Artificial neural network-based sound insulation optimization design of composite floor of high-speed train

Abstract

Increasing the speed of high-speed trains requires the lightweight design of vehicles to meet the economic and ecological efficiency requirements of such trains. However, these objectives conflict with the interior noise control in high-speed trains because the sound insulation of panel structures follows the mass law principle. The train floor, the main train body structure of the high-speed train, is vital for interior noise control because its sound insulation performance directly affects the interior noise levels. Owing to the complexity of the composite floor system, reliable measurement and accurate estimation of its sound insulation performance are often time-consuming and laborious. To address this situation, this study proposes an artificial neural network (ANN)-based model to predict the sound insulation characteristics of a composite floor. First, a sound insulation model of the composite floor is built based on statistical energy analysis (SEA). The sound insulation performance of 200 cases of composite floors is calculated by varying the dimensions of the extruded floor, thickness of the webs, sound-absorbing material, and wooden floor to formulate a sound insulation database of composite floors. Next, an ANN model is introduced and trained on the sound insulation database. The sound insulation prediction results obtained using the ANN model are compared to the prediction results obtained using the experiment to validate its effectiveness. Subsequently, the NSGA-II optimization method is used to optimize the sound insulation structure of the composite floor. Compared with the regular composite floor structure, the optimized structure reduced the mass of the composite floor by 10.93 kg and increased the weight of the sound insulation (R_w) by 6.3 dB. The proposed method can be an effective, economical, and efficient tool for vehicle designers and can help promote the sound insulation optimization design of high-speed train composite floors.

Keywords

High-speed train composite floor sound insulation artificial neural network lightweight design

Get full access to this article

View all access options for this article.

References

Forssén

Tober

Corakci

, et al. Modelling the interior sound field of a railway vehicle using statistical energy analysis. Appl Acoust 2012; 73(4): 307–311.

Kim

Ryue

Thompson

, et al. Prediction of radiation ratio and sound transmission of complex extruded panel using wavenumber domain unite element and boundary element methods. J Phys Conf Ser 2016; 744(1): 012144.

Zhang

Yao

Wang

, et al. Vibro-acoustic modelling of high-speed train composite floor and contribution analysis of its constituent materials. Compos Struct 2021; 256: 113049.

Yao

Zhang

Wang

, et al. Lightweight design and sound insulation characteristic optimisation of railway floating floor structures. Appl Acoust 2019; 156: 66–77.

Yao

Zhang

Wang

, et al. Vibroacoustic damping optimisation of high-speed train floor panels in low- and mid-frequency range. Appl Acoust 2021; 174: 107788.

Bianco

Gerstoft

Traer

, et al. Machine learning in acoustics: theory and applications. J Acoust Soc Am 2019; 146(5): 3590–3628.

Garg

Dhruw

Gandhi

Prediction of sound insulation of Sandwich partition panels by means of artificial neural networks. Arch Acoust 2017; 42(4): 643–651.

Buratti

Barelli

Moretti

Wooden windows: sound insulation evaluation by means of artificial neural networks. Appl Acoust 2013; 74(5): 740–745.

Bader Eddin

Menard

Bard

, et al. A sound insulation prediction model for floor structures in wooden buildings using neural networks approach. In: Inter-Noise21, Washington, DC, USA, 1–5 August 2021, Paper no. 2, pp. 4166–4176. Institute of Noise Control Engineering.

10.

Bader Eddin

Ménard

Bard Hagberg

, et al. Prediction of sound insulation using artificial neural networks—part I: lightweight wooden floor structures. Acoust 2022; 4(1): 203–226.

11.

Luo

Yan

, et al. Prediction of sound insulation performance of aramid honeycomb sandwich panel based on artificial neural network. Appl Acoust 2022; 190: 108656.

12.

Craik

RJM

Smith

. Sound transmission through lightweight double walls using statistical energy analysis. In: Inter-Noise 96, UK, 30 July–2 August 1996, pp. 2933-2938. Institute of Noise Control Engineering.

13.

Shorter

Langley

RS.

Vibro-acoustic analysis of complex systems. J Sound Vib 2005; 288(3): 669–699.

14.

Langley

Cordioli

JA.

Hybrid deterministic-statistical analysis of vibro-acoustic systems with domain couplings on statistical components. J Sound Vib 2009; 321(3-5): 893–912.

15.

Cotoni

Langley

Shorter

PJ.

A statistical energy analysis subsystem formulation using finite element and periodic structure theory. J Sound Vib 2008; 318(4-5): 1077–1108.

16.

Delany

Bazley

EN.

Acoustical properties of fibrous absorbent materials. Appl Acoust 1970; 3(2): 105–116.

17.

Brouard

Lafarge

Allard

JF.

A general method of modelling sound propagation in layered media. J Sound Vib 1995; 183(1): 129–142.

18.

Form-X 2018 User’s Guide. ESI Group; 2018.

19.

The MathWorks Inc. MATLAB Version 2014b. 2014.

20.

VA One 2012 User’s Guide, The ESI Group, September, 2012.

21.

Loh

. On Latin hypercube sampling. Ann Stat 1996; 24(5): 2058–2080.

22.

Beck

MW.

NeuralNetTools: visualization and analysis tools for neural networks. J Stat Softw 2018; 85(11): 1–20.

23.

ISO 717-1. 2013. Acoustics—rating of sound insulation in buildings and of building elements—part 1: airborne sound insulation.

24.

Deb

Pratap

Agarwal

, et al. A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Trans Evol Comput 2002; 6(2): 182–197.