Experimental Uncertainty in the Validation of Fire Models for Nuclear Power Plant Applications

摘要

The current V&V study highlights the use of experimental uncertainty as a criterion to assess the level of agreement between models and measurements. Although other means to judge the reasonableness of model validation may be possible, the method developed here provides a quantitative and rigorous approach that emphasizes the importance of experimental quality and measurement accuracy in the evaluation of fire models. A consequence of this approach is to place a burden on experimentalists to document and reduce the uncertainty associated with the measurements used for model validation, and on modelers to improve the accuracy of their computations in concert with the reduction in measurement uncertainty. If a key goal of modeling is to remove, or at least reduce, the need for full-scale experimentation, then this study illustrates how modelers and experimentalists can drive each other towards this goal through the model validation process.