Time-dependent model calibration through consistent data assimilation

Reactor analysis and design needs experimentally validated numerical simulation tools, capable of handling the inherently time-dependent and nonlinear nature of the underlying physical phenomena. The objective of this paper is to present applications of a comprehensive and efficient methodology for the experimental validation and calibration of complex time-dependent numerical simulation models, incorporating both computational and experimental uncertainties. The resulting validated and calibrated models are capable of yielding bona fide 'best-estimate' values for the quantities computed by the respective simulation tools. This general purpose methodology for experimental validation and calibration of time-dependent simulation tools is currently being incorporated into a software module called 'BEST-EST' (for 'best estimate'), to be used as a general-purpose software module for validating and calibrating both time-dependent and time-independent models for numerical simulation of nonlinear systems. The 'BEST-EST' methodology is applied to an illustrative time-dependent depletion problem (radioactive decay) and also to a transient thermal-hydraulic model of the BFBT benchmark, simulated by the FLICA4-code system, and calibrated with experimental data provided by the BFBT benchmark. (authors)