Aerodynamic Shape Optimization Design Of Vehicle Based On Adjoint Method And Intelligent Optimization Algorithm

Vehicle aerodynamic shape optimization is an important way to improve the performance of automobile,such as power,fuel economy and operation stability.The research on this technology has important engineering practical significance and theoretical value.In the traditional aerodynamic shape optimization process,the selection of optimization parts and optimization methods depends on the engineering experience of engineers.However,due to the complexity of vehicle flow field,the optimization variables found only by engineering experience are often not the most sensitive,and it is easy to ignore those potential effective optimization parts,resulting in large trial and error costs.In addition,the local iteration and accumulation method used in the traditional optimization method is inefficient and can not obtain the global optimal solution that fully considers the interaction of various variables and the contradiction between multiple objectives.In order to avoid the blindness of variable selection,this paper introduces the adjoint method as a fast gradient solution tool,and uses the gradient information to guide the selection of optimization variables.After finding the effective optimization variables,the method based on mesh morph,design of experiment,surrogate model,intelligent optimization algorithm and other intelligent optimization technologies based on CFD(Computational Fluid Dynamics)for multivariable global optimization,so as to form a set of aerodynamic shape optimization process that can be used in the local modification stage of vehicle.Firstly,taking the simplified model of Ahmed vehicle-like body as the research object,the single objective drag reduction optimization research is carried out to verify the effectiveness of the process and lay a theoretical foundation for the follow-up real vehicle application research.In this part,firstly,the numerical simulation model is established,and the simulation results of this paper are compared and verified by using the wind tunnel test results.The comparison results show that the simulation method in this paper has high accuracy.On this basis,the adjoint calculation is carried out with the drag coefficient as the cost function,and the three most sensitive optimization variables are determined according to the adjoint sensitivity information.The mesh morph technology is used to parameterize the variables and realize the rapid geometric reconstruction,and build an automatic computing platform to improve the computing efficiency of sample points.Finally,the EGO(Efficient Global Optimization)algorithm based on the dynamic Kriging surrogate model performs global optimization in the variable space,and the final drag reduction rate is 33.9%.The effect is remarkable,which preliminarily verifies the effectiveness of the optimization process.Secondly,in order to improve the practicability of EGO algorithm,the algorithm improvement research is carried out.The results show that the hybrid point adding strategy combining the respective advantages of EI(Expected Improvement)and MSP(Minimum Surrogate Prediction)point adding criterion can improve the probability of finding a better solution within a limited number of iterations,the parallel hybrid addition strategy can greatly reduce the number of iterations on the premise that the amount of computation remains unchanged,which provides a way for engineering applications to exchange computing resources for computing time.Then,because there is still a certain gap between the simplified model and the real vehicle model,in order to explore the effectiveness of the optimization process applied to the real vehicle model,taking the fastback Aero SUV model as the research object and minimizing the aerodynamic drag and reducing the aerodynamic lift as the optimization goal,the aerodynamic shape optimization design of the upper body and chassis is carried out respectively.Firstly,the initial simulation model conforming to the accuracy is established.Through the analysis of the adjoint sensitivity of drag coefficient and lift coefficient,the mesh morph scheme and chassis spoiler addition scheme are determined,and combining Kriging surrogate model and NSGA-Ⅱ(Non-dominated Sorting Genetic Algorithm II)multiobjective optimization algorithm for optimization.Finally,the drag coefficient decreases from 0.2809 to 0.2527,with a decrease of about 10%,and the lift coefficient decreases from0.145 to 0.0708,with a decrease of about 51%,which realized the comprehensive improvement of the aerodynamic characteristics of the vehicle.The results show that the adjoint method,mesh morph technology,surrogate model and intelligent optimization algorithm can be combined to optimize the aerodynamic shape of the vehicle quickly and effectively,and the work of this paper can provide a reference for the future optimization of the aerodynamic shape of the vehicle.