Surrogate Model Based Liner Optimization for Aeroengines and Comparison With Finite Elements

Abstract

Numerical optimizations are very useful in liner designs for low-noise aeroengines. Although modern computational tools are already very efficient for a single aeroengine noise propagation simulation run, the prohibitively high computational cost of a broadband liner optimization process which requires hundreds of thousands of runs renders these tools unsuitable for such task. To enable rapid optimization using a desktop computer, an efficient analytical solver based on the Wiener–Hopf method is proposed in the current study. Although a Wiener–Hopf-based solver can produce predictions very quickly (order of a second), it usually assumes an idealized straight duct configuration with a uniform background flow that makes it arguable for practical applications. In the current study, we employ the Wiener–Hopf method in our solver to produce an optimized liner design for a semi-infinite annular duct setup and compare its noise-reduction effect with an optimized liner designed by the direct application of a numerical finite element solver for a practical aeroengine intake configuration with an inhomogeneous background flow. The near-identical near- and far-field solutions by the Wiener–Hopf-based method and the finite element solvers clearly demonstrate the accuracy and high efficiency of the proposed optimization strategy. Therefore, the current Wiener–Hopf solver is highly effective for liner optimizations with practical setups and is very useful to the preliminary design process of low-noise aeroengines.