Multi-Objective Aerodynamic Shape Optimization of High Speed Train Nose Using Adaptive Surrogate Model

Abstract

The objective of this work is to demonstrate the possibility of using adaptive surrogate models for optimization problems which require expensive computations. A hybrid GA PSO algorithm is combined with a kriging based surrogate model. The suggested method was used to nd the optimum shape of a two dimensional nose shape of a high speed train traveling at 350 Km/hr considering both the induced aerodynamic drag and the generated aerodynamic noise. Since the prediction of aerodynamic drag and aerodynamic noise requires computational uid dynamic simulations, to limit the number of computer simulations required for optimization, a surrogate model identical to the kriging model was used. The accuracy of the surrogate model is checked using the parameter EIV there by updating the surrogate model whenever necessary. The results show that the combined shape optimization algorithm requires a small number of simulations to identify the optimum shape compared with other methods. The suggested method not only requires a small number of simulations but is also robust. This makes the study on the eect of different weights on the optimum shape feasible without the need for additional simulations. The results show that the nose shape should be slightly short and pointed to get the best aerodynamic performance in terms of induced drag and the nose shape should be slightly long and little bit blunt for the least aerodynamic noise generated. The optimum nose shapes fall between these two shapes based upon the choice of the weights. Regarding the choice of the weights for the given two dimensional test geometry the best compromise would be to choose 50% drag and 50 % noise.