Optimization Study On Crashworthiness Of Automobile Thin-walled Structures With Functionally Graded Properties

The 22MnB5 high-strength steel structure with functionally graded properties is simultaneously customized for variable strength and variable thickness according to the working conditions requirements to make full use of the material.So it possesses excellent crashworthiness and light-weight potential.The manufacture can be achieved by controlling the cooling rate of the tailored rolling blank(TRB)in hot forming(HF)process.Therefore,the structure has certain application value due to its great advantages in performance and cost.Strength and thickness factors are simultaneously designed to optimize the crashworthiness leads to more complicated problems.Therefore,it is essential to ensure the accuracy of design,efficiency of optimization,and applicability of actual engineering.So the optimization strategy has certain research significance.In this paper,the 22MnB5 steel sheet with different quenched mechanical properties is prepared by adjusting the process parameters during the customized HF process.The Voce constitutive model which describes the elastoplastic mechanical behavior of different hardened 22MnB5 steels is introduced and its applicability is verified by comparing the mechanical properties obtained through tensile tests.Subsequently,the samples with different stress triaxial dimensions are designed and manufactured to carry out tensile tests and simulations.Combined with the tests and simulations,the GISSMO damage failure criterion was established to describe the mechanical behavior of the failure stage.Then the crashing model of the functionally graded properties single-hat thin-wall structure is established on the Hypermesh platform.And the reliability of the established CAE model is verified by comparing simulation and experiment.Finally,the crashworthiness of functionally graded properties thin-walled structures is systematically optimized.In the optimization process,the functionally graded properties distribution function is established,and the optimization problem is defined.The optimal Latin square design(OLHD)method is used for experimental design,and the radial basis neural network(RBFNN)meta-model is established.The multi-objective particle swarm optimization algorithm(MOPSO)is used to globally carry out multi-objective optimization.Then the engineering error is introduced into the optimization design in normal probability distribution,and the reliability of the optimal design of the functionally graded properties structure is evaluated.The optimal design scheme to meet the requirements of different working conditions is further given.It is found that the crashworthiness and lightweight performance of the functionally graded properties structure are better than the functionally graded strength or functionally graded thickness structure.The optimal design is that the strehgth exhibits a three-stage graded distribution and the thickness exhibits a uniform graded distribution.The optimized design method in the process has high efficiency,accuracy and reliability.