GNF2 Counter-current flow limitation testing

A counter-current flow limitation (CCFL) test loop was constructed to support GNF's current and future CCFL testing needs. The test loop is capable of the adiabatic full-scale testing of BWR fuel assemblies with injected steam and water at atmospheric pressure. Tests were performed in July and August 2007 with this loop to confirm the expected CCFL characteristics of various components in the GNF2 bundle, GNF's latest fuel design. All tests injected steam through the lower plenum and subcooled water through the upper plenum. A sequential approach was used for testing the fuel bundle internal components for CCFL. Using the full bundle with prototypical flow, these tests closely simulated the actual in-reactor CCFL phenomenon. Data were analyzed to determine whether the system was in breakdown, or holdup (i.e. CCFL), or neither. CCFL breakdown is identified as the situation in which the injected steam's superheat and enthalpy of vaporization is not enough to overcome the liquid subcooling and test section heat loss. When this happens, the counter-current phenomenon breaks down, condensing all of the injected lower plenum steam, and allowing liquid to flow uninhibited through the restrictive areas of the fuel bundle components. For each tested component, investigations of the locations of liquid inventory throughout the bundle ensured that holdup occurred at the specific restrictive area of interest. The resultant test points represent the average of 120 seconds of data taken at 10 Hz. A minimum holdup rate (the difference between injected and drained liquid flow rate) was used to determine whether CCFL had occurred for each test point. Steam superheat, liquid subcooling and heat losses were taken into account in determining the vapor and liquid flow rates at the CCFL location. Water properties at the local pressure were then used to calculate the non-dimensional liquid and vapor velocities. All resultant CCFL points for each component were used to correlate the non-dimensional velocities to determine the best-fit in the form of a modified Wallis correlation for that component. Despite significantly different flow geometries, the data showed all tested components to have comparable modified Wallis constants, allowing all data to be combined to create a single correlation. (authors)