HergNet: a Fast Neural Surrogate Model for Sound Field Predictions via Superposition of Plane Waves

TL;DR

HergNet is a neural network model that efficiently predicts sound fields by inherently satisfying physical laws, outperforming existing methods especially at mid to high frequencies in acoustics simulations.

Contribution

The paper introduces HergNet, a neural network architecture that automatically enforces the Helmholtz equation, providing physically valid solutions for wave phenomena in a computationally efficient manner.

Findings

Outperforms state-of-the-art room acoustics simulation methods

Effective in mid to high frequency ranges

Ensures physical validity of predictions

Abstract

We present a novel neural network architecture for the efficient prediction of sound fields in two and three dimensions. The network is designed to automatically satisfy the Helmholtz equation, ensuring that the outputs are physically valid. Therefore, the method can effectively learn solutions to boundary-value problems in various wave phenomena, such as acoustics, optics, and electromagnetism. Numerical experiments show that the proposed strategy can potentially outperform state-of-the-art methods in room acoustics simulation, in particular in the range of mid to high frequencies.