[ajsarode]

Hello Everyone,

I am currently using SolidWorks Simulia for transient thermal simulations. There is a solid object that needs to be heated with a hot liquid. Initial condition of the solid is 300K; fluid is 350K. Inlet BC is mass inlet and outlet is pressure outlet. To validate the simulation, I need to balance the heat taken by the solid object and heat given by the fluid. To calculate the heat given by the fluid, I cannot directly consider (Inlet energy - Outlet energy) as there will be thermal information lag because the fluid particle that gave away heat to solid object at time t=t0 will reach the outlet at t=t0 + dt (depending on the fluid velocity). Therefore, there will always be a lag between data capture.

One solution that I was looking at was to consider the volume weighted average of solid and fluid temperature at any instance and then try to validate the simulation. However, Simulia cannot capture the volume weighted average temperature. Also, I thought of having a cut section just after the solid object and then take the energy difference between inlet and cut-section. This seem to did not work as the wakes formed after the solid object changes the flow dynamics of the fluid.

Thus, I would like to get a help from the community regarding the validation of this transient thermal simulation.

[FliegenderZirkus]

There is an interface between the solid and the fluid regions (this may be called differently in your software). The interface has two sides and you can evaluate hear transfer on each of them and check if they match. If the mesh is conformal, there is little reason to suspect that the heat transfer is different, but if the mesh is not conformal, an error is to be expected. Then it's indeed good practice to check the energy conservation error. Is this your case? Looking at the enthalpy balance on the fluid alone seems unnecessary to me (as long as you have basic trust in the solver). If each time step is converged, the energy should be conserved.

[ajsarode]

Originally Posted by FliegenderZirkus

There is an interface between the solid and the fluid regions (this may be called differently in your software). The interface has two sides and you can evaluate hear transfer on each of them and check if they match. If the mesh is conformal, there is little reason to suspect that the heat transfer is different, but if the mesh is not conformal, an error is to be expected. Then it's indeed good practice to check the energy conservation error. Is this your case? Looking at the enthalpy balance on the fluid alone seems unnecessary to me (as long as you have basic trust in the solver). If each time step is converged, the energy should be conserved.

The mesh is not conformal. However, the heat transfer value on either side of the interface matches with -ve sign. Do you think that this is a sufficient criteria to validate. What is the usual practice to validate such problems if not enthalpy balance.

[FliegenderZirkus]

Interesting that the heat transfer on both sides of your interface matches exactly (with a different sign). This probably means that the mesh is effectively made conformal at runtime by intersecting the mating cells. If a mapping (interpolation) had been used, I think there would always be a small error.
What kind of validation are you looking for? There are many things you can check in the simulation in terms of best practices and sanity checks, but when I hear validation, I think of comparison to measured data. In terms of the energy balance, I don't think too much can be done in the general case. What you can do is to look at special cases to gain confidence in your model, such as:

1. If you let the simulation run long enough to find a steady state, then the heat flux integral over the fluid/solid interface should tend to zero.
2. If your Biot number is small enough (e.g. a small copper sphere in high-speed flow) you can ignore the temperature profile in the solid and write a simplified heat equation as
   m*cp*dT/dt = Q
   where Q is the heat flux from the fluid. In this case the energy balance is useful because it's actually all we need (so we could have saved ourselves meshing the solid in the first place).

One solution that I was looking at was to consider the volume weighted average of solid and fluid temperature at any instance and then...

Be careful about taking averages, in the general case where all of the variables change both in time and space, I'm skeptical you can do much with such an approach. But I could be wrong on that, feel free to try it.