Large eddy simulation of thermal turbulent mixing and reduction of temperature fluctuations with swirl in T-junction

Abstract

Turbulent mixing of two fluids having a temperature difference can lead to temperature fluctuations in the T-junctions, particularly in the primary cooling circuit of nuclear power plants. Temperature changes produce thermal stresses causing thermal fatigue. This is a key topic related to nuclear security. This article presents the numerical results of two thermal turbulent mixing simulations, applied to a T-junction (OECD-NEA-Vattenfall T-junction). These simulations apply, respectively, to the case of a flow with and without rotational movements into the cold side. The numerical predictions were made using the large eddy simulation with the commercial code Fluent. The numerical results obtained in the first case were compared with the available experimental data and were found in an agreement satisfactory to the profiles of average variables and the spectrum of the temperature near the wall. The comparison with RMS profiles was less satisfactory in the center of the main tube, but was of good quality in the near-wall region. In the second simulation, the main movement of the fluid is affected by the spiral paths created by the swirl velocity caused at the cold inlet, which improves the mixing of hot and cold fluids downstream of the T-junction. From the analysis of these results, it clearly appears that the temperature fluctuations decrease near the

Design optimization and CFD evaluation of a volute swirl burner with central gas supply

“…where Q is the rated thermal load (i.e., the design thermal load), kW; q is the heat intensity of burner head, typically (35)(36)(37)(38)(39)(40) × 10 3 kW/m 2 [14].…”

“…The Reynolds-averaged Navier-Stokes (RANS) method for both non-reacting and reacting conditions was conducted by using FLUENT 17.0 and validated by the calculated parameters. For turbulent simulation, although large-eddy simulation (LES) is considered to be a promising analysis tool for design applications [34][35][36][37], it is computationally demanding for problems with high Reynolds number (Re > 140,000) [24], and the three-dimensional geometry of the swirl burner is another challenge. In contrast, RANS turbulent models, such as standard k-ε model [30,38,39], RNG k-ε model [40,41] and realizable k-ε model [24,42,43], have been widely implemented in swirling flow simulations.…”

Numerical Study of Induced Condensation upon Mixing Flows of Water-stream Flow in a Tee-Junction Pipe

“…Furthermore, CFD results were in good agreement with the obtained experimental data. Benyamina et al [4] discussed the turbulent mixing in a T-junction using large eddy simulation (LES). They concluded that to get numerical results close to the experimental data, the required mesh resolution should consider the Taylor micro-scale length.…”