Issue |
JNWPU
Volume 37, Number 1, February 2019
|
|
---|---|---|
Page(s) | 57 - 62 | |
DOI | https://doi.org/10.1051/jnwpu/20193710057 | |
Published online | 03 April 2019 |
Building Sound Absorption Performance Model of Porous Glass Based on GRNN
基于GRNN建立开孔型多孔玻璃吸声性能模型
School of Marine Science and Technology, Northwestern Polytechnical University, Xi’an 710072, China
Received:
1
March
2018
The generalized regression neural network (GRNN) model of sound absorption coefficient of porous glass was built on data from 16 groups gained by experiments, where 12 groups were randomly selected as trained samples and the other 4 groups were as tested ones. This GRNN model which has two parameters, porosity and thickness as the inputs, was set the maximum iteration number 20, getting the optimal trained spread parameter σ=0.1. The results showed that the average error of this model was 0.003, and this model has high precision and the prediction curve of the sound absorption coefficient was very similar to the experiments. The advantages of this method are simple, needing less trained samples, rapid and accurate.
摘要
采用广义回归神经网络(GRNN)方法,在开孔型多孔玻璃16组实验数据基础上,以12组随机数据作为训练样本,4组作为检验样本,建立以多孔玻璃厚度和孔隙率的GRNN模型,得到训练的最佳光滑因子σ=0.1,最大迭代次数为20;结果表明,模型预测值与实验值的平均误差为0.003,建立的模型精度高,预测吸声系数曲线形貌相似度高;该方法有简单、训练样本少、快速、准确等优点。
Key words: generalized regression neural network / sound absorption coefficient / porous glass
关键字 : 广义回归神经网络 / 吸声系数 / 多孔玻璃
© 2019 Journal of Northwestern Polytechnical University
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.