Brain Implant Thermal Simulation Using Ansys Fluent

Hello and welcome to this video on thermal simulation of brain implants using Ansys Fluent. I'm showing here a portion of a geometry of a brain implant within a brain domain. This is the side where we have a brain implant, Ansys bridge geometry, and 100 electrodes that are protruding from it.

The dimensions of this are about 7 millimeters by 11, or another 8 millimeters, and looking in the context, that brain implant would be sitting on the skull here. And how I did this model is I meshed it in two separate parts.

I did the brain implant and the surrounding tissue in one mesh, and then I did the remainder of the geometry in another mesh, and primarily that's because of the large difference in length scales. So with these implant electrodes, the measurements are in the order of microns.

See here the circumference of 52, where with the domain for the brain, surrounding brain tissue, this is on the order of a diameter of 150 millimeters. So looking at the mesh, you see the final results, but I'll just walk through some of the settings.

So for the localizing for the electrodes, I use the body size of 15 microns. For the bridge and the integrated circuit, I had body's curvature. So that mesh was generated for a surface mesh using a min of 15 microns, the max of 800 again.

These are all solid zones, no boundary layers, and then when you get to creating the volume mesh, my maximum cell length size I use one millimeter.

Then looking at the rest of the domain, which did not include the brain implant, I just have a curvature local sizing with sizes from 100 up to 800 microns.

For the skull, which is in the scalp, which are these two layers here, and then I let the mesher increase the surface mesh size up to 2 mm for the brain cylindrical surface as well as the bottom. So that gave me these two volume meshes. So I went to Fluent Flow Solver and put those two together.

I used the domain, and then within the zones, I used append with append case to append those two meshes together, and looking at the solver settings and the boundary settings, this was a steady-state simulation. And I have energy turned on because I'm in desire to solve for the temperature field.

Since everything is solid, there's really no need for viscous, so I just set that to laminar. Then looking at materials, there's no fluid materials, so I just have a group of solid materials here to represent, say, the brain, the integrated circuit, the scalp.

Silicon represents the electrodes, skull material, and this one would be for the bridge. Looking at cell zone conditions, so the brain and the brain window. So the brain window would be the region above the brain, so in case we want to have an open area here, I can turn those off.

I set the material to the same, so down here would be the brain window and down here would be the brain.

The electrodes would be all of the spikes that you see, and then down towards the bottom, we have the integrated circuit, and this is the only cell zone with something specified in terms of these options, and that includes the source term.

So I have a volumetric heat source, use an expression to define that. So this expression named heat load, and I jump down to show what that is.

So looking at, you know, a value here, this would be something in the terms of 17-18 milliwatts, and then after that, I divide it by the volume of that cell set, so that gives me the volumetric heat source. And I'm going back up now to boundary conditions.

So we have interfaces, which I'll touch on later, the internals, which are generated by the mesher. Now looking at the walls, so wherever there is a solid body touching another solid body, I just have the thermal coupled boundary condition type.

But then we're, say, on the top surface, where the top of the scalp, so where the scalp is now interacting with the air above, there should be a boundary condition here, like your top scalp.

And for that, we have a convection boundary condition with a heat transfer coefficient of 5 to represent say natural convection, a free stream temperature of 24 degrees C, and then say the brain, and say the brain bottom down here, we're using a fixed temperature of 37 degrees Celsius.

So when we use the append with the append case files, there's going to be interfaces between those two meshes, and one has to use the mesh interfaces option to generate the and the coupled connection between those.

And what I did was I renamed some of my zones to be things that have a little identifier in them, so that I can use like a filter text. So here I had intf, and that helped me to find the cell zones I needed. So this makes the pair.

So here I've like, and that touched with another face in the other mesh called skull window, but the intf was like a suffix in one and a prefix in the other. So that helps me to quickly find them, and then use create to generate the mesh interfaces.

Touched on the heat load expression methods, being no fluids, the main thing here is to have a second-order upwind for energy.

And then in terms of controls, to have a one for the under-relaxation to make for a nice fast calculation, I've used some report definitions to be able to capture what is the maximum temperature on a surface, and that surface would be touching the brain tissue.

So for in this example, I've got the bridge where it touches the brain window. So I'm looking at what is the maximum facet temperature on that. And so we track that on monitors, both in files and then in plots.

So using the plots, I can see the convergence, how the temperature stabilizes with iteration, and then using the report files, I can export that, say, to a spreadsheet to analyze the temperature, maximum temperature, versus the changing this heat load value.

And I think that takes care of the Fluent setup. One can do some post-processing within Fluent, but one thing I like is to be able to use the compare cases feature within CFTPost.

So one loads in both pairs of case and data files, say for a soak condition, and that soak condition would be zero heat load, and then there'd be another case with a heat load, say like that 18 milliwatts.

And then I can use the, and then the compare cases tool to compare those two simulations, and then in this example, I have temperature plotted on a planar surface.

So I can see in the right-hand corner, I'm sorry, the left top here, this would be with say 18 milliwatts of power, and then this result over here would be with a zero milliwatt heat load, so I can compare what is the impact of having that implant turned on on the local temperature, which is given in the bottom plot here, so we can see for that heat load value, there is a one-degree maximum say temperature rise within the tissue nearby the implant.

That concludes this video. Please contact us at https://ozeninc.com/contact for more information.