Mixing processes of hot and cold fluids in a tee with and without sin- tered copper spheres are simulated by FLUENT using the large-eddy simulation （LES） turbulent flow model and the sub-grid scale （SGS） Smagorinsky-Lilly （SL） model with buoyancy. Comparisons of numerical results of the two cases with and without sintered copper spheres show that the porous medium significantly reduces velocity and temper- ature fluctuations because the porous medium can effectively restrict the fluid flow and enhance heat transfer. The porous medium obviously increases the pressure drop in the main duct. The porous medium reduces the power spectrum density （PSD） of tempera- ture fluctuations in the frequency range from 1 Hz to 10 Hz.
The temperature fluctuation caused by thermal striping phenomena of hot and cold fluids mixing results in cyclical thermal stress fatigue failure of the pipe wall. Mean temperature difference between hot and cold fluids was often used as thermal load in previous analysis of thermal fatigue failure, thereby the influences of the amplitude and frequency of temperature fluctuation on thermal fatigue failure were neglected. Based on the mechanism of flow and heat transfer which induces thermal fatigue, the turbulent mixing of hot and cold water in a tee junction is simulated with FLUENT platform by using the Large-eddy simulation（LES） turbulent flow model with the sub-grid scale（SGS） model of Smagorinsky-Lilly（SL） to capture the amplitude and frequency of temperature fluctuation. In a simulation case, hot water with temperature of 343.48 K and velocity of 0.15 m/s enters the horizontal main duct with the side length of 100 mm, while cold water with temperature of 296.78 K and velocity of 0.3 m/s enters the vertical branch duct with the side length of 50 mm. The numerical results show that the mean and fluctuating temperatures are in good agreement with the previous experimental data, which describes numerical simulation with high reliability and accuracy; the power spectrum density（PSD） on top wall is higher than that on bottom wall（as the frequency less than 1 Hz）, while the PSD on bottom wall is relatively higher than that on top wall （as the frequency of 1-10Hz）. The temperature fluctuations in full mixing region of the tee junction can be accurately captured by LES and can provide the theoretical basis for the thermal stress and thermal fatigue analyses.
Temperature fluctuations in a mixing T-junction have been simulated on the FLUENT platform using the large eddy simulation （LES） turbulent flow model and a sub-grid scale Smagorinsky-Lilly model. The normalized mean and root mean square temperatures for describing time-averaged temperature and temperature fluctuation intensity, and the velocity are obtained. The power spectrum densities of temperature fluctuations, which are key parameters for thermal fatigue analysis and lifetime evaluation, are analyzed. Simulation results are consistent with experimental data published in the literature, showing that the LES is reliable. Several mixing processes under different conditions are simulated in order to analyze the effects of varying Reynolds number and Richardson number on the mixing course and thermal fluctuations.
