Global uncertainty and sensitivity analysis of MELCOR and TRACE critical flow models against MARVIKEN tests

A Sensitivity Study of Critical Flow Modeling with MELCOR 2.2 Code Based on the Marviken CFT-21 Experiment

“…It was earlier discussed by Mosunova in reference [25] and by the code developers in MELCOR Manual [1,2]. The code calculates the subcooled liquid choking flow with the Henry-Fauske model and gas choking flow with reasonable accuracy, comparable to the results achieved by the TRACE code [23,24]. The main problem exists for a critical two-phase flow that is modeled by the Moody model based on the RE-TRAN code implementation [1][2][3]23].…”

“…The code calculates the subcooled liquid choking flow with the Henry-Fauske model and gas choking flow with reasonable accuracy, comparable to the results achieved by the TRACE code [23,24]. The main problem exists for a critical two-phase flow that is modeled by the Moody model based on the RE-TRAN code implementation [1][2][3]23]. Future improvements in the MELCOR code will be necessary and are planned by developers [1,2].…”

“…The selected models were primarily intended to represent the most straightforward approach, ensuring short computation times and good computation stability and applicability for the plant scale studies. The models studied in this work were created from scratch and independently from the MELCOR models developed and discussed in previous papers [23,24]. One of the differences compared to previous models is a detailed height-volume table for all the applied control volumes (see Figure 1).…”

“…MELCOR critical flow modeling was discussed and described in the previous paper [23], and its detailed descriptions are available in the Manual [1,2] and are not repeated in this work. The MELCOR code has issues with predicting the critical flow, and they were discussed in previous papers [23,24]. It was earlier discussed by Mosunova in reference [25] and by the code developers in MELCOR Manual [1,2].…”

“…Another purpose was to benchmark the recent code version with the Marviken test and identify which model parameters and nodalization choices impact the results. The previous work focused on the non-code-related parameters, and this work is focused on the MELCOR specific setup and modeling [23]. The additional motivation of this work was to identify the important parameters for future broader global uncertainty and sensitivity analyses focused on critical flow phenomena.…”