Vehicle cab sound quality evaluation method of PSO optimization delay operator model

Abstract

The invention relates to a vehicle cab sound quality evaluation method of a PSO optimization delay operator model, the PSO algorithm with the global optimization capability is adopted, so that the initial weight and the threshold value of the PSO algorithm optimization network are within a small range at the beginning, the subsequent training cost is reduced, and the function of improving the evaluation precision of the sound quality is achieved. A sound quality prediction model containing the delay operator has the advantages that the convergence speed is high, the training time is short, historical learning samples can be effectively utilized, the accuracy of the evaluation model is high, and the occurrence of local minimum values can be effectively inhibited. According to the method, the problems that the training time complexity of a traditional sound quality evaluation model is large, and the sound quality prediction precision is low due to the fact that a certain minimum value cannot be automatically found are solved.

Description

technical field [0001] The invention belongs to the technical field of fuel cell vehicle cab noise control, and in particular relates to a vehicle cab sound quality evaluation method of a PSO optimized delay operator model. Background technique [0002] With the continuous consumption of petroleum resources, FCEV (Fuel Cell Electric Vehicle, fuel cell electric vehicle), especially hydrogen fuel cell vehicles, as the leader of new energy vehicles, will play an increasingly important role. Hydrogen fuel cell vehicles use drive motors to replace the engines of traditional vehicles, and the vehicle structure of hydrogen fuel cell vehicles has also undergone tremendous changes compared with traditional vehicles, which also leads to changes in the sound characteristics of the cab. The present invention is based on the hydrogen energy vehicle project. After the actual vehicle inspection test, it is found that although the noise sound pressure level of the cab is significantly lower...

AI Technical Summary

Problems solved by technology

[0011] The BP network model used in the sound quality prediction model has a strong nonlinear mapping ability. However, the disadvantage of the BP neural network is that when the error surface is relatively flat, even if the weight value is adjusted greatly, the error will not drop much. Training takes a long time; when there are multiple minimum values ​​on the error surface, it is easy to fall into the process of finding a certain minimum value and cannot be corrected

Images