The approximate reliability analysis methods require accurate sensitivity information. In spite of being exact, the analytical sensitivity analysis is not always feasible. The popular finite difference (FD) method for sensitivity analysis is step-size dependent. Moreover, the calculation of optimal step-size in FD increases computation cost. The complex variable method (CVM) for sensitivity analysis is virtually independent of step-size. Moreover, due to increased computing power, the ease with which calculations can be handled in complex domain makes this method more desirable. Here, CVM is coupled with the first order approximation methods, viz, Saddlepoint and Mean-Value Methods. Examples are presented to demonstrate the application of CVM in RBDO.
Application of Complex Variable Method for Sensitivity Analysis in RBDO Using First Order Methods
Sae Technical Papers
SAE 2005 World Congress & Exhibition ; 2005
2005-04-11
Aufsatz (Konferenz)
Englisch
Application of Complex Variable Method for Sensitivity Analysis in RBDO Using First Order Methods
British Library Conference Proceedings | 2005
|Sensitivity Developments for RBDO With Dependent Input Variable and Varying Input Standard Deviation
Online Contents | 2017
|Design sensitivity method for sampling-based RBDO with varying standard deviation
Online Contents | 2016
|