Abstract:The Helmholtz equation has been used for modelling the sound pressure field under a harmonic load. Computing harmonic sound pressure fields by means of solving Helmholtz equation can quickly become unfeasible if one wants to study many different geometries for ranges of frequencies. We propose a machine learning approach, namely a feedforward dense neural network, for computing the average sound pressure over a frequency range. The data is generated with finite elements, by numerically computing the response of the average sound pressure, by an eigenmode decomposition of the pressure. We analyze the accuracy of the approximation and determine how much training data is needed in order to reach a certain accuracy in the predictions of the average pressure response.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **An efficient method for calculating the average sound pressure in different frequency ranges**. Specifically, the author focuses on how to approximately solve the Helmholtz equation (Helmholtz equation) through machine learning methods (especially feed - forward dense neural networks) to avoid the high computational cost of traditional numerical methods when dealing with complex geometries and multiple frequency ranges. ### Problem Background 1. **Helmholtz Equation and Sound Pressure Field Modeling** - The Helmholtz equation is often used to model the sound pressure field under harmonic loads. - The traditional method for calculating the harmonic sound pressure field is by solving the Helmholtz equation, but this becomes infeasible when studying a variety of different geometries and frequency ranges. 2. **Computational Challenges** - For complex geometries, it is necessary to decompose them into multiple sub - domains and analyze each sub - domain. - Repeatedly solving the Helmholtz equation to obtain responses at different frequencies is very time - consuming. - The most expensive part in the finite element method (FEM) is calculating the eigenfrequencies and eigenfunctions of each sub - domain. ### Solution The author proposes a machine - learning - based method, that is, using a feed - forward dense neural network (feedforward dense neural network, MLP) to calculate the average sound pressure in the frequency range. The specific steps are as follows: 1. **Data Generation** - Use the finite element method to generate training data and calculate the average sound pressure response through eigenmode decomposition. 2. **Neural Network Design** - Construct a feed - forward neural network with three hidden layers and 128 nodes in each layer, and the activation function is ReLU. - Use the ADAM optimizer for training, and the gradient descent step size adopts a polynomial decay schedule. 3. **Performance Evaluation** - Analyze the accuracy of the model and determine the amount of training data required to reach a specific accuracy. - Compare the performance of the linear regression method with the proposed neural network method and verify the superiority of the latter on non - linear sound pressure functions. ### Application Scenarios This method is particularly suitable for standardized frequency ranges in engineering applications, such as the frequency ranges in ISO standards. In addition, this method can be applied to the frequency response analysis of elastic bodies and fluid - structure interaction problems. ### Formula Summary - **Helmholtz Equation** \[ -\Delta p(x)-\frac{\omega^{2}}{c^{2}}p(x) = f(x) \] where \(p(x)\) is the sound pressure, \(\omega\) is the angular frequency, \(c\) is the speed of sound, and \(f(x)\) is the external force. - **Average Sound Pressure** \[ \langle p_{\lambda}\rangle=\frac{1}{|\Omega|}\int_{\Omega}p_{\lambda}\,dx \] - **Objective Function** \[ \Psi=\frac{1}{\lambda_{\text{max}}-\lambda_{\text{min}}}\text{p.v.}\int_{\lambda_{\text{min}}}^{\lambda_{\text{max}}}\langle p_{\lambda}\rangle\,d\lambda \] Through this method, the author aims to provide a more efficient and accurate method for calculating the average sound pressure, especially when dealing with complex geometries and multiple frequency ranges.

A feedforward neural network for modelling of average pressure frequency response

A Data-Driven Method to Identify the Probability Density Expression of Nonlinear System under Gaussian White Noise and Harmonic Excitations

Neural network based approach for solving problems in plane wave duct acoustics

Helmholtz-equation Solution in Nonsmooth Media by a Physics-Informed Neural Network Incorporating Quadratic Terms and a Perfectly Matching Layer Condition

International recruitment of nurses: policy and practice in the United Kingdom.

Induction of sperm head abnormalities by incorporated radionuclides: dependence on subcellular distribution, type of radiation, dose rate, and presence of radioprotectors.

Hard-constrained neural networks for modelling nonlinear acoustics

[Primary B cell lymphoma of the skull following head trauma; a case report].

A feedforward neural network framework for approximating the solutions to nonlinear ordinary differential equations

Neural network-based estimation of biomechanical vocal fold parameters

Multiharmonic analysis for nonlinear acoustics with different scales

Approximation of acoustic black holes with finite element mixed formulations and artificial neural network correction terms

Fast prediction of aerodynamic noise induced by the flow around a cylinder based on deep neural network

Using learned priors to regularize the Helmholtz equation least-squares method

A Deep Neural Network Based Glottal Flow Model for Predicting Fluid-Structure Interactions during Voice Production

Wavefield solutions from machine learned functions

Modelling pressure-Hessian from local velocity gradients information in an incompressible turbulent flow field using deep neural networks

Fourier Neural Operator for Fluid Flow in Small-Shape 2D Simulated Porous Media Dataset

Learning based numerical methods for acoustic frequency-domain simulation with high frequency

Error approximation and bias correction in dynamic problems using a recurrent neural network/finite element hybrid model

A generalized computational approach to predict high-frequency acoustic pressure response of cavity structures for structural health monitoring of wind turbine blades