| With the increasing demands of the vehicle noise performance,the sound pressure level of the vehicle sound control is difficult to meet the needs of users.Especially,some uncomfortable noise with low sound pressure level is difficult to find and improve.As the engine noise is controlled better,the noise of gearbox become the main source of interior noise of heavy commercial vehicles when it is in idle sliding state.Therefore,in this paper,the sound quality research methods are used to evaluate and predict the sound quality of heavy commercial vehicle gearbox.Firstly,the current research situations and achievements about sound quality were analyzed systematically.In this paper,the methods were applied to analyze the noise of heavy gearbox.And it is proved that the method is necessary and reasonable.Secondly,the noise signals of heavy gearbox at different gears,rotational speeds and locations under no-load conditions were measured in the anechoic chamber.And then,the signals were the monitored and played back to screen the steady and undisturbed parts.According to these signals,the objective parameters of sound quality were calculated theoretically.Then,the objective parameters were analyzed by correlation analysis and principal component analysis in order to establish the annoyance model.It has been proved that the annoyance model can better meet the subjective feelings of human.At the same time,it could better evaluate the sound quality of heavy commercial vehicle gearbox.The result provided that the sound quality can be useful for the fault diagnosis of gearbox.Then,the subjective evaluation test of noise samples was carried out with the grouped pair-wise comparison.And the unstable evaluation results were eliminated by checking the consistency of weights.The average subjective evaluation values of all the evaluators are calculated and analyzed.The correlation between subjective evaluation values and objective parameters was also analyzed.Finally,the prediction model between subjective evaluation value and objective parameters was established.Multivariate linear regression model,generalized neural network(GRNN)and particle swarm optimization-generalized neural network(PSO-GRNN)were used to predict the sound quality of heavy transmission.The prediction results showed that the PSO-GRNN model could improve the prediction accuracy and stability which provide a reference for the sound quality design and research engineering application of heavy vehicle gearbox. |
