STOCHASTIC COST ESTIMATING IN REPOSITORY LIFE-CYCLE COST ANALYSIS

The conceptual development, the design, and the final construction and operation of a nuclear repository span many decades. Given this lengthy time frame, it is quite challenging to obtain a good approximation of the repository life-cycle cost. One can deal with this challenge by using an analytic method, the method of moments, to explicitly assess the uncertainty of the estimate. The Gram- Charlier series expansion is used to approximate the uncertainty distribution of the cost estimate. In this paper, the moment methodology is derived and is illustrated through a numerical example. The range of validity of the Gram-Charlier approximation is discussed The method of moments is compared to the traditional stochastic costestimating methods and found to provide more and better information on cost uncertainty. The two methods converge to identical results as the number of convolved variables increases and approaches the range where the the central limit theorem is valid.