math.APNov 3, 2026

Solving the Nonlinear Westervelt Equation in Ultrasound Propagation Using Physics-Informed Neural Networks

Authors: Luke Oluwaseye Joel

Organizations: Department of Mathematics and Applied Mathematics, University of Johannesburg, South Africa

Abstract

Modeling the nonlinear acoustic wave propagation in biological tissues needs accuracy, especially in ultrasound imaging and therapeutic applications. The 1D Westervelt equation, pxx−ptt+γ pttt+η (p2)tt=0,p_{xx} - p_{tt} + γ\, p_{ttt} + η\,(p^{2})_{tt}=0, could be used to describe a finite-amplitude ultrasound propagation in these tissues. This equation accounts for nonlinearity, ηη, and attenuation, γγ. To solve this equation, we observe that traditional solvers may struggle to do so correctly. This is due to the difficulties arising from the nonlinear nature of the equation. We solve the equation by defining three synthetic cases - linear (η=0.001η=0.001, γ=10−6γ=10^{-6}), shock-dominated (η=0.02η=0.02, γ=10−6γ=10^{-6}), and dissipation-dominated (η=0.01η=0.01, γ=5×10−4γ=5\times10^{-4}) - and using Physics-Informed Neural Networks (PINNs) to showcase the solution. A sinusoidal activation function is used to capture the oscillatory behavior of the problem. The results of our study, with PDE residual converging to at least 10−610^{-6}, boundary condition loss to 10−510^{-5}, and the PINN loss converging to 10−510^{-5} for each case at 30,00030,000 epochs, show that γγ scales the rate of amplitude decay, while ηη governs frequency-dependent phase speeds.

Figures & tables

Explore similar work

CardsList
  1. Linearized PINN with pretrained nonlinear layers

    Sep 14, 2026Wenhao Chen, Alexandre M. TartakovskyEfficient Neural Network InferencePDE Solving

  2. A Green-Integral-Constrained Neural Solver with Stochastic Physics-Informed Regularization

    Apr 23, 2026Mohammad Mahdi Abedi, David Pardo, Tariq AlkhalifahNeural PDE SolversPDE Solving