Font Size: 

Reliability-based  design  optimization  (RBDO)  is  an  appropriate  design  optimization  method considering  the  uncertainties.  Accuracy  and  efficiency  are  two  important  issues.  In  this  research, polynomial chaos expansion (PCE) meta-model is used to approximate the expensive computational model  (e.g.  a  finite  element  model).  The  probability  of  failure  is  computed  through  importance sampling  which selects sample from a larger  stationary  sample set.  By using the new importance sampling, the method can not only calculate the accurate failure probabilities, but also avoid a large number  of  model  estimations  as  it  only  needs  one  simulation  run  during  the  optimum  design process. Thus, the trade-off of accuracy and efficiency of RBDO can be obtained for high nonlinear limit-sate function and time-consuming model.

computation; modeling;simulation