| As a kind of emergency power generation equipment,mobile diesel generator set provides great convenience for production and life,but the noise brought by its work has caused serious harm to people’s physical and mental health.Therefore,it is of great significance and practical application value to analyze the noise of the mobile diesel generator set and study the methods to effectively separate noise signals and identify the main noise sources of the mobile diesel generator set,so as to improve the noise of the whole machine and enhance the product performance and market competitiveness.In this paper,a mobile diesel generator set produced by an enterprise is taken as the experimental research object,and the whole noise test and noise characteristics analysis of the unit are carried out.At the same time,the noise source identification technology based on signal analysis method is studied,and EWT-FastICA noise signal blind source separation algorithm is proposed,and CWT is used to identify the noise source of the unit,which provides experimental and theoretical basis for the noise control of the unit.A mobile diesel generator noise analysis system is developed to make noise analysis more convenient.The main research contents of this paper are as follows:(1)Collect the original noise data of the test unit and conduct noise analysis based on the sound pressure level evaluation index.The original noise data of the test unit are analyzed and studied from the aspects of near-field noise sound pressure level,sound insulation volume of exhaust muffler end sound shield and noise spectrum characteristics of the unit.The radiation azimuth plane of the main noise source of the test unit is determined and the frequency components which contribute more to the noise of the whole unit are obtained.Based on the structure and test data of the acoustic shield,the improvement scheme of the acoustic shield is proposed.(2)Study and select the optimal wavelet base and test data analysis based on the optimal wavelet base function.Through simulation analysis,it is verified that wavelet transform has good time-frequency analysis characteristics,and cmorlet3-2 with high analysis accuracy is used for experimental analysis and verification of the test noise signal.The results show that it can be used as the optimal basis function for subsequent time-frequency analysis.(3)The EWT-ICA joint analysis method is proposed and noise source identification is carried out based on wavelet transform.The advantages and disadvantages of EMD,VMD and EWT in signal decomposition were analyzed and verified by the experimental data.The results show that EWT has good anti-aliasing ability.A combined method of EWT-ICA analysis is proposed to satisfy the precondition of ICA analysis.The Gaussian analysis of the test noise signals shows that they all conform to the super Gaussian distribution.EWT-ICA method was used to isolate 6 independent components,and CWT was used to identify the noise source of each component.The analysis results can provide data support for vibration and noise reduction of the unit.(4)The noise analysis system of mobile diesel generator set is developed.Matlab is used to carry out modular programming of the correlation analysis method in the paper,and it is integrated into the noise analysis system,which makes the test data analysis more convenient,and also provides help for noise analysis of other models of units. |
