Aerodynamic Shape Optimization of Train Heads using Adjoint-Based Computational Fluid Dynamics with different Objective Functions

Abstract

Aerodynamic design of modern high-speed trains requires efficient optimization stategies for train specific operating conditions and objective functions. Adjoint-based shape optimization using computational fluid dynamics has become a preferred optimization technique for vehicles, as it is the only method which allows rapid optimization for extremely large numbers of surface parameters. Based on the continuous adjoint formulation for the computation of sensitivities of incompressible, steady-state, ducted flows, we will describe and introduce an iterative, CAD-free, continuous, adjoint-based shape optimization procedure using gaussian filtered sensitivities and mesh morphing with radial basis function interpolation for the optimization of the front part of the model of a conceptual, generic high-speed train with respect to drag and pressure wave. The presented results will prove the efficiency of the developed process chain. We will show that the sidewise extension of the vehicle plays an important role for the intensity of the generated pressure wave. In contrast, the most sensitive areas and most significant modifications during drag optimization could be found in the front part of the nose tip section.