基于样本熵与ELM-Adaboost的悬架减振器异响声品质预测

PDF(3783 KB)

振动与冲击 ›› 2016, Vol. 35 ›› Issue (13) : 125-133.

论文

黄海波，李人宪，黄晓蓉，杨明亮，丁渭平

作者信息 +

Prediction of Power Train Sound Metric Based on Sample Entropy and ELM-Adaboost

Huang Haibo, Li Renxian,Huang Xiaorong , Yang Mingliang, Ding Weiping

Author information +

文章历史 +

摘要

车辆悬架减振器异响严重削弱了车内声品质，针对该异响问题设计并开展了不同路况条件下的整车道路试验，对采集到的车内噪声信号分别计算A计权声压级与心理声学客观参量（响度、尖锐度、语音清晰度、抖动度和粗糙度）以提取减振器异响特征信息，并将其与主观评价进行了相关分析。另一方面，再引入小波包分解与样本熵的概念，对减振器异响特征信息进行了有效地提取，同时提出基于Adaboost的极限学习机(ELM)算法，建立了减振器异响声品质预测改进模型，并将其与支持向量机(SVM)、广义神经网络(GRNN)进行对比。研究结果表明：传统的A计权声压与心理声学指标不能有效地用于减振器异响声品质预测而结合小波包样本熵提取的异响特征与ELM-Adaboost算法能有效地对减振器异响声品质进行预测，并且效果优于SVM与GRNN。

Abstract

The abnormal noise from automobile suspension shock absorber has weakened the interior noise seriously. For this issue, a complete automobile road test has been carried out on different road surfaces to collect the shock absorber noises, and then calculate the A-weighted sound pressure level and the psychoacoustics sound metrics of the interior noise, such as loudness, sharpness, articulation index, fluctuation strength and roughness, to investigate the correlation between those objective parameters and subjective evaluation. And based on these experiments, the concepts of wavelet packet decomposition and sample entropy have been induced to extract the characteristics of abnormal noise from shock absorber. An improved model, ELM-Adaboost, has been built to predict the sound metric of shock absorber abnormal noise, and this algorithm has been compared with Support Vector Machine and Generalized Regression Neural Network. The result shows that the traditional A-weighted sound pressure level and the psychoacoustics indices cannot be used to evaluate the shock absorber abnormal noise sound quality effectively, but the developed model, which combines wavelet packet sample entropy and ELM-Adaboost algorithm, can predict the sound metric of shock absorber noise efficiently. And the root-mean-square errors of the latter is lower than those of SVM and GRNN as well.

导出引用

黄海波，李人宪，黄晓蓉，杨明亮，丁渭平. 基于样本熵与ELM-Adaboost的悬架减振器异响声品质预测[J]. 振动与冲击, 2016, 35(13): 125-133

Huang Haibo, Li Renxian,Huang Xiaorong,Yang Mingliang, Ding Weiping. Prediction of Power Train Sound Metric Based on Sample Entropy and ELM-Adaboost[J]. Journal of Vibration and Shock, 2016, 35(13): 125-133

参考文献

[1] 张立军，余卓平，靳晓雄等. 减振器异常噪声的试验研究与分析[J]. 振动与冲击，2002，21（1）：33-38.
Zhang Lijun，Yu Zhuoping，JinXiaoxiong，et al. EXPERIMENTAL RESEARCH ON THE ABNORMAL NOISE OFHYDRAULIC SUSPENSION SHOCK ABSORBER[J]. Journal of Vibration and Shock, 2002，21（1）：33-38.
[2] 宋睿，丁渭平，杨明亮等. 汽车悬架减振器结构传递异响的试验研究[J]. 汽车技术，2011，39-42
SONG Rui，DING Weiping，YANG Mingliang，et al.Experimental research on structure transfer sound of Automobile suspension shock absorber [J]. Automobile Technology，2011,39-42.
[3] J. Roger, O. Johan, R. Matti. A new test track for automotive squeak and rattle (S&R)detection. AppliedAcoustics 2014; 80: 79–85.
[4] Y. Li, W.K. Jiang. Research on the procedure for analyzing the sound quality contributionof sound sources and its application[J]. Applied Acoustics 2014; 79:75-80.
[5] Y.S. Wang, G.Q. Shen, H. Guo, et.al. Roughness modelling based on human auditory perceptionfor sound quality evaluation of vehicle interior noise. Journal of Sound and Vibration 2013; 332: 3893–3904.
[6] E. Zwicker, H. Fastl. Psychoacoustics: Facts and Models, third edition, Springer-VerlagBerlin, 2006.
[7] Vapnik V.The nature of statistical learning theory[M].New
York:Springer-Verlag,1995.
[8] 姚昌荣. 基于小波与分形理论的斜拉桥健康监测系统研究 [D]. 成都：西南交通大学，2007.
YAO Changrong. Research on structural health monitoring system of cable-stayed bridges based on wavelet and fractal theory [D]. Chengdu: Southwest Jiaotong University, 2007.
[9] Pincus S. M. Assessing serial irregularity and its im- plications for health [J] . Ann. N. Y. Acad. Sci, 2002, 954: 245-267.
[10] 赵晓华，许士丽，荣建等．基于Roc曲线的驾驶疲劳脑电样本熵判定阈值研究[J]．西南交通大学学报，2013，48(1)：178-183．
Zhao Xiaohua, Xu Shili, Rongjian, et.al. Discriminating threshold of driving fatigue based on the electroencephalography sample entropy by receiver operating characteristics curve analysis [J]. Journal of SouthwestJiaotong University, 2013, 48(1): 178-183.
[11] G. B. Huang, Q. Y. Zhu, C. K. Siew. Extreme learning machine: a new learning scheme of feedforward neural networks [C]//Proceedings of 2004 IEEE International Joint Conference on Neural Networks. Budapest, 2004, 2: 985-990.
[12] 黄海波，李人宪，丁渭平等. 基于台架试验的悬架减振器异响辨识研究[J]. 振动与冲击，2015，34(2): 191-196.
Huang Haibo, Li Renxian, Ding Weiping, et al. Rig test for identifying abnormal noise of suspension shock absorber[J]. Journal of Vibration and Shock, 2015, 34(2): 191-196.