Abstract
The MYRRHA reactor is a next generation nuclear reactor currently being designed at SCK-CEN in Belgium cooled by Lead Bismuth Eutectic (LBE). Temperature fluctuations (as a consequence of the oscillatorymixing of coolant jets at the core-outlet region) are transferred to the structural surface, possibly leading to high cycle thermal fatigue otherwise known as thermal striping. Therefore, a reliable prediction of the frequency and amplitude of these temperature fluctuations is paramount for providing a safe design. Currently, the
common approach in thermal hydraulics modelling is to apply the Reynolds’s analogy, directly coupling the turbulent conductivity to the turbulent viscosity through a proportionality constant known as the turbulent Prandtl number. However, this approach has been demonstrated to become less accurate within boundary layer heat transfer when the Prandtl number significantly departs from a value around unity. In those cases, a separate closure of the turbulent heat fluxes is desired. Moreover, the interaction between multiple jets may lead to highly complex flow behaviour. To bridge these gaps in knowledge, the low-Prandtl triple jet has been selected as a fundamental flow-case to be bench-marked within the framework of the SESAME project.
This master’s thesis features a numerical investigation of the low-Prandtl triple jet in forced convection applying the transient RANS (TRANS) technique. The TRANS technique is considered a hybrid technique between RANS and LES, capable of resolving unsteady behaviour within the mean flow whilst modelling the turbulent structures. Hence TRANS is of great interest for industrial applications. In this work, the TRANS technique was proven capable of reproducing the dominant frequency of the temperature fluctuations as well as three harmonics. Convergent solutions for the flow field were obtained with k ¡² and realizable k ¡² turbulence models on a full 3D domain, andwith k¡!on a 2D domain. Of the three models, k¡² was selected as the best choice and was coupled with both the Reynolds analogy and transport equations for the turbulent temperature variance kµ and its dissipation ²µ. Good agreement with the experimental data was obtained for the temperature field independent of the used turbulent-heat-flux model. Therefore, contrary to boundary layer flows, the present work suggests the Reynolds analogy remains applicable as an engineering approach to forced-convection dominated free shear flow, even for low-Prandtl fluids. On the other hand, the rootmeansquare temperature fluctuations were significantly overpredicted with both the Reynolds analogy and
k ¡²¡kµ ¡²µ based models. However, considering only the coherent fluctuations and omitting the predicted turbulent heat fluxes, a near-perfect agreement with the experimental datawas observed. This might indicate that the response time of the thermocouples was too slow, and consequently only the coherent and not the turbulent heat fluxes were measured.
common approach in thermal hydraulics modelling is to apply the Reynolds’s analogy, directly coupling the turbulent conductivity to the turbulent viscosity through a proportionality constant known as the turbulent Prandtl number. However, this approach has been demonstrated to become less accurate within boundary layer heat transfer when the Prandtl number significantly departs from a value around unity. In those cases, a separate closure of the turbulent heat fluxes is desired. Moreover, the interaction between multiple jets may lead to highly complex flow behaviour. To bridge these gaps in knowledge, the low-Prandtl triple jet has been selected as a fundamental flow-case to be bench-marked within the framework of the SESAME project.
This master’s thesis features a numerical investigation of the low-Prandtl triple jet in forced convection applying the transient RANS (TRANS) technique. The TRANS technique is considered a hybrid technique between RANS and LES, capable of resolving unsteady behaviour within the mean flow whilst modelling the turbulent structures. Hence TRANS is of great interest for industrial applications. In this work, the TRANS technique was proven capable of reproducing the dominant frequency of the temperature fluctuations as well as three harmonics. Convergent solutions for the flow field were obtained with k ¡² and realizable k ¡² turbulence models on a full 3D domain, andwith k¡!on a 2D domain. Of the three models, k¡² was selected as the best choice and was coupled with both the Reynolds analogy and transport equations for the turbulent temperature variance kµ and its dissipation ²µ. Good agreement with the experimental data was obtained for the temperature field independent of the used turbulent-heat-flux model. Therefore, contrary to boundary layer flows, the present work suggests the Reynolds analogy remains applicable as an engineering approach to forced-convection dominated free shear flow, even for low-Prandtl fluids. On the other hand, the rootmeansquare temperature fluctuations were significantly overpredicted with both the Reynolds analogy and
k ¡²¡kµ ¡²µ based models. However, considering only the coherent fluctuations and omitting the predicted turbulent heat fluxes, a near-perfect agreement with the experimental datawas observed. This might indicate that the response time of the thermocouples was too slow, and consequently only the coherent and not the turbulent heat fluxes were measured.
| Original language | English |
|---|---|
| Awarding Institution |
|
| Supervisors/Advisors |
|
| State | Published - 3 May 2019 |