Kriging Dimension Reduction Method for Reliability Analysis in Spring Design

스프링 설계문제의 신뢰도 해석을 위한 크리깅 기반 차원감소법의 활용

  • 강진혁 (한국항공대학교 항공우주 및 기계공학과) ;
  • 안다운 (한국항공대학교 항공우주 및 기계공학과) ;
  • 원준호 (한국항공대학교 항공우주 및 기계공학과) ;
  • 최주호 (한국항공대학교 항공우주 및 기계공학부)
  • Published : 2008.04.17

Abstract

This study is to illustrate the usefulness of Kriging Dimension Reduction Method(KDRM), which is to construct probability distribution of response function in the presence of the physical uncertainty of input variables. DRM has recently received increased attention due to its sensitivity-free nature and efficiency that considerable accuracy is obtained with only a few number of analyses. However, the DRM has a number of drawbacks such as instability and inaccuracy for functions with increased nonlinearity. As a remedy, Kriging interpolation technique is incorporated which is known as more accurate for nonlinear functions. The KDRM is applied and compared with MCS methods in a compression coil spring design problem. The effectiveness and accuracy of this method is verified.

Keywords