Optimization Of Vehicle Aerodynamic Drag Reduction Based On Discrete Adjoint Method And Agent Model

The commonly used optimization methods of automobile aerodynamic drag reduction,such as cut-and-try method,approximate model method and discrete adjoint method,all have deficiencies.To solve this problem,the discrete adjoint method and the approximate model method were combined to provide a fast and effective optimization strategy for the aerodynamic drag reduction optimization of automobiles.In this paper,a production SUV was taken as the research object to conduct CFD modeling and simulation,CFD model accuracy verification and simulation results analysis,aerodynamic drag reduction optimization and analysis.The main contents are as follows:1)Reasonably simplified the geometric model of the vehicle,such as removing screws,wire harness and other small structures,and selected the appropriate volume mesh generation strategy to ensure the simulation accuracy and efficiency;According to the wind tunnel test conditions,the simulation boundary conditions and solution parameters were set appropriately.The wind resistance coefficient and the residual value fluctuated little,indicating that the simulation results converged.2)The results of wind tunnel test were used to verify the accuracy of the CFD model,and the accuracy was reliable.The simulation results of flow field variables such as pressure,velocity and turbulent kinetic energy were analyzed in detail,and it was found that there was an abnormal low pressure area on both sides of the rear protection and a large area of high turbulent kinetic energy in the rear of the vehicle.3)Discrete adjoint method was used to identify the sensitivity of automobile exterior surface with respect to wind resistance coefficient.The free form deformation technique was used to parameterize the high sensitivity region to obtain the design variables.The sample space of the design variables was constructed by using the Hammersley experiment design method and the wind drag coefficients of the sample point models were calculated.Kriging interpolation method was used to construct the approximate model of design variables on the wind drag coefficient.Multi-island genetic algorithm was selected to optimize the approximate model globally.The optimization results showed that under the condition of small change of optimal position,the value of wind drag coefficient decreased by 3.29%,and the drag reduction effect was obvious.Combined with the automation application of simulation software and batch processing technology of operating system,the above optimization process was automated,which saved a lot of time for the aerodynamic optimization task.By comparing and analyzing the simulation results before and after optimization,the physical mechanism of wind resistance coefficient reduction was revealed.