Validation of CTF Void Predictions Using the BFBT Database

A quantified validation of CTF void fraction and equilibrium quality predictions is performed using the NUPEC BFBT database. Standard code modeling options are used, and there is no calibration to the experimental data. Each of the 392 steady state BFBT void distribution experiments has a specified geometry, power distribution, and spacer grid design. All experiments are modeled in CTF, but some cases that have oscillations are excluded from the final analyses. Three quantities of interest are compared between the experiment and the code predictions: exit equilibrium quality, average exit void fraction, and subchannel void fractions. The results indicate that, in general, CTF accurately predicts equilibrium quality but overpredicts void content. By splitting the subchannels into different groups, it is shown that subchannels near unheated surfaces are the least accurately modeled, especially for cases which have a high concentration of unheated surfaces. Though the BFBT database is the most complete source of experimental data for fine-mesh void distributions, the measurements are asymmetric even for symmetric geometry and boundary conditions. A method is proposed to approximate the experimental uncertainties which led to these asymmetries.