【摘要】 为了充分提取基于内燃机振动信号形成的振动谱时频图像的二维时频信息,实现基于内燃机振动谱时频图像特征自动提取及识别,提出了一种基于EMD-WVD(EMD-Wign-er-Ville Distributions)振动谱时频图像SVM识别的内燃机故障诊断方法。该方法利用二进制小波对振动信号进行预处理,然后利用EMD-Wigner-Ville时频分布生成不同工况下振动信号的时频图像,并通过提取振动信号的EMD-WVD振动谱时频图像的不变矩特征形成诊断特征向量,利用一种基于类识别率排序的二叉树SVM分类器进行模式识别。在BF4L1011F型内燃机上进行了6种不同工况下气门故障模拟试验,诊断结果表明总体诊断正确率为98.57%。更多还原

【Abstract】 In order to realize automatic extraction and recognition of internal combustion engine vibration spectrum time-frequency image feature,a new fault diagnosis method based on EMD-WVD(EMD-Wigner-Ville Distribution) vibration spectrum time-frequency image recognition by SVM was proposed.In the method,the engine vibration signals were decomposed using the binary wavelet and the vibration signal time-frequency images under various operation conditions were generated from the EMD-Wigner-Ville time-frequency distributions.Through extracting moment invariant feature of the images,the diagnosis eigenvectors were achieved and their modes were recognized by an improved binary tree SUM classifier.Experiment results in model BF4L1011F engine indicate that all valve faults under six conditions can be exactly recognized,the accuracy reaches 98.57%.更多还原