Efficient estimation of structural reliability for problems with uncertain intervals

The uncertain parameters in engineering design may appear as random or intervals based on the information available from the past history or physical experiments. When the random variables are provided with probability distribution functions, suitable methods have to be used in computing the failure probability. Similarly, when there are only bounds or ranges on some of the uncertain variables, failure analysis methods are needed to consider the complete range of uncertain intervals. When dealing with the combination of both random distributions and interval variables, the computational cost involved in estimating the reliability of the structure increases exponentially. Every combination of interval values requires one probabilistic based analysis, which is by itself an expensive procedure. Therefore, to make the problem tractable, use of function approximations is presented in this paper, which reduces the number of actual FEA/computational fluid dynamics (CFD) simulations. The proposed method is demonstrated with structural examples. (C) 2002 Elsevier Science Ltd. All rights reserved.