基于飞蛾扑火优化的MCKD算法在往复压缩机轴承故障诊断中的应用

Application of MCKD Algorithm Based on Moth-flame Optimization in Fault Diagnosis of Reciprocating Compressor Bearing

往复压缩机作为一种高噪声设备,采集到的轴承振动信号存在大量噪声干扰,针对这种具有非线性、非平稳性的信号,提出了一种基于最大相关峭度解卷积的往复压缩机轴承故障诊断方法。使用飞蛾扑火优化算法对最大相关峭度解卷积的参数周期T和滤波器长度L进行自适应选择,用参数优化后的最大相关峭度解卷积算法对往复压缩机轴承振动信号进行解卷积处理,从而有效地提取轴承振动信号中的冲击成分。利用多尺度熵对解卷积后的信号进行量化分析,用熵值来构建表征轴承故障的特征向量,并将特征向量输入到极限学习机对故障特征进行分类识别。最后,用实测数据验证了该方法的有效性。

Reciprocating compressor is a kind of high-noise equipment,and the collected bearing vibration signal has a lot of noise interference.Aiming at this nonlinear and non-stationary signal,a fault diagnosis method for reciprocating compressor bearing based on maximum correlation kurtosis deconvolution is proposed.The parameter period T and filter length L of the maximum correlation kurtosis deconvolution were selected adaptively by the moth-flame optimization algorithm,and the maximum correlation kurtosis deconvolution algorithm with optimized parameters was used to deconvolve the vibration signals of reciprocating compressor bearing,so as to extract the impact components in the vibration signals of bearings effectively.The multiscale entropy is used to quantitatively analyze the deconvolved signal,and the entropy value is used to construct the feature vector that characterizes the bearing fault,and the feature vector is input to the extreme learning machine to classify and identify the fault feature.Finally,the effectiveness of the method is verified by the measured data.