Multi-dimensional modeling of two-phase flow in rod bundles and interpretation of velocities measured in BWRs by the cross-correlation technique

Abstract

Two-phase flow in rod bundles is of the atmost importance in nuclear technology since it is a naturally occurring phenomenon in BWRs under normal operational conditions, or in PWRs undergoing a severe transient. It has recently been shown that by neutron noise analysis (cross-correlation) techniques, in the upper half of a normally operating BWR, one measures two or even three two-phase flow velocities (two or three peaks in the cross-correlation function); this was also found to be the case in measurements performed in simple air-water loops with different stationary and adiabatic two-phase flows, the direct consequence of these findings being that no cross-sectionally averaged two-phase flow models can be successfully employed for interpreting this kind of non-intrusive velocity measurements. It is the aim of this work to present an as precise as possible interpretation of velocity measurements in BWRs by the cross-correlation technique, which is based on the radially non-uniform quality and velocity distribution in BWR type bundles, as well as on our knowledge about the spatial ‘field of view’ of the in-core neutron detectors. After formulating the three-dimensional two-fluid model volume/time averaged equations and pointing out some problems associated with averaging, we expound a little on the turbulence mixing and void drift effects, as well as on the way they are modilled in advanced subchannel analysis codes like THERMIT or COBRA-TF. Subsequently, some comparisons are made between axial velocities measured in a commercial BWR by neutron noise analysis, and the steam velocities of the four subchannels nearest to the instrument tube of one of the four bundles as predicted by COBRA-III and by THERMIT. Although as expected, for well-known reasons, COBRA-III predicts subchannel steam velocities which are close to each other, THERMIT correctly predicts in the upper half of the core three largely different steam velocities in the three different types of BWR subchannels (corner, edge and interior). In the upper part of the core where a pronounced radial steam velocity and quality profile exists in the bundles, we associate the main peak of the cross-correlation function with the steam velocities in the edge subchannels, the second peak with the steam velocities in the corner subchannel, and the third small peak with the steam velocities in the interior subchannels. This interpretation is verified by a computer simulation with synthetic signals as well as by a simple phenomenological analytical model, and it opens the way for utilizing this kind of measurements (to a certain degree and within certain error bounds) for verification of advanced subchannel analysis codes like THERMIT-2 or COBRA-TF and in particular, for improving the two-phase mixing correlations employed in these codes.