Noise Souce Identification Of I.C.Engine Based On Independent Component Analysis And Wavelet Transform Technology

With the increasing of urban vehicles, the noise pollution from motor vehicles has become increasingly serious one of the major environmental problems influencing people's living and working. The noise from internal combustion engine is the main noise source, so the noise control research has become a hot research problem at home and abroad. Due to the complex structure, internal combustion engine vibration signal is a kind of non-stationary periodic signals. It is difficult to find the largest radiation sites of the internal combustion engine noise by general methods. Therefore, how to accurate and fast identify the internal combustion engine noise is the important prerequisite of internal combustion engine noise control. The traditional internal combustion engine noise source identification methods of application have certain limitation, so relevant research and application were constrained.Aiming at the characteristics of internal combustion engine radiation noise signal, the paper researches modern testing technology and signal analysis method using the noise test technology, independent component analysis, and the wavelet transform technology and blind separation uncertainty eliminate. Separation and identify of the main noise source of internal combustion engine are calculated from the complex noise signals, and the noise contribution of each independent component to the whole machine are calculated. The main research contents are as follows:1. Firstly introduces the internal combustion engine noise research developing situation at home and abroad, and provides theoretical basis for separation and identification methods of sound sources selection and related technical improvement. The characteristics of the internal combustion engine and its main noise sources of generating mechanism are systematic analyzed based on literature retrieval. The internal combustion engine noise source and the main incentive the flow path in internal combustion engine inside the structure are revealed as knowledge preparation to accurately identify noise source and control noise effectively.2. Introduces the principle of independent component analysis method and its application in engineering. Taking a six cylinder diesel engines as the research object, the noise signal under different working conditions statistical independence and Gaussian analysis has been done. The results showed that the noise signals meet the basic ICA condition. The measured noise signal is decomposed into a series of independent component based on negative entropy greatly fast independent component analysis (FastICA) algorithm.3. Using Fourier transform and wavelet transform technology to each component feature analysis to identify the independent component and the corresponding relation of internal combustion engine noise. Combining time-frequency analysis results and prior knowledge, the main independent component was determined.4. Based on the fast Fourier change and the spectrum similarity ICA uncertainty elimination method eliminates ICA (BSS) estimates uncertainty and realized the noise signal waveform recovery.5. The sound power of independent component noise signals and the sound power contribution to the radiation noise of the whole engine are calculated. The main noise source control measures are put forward to lay a foundation for further control the machine noise.