fault diagnosis method based on EMD-AR and improved broad learning system(BLS)is proposed.First,empirical mode decomposition(EMD)is performed on the rolling bearing vibration signal to obtain a set of intrinsic mode functions(IMF);an autoregressive model is developed for each IMF,and the AR model parameters and residual variance are obtained as the feature matrix for each type of state signal;inputting this feature matrix to the improved BLS to determine the fault location and fault size of rolling bearings and to realize the fault diagnosis classification of rolling bearings;the method is also applied to the diagnosis and location of rolling bearing faults under variable load and noisy conditions.Two comparison methods are designed at the same time,and then the prediction results of the two contrasting ways with the method we devised are compared.The experimental results show that the proposed method has better diagnostic recognition ability,which provides reference for the diagnosis and location of rolling bearing faults under variable load and noisy conditions.