On Machine-Learning-Driven Surrogates for Sound Transmission Loss Simulations
Source abstract
Abstract
Surrogate models are data-based approximations of computationally expensive simulations that enable efficient exploration of the model’s design space and informed decision making in many physical domains. The usage of surrogate models in the vibroacoustic domain, however, is challenging due to the non-smooth, complex behavior of wave phenomena. This paper investigates four machine learning (ML) approaches in the modelling of surrogates of sound transmission loss (STL). Feature importance and feature engineering are used to improve the models’ accuracy while increasing their interpretability and physical consistency. The transfer of the proposed techniques to other problems in the vibroacoustic domain and possible limitations of the models are discussed. Experiments show that neural network surrogates with physics-guided features have better accuracy than other ML models across different STL models. Furthermore, sensitivity analysis methods are used to assess how physically coherent the analyzed surrogates are.
Source: crossref · Quality: baseline
Affiliations
Laboratory of Tribology and Dynamics of Systems, Ecole Centrale de Lyon, 69134 Ecully, France
Compredict GmbH, 64283 Darmstadt, Germany
Institut Camille Jordan, Ecole Centrale de Lyon, 69134 Ecully, France
COSYS/SII, I4S Team, University Gustave Eiffel, Inria, 35042 Rennes, France
Funding
European Union’s Horizon 2020 research and innovation program
References (45)
Show source-provided references
1. Sobester, A., Forrester, A., and Keane, A. (2008). Engineering Design via Surrogate Modelling: A Practical Guide, John Wiley & Sons.10.1002/9780470770801 ↗
2. Surrogate-based analysis and optimization10.1016/j.paerosci.2005.02.001 ↗
3. Advances in surrogate based modeling, feasibility analysis, and optimization: A review10.1016/j.compchemeng.2017.09.017 ↗
4. Structural-borne acoustics analysis and multi-objective optimization by using panel acoustic participation and response surface methodology10.1016/j.apacoust.2016.09.013 ↗
5. Acoustic analysis of damping structure with response surface method10.1016/j.apacoust.2006.05.021 ↗
6. Research on aerodynamic performance and noise reduction of high-voltage fans on fuel cell vehicles10.1016/j.apacoust.2021.108454 ↗
7. Moustapha, M. (2016). Adaptive Surrogate Models for the Reliable Lightweight Design of Automotive Body Structures. [Ph.D. Thesis, Université Blaise Pascal-Clermont-Ferrand II].
8. A comparative study of non-traditional methods for vehicle crashworthiness and NVH optimization10.1007/s11831-015-9155-y ↗
9. The optimal design for low noise intake system using Kriging method with robust design
10. Machine-learning techniques for the optimal design of acoustic metamaterials10.1007/s10957-019-01614-8 ↗
11. Diestmann, T., Broedling, N., Götz, B., and Melz, T. (2021, January 7–8). Surrogate Model-Based Uncertainty Quantification for a Helical Gear Pair. Proceedings of the International Conference on Uncertainty in Mechanical Engineering, Darmstadt, Germany.10.1007/978-3-030-77256-7_16 ↗
12. Comparison of uncertainty quantification process using statistical and data mining algorithms10.1007/s00158-019-02381-w ↗
13. Sensitivity analysis of generalised eigenproblems and application to wave and finite element models10.1016/j.jsv.2020.115345 ↗
14. Membrane-type acoustic metamaterial using cork sheets and attached masses based on reused materials10.1016/j.apacoust.2021.108605 ↗
15. Casaburo, A., Magliacano, D., Petrone, G., Franco, F., and De Rosa, S. (2021). Gaussian-Based Machine Learning Algorithm for the Design and Characterization of a Porous Meta-Material for Acoustic Applications. Appl. Sci., 12.10.3390/app12010333 ↗
16. Li, M., Zhou, W., Liu, J., Zhang, X., Pan, F., Yang, H., Li, M., and Luo, D. (2021). Vehicle Interior Noise Prediction Based on Elman Neural Network. Appl. Sci., 11.10.3390/app11178029 ↗
17. Surrogate-based acoustic noise prediction of electric motors10.1109/TMAG.2019.2945407 ↗
18. Virtual reality of interior noises of vehicles based on boundary element and neural networks10.1007/s00521-016-2836-0 ↗
19. Bárkányi, Á., Chován, T., Németh, S., and Abonyi, J. (2021). Modelling for Digital Twins—Potential Role of Surrogate Models. Processes, 9.10.3390/pr9030476 ↗
20. Cunha, B., Droz, C., Zine, A., Foulard, S., and Ichchou, M. (2022). A Review of Machine Learning Methods Applied to Structural Dynamics and Vibroacoustic. arXiv.
21. A high-bias, low-variance introduction to machine learning for physicists10.1016/j.physrep.2019.03.001 ↗
22. A few useful things to know about machine learning10.1145/2347736.2347755 ↗
23. A review of the different approaches to predict the sound transmission loss of building partitions10.1177/1351010X20911599 ↗
24. Fahy, F.J., and Gardonio, P. (2007). Sound and Structural Vibration: Radiation, Transmission and Response, Academic Press. Chapter 5.10.3397/1.2741307 ↗
25. Theorie der Schalldämmung dünner Wände bei schrägem Einfall
26. Wave finite element formulation of the acoustic transmission through complex infinite plates10.3813/AAA.919013 ↗
27. On the modeling of sound radiation from poroelastic materials10.1121/1.2261244 ↗
28. A simple method to account for size effects in the transfer matrix method10.1121/1.3280237 ↗
29. Sound transmission through simply supported finite double-panel partitions with enclosed air cavity10.1115/1.4000466 ↗
30. Modal sound transmission loss of a single leaf panel: Asymptotic solutions10.1121/1.4938222 ↗
31. COMSOL AB (2020). COMSOL Multiphysics®, COMSOL AB. Version 5.6.
32. COMSOL AB (2020). COMSOL Multiphysics Reference Manual, COMSOL AB.
33. Murphy, K.P. (2012). Machine Learning: A Probabilistic Perspective, MIT Press.
34. Rasmussen, C.E. (2003, January 4–16). Gaussian processes in machine learning. Proceedings of the Summer School on Machine Learning, Tübingen, Germany.
35. Louppe, G., Wehenkel, L., Sutera, A., and Geurts, P. (2013). Understanding variable importances in forests of randomized trees. Adv. Neural Inf. Process. Syst., 26.
36. Sudret, B., Marelli, S., and Wiart, J. (2017, January 19–24). Surrogate models for uncertainty quantification: An overview. Proceedings of the 2017 11th European Conference on Antennas and Propagation (EUCAP), Paris, France.10.23919/EuCAP.2017.7928679 ↗
37. Casalicchio, G., Molnar, C., and Bischl, B. (2018, January 16–20). Visualizing the feature importance for black box models. Proceedings of the Joint European Conference on Machine Learning and Knowledge Discovery in Databases, Würzburg, Germany.10.1007/978-3-030-10925-7_40 ↗
38. Pizarroso, J., Portela, J., and Muñoz, A. (2020). NeuralSens: Sensitivity analysis of neural networks. arXiv.
39. Cao, M., Alkayem, N.F., Pan, L., Novák, D., and Rosa, J. (2016). Advanced methods in neural networks-based sensitivity analysis with their applications in civil engineering. Artificial Neural Networks: Models and Applications, IntechOpen.10.5772/64026 ↗
40. Calculations of Sobol indices for the Gaussian process metamodel10.1016/j.ress.2008.07.008 ↗
41. Multiple additive regression trees with application in epidemiology10.1002/sim.1501 ↗
42. Scornet, E. (2020). Trees, forests, and impurity-based variable importance. arXiv.
43. Imaizumi, M., and Fukumizu, K. (2019, January 16–18). Deep neural networks learn non-smooth functions effectively. Proceedings of the 22nd International Conference on Artificial Intelligence and Statistics, PMLR, Naha, Japan.
44. Hastie, T., Tibshirani, R., Friedman, J.H., and Friedman, J.H. (2009). The Elements of Statistical Learning: Data Mining, Inference, and Prediction, Springer.10.1007/978-0-387-84858-7 ↗
45. Williams, C.K., and Rasmussen, C.E. (2006). Gaussian Processes for Machine Learning, MIT Press.10.7551/mitpress/3206.001.0001 ↗