Advances in Fuel Modeling and Design/Licensing Methodologies by Improved Knowledge and Uncertainty Quantification – AREVA Contribution

Statistical methods involving direct applications of the Monte Carlo sampling technique have been developed in recent years as an alternative to the traditional conservative bounding calculations for fuel rod design and licensing analyses. The LWR best-estimate methodology for mechanical licensing analyses (first of its kind) that was recently developed by AREVA NP and, which was approved by US-NRC for BWR applications, is presented herein. In a nutshell, the new approach consists in quantifying the uncertainties of the relevant analysis outcomes, while the previous methods relied on conservative bounding assumptions and models, in order to compensate for the lack of either phenomenological knowledge or experimental data. In addition, a critical requirement is employing a best-estimate fuel performance code that is well qualified on a comprehensive experimental database, which allows adequate quantification of code model uncertainties. The methodology basis is an uncertainty propagation method. Employing a best-estimate fuel code and a proper definition of uncertainties associated with all significant input variables, which are then propagated through the code, the uncertainties of relevant output parameters are then determined. Non-parametric order statistics is used to estimate a certain quantile of the interest outputs distribution, which is then compared to preset licensing acceptance limit. This technique has the advantage of being less time-consuming than other methods and is also independent of the number of input variables. Finally, the outline of the methodology is presented focusing on the manner in which the uncertainties of different input parameters are accounted for. Special attention is given to the methodology verification/validation and a simplified numerical example is presented, which also demonstrates the level of conservatism associated with the statistical procedure itself. (author).