基于声发射信号的蜗轮磨损程度在线监测

In-situ monitoring of worm gear wear using acoustic emission signal

为实现对蜗轮蜗杆减速器工作过程中蜗轮磨损程度的精确监测,利用多通道声发射检测仪器对不同磨损程度的蜗轮声发射信号进行在线采集。采用小波分析方法对信号进行去噪处理,提取声发射特征信号值,根据最小模糊熵优化模型构造出不同磨损程度蜗轮的模糊隶属度函数。采用ANFIS多维模糊神经网络实现多通道声发射信号的决策融合,提高了蜗轮磨损程度识别结果的准确性。通过对随机磨损程度的蜗轮进行实际验证,实验结果验证了系统的有效性和可靠性。

In order to accurately monitor the wear degree of worm gear in the working processing of worm gear reducer, acoustic emission signals of worm gear with different wear degrees were collected online by the multi-channel acoustic emission detecting instrument. The signals were denoised by wavelet analysis method, and the eigenvalues of acoustic emission signals were extracted. According to the minimum ambiguity optimization model, the fuzzy membership function of worm gear with different wear degrees was constructed. The decision fusion of multi-channel acoustic emission signals was realized by using ANFIS multi-dimension fuzzy neural network, which can improve the accuracy of the identification results of worm gear wear. Experimental results validate the effectiveness and reliability of the worm gears with random wear degrees.