Adjoint-based Optimization of Liners for Noise Reduction

在这个项目中，我们的目标是开发一个高效和创新的框架来优化声学衬垫，以降低喷气发动机的噪音。特别是，本研究旨在通过采用涉及大量设计参数的空间可变线性参数化来提高线性效率。由于尾管效率在很大程度上取决于源的位置和流动特性，我们的目标是执行多点优化，以考虑设计点的变化。为此，将使用自动微分技术开发基于底层高阶线性欧拉求解器的离散伴随求解器。然后将离散伴随解算器与基于梯度的优化算法相结合，从而能够对声学衬垫进行大规模优化。

In this project, we aim to develop an efficient and innovative framework to optimize acoustic liners in reducing the jet engine noise. In particular, this research is aimed at improving the liner efficiency by employing a spatially variable liner parameterization involving a large number of design parameters. Since the liner efficiency strongly depends on the location of the source and the flow properties, we aim at performing multi-point optimization to account for the changes in the design point. For this purpose, a discrete adjoint solver based on the underlying high-order linearized Euler solver will be developed using Automatic Differentiation techniques. The discrete adjoint solver will then be combined with a gradient-based optimization algorithm to able to perform large-scale optimization of acoustic liners.