Research On Noise Source Identification Of Diesel Engine Based On Wavelet Analysis And Psychoacoustical Evaluation

Auto industry has become a huge industry through the hundred years of development, which is an important pillar of global economy. Thus, more and more attentions have been paid to the traffic noise. The Internal Combustion Engine is one of the vehicle's main sources of vibration excitation and noise radiation. The issue of noise control has become the focus of the research area of power machinery. Meanwhile, with the improvement of international emission standards and the increase of the customer demands, the issue of the sound quality has occupied the important position of power machinery. It not only requires a low noise power of the acoustic environment, but also wants a high quality environment.In this paper, the Independent Component Analysis theory was used to separate the noise source of the diesel engine.The method of Fourier transform and wavelet transform were used to extract the feature of the noise source signal of diesel engine. The objective analysis of the noise source signal was compeleted by the psychoacoustics method, which was verified by the subjective evaluation later. Then the noise control strategy of diesel engine was proposed, which was met the people's subjective hearing. The contents of this paper are as follows:First, multi-condition noise signals of a six-cylinder diesel engine were acquired through experimental method, one of which was used to the research. Analysing and verifing whether the noise signals satisfied the Gauss distribution by using Gauss theory. If the result meet the Non-Gauss distribution, the blind source separation of noise signal would use the Independent Component Analysis, and the noise source would be identified by the Fourier transform and wavelet transform. Eventually three independent noise source signals were identified, which were combustion noise, air compressor noise and piston knock noise. Then the validity of the result of the noise source identification was verified by the spectrum analysis and time-frequency analysis of the noise signals from the near-field experiment. Finally, the noise source signals'waveform was recoveried by the method of FFT-MCC.Second, the psychoacoustics parameters and satisfaction factor of the independent noise source signals were calculated and analysed according to the psychoacoustics theory and Zwicker satisfaction model. The result was that the value of the satisfaction of the combustion noise was much higher than the air compressor noise and piston knocking noise. Meanwhile, the effects of each independent noise source on the overall engine noise were also studied. The noise signals containing of rich of frequency components were processed by filtering, and then 8 different noise samples were obtained. Through calculate and analysis the psychoacoustics parameters and satisfaction factor of the 8 sound samples, the effect of each independent noise source and each psychoacoustics parameter on the overall noise quality was obtained. Form above analysis, the noise optimized measures were proposed for the test diesel engine.Finally, 8 different noise samples were choosen for the test of the subjective evaluation by using the A/B comparison method and the professional jury. The validity of the experimental data was verified by the statistical method. The result of the subjective evaluation was in a good agreement with the result calculated by the Zwicker satisfaction model, which further confirmed the validity of diesel sound quality optimized strategy based on psychoacoustics.