Abstract:Physics-informed Neural Networks (PINNs) is a method for numerical simulation that incorporates a loss function corresponding to the governing equations into a neural network. While PINNs have been explored for their utility in inverse analysis, their application in acoustic analysis remains limited. This study presents a method to identify loss parameters in acoustic tubes using PINNs. We categorized the loss parameters into two groups: one dependent on the tube's diameter and another constant, independent of it. The latter were set as the trainable parameters of the neural network. The problem of identifying the loss parameter was formulated as an optimization problem, with the physical properties being determined through this process. The neural network architecture employed was based on our previously proposed ResoNet, which is designed for analyzing acoustic resonance. The efficacy of the proposed method is assessed through both forward and inverse analysis, specifically through the identification of loss parameters. The findings demonstrate that it is feasible to accurately identify parameters that significantly impact the sound field under analysis. By merely altering the governing equations in the loss function, this method could be adapted to various sound fields, suggesting its potential for broad application.

What problem does this paper attempt to address?

This paper discusses the problem of identifying the physical properties of acoustic pipes using Physics-Informed Neural Networks (PINNs). The researchers divided the loss parameters into two categories: one depends on the pipe diameter, and the other is a constant. They set the latter as trainable parameters of the neural network and formalized the problem of identifying the loss parameters as an optimization problem. The proposed method is based on the previously proposed ResoNet architecture and is suitable for analyzing acoustic resonances. PINNs are a numerical simulation method that integrates the loss function of partial differential equations into neural networks and can be used for forward and inverse analyses. Although widely used in inverse analysis, their application in acoustic analysis is relatively limited. The study evaluated the effectiveness of the proposed method through forward and inverse analysis, particularly in identifying key parameters that affect the sound field. The paper introduces the application of PINNs in acoustic pipes, especially in the areas of automotive muffler design, active noise control, duct analysis, and brass wind instrument design. In the research, the sound source is located at one end of the pipe, and the sound pressure is measured at the other end using a microphone to identify the loss parameters. The authors conducted parameter identification through the PINNs structure, including predictions of sound pressure and volume velocity, while considering boundary conditions and radiation conditions. The paper also describes the training process of PINNs, including the composition of the loss function, such as PDE loss, boundary condition loss, coupling loss, and periodicity loss. The experimental results show that PINNs can accurately identify parameters that significantly affect the sound field, but the identification accuracy is lower for certain parameters (such as loss parameters related to viscosity). This may be because different parameters have different degrees of influence on the sound pressure waveform. Despite these limitations, the paper believes that PINNs have broad potential applications in solving inverse problems related to acoustic pipes. In summary, this paper aims to improve the identification of physical properties of acoustic pipes using the PINNs technique, providing new tools and insights for research in related fields.

Identification of Physical Properties in Acoustic Tubes Using Physics-Informed Neural Networks

Physics-informed neural network for acoustic resonance analysis in a one-dimensional acoustic tube

Acoustic Field Visualization and Source Localization Via Physics-Informed Learning of Sparse Data with Adaptive Sampling

Spatial acoustic properties recovery with deep learning

Deep Learning based Spatially Dependent Acoustical Properties Recovery

A physics-informed neural network for sound propagation in the atmospheric boundary layer

Sensitivity analysis using Physics-informed neural networks

Practical Aspects of Physics-Informed Neural Networks Applied to Solve Frequency-Domain Acoustic Wave Forward Problem

Physics-Informed Neural Network for Volumetric Sound field Reconstruction of Speech Signals

On Acoustic Fields of Complex Scatters Based on Physics-Informed Neural Networks

k-space Physics-informed Neural Network (k-PINN) for Compressed Spectral Mapping and Efficient Inversion of Vibrations in Thin Composite Laminates

The Potential for Material Property Characterization Using Physics-Informed Neural Networks and Ultrasonic Wave Data

Spatial domain decomposition-based physics-informed neural networks for practical acoustic propagation estimation under ocean dynamics

Point Neuron Learning: A New Physics-Informed Neural Network Architecture

Deep learning for full-field ultrasonic characterization

Adaptive-Sampling Physics-Informed Neural Network for Viscoacoustic Wavefield Simulation

Physics-informed neural networks for one-dimensional sound field predictions with parameterized sources and impedance boundaries

Loss-attentional physics-informed neural networks

Novel Physics-Informed Artificial Neural Network Architectures for System and Input Identification of Structural Dynamics PDEs

MPIPN: A Multi Physics-Informed PointNet for solving parametric acoustic-structure systems

Physics-informed neural networks in support of modal wavenumber estimation