Radiated Noise Source Identification And Weight Analysis Of Diesel Engine

Noise has become one of the major pollution problems affecting people's lives and work.There are a large number of complaints about traffic noise in urban noise complaints,and vehicle noise accounts for the biggest proportion of traffic noise.As Internal combustion engine has become the main noise source of vehicle,it would be an important measure to reduce traffic noise and improve the living environment of the city by controlling the noise of the internal combustion engine.However,the sources of internal combustion engine noise are numerous and complex,and controlling such noise will become extremely difficult.Research on the identification method of noise sources of the internal combustion engine and the analysis of the noise source characteristics of the internal combustion engine would not only have scientific significance for the research of engineering acoustics,but also have practical application value for the noise control of internal combustion engines.The main sources and mechanisms of the internal combustion engine noise are analyzed systematically in this dissertation.It was clear that the basis of the noise control of the internal combustion engine was to accurately identify the main noise sources,and to analyze and determine their weights,so as to prepare for specific noise optimization control schemes.By analyzing and comparing different noise source identification methods,it was found that the noise source identification technology based on signal processing algorithm has obvious advantages in cost and recognition accuracy.In order to find a signal processing algorithm suitable for the identification of internal combustion engine noise sources,we studied existing algorithms,and it was found that the variational mode decomposition algorithm has obvious advantages compared with the traditional algorithm in analyzing the unsteady impact signal.But there are still problems such as difficulty in selecting the number of decomposition layers and failure to select the second penalty in actual use with this algorithm.This dissertation proposes an algorithm optimization path based on the reconstructed signal energy ratio to determine the number of decomposition layers,the optimal quadratic penalty based on the maximum margin factor,and the initial center frequency based on the spectral peak.Taking the simulation signal as the verification case,it was found that the proposed optimization algorithm was superior to the traditional variational mode decomposition algorithm both in decomposition speed and decomposition accuracy.In order to identify noise sources,the vibration and noise experiments of the diesel engine were carried out.The surface radiation noise of the diesel engine and the surface vibration of the components were also collected.The noise signal was first preprocessed,then the variational mode decomposition algorithm was used to decompose the noise signal.The main noise component was filtered by mutual information,and the noise sources of the internal combustion engine was separated by robust independent component analysis.The wavelet transform performs time-frequency analysis on the signal to initially determine the noise source.Finally,the results of noise sources identification were verified by the reverse drag test and the near-field noise test.The combustion noise,piston knock noise and air compressor noise were successfully separated and identified.Finally,in order to determine the noise weight of the diesel engine surface components,the frequency bands of the collected noise signals are divided.After that,the coherent power spectrum analysis of each noise component as well as vibration of engine parts was carried out to obtain the noise contribution of different components to varied frequency bands.Then,through the analytic hierarchy process,a pairwise comparison judgment matrix was established,and the contribution of the surface parts to the noise of the diesel engine was obtained.It is found that the oil sump of the parts has the largest proportion of noise.These results can be used to help the noise control of the whole machine.