自适应MCKD结合Autogram的矿用滚动轴承故障特征提取

Fault Feature Extraction of Mining Rolling Bearing Based on PSO-MCKD-Autogram

为辨析矿用滚动轴承运行状态,有效地提取矿用滚动轴承故障特征,提出了一种基于粒子群优化算法(Particle Swarm Optimization,PSO)的参数自适应优化最大相关峭度解卷积算法(Maximum Correlated Kurtosis Deconvolu-tion,MCKD)与自相关谱峭度图法(Autogram)相结合的矿用滚动轴承故障特征提取算法.首先,在考虑振动信号的强周期性的基础上,采用MCKD对原始信号进行预处理以实现信号的降噪与增强;同时,针对MCKD参数选择问题,构造利用PSO对适应度函数进行寻优得到合适的参数组合[滤波长度L,解卷积周期T];此后,利用Autogram对处理后信号进行特征提取;最后,通过仿真信号及公开数据集试验信号对该算法进行验证.结果表明:PSO-MCKD-Autogram算法能够较好地克服噪声影响,可有效提取矿用滚动轴承故障特征且具有一定的鲁棒性.

In order to analyze the operation state of mining rolling bearings and effectively extract the fault characteristics of mining rolling bearings,a parameter adaptive optimization Maximum Correlated Kurtosis De-convolution(MCKD)combined with Autogram is proposed based on Particle Swarm Optimization(PSO).MCKD combined with Autogram as a fault feature extraction algorithm for mining bearings.Firstly,based on the strong periodicity of the vibration signal,MCKD is used to preprocess the original signal to realize the noise reduction and enhancement of the signal;At the same time,in view of the MCKD parameter selection problem,PSO is constructed to optimize the fitness function to obtain the suitable parameter combination[filter length L,deconvolution period T];Thereafter,Autogram is used to extract the features of the processed signal.Finally,the algorithm is validated by simulation signals and experimental signals from public datasets.The results show that the PSO-MCKD-Autogram algorithm can better overcome the influence of noise,and can effectively extract the fault features of mining bearings with certain robustness.The results can provide theoretical basis for condition monitoring and fault analysis of rolling bearings in mining.