| With the improvement of NVH technology,the noise evaluation of diesel engine based on sound pressure level and sound power level can not fully meet the market demand,and the sound quality has attracted more and more attention.Compared with the sound quality of steady-state condition,the research on the sound quality of unsteady condition noise(such as acceleration noise)is more important for the traditional power commercial vehicle.Based on an in-depth understanding of the current research status of sound quality at home and abroad,the thesis carried out the subjective and objective evaluation of the sound quality of diesel engines.A diesel engine accelerated noise test was performed and a noise sample database was established.The subjective preference evaluation of accelerated noise was completed by organizing the hearing test,and a new objective evaluation index of diesel engine sound quality was put forward.The multivariate regression analysis method and BP neural network method were used to establish the accelerated sound quality prediction model of diesel engine.The accuracy verification and model comparison analysis were completed.The factors affecting the accelerated sound quality of diesel engine were analyzed.Accelerated noise acquisition test of commercial vehicle diesel engine was carried out.According to national standards,the arrangement and selection of test environment and test equipment are completed,and the appropriate test conditions were designed.The sound samples were screened,segmented and analyzed,and 39 diesel engine acceleration noise samples with a length of 5 seconds were obtained,which provides data support for the subjective and objective evaluation of sound quality.The subjective evaluation test of diesel engine acceleration noise was carried out,and the problems about selection of evaluators,design of evaluation experiments and screening of evaluation results were solved.One sound sample with a coincidence degree less than 80% was eliminated and 38 acceleration noise samples with an average coincidence degree of 92.699% were left.A new two-level objective evaluation system was proposed.The average value of traditional objective evaluation parameters such as loudness and sharpness are used as the first evaluation index,and the change rate of parameters with time are used as the second evaluation index.Finally,the evaluation index system is constructed.The prediction model of accelerated sound quality of diesel engine was established by multivariate linear regression method.The multiple linear regression model was solved by Matlab programming and the model was verified.The result shows that the accuracy of the multivariate linear regression prediction model can meet the requirements.The BP neural network prediction model was built by using Matlab,and the training and accuracy verification of the model were completed.Comparing two models,the multivariate linear regression model is helpful to analyze the mapping relationship between subjective and objective results,and the BP neural network model is more suitable for the prediction of accelerated sound quality of diesel engine.The better prediction results show that the two-level objective evaluation index is suitable for the prediction of accelerated sound quality of diesel engine,and it has practical application value.The weights of objective indicators were analyzed.The results show that besides loudness,loudness change rate,roughness change rate,fluctuation strength and articulation index also have a great influence on the acceleration sound quality.The factors affecting the accelerating sound quality are analyzed.It is found that the increase of speed and torque will lead to the deterioration of the sound quality of the diesel engine.At low torque,the slower acceleration of the diesel engine will have better sound quality.In the case of medium-high torque,medium-high rotational speed,diesel engine acceleration faster sound is more popular with reviewers. |
PDF Full Text Request