Abstract
The Oyster Creek Generating Station is a 1930-MW(thermal) boiling water reactor 2 plant. During the past year, a program to qualify the Oyster Creek RETRAN model against plant data was in effect at GPU Nuclear. As part of this program, a major transient that occurred on May 2, 1979, was chosen for analysis comparison. While operating at 100% power, a spurious high-pressure scram occurred, coupled with a simultaneous trip of the recirculation pumps. Other events resulted in a loss of feedwater flow and the inadvertent closure, by the operator, of the recirculation pump discharge valves, which limited recirculation flow to only five 0.0508-m (2-in.) bypass lines. The operator proceeded to isolate the vessel and use the emergency condensers for decay heat removal until feed flow was restored 45 min later. The plant RETRAN model was benchmarked against this transient for the first 45 min, using 39 volumes, 54 junctions, 25 heat conductors, and a bubble rise model for the separator/upper downcomer regions.
The RETRAN results showed good agreement with plant data for downcomer level and dome pressure. The unique coupling between the downcomer and core zone liquid levels during the cyclic operation of the emergency condensers was simulated quite well. The use of the bubble rise model for the separator/upper downcomer, however, resulted in a higher dome pressure given by RETRAN, which is believed to be due to the 100% separation efficiency of the model as compared to the degraded separator efficiencies at off-optimum operating conditions. The fuel zone liquid level was an outstanding issue at the time where a conservative simple calculation showed that the core remained covered during the transient. The RETRAN model confirmed that, but also showed that the fuel zone liquid mass during the transient was more than that at steady state. The good agreement obtained against plant data verifies the adequacy of the RETRAN code and the Oyster Creek model for performing transient and accident analyses. Recently, a RELAP5 model has also provided a benchmark for the same transient, and a good comparison with RETRAN and plant data was obtained.