基于人工神经网络的共振吸声超材料声学性能快速预测及结构优化设计

高兆瑞; 李铮; 姜永烽; 沈承; 孟晗

doi:10.21656/1000-0887.450170

基于人工神经网络的共振吸声超材料声学性能快速预测及结构优化设计

doi: 10.21656/1000-0887.450170

高兆瑞^{1, 2,},
李铮^{1, 2},
姜永烽^{1, 2},
沈承^{1, 2, ,},
孟晗^{1, 2, ,}

1.
南京航空航天大学航空航天结构力学及控制全国重点实验室，南京 210016
2.
南京航空航天大学多功能轻量化材料与结构工信部重点实验室，南京 210016

(我刊青年编委孟晗来稿)

基金项目:

国家自然科学基金 12202183

国家自然科学基金 12202188

国家自然科学基金 52361165626

国家重点研发计划 2023YFB4604800

详细信息

作者简介:
高兆瑞(2000—)，男，硕士生(E-mail: gaozr123@nuaa.edu.cn)

通讯作者:
沈承(1986—)，男，副教授，博士，硕士生导师(通讯作者. E-mail: cshen@nuaa.edu.cn)

孟晗(1989—)，女，教授，博士，博士生导师(通讯作者. E-mail: menghan@nuaa.edu.cn)

中图分类号: TB535; TP183
计量
- 文章访问数: 153
- HTML全文浏览量: 59
- PDF下载量: 41
- 被引次数: 0
出版历程
- 收稿日期: 2024-06-11
- 刊出日期: 2024-08-01

Acoustic Performance Rapid Prediction and Structural Optimization for Resonant Sound-Absorbing Metamaterials Based on Artificial Neural Networks

GAO Zhaorui^{1, 2
,},
LI Zheng^{1, 2},
JIANG Yongfeng^{1, 2},
SHEN Cheng^{1, 2
, ,},
MENG Han^{1, 2
, ,}

1.
State Key Laboratory of Mechanics and Control for Aerospace Structures, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, P.R.China
2.
MIIT Key Laboratory of Multifunctional Lightweight Materials and Structures, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, P.R.China

(Contributed by MENG Han, M.AMM Youth Editorial Board)

摘要

摘要: 针对共振吸声超材料声学性能快速预测及结构优化设计需求，提出了一种基于人工神经网络的共振吸声超材料性能预测方法. 首先, 建立了由微穿孔板和Helmholtz共振腔组成的多层穿孔型共振吸声超材料的理论模型，并通过仿真与实验验证其正确性；随后，通过理论模型生成数据集，并以此为基础，采用BP(back propagation)神经网络原理，搭建了结构特征参量与声学性能的人工神经网络模型；之后，将训练后的人工神经网络模型与遗传算法相结合，对共振吸声超材料进行声学性能最优化设计. 结果表明：训练后的人工神经网络模型可以对目标结构的吸声性能进行准确预测，并且预测效率相较理论模型提高50%以上；人工神经网络模型与优化算法的结合不仅能提高优化效率，优化后的结构也具有良好的低频宽带吸声性能. 人工神经网络为大规模结构性能预测计算提供了便利，在超材料等结构设计及优化领域具有广阔的应用前景.
- 共振吸声超材料 /
- 人工神经网络 /
- 吸声系数 /
- BP神经网络 /
- 遗传算法
Abstract: A sound performance prediction method based on the artificial neural network (ANN) was proposed to meet the requirements of rapid prediction and optimization design of resonant sound-absorbing metamaterials. Firstly, a theoretical model was established for multilayer perforated resonant sound-absorbing metamaterials (MPRSMs) composed of microperforated panels and Helmholtz resonators, which was then verified through simulation and experiments; subsequently, a dataset was generated with the theoretical model, and in turn an ANN model was constructed by means of the back propagation (BP) neural network to build the mapping relationship between structural parameters and acoustic performances; afterwards, the trained ANN model was combined with the genetic algorithm to optimize the acoustic performance of the MPRSMs. The results show that, the trained ANN model can accurately predict the sound absorption performance of the MPRSMs, and the prediction efficiency improves by more than 50% compared to the theoretical model; the combination of the ANN model and the optimization algorithm can not only improve the optimization efficiency, but bring good low-frequency broadband sound absorption performance of the optimized structure. The ANN provides convenience for large-scale structural performance prediction calculations and has broad application prospects in structural design and optimization of metamaterials.
- resonant sound-absorbing metamaterial /
- artificial neural network /
- sound absorption coefficient /
- BP neural network /
- genetic algorithm
edited-by

edited-by
注释:

1) (我刊青年编委孟晗来稿)

HTML全文

图 1 MPRSM元胞结构剖视图

Figure 1. The MPRSM sectional view

下载: 全尺寸图片幻灯片

图 2 MPRSM有限元仿真模型

注为了解释图中的颜色，读者可以参考本文的电子网页版本，后同.

Figure 2. The finite element simulation model for the MPRSM

下载: 全尺寸图片幻灯片

图 3 阻抗管声学测试实验设备以及阻抗管内部实验样件安装位置

Figure 3. The impedance tube acoustic experimental equipment and the installation location of the experimental specimen

下载: 全尺寸图片幻灯片

图 4 MPRSM实验样件图

Figure 4. The MPRSM specimen photo

下载: 全尺寸图片幻灯片

图 5 MPRSM理论模型、有限元仿真与实验的吸声系数曲线对比

Figure 5. Comparison of sound absorption coefficients by the theoretical model, the finite element simulation, and the experiment of the MPRSM

下载: 全尺寸图片幻灯片

图 6 BP神经网络的神经元处理信息与反向传播示意图

Figure 6. Schematic diagram of neural information processing and backpropagation in the BP neural network

下载: 全尺寸图片幻灯片

图 7 人工神经网络预测模型

Figure 7. The ANN prediction model

下载: 全尺寸图片幻灯片

图 8 人工神经网络模型训练误差分析

Figure 8. Analysis of training errors in the ANN model

下载: 全尺寸图片幻灯片

图 9 人工神经网络模型预测MPRSM结构吸声性能结果与理论模型结果对比

Figure 9. Comparison between the predicted sound absorption performance results of the MPRSM structure with the ANN model and the theoretical model

下载: 全尺寸图片幻灯片

图 10 优化后MPRSM结构吸声系数曲线

Figure 10. Sound absorption coefficients of the optimized MPRSM

下载: 全尺寸图片幻灯片

表 1 MPRSM结构模型几何参数

Table 1. Geometric parameters of the MPRSM structural model

geometric parameter	value
the 1st perforated plate thickness t₁/mm	1.1
the 1st perforated plate hole diameter d₁/mm	1.2
the 1st perforated plate back cavity thickness h₁/mm	35
Helmholtz resonator back cavity thickness d_c/mm	8
Helmholtz resonance cavity neck length d_n/mm	2
Helmholtz resonant cavity neck radius r_n/mm	1
the 2nd perforated plate thickness t₂/mm	2.9
the 2nd perforated plate hole diameter d₂/mm	1
the 2nd perforated plate back cavity thickness h₂/mm	25
the 1st perforated plate perforation rate p₁	0.024 7
the 2nd perforated plate perforation rate p₂	0.020 4
structural unit cell length L/mm	31

下载: 导出CSV

表 2 用于人工神经网络模型训练的MPRSM结构参数取值范围

Table 2. MPRSM structural parameter values for the artificial neural network (ANN) model training

geometric parameter	value range
the 1st perforated plate thickness t₁/mm	0.5~2
the 1st perforated plate hole diameter d₁/mm	0.3~1
the 1st perforated plate back cavity thickness h₁/mm	5~30
Helmholtz resonator back cavity thickness d_c/mm	5~30
Helmholtz resonance cavity neck length d_n/mm	1~5
Helmholtz resonant cavity neck radius r_n/mm	0.5~2
the 2nd perforated plate thickness t₂/mm	0.5~2
the 2nd perforated plate hole diameter d₂/mm	0.3~1
the 2nd perforated plate back cavity thickness h₂/mm	5~30

下载: 导出CSV

表 3 验证人工神经网络模型预测能力的四组结构参数

Table 3. Four sets of structural parameters for verifying the predictive ability of the ANN model

optimization parameter	set 1	set 2	set 3	set 4
the 1st perforated plate thickness t₁/mm	0.51	1.60	1.41	0.70
the 1st perforated plate hole diameter d₁/mm	0.50	0.81	0.90	0.71
the 1st perforated plate back cavity thickness h₁/mm	13.89	16.80	25.11	13.20
Helmholtz resonator back cavity thickness d_c/mm	28.20	8.21	17.90	25.10
Helmholtz resonance cavity neck length d_n/mm	4.10	3.10	3.50	3.70
Helmholtz resonant cavity neck radius r_n/mm	0.50	1.21	0.80	0.90
the 2nd perforated plate thickness t₂/mm	0.80	1.01	1.50	1.50
the 2nd perforated plate hole diameter d₂/mm	0.40	0.80	0.70	0.60
the 2nd perforated plate back cavity thickness h₂/mm	24.80	14.19	10.30	28.10

下载: 导出CSV

表 4 MPRSM结构优化参数取值范围以及优化结果

Table 4. Range of MPRSM structural optimization parameters and optimization results

optimization parameter	value range	optimization result
the 1st perforated plate thickness t₁/mm	0.5~2	0.51
the 1st perforated plate hole diameter d₁/mm	0.3~1	0.89
the 1st perforated plate back cavity thickness h₁/mm	5~30	27.90
Helmholtz resonator back cavity thickness d_c/mm	5~30	29.90
Helmholtz resonance cavity neck length d_n/mm	1~5	4.81
Helmholtz resonant cavity neck radius r_n/mm	0.5~2	0.50
the 2nd perforated plate thickness t₂/mm	0.5~2	0.50
the 2nd perforated plate hole diameter d₂/mm	0.3~1	0.41
the 2nd perforated plate back cavity thickness h₂/mm	5~30	15.88

下载: 导出CSV

参考文献(27)

[1]	郭梦媛, 刘崇锐, 苏文斌, 等. 高阶微穿孔型超材料低频宽带吸声机理[J]. 西安交通大学学报, 2024, 58(4): 192-199, 220. doi: 10.7652/xjtuxb202404018 GUO Mengyuan, LIU Chongrui, SU Wengbin, et al. Low-frequency broadband absorption mechanism of high-order micro-perforated meta-materials[J]. Journal of Xi'an Jiaotong University, 2024, 58(4): 192-199, 220. (in Chinese) doi: 10.7652/xjtuxb202404018
[2]	MAA D Y. Theory of microslit absorbers[J]. Acta Acustica, 2000, 25(6): 481-485. doi: 10.3321/j.issn:0371-0025.2000.06.001
[3]	MAA D Y, LIU K. Sound absorption characteristics of microperforated absorber for random incidence[J]. Acta Acustica, 2000, 25(4): 289-296.
[4]	LALY Z, ATALLA N, MESLIOUI S A. Acoustical modeling of micro-perforated panel at high sound pressure levels using equivalent fluid approach[J]. Journal of Sound and Vibration, 2018, 427: 134-158. doi: 10.1016/j.jsv.2017.09.011
[5]	GAI X L, XING T, LI X H, et al. Sound absorption of microperforated panel with L shape division cavity structure[J]. Applied Acoustics, 2017, 122: 41-50. doi: 10.1016/j.apacoust.2017.02.004
[6]	LEE D H, KWON Y P. Estimation of the absorption performance of multiple layer perforated panel systems by transfer matrix method[J]. Journal of Sound and Vibration, 2004, 278(4): 847-860.
[7]	GAI X L, XING T, LI X H, et al. Sound absorption properties of microperforated panel with membrane cell and mass blocks composite structure[J]. Applied Acoustics, 2018, 137: 98-107. doi: 10.1016/j.apacoust.2018.03.013
[8]	LIU X W, YU C L, XIN F X. Gradually perforated porous materials backed with Helmholtz resonant cavity for broadband low-frequency sound absorption[J]. Composite Structures, 2021, 263: 113647. doi: 10.1016/j.compstruct.2021.113647
[9]	ZHANG H J, WANG Y, LU K Y, et al. SAP-net: deep learning to predict sound absorption performance of metaporous materials[J]. Materials & Design, 2021, 212: 110156.
[10]	YANG H T, ZHANG H J, WANG Y, et al. Prediction of sound absorption coefficient for metaporous materials with convolutional neural networks[J]. Applied Acoustics, 2022, 200: 109052. doi: 10.1016/j.apacoust.2022.109052
[11]	PAN B R, SONG X, XU J J, et al. Accelerated inverse design of customizable acoustic metaporous structures using a CNN-GA-based hybrid optimization framework[J]. Applied Acoustics, 2023, 210: 109445.
[12]	IANNACE G, CIABURRO G, TREMATERRA A. Modelling sound absorption properties of broom fibers using artificial neural networks[J]. Applied Acoustics, 2020, 163: 107239.
[13]	SHEN X M, BAI P F, YANG X C, et al. Low frequency sound absorption by optimal combination structure of porous metal and microperforated panel[J]. Applied Sciences-Basel, 2019, 9(7): 1507.
[14]	TANG Y F, LI F H, XIN F X, et al. Heterogeneously perforated honeycomb-corrugation hybrid sandwich panel as sound absorber[J]. Materials & Design, 2017, 134: 502-512.
[15]	TANG Y F, XIN F X, HUANG L X, et al. Deep subwavelength acoustic metamaterial for low-frequency sound absorption[J]. Europhysics Letters, 2017, 118(4): 44002.
[16]	WANG S B, WANG B, FAN J, et al. Inversion of equivalent parameters of acoustic coating based on genetic algorithm[J]. Journal of Ship Mechanics, 2023, 27(3): 456-469.
[17]	JIANG Y F, SHEN C, MENG H, et al. Design and optimization of micro-perforated ultralight sandwich structure with N-type hybrid core for broadband sound absorption[J]. Applied Acoustics, 2023, 202: 109184.
[18]	王飞萌, 王良模, 王陶, 等. 微穿孔板-三聚氰胺吸音海绵-空腔复合结构声学性能优化设计[J]. 北京化工大学学报(自然科学版), 2022, 49(1): 113-121. WANG Feimeng, WANG Liangmo, WANG Tao, et al. Optimization of the acoustic performance of micro-perforated panel-melamine sound-absorbing sponge-cavity composite structures[J]. Journal of Beijing University of Chemical Technology (Natural Science), 2022, 49(1): 113-121. (in Chinese)
[19]	LI H X. Fuzzy logic systems are equivalent to feedforward neural networks[J]. Science in China (Series E): Technological Sciences, 2000, 43(1): 42-54.
[20]	LING S H, LAM H K, LEUNG F H F, et al. A genetic algorithm based neural-tuned neural network[C]//Proceedings of the 29 th Annual Conference of the IEEE Industrial Electronics Society. Roanoke, America, 2003: 2423-2428.
[21]	WANG X Y, ZHANG Y. Chaotic diagonal recurrent neural network[J]. Chinese Physics B, 2012, 21(3): 038703.
[22]	WEN L, QIU Z W, QI R N. Passenger capacity prediction based on genetic neural network[C]//Proceedings of the 1 st International Symposium on Information Engineering and Electronic Commerce. Ternopil, Ukraine, 2009: 696-700.
[23]	姚浩, 夏桂然, 刘泽佳, 等. 基于机器学习的黏钢构件黏接层缺陷识别方法研究[J]. 应用数学和力学, 2024, 45(4): 429-442. YAO Hao, XIA Guiran, LIU Zejia, et al. A defect identification method for bonding layers of adhesive steel members based on machine learning[J]. Applied Mathematics and Mechanics, 2024, 45(4): 429-442. (in Chinese)
[24]	CIABURRO G, IANNACE G. Modeling acoustic metamaterials based on reused buttons using data fitting with neural network[J]. Journal of the Acoustical Society of America, 2021, 150(1): 51-63.
[25]	CIABURRO G, IANNACE G. Membrane-type acoustic metamaterial using cork sheets and attached masses based on reused materials[J]. Applied Acoustics, 2022, 189: 108605.
[26]	LUO Z H, LI T, YAN Y W, et al. Prediction of sound insulation performance of aramid honeycomb sandwich panel based on artificial neural network[J]. Applied Acoustics, 2022, 190: 108656.
[27]	DOUTRES O, ATALLA N, OSMAN H. Transfer matrix modeling and experimental validation of cellular porous material with resonant inclusions[J]. Journal of the Acoustical Society of America, 2015, 137(6): 3502-3513.

施引文献

资源附件(0)

访问统计

图(10) / 表(4)

计量

文章访问数: 153
HTML全文浏览量: 59
PDF下载量: 41
被引次数: 0

姓名
邮箱
手机号码
标题
留言内容
验证码

留言板

基于人工神经网络的共振吸声超材料声学性能快速预测及结构优化设计

doi: 10.21656/1000-0887.450170

作者简介:
高兆瑞(2000—)，男，硕士生(E-mail: gaozr123@nuaa.edu.cn)

通讯作者:
沈承(1986—)，男，副教授，博士，硕士生导师(通讯作者. E-mail: cshen@nuaa.edu.cn)

孟晗(1989—)，女，教授，博士，博士生导师(通讯作者. E-mail: menghan@nuaa.edu.cn)

计量

Acoustic Performance Rapid Prediction and Structural Optimization for Resonant Sound-Absorbing Metamaterials Based on Artificial Neural Networks

计量

目录

留言板

基于人工神经网络的共振吸声超材料声学性能快速预测及结构优化设计

doi: 10.21656/1000-0887.450170

作者简介: 高兆瑞(2000—)，男，硕士生(E-mail: gaozr123@nuaa.edu.cn)

通讯作者: 沈承(1986—)，男，副教授，博士，硕士生导师(通讯作者. E-mail: cshen@nuaa.edu.cn) 孟晗(1989—)，女，教授，博士，博士生导师(通讯作者. E-mail: menghan@nuaa.edu.cn)

计量

出版历程

Acoustic Performance Rapid Prediction and Structural Optimization for Resonant Sound-Absorbing Metamaterials Based on Artificial Neural Networks

计量

出版历程

目录

作者简介:
高兆瑞(2000—)，男，硕士生(E-mail: gaozr123@nuaa.edu.cn)

通讯作者:
沈承(1986—)，男，副教授，博士，硕士生导师(通讯作者. E-mail: cshen@nuaa.edu.cn)

孟晗(1989—)，女，教授，博士，博士生导师(通讯作者. E-mail: menghan@nuaa.edu.cn)