Ultrathin acoustic absorbing metasurface based on deep learning approach

Abstract

Acoustic metasurface has become one of the most promising platforms for manipulating acoustic waves with the advantage of ultra-thin geometry. The conventional design method of acoustic metasurface relies on numerical, trial-and-error methods to retrieve effective properties of the locally resonant unit cells. It is often inefficient and requires significant efforts to investigate the enormous number of possible structures with different physical and geometric parameters, which demands huge computational resources. This is especially when modeling narrow cavities where thermoviscous loss has to be considered. In this paper, a deep learningbased acoustic metasurface absorber modeling approach is introduced to significantly reduce the characterization time while maintaining accuracy. Based on a convolution neural network (CNN), the proposed network can model wide absorption spectrum response in the timescale of milliseconds. The performance of the implemented network is compared with other classical machine learning methods. Using CNN, we have demonstrated an ultrathin metasurface absorber having perfect absorption at an extremely low frequency of 38.6Hz with an ultrathin thickness down to λ/684 (1.3cm). The total path length for the propagating waves inside the channel is about λ/5.7 which breaks the quarter-wavelength resonator theory. The network prediction is validated using the experiments to demonstrate the effectiveness of this physical mechanism. Furthermore, we propose a broadband low-frequency metasurface absorber by coupling unit cells exhibiting different properties based on the supercell concept. This approach is attractive for applications necessitating fast on-demand design and optimization of a metasurface acoustic absorber.

A machine learning accelerated inverse design of underwater acoustic polyurethane coatings

“…Traditionally, this gap is minimized using empirical trial-and-error methods in conjunction with prior domain knowledge. With technological advancements and increased computational power, efficient optimization algorithms and data-driven techniques, such as machine learning (ML) are employed to automate the learning process while utilizing physics-based understanding (Bacigalupo et al 2020;Donda et al 2021;Ahmed et al 2021;Sun et al 2021;Gurbuz et al 2021;Zheng et al 2020;Wu et al 2021;Bianco et al 2019;Wu et al 2022) through physical models.…”

A machine learning accelerated inverse design of underwater acoustic polyurethane coatings with cylindrical voids

“…Traditionally, this gap is minimized using empirical trial-and-error methods in conjunction with a prior domain knowledge. With technological advancements and increased computational power, efficient optimization algorithms and data-driven techniques, such as machine learning (ML) can be employed to automate the learning process while utilizing the physics-based understanding [19,20,21,22] through physical models.…”

Metahouse: noise-insulating chamber based on periodic structures

“…In addition to the realization of negative refractive index [5], superlenses [6,7], holograms [8], and acoustic cloaks [9], recent advances include the development of non-reciprocal systems [10], topological insulators [11,12], nonlinear [13], tunable [14], coding [15], and programmable metasurfaces [16]. Acoustic metasurfaces were also explored as potential platforms for analog computing [17] and, vice versa, advances in computer science and artificial intelligence boost design procedures to achieve desired properties of metamaterials and metasurfaces [18][19][20][21]. Metamaterials can be also used as a platform to explore analogies of the quantum concepts, such as Hall effect [22,23], spin properties [24][25][26][27], skyrmions [28], and twistronics [29].…”