Genetically aerodynamic optimization of the nose shape of a high-speed train entering a tunnel

The optimization of the nose shape of a high-speed train entering a tunnel has been performed using genetic algorithms (GA). This optimization method requires the parameterization of each optimal candidate as a design vector. The geometrical parameterization of the nose has been defined using three design variables that include the most characteristic geometrical factors affecting the compression wave generated at the entry of the train and the aerodynamic drag of the train. A large set of three-dimensional, turbulent, compressible, unsteady simulations of realistic train models have been done, and this information has been used to fit a metamodel. The metamodel is used by the GA to evaluate each optimal candidate in a more efficient way. The optimal designs that minimize the maximum pressure gradient and the aerodynamic drag are in good agreement with the literature. To complete this single-objective optimization, a multi-objective optimization has been developed, and a Pareto front has been obtained. The use of metamodels has permitted to analyze the influence of each design variable.