Research On Frontal Crashworthiness Optimization Of Autobody

Research on Frontal Crashworthiness Optimization of AutobodyTraffic accidents are severe social problems all around the world and especially in China. Although the number of cars per person is quite low in China, the number of fatalities is always high. Vehicle passive safety is one of the key issues in automobile industry. Simulation technology of impact and crash has been widely used in crashworthiness analysis and design, which help to improve the safety of vehicle significantly. With the rapid development of impact and crash simulation, seeking optimal design for automotive crashworthiness and safety shows particular importance and practical significance. The autobody structure, which mainly influences the total crashworthiness, should have the capability to absorb enough crash energy so that the passengers are not fatally wounded.A systematic study is performed to understand the dynamics of collisions involving a bumper beam structure and a curved frontal rail. To design an improve crash energy management system to reduce the severity of collisions, the optimization methods and processes are discussed in the dissertation:First of all, the nonlinear finite element theories are introduced and investigated, including governing equations for lagrangian formulation, elements of automotive crash models, time integration and single surface contact algorithms. They are very important for FEA in vehicle crash, and follow-up works in this paper are based on these theories.When topological modification the number of DOFs is increased, it is necessary to establish the condensed equation by the Guyan reduction such that the new degrees of freedom are included in the analysis model. A new method for the static reanalysis based on Lanczos algorithm is presented. In order to accelerate the speed of convergence, the existing decomposed stiffness matrix of initial design as the preconditioner is used to solve the equations. The computational cost is significantly reduced by the proposed method. The numerical results also show the exactitude and effectiveness of the reanalysis algorithm. This work is the foundation of the follow works on the nonlinear analysis and topology optimization problems.Despite the advances in computer technology, the enormous computational cost associated with the complex nonlinear crashworthiness analysis renders it to be impractical to rely exclusively on computer simulations for crashworthiness design. Therefore, for the crashworthiness optimization design problems, how to avoid the sensitivity analysis of large complex systems becomes the key step. In the dissertation, the SED (strain-energy-density) method is put forward to solve the complexity problem of sensitivity analysis. This method is based on static analysis of strain energy analysis. By transforming the dynamic loads into different static loads, the strain energy density can be calculated to measure the load-bearing efforts of local subdomains. Based on this method, the subdomains of the rail with different stiffness characteristics can be separated feasibly. Also, a multidomain topology optimization technology based on the standard topology optimization method is discussed in the dissertation. This kind of topology technique enables the effective design of a complex engineering structure by allowing the designers to control the material distribution among the subdomains during the optimal design process. The frontal rail structure, which mainly affects the frontal crashworthiness characteristic, is analyzed by the proposed SED and multidomain topology methods to obtain the reasonable material distribution.The fully crush deformation of straight part on the frontal rail is the main influence on the energy absorption behavior. The trigger size, trigger shape, and trigger orientation, they both affect the energy absorption of the rail. In this research, small size and square-shape triggers are located at the corners of the frontal straight part. This type of triggers has the highest manufacturing feasibility, and most of all, this location is conformed to the analysis results of the material distribution. The reinforcement plates are usually placed inside of the reinforced components. Also, their shapes are similar and the thicknesses are almost the same. Thereby, by considering the lightweight demands, the reinforcement plates which placed on the positions of local plastic hinges could avoid the excessive bending deformation and help the axial collapse performing. The final analysis results are also testified the advantages and efficiency of the proposed methods and processes on improving the crashworthiness characteristics of the frontal rail.Another preferable strategy on nonlinear optimal problem is to employ the computational efficient meta-model in lieu of the expensive simulations to facilitate the optimization process and design concept exploration. During the process of meta-model construction, choosing the proper design samplings is very important to the accuracy of meta-model. Design of experiment (DOE) provides a mean to the selection of samplings in the design domain, and the fundamental aim of this approach is to furthest improve the efficiency and accuracy of design. To develop a perfect meta-model, a reasonable design of experiment to sample the region of interest must be chosen. Therefore, central composite design is applied to distribute the experimental samplings. Central composite design allows the engineers to build a regressive function with more accuracy and better statistical property, just needs fewer experiments by choosing proper design samplings. To construct the meta-model from the observed samplings'data, least squares method and Kriging are utilized, respectively. Finally, for the design characteristics of bumper beam structure, and the error tests, the meta-model constructed by least squares method are chosen for the optimal design.In addition, by obtaining the meta-model, PSO (Particle Swarm Optimization) algorithm is used to solve the crashworthiness optimization problem. The meta-model has been optimized by PSO method after giving initial parameters and transforming the constraint problem into unconstraint problem based on penalty function method. Finally, the optimized results show that the internal energy of structure increases. Meanwhile, the maximal crash force and total mass of structure decrease. Specially, there is no violation of constraint conditions which can prove that the crashworthiness optimization problem has been solved well.