Application of Response Surface Model and Kriging Model to ADI Lower Control Arm Optimization Xue Guan Song School of Mechanical Engineering, Dong-A University, Busan 604-714, Korea songxguan@gmail.com Han Seok Park School of Mechanical Engineering, Dong-A University, Busan 604-714, Korea hazelut83@nate.com Kwon Hee Lee Department of Mechanical Engineering, Dong-A University, Busan 604-714, Korea leekh@dau.ac.kr Young Chul Park Department of Mechanical Engineering, Dong-A University, Busan 604-714, Korea parkyc67@dau.ac.kr Abstract Simulation-based surrogate models have been used for a variety of application in automotive industry. In this paper, based on the FEM analysis, both of the response surface model (RSM) and kriging model are used to optimize an ADI control arm, where the weight of the lower control arm is considered as the design objective, and the maximum allowable von-Mises stress the constraint objective. The initial FEM analysis shows the stress distribution and maximum stress on the lower control arm under a very severe loading condition, by virtue of the result of FEM analyses, fifty simulations with six design variables are performed for RSM and kriging model to construct the approximation of the weight and maximum stress to obtain the optimum result. In addition, the optimized results obtained by using RSM and KRG are confirmed by the verified FEM analysis. The result shows that both of them can get the optimum result well. 1. Introduction In automotive suspension, a control arm or sometimes known as wishbone or A-arm, is a car suspension component that is located between the frame and rear axle housing to carry brake and driving torque. The control arms primary function is to manage the wheel’s motion in relation to the vehicle’s body. With the growth of lightweight demand of subassembly for cost saving and fuel efficiency, the design of control arm with less weight and stronger mechanical performance becomes necessary and popular. To this end, finite element method is performed extensively in the design of control arm to predict its mechanical performance. However, for the optimization of complex subassembly unit such as a control arm, using high fidelity FEM analysis becomes too computationally expensive, due to the complex shape and many design variables. One alternative method is to construct a simple approximation model of complicated FEM analyses. The approximation model which is obtained in approximation with the statistical method is also called as the metamodel. The popularly used approximation model/metamodels are the response surface method and kriging model due to their simplicity and ease of use. In 2007, Liao et al. [1] presented a multiobjective optimization procedure for crash safety design of vehicles using response surface method. In the same year, Lee and Kang [2] presented a structural optimization of an automotive door using kriging model. The researchers adopted different types of metamodels to solve the optimization problems. However, Simpson (1998) el at. [3], Yang el at. (2001) [4] and F. A. C. Viana (2008) el at. [5] have revealed that the best surrogate does not necessarily leads to the best solution, multiple surrogates rather than a single one will works well for engineering optimization. This paper presents an engineering optimization of control arm made from ADI material using both RSM and kriging model. A control arm with complex shape under a typical loading condition is investigated using FEM analysis initially. And then RSM and kriging model are performed to predict the optimum result and corresponding values of design variables. The verification analysis is conducted to show the accuracies of these two metamodels. 2009 International Joint Conference on Computational Sciences and Optimization 978-0-7695-3605-7/09 $25.00 © 2009 IEEE DOI 10.1109/CSO.2009.24 1013 2009 International Joint Conference on Computational Sciences and Optimization 978-0-7695-3605-7/09 $25.00 © 2009 IEEE DOI 10.1109/CSO.2009.24 1013 2009 International Joint Conference on Computational Sciences and Optimization 978-0-7695-3605-7/09 $25.00 © 2009 IEEE DOI 10.1109/CSO.2009.24 1013 2009 International Joint Conference on Computational Sciences and Optimization 978-0-7695-3605-7/09 $25.00 © 2009 IEEE DOI 10.1109/CSO.2009.24 1013