基于内燃机振动信号的可视化识别诊断

Visual identification and diagnosis based on IC engine vibration signals

为提高故障识别诊断的精确度和实时性,有效解决内燃机多分量、非平稳振动信号特征提取困难的问题,提出一种基于改进局部二值模式(ILBP)与双向二维主成分分析(TD-2DPCA)的内燃机振动信号可视化故障识别诊断方法。针对传统时频方法在分析内燃机振动信号中,存在时频分辨率低及交叉干扰项的问题,将经验小波变换(EWT)与同步压缩小波变换(SST)应用到内燃机振动信号的时频图表征中;利用ILBP提取图像的纹理特征,并对ILBP图谱采用TD-2DPCA降维,将降维后的编码矩阵向量化后得到图像的特征参数;通过支持向量机(SVM)和最近邻分类器(NNC)分别特征向量进行训练、测试,实现内燃机的故障识别诊断。在内燃机气门间隙故障8种工况下缸盖振动信号的识别诊断试验中,均得到较高的分类精度;通过参数的合理优化,在保证了分类速率的同时,最高识别率达到96.67%,对比其他方法,充分表明该方法在内燃机故障诊断中的有效性。

In order to improve the accuracy and real-time of IC engine fault identification and diagnosis,and solve the difficulty of feature extraction for IC engine multi-component and non-stationary vibration signals effectively,a visual fault diagnosis method for IC engine vibration signals based on improved local binary pattern(ILBP)and two directional-two dimensional principal component analysis(TD-2DPCA)was proposed.Firstly,aiming at the problem of low time-frequency resolution and cross-interference in the analysis of IC engine vibration signals by the traditional method,the empirical wavelet transform(EWT)and synchro-squeezing wavelet transform(SST)were applied to the time-frequency representation of IC engine vibration signals.Secondly,the texture feature of the image was extracted by ILBP and TD-2DPCA was used to reduce the dimension of an ILBP texture image.The feature parameters of the image were obtained by vectorizing the coding matrix.Last,the feature vectors were trained and tested by support vector machine(SVM)and nearest neighbor classifier(NNC)respectively to realize the fault diagnosis of the IC engine.In the identification and diagnosis test of cylinder head vibration signals under 8 working conditions of the IC engine,higher classification accuracy was obtained.Through reasonable optimization of the parameters,the classification rate is guaranteed and the highest recognition rate reaches 96.67%.Compared with other methods,the effectiveness of this method in IC engine fault diagnosis was fully demonstrated.