Aerodynamic noise optimization at the front side window area of DrivAer model based on firefly algorithm

Abstract

Vehicle shape optimization through parametric modeling is an important means to reduce vehicle interior aerodynamic noise. A feasible and efficient simulation and optimization method to reduce vehicle interior aerodynamic noise was developed in this paper. A DrivAer model was taken as the research object with a focus on its front left side window area, and a corresponding subdomain model of this area was constructed. The incompressible Detached Eddy Simulation method was applied to obtain the convective pressure fluctuation on the side window surface, the acoustic pressure fluctuation was calculated with the acoustic perturbation equation, then the statistical energy analysis was applied to calculate the interior noise. The simulated result shows good agreement with the experimental data. Afterward, the rearview mirror and A-pillar section around the front left side window was selected as the optimization area, vehicle interior noise as the optimization objective, then the base model was constructed. The optimal Latin hypercube design was used to sample, and the free mesh deformation technique to parameterize the sampling model and the corresponding sound pressure level of vehicle interior was calculated. The Response Surface Method was applied to create an approximation model and the firefly algorithm for global optimization. With this optimization method, a noise reduction of 1.5 dB was achieved at the high frequency range of 4–5 kHz in the vehicle interior, while the overall sound pressure level decreases of 0.45 dB. Through comparison of the optimized result with which optimized based on particle swarm algorithm, the firefly algorithm shows a definite advantage in terms of interior overall sound pressure level and noise level at the specific critical frequency band 2–5 kHz.