RELAP5/MOD3.4 Calculation and Model Evaluation Based on Upper Plenum Entrainment Experiment in AP1000
Gongle Song,Dalin Zhang,Qiang Liu,Yan Xiang,G. H. Su,Wenxi Tian,Suizheng Qiu
DOI: https://doi.org/10.1016/j.anucene.2019.107143
IF: 1.9
2019-01-01
Annals of Nuclear Energy
Abstract:After a small break in the AP1000, the coolant would be entrained to the hot leg by the steam when the liquid level of the pressure vessel is lower than the bottom of the hot leg, resulting in the loss of coolant, which would cause the core to be exposed and melted. Accurately calculating the entrainment rate is the key to determining if the core is exposed. Due to the lack of the upper plenum entrainment module, the RELAP5/MOD3.4 code is not conservative in predicting small break accidents. Based on the results of the upper plenum entrainment experiment in AP1000, the calculation results of RELAP5/MOD3.4 was compared to the experimental results of the upper plenum entrainment rate. And the error of RELAP5/MOD3.4 calculation was within 30%. In addition, three upper plenum entrainment models were evaluated. The first model was the dimensionless analysis model (DAM), which classified and integrated the factors affecting the upper plenum entrainment rate and obtained the prediction relationship, and the error was within 50%. The second model was the upper plenum entrainment mechanistic model (UPEMM), which analyzed the initial dynamic process of droplets, the flow process of the vapor phase, and the fly-up process of the liquid phase, and eventually the mechanistic relationship of the entrainment rate was obtained, the model error was within 40%. The third model was the liquid mass conservation model (LMCM), and the entrainment rate was calculated based on the liquid mass conservation equation, and the model error was less than 5%. The calculation comparison of the three models illustrated that LMCM was not only simple but also had higher precision when calculating the upper plenum entrainment rate. (C) 2019 Elsevier Ltd. All rights reserved.