Research On The Characteristic Of Vibration And Acoustic Signals From I.C.Engine Based On Wavelet Transform And Independent Component Analysis

The study on noise and vibration of internal combustion engine were paid more attention recently by internal researchers for the improving requirement of vehicles' NVH (noise vibration and harshness) performance. .Finite element method (FEA), Boundary element method (BEM) and multi-body dynamic analysis (MSS) were applied to predict and solve the engine vibration and the noise emission problem, and these methods had been proved to be more efficient then the traditional ways. On the other hand, the identifications of the main noise source and the research on mechanism of engine noise emission are also critical for both noise control engineer and fault diagnose. The vibration and noise emission from engine is complicated for there are too many internal excitations of engine, and the excitations are also time variable. The traditional analysis methods of acoustic and vibration signals from I.C.Engine show their limitation in describing the complexity.The digital signal processing (DSP) technique had got great achievement in recent twenty years. And the achievement of digital signal processing was continuously applied to analyze the vibration and noise signals from machines. These researches also enhance the ability of noise source identification and fault diagnosis of I.C.Engine.Base on present research foundation from worldwide, the application of sound intensity measurement, wavelet transform and independent component analysis to the research of vibration and noise signals from I.C.Engine was deeply investigated here:1. The mechanism of noise emission for engine was studied from the following aspects: the main excitations as important factors that caused the noise radiation; the path of vibration transmission in the internal parts of the engine .the noise and vibration signal was no-stationary induced by both the time-variable excitations and time-dependent transmission path. The methods of noise identification of internal combustion engine were reviewed.2. The physical meaning of sound intensity and its application in noise source field analysis were discussed. Based on Fourier transform, the characteristic of acoustic signals in frequency domain from WD415 diesel engine, ZH1130 diesel engine and a motorcycle was studied by sound intensity measurement. The conclusion from the sound intensity measurement was proved to be reliable in the succeeding noise control engineering.3 Compared with Fourier transform, the advantage of wavelet transform in non-stationary signal processing was discussed. And Complex Morlet was induced here to process the acoustic signals from WD415 diesel and ZH1130 diesel engine. The results revealed signal's time-frequency property and the spatial distribution. Averaged scale—wavelet power spectrum and Euclidean distance here were used here to identify the main noise sources of ZH1130 diesel engine. The method to study the acoustic characteristic based on the wavelet transform of instantaneous sound intensity was proposed here. This method was applied to analyze the acoustic emission from WD415 diesel engine.4 Independent component analysis was applied to analyze the vibration and acoustic signals from diesel engine. By the FactICA algorithms, the raw vibration and acoustic signals were decomposed to different components. By wavelet transform and correlation analysis, the main noise sources were identified.5 The relationship between the combustion component of vibration, acoustic signals and internal combustion pressure was modeled by ARMA. And by the relationship, the internal pressure of ZH1130 diesel engine was reconstructed successfully.