Analysis of Engine Sound by using Wavelet Transform

本研究将小波分析应用于四冲程发动机及其驱动系统的监测，首先，我们在正常情况下空载驱动发动机，测量发动机发出的声音和发动机缸体顶部的加速度。然后，我们将这些信号进行小波变换，并将这些信号与每个工作循环（进气、压缩、点火、膨胀、排气）相关联。(i)缺火，（ii）大间隙，（iii）凸轮链熔化。在小波分析的计算中，我们发现正常的发动机信号和异常的发动机信号之间存在着显著的差异。最后，我们得到了以下结论，结果表明，我们可以利用小波分析来分析发动机的声音和发动机的加速度，从而检测发动机的运行状态和发生的异常。(1)通过小波分析计算出声音和振动的各频率分量随时间的变化，找出发动机的一个冲程（爆炸、进气门或排气门关闭运动）和一个振动部位。(2)基于上述（1），我们可以通过寻找振动发生的幅度（峰值幅度）和振动发生的定时来检测异常。

In this study, we applied Wavelet analysis to monitor a four-cycle engine and its driving system.First, we drove an engine in normal condition with no load and measured sound emitted from engine and acceleration on the top of the engine block. Then we transformed these signals by Wavelet and related these to each operating cycle (intake, compression, gnition, expansion, exhaust).Nex, we set up an engine in abnormal condition, i.e., (i) miss fire, (ii) larget tappet clearance, (iii) melt of cam chain. In the calculation of Wavelet analysis, we found a significant difference between normal engine signal and abnormal one.Finally, we got following conclusions as a result that we analyzed sound emitted from an engine and acceleration of an engine by using Wavelet analysis to detect operating condition of engine and abnormality occurring to an engine.(1) We calculate change of each frequency component of sound and vibration with time by Wavelet analysis and find a stroke of engine (explosion, closing motion of intake or exhaust valve) and a vibrating part.(2) Based on above-mentioned (1), we can detect abnormalities by seek magnitudes of occurring vibration (peak magnitude) and the timing when vibrations occur.