Guidelines for Exchanging Thermal Data

The following guidelines should govern the exchange of thermal data in a file-based coupling of Simcenter STAR-CCM+ with another CAE code:

  1. When explicitly coupling a fluid and a solid simulation, the suggested practice is for:
    • the fluid model to provide spatially varying heat transfer coefficients and their corresponding reference temperatures to the solid model, and,
    • the solid model to return the local wall temperatures to the fluid model.

    In order to avoid issues with numerical stability, you are advised to avoid alternative techniques of exchanging thermal boundary data between Simcenter STAR-CCM+ and other applications.

    Two specific scenarios are:

    1. For the case of fluid flow with turbulence, use the Specified y+ Heat Transfer Coefficient together with the Specified y+ Heat Transfer Reference temperature. A relatively large value of y+ such as 100 should be specified, in which case, the local temperature difference in Eqn. (548) will be relatively large due to the relatively large wall distance. This in turn should allow the solid temperatures room to change significantly, and so allow the models to converge to a steady state in fewer exchanges between them.
    2. For the case of fluid flow without turbulence, use the Local Heat Transfer Coefficient together with the Local Heat Transfer Reference Temperature.
  2. Exchanging local heat fluxes in one direction (whether fluid to solid, or solid to fluid), and receiving local wall temperatures in return, is not a recommended practice for reasons of numerical stability.
  3. More than a single round of data exchange between the fluid and solid models will be required. A minimum of two rounds is recommended. This is because the initial guess of wall temperature for the fluid model may have been significantly different from the final solution.
  4. During the first round, the initial temperature distribution in the solid model should be the same as the initial wall temperature used in the fluid model.
  5. As either the fluid or the solid model must be solved first in the initial round, it is suggested that the fluid model be solved first. An initial guess for the value of wall temperature is required.
  6. When comparing the heat transfer coefficient predictions with benchmarks (either measurements or model predictions), it is important to ensure that the heat transfer reference temperature is the same as in the benchmark. Otherwise the comparisons of heat transfer coefficient can be significantly different.
  7. In a thermo-fluids CFD analysis, the energy equation should be solved (which is usually relatively cheap, and should always be done when circumstances allow, especially when temperature-dependent properties are involved as in buoyant flows).
  8. When the energy equation is not solved, the local heat transfer coefficient is called the Virtual Local Heat Transfer Coefficient.
  9. When using a low-y+ mesh for turbulent flows, you should reduce the thickness of the prism cells adjacent to walls until the fluxes at the walls become grid independent.
  10. For laminar flows, you should reduce the thickness of the prism cells adjacent to walls until the fluxes at the walls become grid independent.
  11. As the near wall region is coupled to the flow in the interior of the domain, it is important that the interior solution is also grid independent within practical constraints.