Design for Vehicle Structural Crashworthiness Via Crash Mode Matching.

Vehicle Crashworthiness;Design Optimization;Crash Modes;Crash Mode Matching;Mechanical Engineering;Engineering;Mechanical Engineering

Hamza, Karim T.Washabaugh, Peter D. ;

University of Michigan

关键词: Vehicle Crashworthiness;    Design Optimization;    Crash Modes;    Crash Mode Matching;    Mechanical Engineering;    Engineering;    Mechanical Engineering;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/60666/khamza_1.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Vehicle crashworthiness is an important design attribute which designers strive to improve. However, design for structural crashworthiness is a difficult task. A vehicle structure must have strength to shield the passenger compartment, as well as compliance to cushion the impact energy. The best known analysis method for crashworthiness performance prediction is nonlinear finite element (FE). FE analysis of detailed vehicle models requires enormous computational resources thereby hindering the success of general-purpose optimization approaches. An approach which is more of an art than a formal procedure is that of crash mode matching. Qualitatively, the crash mode is the observed gross-motion of the structure and its time history of deformation in various zones. Crash mode matching involves adjusting the design variables of the structure in order to achieve a desirable structural deformation history, which an experienced designer can typically do while requiring only a few trial FE runs. This dissertation to develops an algorithmic methodology by formalizing this crash mode matching approach.A quantitative representation of the crash mode is introduced as a matrix of time series of the structural deformation history, with dimensions of the matrix being the structural location and type of deformation. A comparison metric is then introduced for the degree of matching between crash modes as the integral of the error between the time series. An automated algorithm for crash mode matching heuristically directs stochastic sampling of the design space by adjusting the mean and standard deviation of normal distributions on the design variables. Adjustment of the mean and standard deviation is performed via Fuzzy logic rules that are defined by the algorithm user in analogy to the type of decisions that an experienced designer would make. Stochastic search allows for global convergence properties, as well as accounting for different expert designers sometimes having different opinions on how to modify a design. Implementation of the proposed framework is applied to two real-life case studies involving front half of a vehicle, as well as full vehicle models. The studies show success in attaining high performance designs, while requiring a modest number of FE runs, hence reasonable computational resources.