基于CEEMD样本熵和GA-BP的排气噪声声品质预测  

Sound Quality Prediction for Exhaust Noise Based on CEEMD Sample Entropy and GA-BP

作　　者：曾发林 蔡嘉伟[2] 孙苏民 ZENG Falin;CAI Jiawei;SUN Sumin(Jiangxi Province Key Laboratory of Vehicle Noise and Vibration,Nanchang 330013,China;Automobile Engineering Research Institute,Jiangsu University,Zhenjiang 212013,China)

机构地区：[1]江西省汽车噪声与振动重点实验室,江西南昌330013 [2]江苏大学汽车工程研究院,江苏镇江212013

出　　处：《郑州大学学报（工学版）》2020年第6期19-25,共7页Journal of Zhengzhou University（Engineering Science）

基　　金：国家自然科学基金资助项目(51575238);江西省汽车噪声与振动重点实验室开放基金资助项目(JXNVHKB-KFKT-201802)。

摘　　要：为了预测汽车非稳态排气噪声声品质,进行了1G WOT和2G WOT试验,并结合主、客观评价分析非稳态排气噪声满意度的主要影响因素。同时,通过相关性分析发现心理声学客观参量与满意度之间的内在关系。运用互补总体经验模态分解(complementary ensemble empirical mode decomposition,CEEMD)方法对非稳态排气噪声进行分解,得到多个本征模态函数(intrinsic mode function,IMF),并用样本熵计算IMF分量,完成噪声信号特征提取。采用主成分分析(principal component analysis,PCA)对数据进行降维处理,减少冗余并保留原始数据的主特征,得到新参量SQP-CSP(sound quality parameter base on CEEMD and then proceed SE-PCA)值S。同时,运用遗传算法(genetic algorithm,GA)优化BP神经网络的权值和阈值,建立GA-BP预测模型。将心理声学参量也作为模型的输入进行预测。模型对比结果表明,根据新参量建立的模型对非稳态信号声品质预测具有更高的精度,更能体现非稳态信号的特征。In order to predict the sound quality of automobile unsteady exhaust noise,2G WOT and 3G WOT tests were conducted.The main factors of satisfaction level were analyzed.Via the correlation analysis,internal relations between the psychoacoustic objective parameters and the subjective evaluation were revealed.The complementary ensemble empirical mode decomposition(CEEMD)was used to decompose the signals of the accelerated exhaust noise,obtaining multiple IMF components and calculating sample entropy of the IMF components.In order to reduce the possibility of redundancy and retain the main features of the original data,principal component analysis(PCA)was applied to reduce dimension of data,so as to establish a new sound quality parameter SQP-CSP(sound quality parameter base on CEEMD and then proceed SE-PCA).Meanwhile,the genetic algorithm(GA)was used to optimize the weights and thresholdsin BP neural network,so that a GA-BP was developed to predict the sound quality of the accelerated exhaust noise.In order to validate the newly extracted unsteady exhaust noise features,the psychoacoustic parameters were also taken as the model’s inputs to predict the sound quality.The results showed that the model based on the new parameters had higher accuracy for predicting the sound quality of unsteady exhaust noise.

关 键 词：非稳态排气噪声 GA-BP 声品质 CEEMD 样本熵 主成分分析 

分 类 号：U467.4[机械工程—车辆工程]