Refinement and testing of CTF for annular flow regime and incorporation of fluid properties

Abstract

The current study focuses on improving and testing the CTF thermalhydraulics computer code. CTF is a thermalhydraulic code used for subchannel analysis of nuclear power reactors developed as part of the US DOE CASL program and distributed by North Carolina State University. Subchannel analyses are used to predict the local fuel temperatures and coolant conditions inside a complex nuclear fuel assembly. Such calculations are used to improve designs of nuclear fuel, improve operating margins, or perform safety analysis. An important part of the code development process is the verification and validation for its intended use. In this work validation activities are performed using the RISO experiments are modeled in CTF for adiabatic and diabatic cases in annular flow regimes and a limited set of tests in CANDU geometries. The CTF predictions significantly overpredicted the pressure drop for cases involving annular flow conditions. Depending on the application, such overprediction can result in significant errors in the computation of fuel element dryout and other figures of merit. For example, an analysis using fixed pressure boundary conditions CTF predicts much lower subchannel flows and hence fuel element temperatures may be overestimated. On the other hand, for a scenario with mass flux and inlet pressure as boundary conditions, the impact of pressure drop discrepancies on dryout predictions may be lower. Therefore, there is a particular focus in this thesis on the two-phase pressure drop models and the RISO experiment specifically, since the RISO tests involve a range of annular flow conditions which is prototypical of many CANDU accident analysis conditions. In addition to the RISO experiments, 28-element CANDU full scale rod bundle experiments are modeled in CTF for single-phase and two-phase flow conditions. Cases are modeled for crept and uncrept conditions with different bearing pad heights i.e., 1.17 mm and 1.35mm. Pressure drop predictions are compared with the experimental results where single-phase comparisons are in good agreement while an overprediction of ~25% is observed for two-phase conditions. The effect of bearing pads on the subchannel local parameters, like mass flow rate, are also studied. Furthermore, the effect of turbulent mixing rate on subchannel enthalpy distribution in the bundle and CHF in different subchannels is also analyzed. Based on the comparison to the RISO and CANDU 28 element test databases, the overprediction of pressure drop in the annular flow regime needs improvement in the current version of CTF. This overprediction of the frictional pressure drop results from either wall drag or interfacial shear stress phenomena. In this study, it is demonstrated that the issue occurs mostly as a result of interfacial friction factor modelling this work examines several alternative approaches. The results show the Ju’s and Sun’s interfacial friction factor better predicts the results among all the other six correlations implemented in CTF. The major impediment in further testing of CTF is that it lacks the capability to simulate R-134a fluids. Given there is a large database of R-134a two-phase tests, another aspect of this thesis is to extend CTF for application and validation using refrigerants. The current CTF version only supports fluid properties for water and FLiBe salts. By adding R-134a fluid properties the testing and validation range of CTF is broadened for different experiments performed using R-134a fluids. CHF experiments are modeled in CTF and results are compared with experimental data. For local conditions correlation, 2006 water LUT are used to predict CHF and DNBR. The fluid-to-fluid scaling method is applied in CTF when using CTF with R-134a fluid properties for CHF and DNBR predictions to account for the difference in fluid properties between R-134a and the CHF look-up table.