图像处理在内燃机故障诊断中的应用研究

An Application of Image Processing to Fault Diagnosis for Internal Combustion Engines

在应用图像处理进行故障诊断的基础上 ,探讨了应用这一技术进行内燃机故障诊断与状态监测时应注意的一些问题。指出利用缸盖振动信号小波包分解后的时 -频分布图的灰度直方图进行故障诊断的效果并不好 ,并分析了原因 ;进行了信号加噪声与不加噪声的诊断效果对比 ,发现噪声对基于图像处理的气阀机构故障诊断的影响不大 。

Based on image processing, we present an advance d fault diagnosis approach for internal combustion engines. The cylinder head vi bration signals were decomposed with wavelet packet algorithm and the time-freq uency images of the vibration signals were obtained. Features of the vibration s ignals were extracted in the time-frequency images. The fault diagnosis model based on the time-frequency image's matching can effectively carry out diagnosi s for the valve train on a four-stroke and six-cylinder diesel engine. It is n ot effective to diagnose with gray histograms of the time-frequency images and the reason is given. The deviation of vibration signal's phase will lead to dete rioration of diagnosis. Therefore attention must be paid to the phase informatio n of the vibration signal when one uses the diagnosis model. Comparisons of diag nosis are made between noise-added signals and original signals. There is littl e effect on the diagnosing after signals were added with noise, so this model is good for engineering application.