Best-Estimate Model Calibration and Prediction Through Experimental Data Assimilation - II: Application to a Blowdown Benchmark Experiment

This work presents a paradigm application of a new methodology for simultaneously calibrating (adjusting) model parameters and responses, through assimilation of experimental data, to the benchmark transient thermal-hydraulic experiment IC1, performed at London's Imperial College. Following the description of the experimental setup, the corresponding mathematical model is developed and solved numerically. The sensitivities of typically important responses (e.g., temperatures, pressures) to model parameters are computed by applying both the forward and the adjoint sensitivity analysis procedures. These sensitivities not only identify the most important model parameters but also propagate, within the data assimilation procedure, parameter uncertainties for obtaining predictive best-estimate quantities, with reduced best-estimate uncertainties (i.e., “smaller” values for the variance-covariance matrices). This assimilation procedure also provides a quantitative indication of the degree of agreement between computations and experiments. In particular, the paradigm application presented in this work indicates the path for validating and calibrating thermal-hydraulic computational models used for reactor safety analyses. The concluding remarks highlight several important open issues, the resolution of which would significantly advance the area of predictive best-estimate modeling, while opening new avenues for applications in nuclear reactor engineering and safety.