chtMultiRegionFoam and the "maxDi" parameter for solid-solid heat transfer

Adam_K · April 25, 2019, 07:44

Hello Foamers,

I am working with OpenFOAM version 5 (blueCFD) in order to simulate heat transfer in a composite slab and have come upon an issue that I can't explain

. I'm hoping that someone can shed some light on this.

Here is a figure of my geometry. The vertical direction repeats using the cyclicAMI BC. The system is at 300 K and at t = 0 s, the left BC becomes 310 K. The right BC is zeroGradient as it corresponds to the center of the slab in order to take advantage of the symmetry of the problem (slab heating).

Here is a more clear view of the mesh.

For the moment, the thermal conductivity in both regions is the same (that of Aluminum) because I want to make sure that OpenFOAM can reproduce the analytical response of a pure slab. For small domains (say 200 µm by 200 µm with a few fibers inside) the results align perfectly with what I would expect. However, as I extend the domain to longer domains (200 µm by 2 cm) there are problems with the simulation. The intermediary domain of 2.5 mm seems to be roughly where I start to see problems, as shown in the following figures.

The top left figure shows the matrix temperature on the right hand side of my domain (corresponding to the center of the physical slab). When I use a maxDi = 5 or 10, the results align pretty well with the analytical solution (the difference is in the top right figure). The difference between the matrix and the fiber temperature is shown in the bottom left, ideally it is zero in this case. The time step size vs time is shown in the bottom right.

For lower maxDi values, the temperature never seems to rise. When maxDi is higher, there is a clear delay in temperature of the center of the slab. This implies to me that the simulation is failing at both ends and for different reasons. I can see why it won't work with an overly large time step (higher maxDi), but I'm surprised that the lower maxDi simulations lead to no temperature increase.

For this 2.5 mm thick slab I could just use maxDi = 10 for my simulations, however I would like to understand what the problem is and also extend my work to a larger domain (a few cm). So far in my tests with a longer domain, there is no longer such "sweet spot" of maxDi values where the results align with the analytical solution.

The case file and .msh file can be found on my DropBox.

Has anyone faced a similar issue before with OpenFOAM?

April 25, 2019, 07:44	chtMultiRegionFoam and the "maxDi" parameter for solid-solid heat transfer	#1
Adam_K Member Adam Join Date: Nov 2018 Posts: 36 Rep Power: 8	Hello Foamers, I am working with OpenFOAM version 5 (blueCFD) in order to simulate heat transfer in a composite slab and have come upon an issue that I can't explain. I'm hoping that someone can shed some light on this. Here is a figure of my geometry. The vertical direction repeats using the cyclicAMI BC. The system is at 300 K and at t = 0 s, the left BC becomes 310 K. The right BC is zeroGradient as it corresponds to the center of the slab in order to take advantage of the symmetry of the problem (slab heating). Here is a more clear view of the mesh. For the moment, the thermal conductivity in both regions is the same (that of Aluminum) because I want to make sure that OpenFOAM can reproduce the analytical response of a pure slab. For small domains (say 200 µm by 200 µm with a few fibers inside) the results align perfectly with what I would expect. However, as I extend the domain to longer domains (200 µm by 2 cm) there are problems with the simulation. The intermediary domain of 2.5 mm seems to be roughly where I start to see problems, as shown in the following figures. The top left figure shows the matrix temperature on the right hand side of my domain (corresponding to the center of the physical slab). When I use a maxDi = 5 or 10, the results align pretty well with the analytical solution (the difference is in the top right figure). The difference between the matrix and the fiber temperature is shown in the bottom left, ideally it is zero in this case. The time step size vs time is shown in the bottom right. For lower maxDi values, the temperature never seems to rise. When maxDi is higher, there is a clear delay in temperature of the center of the slab. This implies to me that the simulation is failing at both ends and for different reasons. I can see why it won't work with an overly large time step (higher maxDi), but I'm surprised that the lower maxDi simulations lead to no temperature increase. For this 2.5 mm thick slab I could just use maxDi = 10 for my simulations, however I would like to understand what the problem is and also extend my work to a larger domain (a few cm). So far in my tests with a longer domain, there is no longer such "sweet spot" of maxDi values where the results align with the analytical solution. The case file and .msh file can be found on my DropBox. Has anyone faced a similar issue before with OpenFOAM?