Assessment of TRACE for predicting startup transients in natural circulation boiling water reactors

Natural circulation is a key passive safety feature in several light water small modular reactor (SMR) designs, but flow instabilities during natural circulation startup transients pose critical challenges for safety analysis. To ensure accurate prediction of key thermal–hydraulic phenomena in startup transients, system-level analysis codes must be rigorously assessed. Among these, TRACE is recognized as the flagship best-estimate reactor system code developed by the U.S. Nuclear Regulatory Commission (NRC). This study evaluates TRACE (V5p8) for predicting two-phase flow behavior in natural circulation using startup transient data from the Purdue University Multidimensional Integral Test Assembly (PUMA). Four startup tests (EQU, EQU-NU, SUP, SUP-NU) were analyzed, simulating natural circulation boiling water reactor (BWR) behavior under different pressure conditions, both with and without neutronic feedback. A systematic refinement study was conducted to determine the optimal TRACE simulation settings, including nodalization, timesteps, and spatial discretization methods. Results indicate that TRACE successfully handles input power oscillations without significant numerical instabilities and generally provides reasonable predictions. While TRACE accurately predicts system pressure and liquid temperature, it significantly underestimates core void fractions, failing to capture geysering phenomena due to underprediction of energy allocated for void generation in subcooled boiling flow. The flashing phenomenon in the chimney is well captured, but chimney void fractions are generally overestimated. Additionally, TRACE overestimates downcomer flow rates in most cases, potentially due to the overestimation of chimney void fraction, which influences natural circulation gravity head. Future work will focus on further validation of TRACE models using an extended natural circulation database from separate effect tests (SETs) and integral effect tests (IETs) to enhance modeling accuracy for complex startup transients in natural circulation systems.