比翱工程实验室丨基于高斯过程模型的机器学习算法用于多孔声学超材料的设计与表征
当今的城市化和交通的快速增长会严重威胁人类健康,引起烦躁、睡眠障碍,甚至缺血性心脏病,因此对环境降噪的需求正在快速增长。在这种情况下,用于声学目的的多孔介质是由通道、缝隙或空腔制成的材料,其中声波穿过泡沫并由于粘性和热效应而损失能量。然而,多孔介质在低频下的效率不如在高频率下那么有效。这种限制通常通过使用多层结构来实现;在任何情况下,这些解决方案的效果始终取决于允许的厚度或隔音配置的总质量。为了克服这种限制,声学封装的低频性能可以通过使用带有嵌入式周期性夹杂物的多孔介质作为局部谐振器来显著增强。本研究旨在开发和验证用于设计和表征基于具有嵌入式周期性结构的泡沫芯层的全局振动声处理的工具,这些泡沫芯层允许在分层概念中被动控制声学路径。
这项工作的目的是研究更易于管理,实用和可解释的机器学习方法(例如高斯过程)的适用性,以表征基于泡沫的超材料,同时证明预测性能的改进,即使涉及更复杂的现象学材料建模。此外,还强调了表征任务的有效性,这可能会提高材料设计、表征和优化的初步阶段,从而减少与实验测试和数值模拟相关的费用与计算时间。图文速览
结论
这项工作的范围是开发和验证工具,通过实施机器学习算法,提高基于具有嵌入式周期性图案的泡沫芯的振动声学封装的表征。
首先,提供了与超材料表征和所选机器学习方法相关的理论框架。然后,在指定的约束下,说明了一个允许分析计算泡沫流阻率和单元尺寸所需值的程序,以便在周期性共振频率方面达到所需的目标。先后介绍所研究的声学包的特性以及三维有限元几何形状,并对多个设置进行参数化测试,每个设置具有不同的泡沫流阻率和单元尺寸值。综上所述,开发并应用了基于高斯的机器学习算法,以预测当使用DBM模型描述元材料时,误差小于5%时,传输损耗在共振频率处增加。当考虑JCA模型时,高斯过程可以通过增加表征观测的特征数量来预测共振频率和TL峰百分比增加,而训练示例要少得多。因此可以提供更好的结果。 因此,易于处理和解释的机器学习方法(例如高斯过程)在超材料包需要表征时具有几个优点。事实上,从DBM模型到JCA模型的切换表明,输入特征的增加导致训练示例数量的明显减少。因此,如果想要生成实验训练集,这种减少可以转化为更少的样本来生成和测试,从而减少支出和节省更多的时间。此外,在数值模拟中引入了重要的时间节省,因为这意味着更少的建模和模拟运行。所有这些元素都有助于提供声学指标的自动和快速估计,这对于在初步设计阶段研究新配置特别有用。 目前工作的未来扩展可能涉及开发和实施更复杂的材料描述公式,以及先进的机器学习技术,不仅用于估计超材料的声学性能,甚至用于根据感兴趣的应用进行选择和优化。最后,已经表明,增加特征的数量将会使结果达到惊人的准确性水平,同时,训练示例的数量明显减少。这些结果为创建实验训练集开辟了道路,因为有限数量的训练样本意味着更少的实验,从而节省时间和费用。1. Cao, L.; Fu, Q.; Si, Y.; Ding, B.; Yu, J. Porousmaterials for sound absorption. Compos. Commun. 2018, 10, 25–35. [CrossRef]
2. Berardi, U.; Iannace, G. Acoustic characterization ofnatural fibers for sound absorption applications. Build. Environ. 2015,94,840–852. [CrossRef]
3. Xinzhao, X.; Guoming, L.; Dongyan, L.; Guoxin, S.;Rui, Y. Electrically conductive graphene-coated polyurethane foam and its epoxycomposites. Compos. Commun. 2018, 7, 1–6. [CrossRef]
4. Groby, J.P.; Nennig, B.; Lagarrigue, C.; Brouard, B.;Dazel, O.; Tournat, V. Using simple shape three-dimensional inclusions toenhance porous layer absorption. J. Acoust. Soc. Am. 2014, 136,1139–1148. [CrossRef]
5. Yang, Y.; Mace, B.; Kingan, M. Wave and finite elementmethod for predicting sound transmission through finite multi-layeredstructures with fluid layers. Comput. Struct. 2018, 204, 20–30. [CrossRef]
6. Weisser, T. Acoustic behavior of a rigidly backedporoelastic layer with periodic resonant inclusions by a multiple scatteringapproach. J. Acoust. Soc. Am. 2016, 139, 617–629. [CrossRef] [PubMed]
7. Gaborit, M.; Dazel, O.; Göransson, P. A simplifiedmodel for thin acoustic screens. J. Acoust. Soc. Am. 2018, 144, 76–81. [CrossRef][PubMed]
8. Groby, J.P.; Wirgin, A.; De Ryck, L.; Lauriks, W.;Gilbert, R.; Xu, Y. Acoustic response of a rigid-frame porous medium plate witha periodic set of inclusions. J. Acoust. Soc. Am. 2009, 126, 685–693. [CrossRef]
9. Xiong, L.; Nennig, B.; Aurégan, Y.; Bi, W. Soundattenuation optimization using metaporous materials tuned on exceptionalpoints. J. Acoust. Soc. Am. 2017, 142, 2288–2297. [CrossRef] [PubMed]
10. Yang, M.; Sheng, P. Sound absorption sctructures:From porous media to acoustic metamaterials. Annu. Rev. Mater. Res. 2017,47, 83–114. [CrossRef]
11. Veselago, V.G. The electrodynamics of substances withsimultaneously negative values of e and m. Sov. Phys. Uspekhi 1968,10, 509–514. [CrossRef]
12. Ma, G.; Sheng, P. Acoustic metamaterials: From localresonances to broad horizons. Sci. Adv. 2016, 2, e1501595. [CrossRef]
13. Kadic, M.; Bückmann, T.; Schittny, R.;Wegener, M.Metamaterials beyond electromagnetism. Rep. Prog. Phys. 2013, 76,126501.[CrossRef] [PubMed]
14. Yang, Z.; Mei, J.; Yang, M.; Chan, N.; Sheng, P.Membrane-type acoustic metamaterial with negative dynamic mass. Phys. Rev.Lett.2008, 101, 204301. [CrossRef] [PubMed]
15. Mei, J.; Ma, G.; Yang, M.; Yang, Z.; Wen, W.; Sheng,P. Dark acoustic metamaterials as super absorbers for low-frequency sound.Nat.Commun. 2012, 3, 756. [CrossRef] [PubMed]
16. Yang, M.; Li, Y.; Meng, C.; Fu, C.; Mei, J.; Yang,Z.; Sheng, P. Sound absorption by subwavelength membrane structures: Ageometricperspective. Comptes Rendus Mec. 2015, 343, 635–644. [CrossRef]
17. Chen, Y.; Huang, C.; Zhou, X.; Hu, G.; Sun, C.Analytical coupled vibroacoustic modeling of membrane-type acusticmetamaterials:Membrane model. J. Acoust. Soc. Am. 2014, 136, 969–979. [CrossRef] [PubMed]
18. Chen, Y.; Huang, C.; Zhou, X.; Hu, G.; Sun, C.Analytical coupled vibroacoustic modeling of membrane-type acusticmetamaterials:Plate model. J. Acoust. Soc. Am. 2014, 136, 2926–2934. [CrossRef] [PubMed]
19. Fang, N.; Xi, D.; Xu, J.; Ambati, M.;Srituravanich,W.; Sun, C.; Zhang, X. Ultrasonic metamaterials with negativemodulus. Nat.Mater. 2006, 5, 452–456. [CrossRef]
20. Wang, X.; Luo, X.; Huang, Z. Ultrathin and durableopen metamaterials for simultaneous ventilation and sound reduction. Appl.Phys.Lett. 2019, 115, 171902-1–171902-5. [CrossRef]
21. Kumar, S.; Xiang, T.; Lee, H. Ventilated acousticmetamaterial window panels for simultaneous noise shielding and aircirculation.Appl. Acoust. 2020, 159, 107088. [CrossRef]
22. Wei, P.; Croënne, C.; Chu, S.; Li, J. Symmetrical andanti-symmetrical coherent perfect absorption for acoustic waves. Appl. Phys.Lett.2014, 104, 121902. [CrossRef]
23. Merkel, A.; Theocharis, G.; Richoux, O.;Romero-Garcia, V.; Pagneux, V. Control of acoustic absorption inone-dimensional scattering by resonant scatterers. Appl. Phys. Lett. 2015,107, 244102. [CrossRef]
24. Yang, M.; Meng, C.; Fu, C.; Li, Y.; Yang, Z.; Sheng,P. Subwavelength total acoustic absorption with degenerate resonators.Appl.Phys. Lett. 2015, 107, 104104. [CrossRef]
25. Magliacano, D.; Ouisse, M.; Khelif, A.; De Rosa, S.;Franco, F.; Atalla, N.; Collet, M. Computation of dispersion diagrams forperiodic porous materials modeled as equivalent fluids. Mech. Syst. SignalProcess. 2020, 142, 106749. [CrossRef]
26. Magliacano, D.; Ouisse, M.; De Rosa, S.; Franco, F.;Khelif, A. Computation of acoustic properties and design guidelines of periodicBiot-modeled foams. Appl. Acoust. 2020, 168, 107428. [CrossRef]
27. Magliacano, D.; Ashani, S.; Ouisse, M.; Deckers, E.;Petrone, G.; Desmet,W.; De Rosa, S. Formulation and validation of the shiftcell technique for acoustic applications of poro-elastic materials described bythe Biot theory. Mech. Syst. Signal Process. 2021, 147,107089. [CrossRef]
28. Catapane, G.; Magliacano, D.; Petrone, G.; Casaburo,A.; Franco, F.; De Rosa, S. Transmission Loss Analyses on Different AngularDistributions of Periodic Inclusions in a Porous Layer. Aerotec. Missili Spazio2021, 100, 363–373. [CrossRef]
29. Magliacano, D.; Petrone, G.; Franco, F.; De Rosa, S.Numerical investigations about the sound transmission loss of fuselage panelsection with embedded periodic foams. Appl. Acoust. 2021, 182, 108265. [CrossRef]
30. Mitchell, T. Machine Learning; McGraw Hill Education:New York, NY, USA, 1997.
31. Bianco, M.; Gerstoft, P.; Traer, J.; Ozanich, E.;Roch, M.; Gannot, S.; Deledalle, C. Machine learning in acoustics: Theory andapplications. J. Acoust. Soc. Am. 2019, 146, 3590–3628. [CrossRef]
32. Michalopoulou, Z.; Gerstoft, P.; Kostek, B.; Roch, M.Introduction to the special issue on machine learning in acoustics. J. Acoust.Soc. Am. 2021, 150, 3204–3210. [CrossRef]
33. Bacigalupo, A.; Gnecco, G.; Lepidi, M.; Gambarotta,L. Machine-Learning Techniques for the Optimal Design of AcousticMetamaterials. J. Optim. Theory Appl. 2020, 187, 630–653. [CrossRef]
34. Gurbuz, C.; Kronowetter, F.; Dietz, C.; Eser, M.;Schmid, J.; Marburg, S. Generative adversarial networks for the design ofacoustic metamaterials. J. Acoust. Soc. Am. 2021, 149, 1162–1174. [CrossRef]
35. Shah, T.; Zhuo, L.; Lai, P.; De La Rosa-Moreno, A.;Amirkulova, F.; Gerstoft, P. Reinforcement learning applied to metamaterialdesign. J. Acoust. Soc. Am. 2021, 150, 321–338. [CrossRef]
36. Wu, L.; Liu, L.; Wang, Y.; Zhai, Z.; Zhuang, H.;Krishnaraju, D.; Wang, Q.; Jiang, H. A machine learning-based method to designmodular metamaterials. Extrem. Mech. Lett. 2020, 36, 100657. [CrossRef]
37. Hou, Z.; Tang, T.; Shen, J.; Li, C.; Li, F.Prediction Network of Metamaterial with Split Ring Resonator Based on DeepLearning. Nanoscale Res. Lett. 2020, 15, 83. [CrossRef]
38. Ciaburro, G.; Iannace, G. Modeling acousticmetamaterials based on reused buttons using data fitting with neural network.J. Acoust. Soc. Am. 2021, 150, 51–63. [CrossRef]
39. Stender, M.; Adams, C.; Wedler, M.; Grebel, A.;Hoffmann, N. Explainable machine learning determines effects on the soundabsorption coefficient measured in the impedance tube. J. Acoust. Soc. Am. 2021,149, 1932–1945. [CrossRef]
40. Rasmussen, C.E.;Williams, C.K.I. Gaussian Processesfor Machine Learning; MIT: Cambridge, MA, USA, 2006.
41. Nabney, I. Netlab: Algorithms for PatternRecognition, 1st ed.; Springer: Berlin, Germany, 2002.
42. Stinson, M. The Propagation Of Plane Sound Waves InNarrow And Wide Circular Tubes, And Generalization To Uniform Tubes OfArbitrary Cross-Sectional Shape. J. Acoust. Soc. Am. 1991, 89, 550–558.[CrossRef]
43. Allard, J.; Atalla, N. Propagation of Sound in PorousMedia: Modelling Sound Absorbing Materials, 2nd ed.; John Wiley and Sons, Ltd.:Hoboken, NJ, USA, 2009; pp. 1–358.
44. Delany, M.; Bazley, E. Acoustical properties offibrous absorbent materials. Appl. Acoust. 1970, 3, 105–116. [CrossRef]
45. Miki, Y. Acoustical properties of porous materials:Modifications of Delany-Bazley models. J. Acoust. Soc. Jpn. (E) 1990,11, 19–24. [CrossRef]
46. Johnson, D.L.; Koplik, J.; Dashen, R. Theory ofdynamic permeability and tortuosity in fluid-saturated porous media. J. FluidMech. 1987, 176, 379. [CrossRef]
47. Champoux, Y.; Allard, J.F. Dynamic tortuosity andbulk modulus in air-saturated porous media. J. Appl. Phys. 1991, 70,1975–1979.[CrossRef]
48. Gibbs, M.; MacKay, D. Variational Gaussian processclassifiers. IEEE Trans. Neural Netw. 2000, 11, 1458–1464. [CrossRef]
49. Jaakkola, T.; Haussler, D. Probabilistic kernelregression models. In Proceedings of the Seventh International Workshop onArtificial Intelligence and Statistics, Fort Lauderdale, FL, USA, 3–6 January1999; Heckerman, D., Whittaker, J., Eds.; PMLR:Cambridge, MA, USA, 1999.
50. Neal, R. Regression and classification using Gaussianprocess priors. In Bayesian Statistics 6; Bernardo, J., Berger, J., Dawid,A.,SMith, A., Eds.; Oxford University Press: Oxford, UK, 1998; pp. 475–601.
51. Seeger, M.; Jordan, M. Sparse Gaussian ProcessClassification with Multiple Classes; Technical Report TR 661; Department ofStatistics,University of California: Berkeley, CA, USA, 2004.
52. Stein, M.L. Interpolation of SpatialData; Springer: Berlin, Germany, 1999.
● 比翱工程实验室丨从天然生态材料到人工结构,现代降噪材料综述● 比翱工程实验室丨主动与被动控制声学超材料的最新进展:综述
● 比翱工程实验室丨基于微观结构和本构特性的开孔泡沫材料声学模拟● 比翱工程实验室丨机器学习方法确定对阻抗管吸声系数测量的影响
● 比翱工程实验室丨增材制造的晶格结构的声学和热声特性● 比翱工程实验室丨“声学中的机器学习” JASA.特刊综述● 比翱工程实验室丨基于人工智能预测纤维材料比翱(BIOT)特性参数的方法
● 比翱工程实验室丨周期性局部共振结构的阻带效应与传输损耗● 比翱工程实验室丨虚拟单元法在控制复杂多孔弹性介质的声传播中的应用● 比翱工程实验室丨声学多孔材料拓扑优化的启发式和元启发式方法比较● 比翱工程实验室丨超材料防爆面板的创新设计与性能研究● 比翱工程实验室丨声波作用下圆柱形类骨多孔材料的三维生物力学建模
● 比翱工程实验室丨热粘性声学超材料用于抑制复杂形状结构的低频声模态● 比翱工程实验室丨使用传递矩阵法表征和开发周期性声学超材料
免责声明:部分资料来源网络,转载目的在于传递信息及分享,并不意味赞同其观点或其真实性,也不构成其他建议。仅提供交流平台,不为其版权负责。如涉及侵权,请联系我们及时删除。