基于非高斯AR模型双谱估计的发动机气门机构故障诊断

Fault Diagnosis for Valve Train of Diesel Engine Based on Bispectrum Estimation via Non-Gaussian AR Model

将高阶谱引入故障诊断领域,利用基于非高斯AR模型的双谱估计方法描述了缸盖振动响应信号的非线性和非高斯特性。基于非高斯AR模型的双谱估计,能够在数据较短的情况下,提供较高的频率分辨率且能提取信号的相位信息,可用于非平稳信号的数据处理和故障诊断。研究结果表明:当柴油机气门机构处于不同工作状态时,缸盖振动信号的双谱及其等高线具有明显的可分性。对双谱进行数据降维处理,从其主对角切片中提取了比较稳定的故障特征,为气门机构的状态监测与故障诊断提供了一种有效手段。

The theory of higher-ordered statistics was applied to the vibration diagnosis of diesel engine, and the nonlinear behavior of cylinder head in different conditions was described by the bispectrum estimation based on non-gaussian AR model. The results show that the bispectrum of cylinder head vibration signals and its contour line has an obvious difference, thus the bispectrum can be used to process and diagnose the nonlinear signals. The diagnosis feature can be extracted from the main cross slice through dimension descent of bispectrum, which can provide an effective method for the condition monitoring and fault diagnosis of diesel engine valve train.