Mechanical and Aerospace Engineering
Department Seminar Series
8/11 (Tuesday) Noon – 1:00 pm RH200
Introduction to Simulation-Based Design under Uncertainties
Ikjin Lee, Ph.D.
Department of Mechanical Engineering
Korea Advanced Institute of Science and Technology (KAIST)
Daejeon, South Korea
In recent years, various methods have been proposed to treat uncertainties in engineering analysis and, more recently, to carry out design optimization with the reliability & robustness requirement, which is called reliability-based design optimization (RBDO) and robustness design optimization (RDO). In this presentation, it is briefly explained what has been researched for design under uncertainties.
In the early stage of reliability analysis and design optimization, there have been many assumptions and approximations such as the linear approximation of the performance functions, fully known input statistical information, statistically independent random variables, no simulation model uncertainty, etc. To resolve the weaknesses of the methods proposed in the early stage of reliability analysis and design optimization, three methods are presented, which are the most probable point (MPP)-based dimension reduction method (DRM) to estimate the probability of failure accurately for highly nonlinear performance functions, Bayesian method to identify input distribution types when only experimental data are given, and copula to model the joint distribution for correlated random variables. In addition, the presentation will discuss robustness estimation methods and recent developments.
Dr. Ikjin Lee is an assistant professor of Mechanical Engineering at Korea Advanced Institute of Science and Technology (KAIST). He received his Engineer and M.S. Diploma from Seoul National University in 2001 and 2003, respectively, and Ph.D. (2008) in Mechanical Engineering from The University of Iowa. From 2011 to 2013, he was at the University of Connecticut as an assistant professor, and joined KAIST in 2013. Dr. Lee’s research focuses on reliability-based design optimization (RBDO), reliability-based robust design optimization (RBRDO), design and analysis using most probable point (MPP)-based dimension reduction method (DRM), system reliability analysis and design optimization, and design under uncertainties with lack of statistical information and correlated input variables. In 2007, he received the International Society of Structural and Multidisciplinary Optimization (ISSMO)/Springer prize for a young scientist.