The most common shortcut in thermal analysis is to leave the solid out of the model and apply a heat transfer coefficient (h) at the wall. It is fast, and sometimes it is right. The problem is that in most cases h is the result. Supplying it as an input means assuming the answer up front [1] [2].
When the solid can be left out
- When the thermal resistance of the solid is negligible against the fluid side (thin metal wall, forced convection on one side)
- When h comes from a reliable correlation or from measurement, and the geometry belongs to the family that correlation was derived for
- When the quantity you want is on the fluid side and the solid temperature is only a boundary condition
When the solid must be solved
When conduction within the solid governs. Thick walls, low-conductivity material, embedded heat sources. Here the temperature distribution follows from the geometry of the solid itself.
When heat spreads laterally. On an electronic board or a heat sink fin, heat spreads sideways from under the source. A one-dimensional resistance model cannot see this and overpredicts the hot spot.
When h varies strongly with position. Impinging jets, separation regions, reattachment points. Applying a single average h erases exactly that variation.
When contact resistances matter. Thermal contact resistance between surfaces is the governing term in many assemblies.
The time scales do not match
The fluid reaches equilibrium in seconds; the solid takes minutes or hours. In a transient conjugate solution, solving the fluid fully at every step is needlessly expensive. Where the thermal inertia of the solid dominates, the fluid side can be updated less often and the solution scaled accordingly.
What belongs in the report
Which regions were solved as solids, the material conductivities, any contact resistances, and where those numbers came from. Material data is the largest source of uncertainty in most thermal analyses; a conductivity value with no stated source makes the result impossible to argue about.
References
- Ferziger, J. H., Perić, M. & Street, R. L. (2020). Computational Methods for Fluid Dynamics, 4th ed. Springer. DOI: 10.1007/978-3-319-99693-6
- JEDEC Solid State Technology Association. JESD51 — Methodology for the Thermal Measurement of Component Packages (Single Semiconductor Device).