Aiming at the difficulty of extracting the characteristic of vibration signal of internal combustion engine, internal combustion engine fault diagnosis problem is transformed into image recognition problem. A novel internal combustion engine visualization fault diagnosis method based on recurrence plots (RP) and improved local binary pattern (ILBP) is proposed. To characterize the internal combustion engine signal in different fault states, RP analysis method is used in the processing of the cylinder head vibration signal. Then the ILBPcoding method is used to extract the texture characteristics of internal combustion engine RP images. The ILBP coding gray orthogonal sequence of RP is used as characteristic parameters, and the support vector machine (SVM) is used to recognize the fault pattern recognition. Through the analysis of four different status of IC engine valve fault signal, a high fault recognition rate is arrived, which shows that this method has strong ability of fault feature description, which can be used for accurate diagnosis of IC engine valve fault.%针对内燃机振动信号特征提取困难的问题,将内燃机故障诊断问题转化为图像的识别问题,提出一种基于递归图(RP)和改进局部二值模式(ILBP)的内燃机可视化故障诊断方法.将递归图分析方法引入内燃机缸盖振动信号的处理中,用以表征内燃机不同故障状态信号;然后对局部二值模式(LBP)的编码方式进行了改进,利用改进后的ILBP算子提取内燃机递归图的纹理特征,将ILBP编码图谱的灰度直方序列作为特征参数,利用支持向量机(SVM)对故障进行模式识别.在4种不同气门状态的内燃机故障诊断实验中,故障识别精度高.该方法利用递归图代替振动谱图像,突破了传统时频分析方法的思路,递归图的纹理特征具有较强的故障特征描述能力,结合递归图与ILBP的方法可用于准确诊断内燃机气门故障.
展开▼
