Meta-model based optimization of spot-welded crash box using differential evolution algorithm - Optimization study

Aug 26, 2026

Uploaded by Liam

This study improves the energy absorption performance of a spot-welded crash box through optimization. Two design parameters are selected: the number of crush initiators and the number of spot welds. The objective maximizes absorbed energy, while a constraint limits the upper bound of the maximum initial peak force to support occupant safety. A DOE full-factorial approach builds polynomial surrogate functions, and a Pascal-based Differential Evolution method solves the optimization problem with the finite-element responses.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages7

Reading Time18 min

AccessFree to download

Introduction

Finite Element Model

Optimization Methodology

DOE and Polynomial Surrogates

Differential Evolution Solver

Results and Conclusions

Meta-model based optimization of spot-welded crash box using differential evolution algorithm Ahmet Serdar Önal1 , Necmettin Kaya2 1Beyçelik Gestamp Kalip ve Oto Yan San. Paz. ve Tic. A.Ş, Bursa, Turkey. 2 Uludağ University, Mechanical Engineering Department, Görükle 16059 Bursa, Turkey Abstract The objective of this study is to improve the energy absorption performance of spot-welded crash box using optimization techniques. The number of crush initiators and the number of spot welds were selected as design parameters for optimization study. Maximization of absorbed energy is selected as objective function and the upper bound of maximum initial peak force is defined as constraint. DOE study has been performed with full factorial sampling method, and then polynomial objective and constraint functions were defined to approximate the responses, such as absorbed energy and initial reaction force. The geometry was prepared using CATIA software. The finite element model of crash box and boundary conditions were prepared using Hypermesh software and the model was then transferred to Ls-Dyna in order to solve the finite element model. A Pascal code based Differential Evolution method was developed for solving optimization problem. Developed optimization software was tested with two test functions, then, optimization problem has been solved. It is concluded that, by introducing crush initiator geometries and optimization studies, performance of energy absorption of thin-walled and spot welded crash box has been improved. Not only using crush initiator geometry to control the crushing process is enough but also optimization is necessary for improving energy absorbing of thin-walled structures. 1 Introduction Thin-walled tubular structures behind the bumpers of vehicles protect passengers and the structure during the impact. The bumper deforms first, then the following component deforms until the all energy is absorbed. Thin-walled structures absorb most of the crash energy with a progressive folding deformation. Generally, one or more crash initiators are used to initiate progressive folding of the tube. After first contact during collision, the axial force first reaches an initial high peak, followed by a drop and then fluctuates. The process is known as crushing. The reaction force has a typical plot with regular peak and valleys, each of them corresponding to the formation of a fold [1] . This force peak induces a high deceleration value that should not exceed a survival limit and that should be kept in any case as low as possible to improve the passenger safety. The mean value of the reaction force is related to the total absorbed energy by means of the total crushing length. It should be kept higher than possible to increase the energy absorbing capability of the analyzed component. The initial peak force is less important from the viewpoint of energy absorption but it should be avoided for the sake of occupant safety. For improved frontal car safety, it is necessary to design a structure that absorbs enough energy but at the same time maximum reaction force should be kept as minimum as possible. The main goal of this study is to maximize the energy absorption performance of spot-welded crash box using optimization techniques. Maximization of absorbed energy is selected as objective function and the upper bound of maximum initial peak force is defined as constraint function. Design of experiment (DOE) study has been performed with full factorial sampling method, and then polynomial objective and constraint functions were defined to approximate the responses, such as absorbed energy and initial reaction force. A Pascal code based Differential Evolution method was developed for solving optimization problem. Developed optimization software was tested with two test functions, then, optimization problem has been solved. 2 Finite Element Model In order to find optimal crashworthiness of structure, nonlinear finite eleme

FAQs

What design parameters are optimized in the crash box study?

The optimization uses the number of crush initiators and the number of spot welds as the design parameters.

How are objective and constraint functions defined?

Absorbed energy is maximized as the objective, while the upper bound of the maximum initial peak force is imposed as a constraint.

How is the optimization problem solved?

A meta-model approach generates polynomial approximations from a full-factorial DOE, and a Pascal-based Differential Evolution method is developed to solve the resulting optimization problem.

Aug 26, 2026

Uploaded by Liam