An NEA/OECD benchmark-experiment for the validation of CFD for mixing and thermal fatigue in T-junction dead leg flows

Abstract

The current screening criteria for thermal fatigue used in isolated branch lines at nuclear power plants are based on simplified approaches, which, on one hand, tend to be overly conservative, and on the other hand, appear to miss important physics. This is demonstrated by the fact that a number of affected pipes were missed by the current screening methods. The purpose of the present work is to produce novel and accurate experimental data, which can be used for a benchmark of Computational Fluid Dynamics (CFD) in order to advance the prediction capabilities of thermal mixing and fatigue in dead leg T-junctions and improve the accuracy of the screening criteria. The ultimate value creation is increased safety and availability at nuclear power plants. This is of great interest and utmost importance to the nuclear reactor safety community. A new pressurized experimental setup of a T-junction has been designed and constructed for the purpose. The experimental setup, with a temperature difference of 120 degree celsius, more closely resembles plant conditions than in earlier experiments. The boundary conditions in terms of the flowrates and temperatures are controlled and stable within their measurement uncertainty - a prerequisite for CFD-grade experiments. The novelty of the present work is measurements of the temperature fluctuations both in the near wall fluid mixing region and throughout the solid pipe wall material. It has been shown that the interaction between the swirl penetration of hot water and the stratified layer in the cold dead leg has an intermittent character and that random clusters of bursts of hot water penetrate into the stratification and the stagnant cold water below. The average time scale of this process is of the order of hundreds of seconds. The stratification between the hot and the cold water in the dead leg has also shown to have a low frequency oscillation of the order of 0.1 Hz, which wiggles in the circumferential direction. After hot water has penetrated downward and excited the instability, the temperature falls back to the initial one in a manner similar to a damped oscillation. The vertical temperature profiles have shown to be self-similar, for moderate temperatures, when scaled with the temperature difference between the hot and the cold water and with the penetration length. The penetration length increases with the velocity in the hot leg confirming earlier reported correlations. A small (less than 0.1 % of the hot main flow) cold leakage flow into the T-junction is sufficient to suppress the swirl penetration and substantially reduces the penetration length. However, the temperature fluctuations have a larger amplitude and range across the whole mixing zone. The fluid temperature fluctuations are shifted towards a more high frequency content, however, the low frequency nature of the swirl penetration is still present. The obtained novel dataset will be provided for a blind CFD benchmark exercise as part of an OECD/NEA project in 2023 and 2024.