基于多传感数据融合的变速运行齿轮异常振动故障诊断

变速运行齿轮异常振动故障诊断性能过差会增加汽车维护成本，缩短齿轮使用寿命。为了及时识别齿轮故障，保证汽车变速器总成具有良好的振动特性，提出基于多传感数据融合的变速运行齿轮异常振动故障诊断方法。通过分析多传感器数据融合技术，掌握变速运行齿轮异常振动故障诊断的理论框架，并以此为基础，参考传感器融合模块、特征级并行多神经网络局部诊断模块和终端分类模块，结合变分模态分解、多通道加权融合和单隐层前馈神经网络训练算法，从信号采集、信号特征提取和信号特征分类3个步骤实现变速运行齿轮异常振动故障诊断。实验结果表明：在齿轮发生轻度磨损时，磨损振动信号的幅值在20~40 mV之间，磨损振动信号的频率在0~4 000 Hz区间；中度磨损时，信号的幅值在30~55 mV之间，信号频率在3 000~7 000 Hz区间；重度磨损时，信号幅值在50~70 mV之间，信号频率在6 000~12 000 Hz区间，且各阶段诊断结果均与故障程度的实际转折点吻合。由此可知在各样本数量均相同的情况下，提出的故障诊断方法预测值与真实值均相同，故障程度和故障类型的诊断性能均较好。

Fault Diagnosis of Abnormal Vibration of Variable Speed Running Gear Based on Multi-sensor Data Fusion

The poor diagnosis performance of abnormal vibration fault of variable speed running gear will increase the vehicle maintenance cost and shorten the service life of the gear.In order to identify the gear fault in time and ensure that the vehicle transmission assembly has good vibration characteristics,a method for diagnosing abnormal vibration fault of variable speed running gear based on multi-sensor data fusion was proposed.The multi-sensor data fusion technology was analyzed,the theoretical framework of abnormal vibration fault diagnosis of variable speed running gear was grasped.On this basis,referring to the sensor fusion module,feature level parallel multi neural network local diagnosis module and terminal classification module,and combining the variational mode decomposition,multi-channel weighted fusion and single hidden layer feedforward neural network training algorithm,the fault diagnosis of abnormal vibration of variable speed running gear was realized from signal acquisition,signal feature extraction and signal feature classification.The experimental results show that when the gear is slightly worn,the amplitude of wear vibration signal is from 20 mV to 40 mV,and the frequency of wear vibration signal is from 0 to 4 000 Hz; in case of moderate wear,the signal amplitude is from 30 mV to 55 mV and the signal frequency is from 3 000 Hz to 7 000 Hz; in case of severe wear,the signal amplitude is from 50 mV to 70 mV,the signal frequency is from 6 000 Hz to 12 000 Hz,and the diagnosis results at each stage are consistent with the actual turning point of the fault degree.It can be seen that in the case of the same number of samples,the predicted value and the true value of the proposed fault diagnosis method are the same,and the diagnosis performance to fault degree and fault type is good.