On CCFL at a PWR Hot-Leg Pipe Geometry and Comparison Between Results in COLLIDER 1/4th- and Kobe 1/15th-Scaled Facilities

Countercurrent flow limitation (CCFL) in a pressurized water reactor hot-leg pipe geometry with a 190-mm pipe diameter was investigated experimentally and numerically at the COLLIDER test facility of the Technical University Munich in the past 3 years. This paper summarizes the most important CCFL findings learned from the COLLIDER test facility and tries to explain the reasons for obtaining different descriptions, results, and conclusions at different CCFL experimental investigations. The factors that can affect CCFL experimental results are explained in detail including some scale effects. The necessary preconditions to compare two sets of data from different CCFL experimental investigations are discussed in detail. The difference among CCFL-related limits/curves is clarified taking data at the COLLIDER as an example. The limits included the limit of the transition from a supercritical into a subcritical flow (SSTL); the onset of CCFL limit (iCCFL) inside the hot-leg pipe; the onset of CCFL limit (eCCFL) at the entrance of the steam generator; the deflooding limit (CCFLd); the CCFL characteristics curve (CCFLch), which predicts the water delivery rate after the onset of iCCFL; and the onset of hysteresis limit. It will be shown that among these limits only SSTL, CCFLch, and eCCFL are original limits while the rest are derivatives of them. In particular, it will be shown that the iCCFL limit is a combination of the SSTL and CCFLch limits. The effect of scale upon the eCCFL’s mechanism (whether a water accumulation or droplet entrainment at the entrance to the steam generator) is clarified via a comparison to a 50-mm CCFL facility at Kobe University. This paper tests the scalability of interface distribution at quasi-stationary conditions (i.e., points along the CCFLch curve) via a comparison of time-averaged interface distributions obtained at similar inlet conditions ( at the COLLIDER 190-mm and Kobe 50-mm channels. The comparison will show that interface distributions (which are directly linked to the pressure drop and interfacial momentum transfer) cannot be scaled at the bend/riser/entrance region because of the influence of the channel diameter upon occurring CCFL mechanism. Meanwhile, the water level gradient can be similar at the horizontal part, but not the relative water depth.