Acoustic shape optimization using energy stable curvilinear finite differences

Gustav Eriksson,Vidar Stiernström

2024-05-08

Abstract:A gradient-based method for shape optimization problems constrained by the acoustic wave equation is presented. The method makes use of high-order accurate finite differences with summation-by-parts properties on multiblock curvilinear grids to discretize in space. Representing the design domain through a coordinate mapping from a reference domain, the design shape is obtained by inverting for the discretized coordinate map. The adjoint state framework is employed to efficiently compute the gradient of the loss functional. Using the summation-by-parts properties of the finite difference discretization, we prove stability and dual consistency for the semi-discrete forward and adjoint problems. Numerical experiments verify the accuracy of the finite difference scheme and demonstrate the capabilities of the shape optimization method on two model problems with real-world relevance.

Numerical Analysis

What problem does this paper attempt to address?

### What problem does this paper attempt to solve?

This paper aims to solve the **acoustic shape optimization problem**, specifically the shape optimization problem constrained by the acoustic wave equation. The following are the core objectives and methods of the paper:

1. **Problem background**:
- Shape optimization problems are very important in research and engineering applications, especially in the field of computer - aided optimal design.
- The acoustic wave equation is a partial differential equation (PDE), and its solution depends on the geometric shape of the solution domain.

2. **Specific problems**:
- The paper proposes a gradient - based method to solve the shape optimization problem constrained by the acoustic wave equation.
- The main task is to minimize a loss function \( J(u, p) \) by adjusting the shape of the solution domain under given initial and boundary conditions, where \( u \) is the solution of the acoustic wave equation and \( p \) is the control variable that determines the shape of the solution domain.

3. **Method overview**:
- Use high - order accurate finite differences to discretize space. These differences have summation - by - parts properties and can maintain stability on multi - block curvilinear grids.
- Simplify the calculation by transforming the design domain from the physical domain to the reference domain through coordinate mapping.
- Efficiently calculate the gradient of the loss function using the adjoint state framework.
- Prove the stability and dual consistency of the semi - discrete forward and adjoint problems.

4. **Innovations**:
- Propose the use of dual - consistent discretization methods, making the discretized optimization problem equivalent to the continuous optimization problem.
- Avoid the remeshing problem caused by poor grid quality and ensure the accuracy of numerical results through regularization methods.

5. **Application scenarios**:
- This method can be applied to practical problems, such as determining the water depth distribution of a lake or ocean section, given the time - series pressure data at the receiver.

### Summary
The main purpose of this paper is to develop an efficient and stable gradient - based shape optimization method for solving optimization problems constrained by the acoustic wave equation. By introducing high - order accurate finite differences and the adjoint state framework, the stability and accuracy of the method are ensured, and its application potential in practical problems is demonstrated.