The purpose of this study was to evaluate the effects of two bodies in contact, in relation to the amount of heat transferred from one to another, using Autodesk Simulation Multiphysics. The solution comes through Simulation’s Body to Body radiation in a Steady State Heat Transfer. Getting the model to converge was quite a pain, until I determined where the problems were in the setup.
The study takes a simple steel block heated to 200°C, placed upon a HDPE tray, and then predicts how much and how far the heat is transferred to the tray. There were some points using the Steady State Heat Transfer that I thought I would point out for everyone’s benefit.
Body to Body Radiation
Body – Body Radiation is the function that causes Simulation to transfer the heat from one object to another. Without this feature set applied, the application will ignore these effects. Once inside, there are three concepts that need to be addressed.
Body to body radiation analysis is encapsulate into enclosures, which Simulation uses to determine and limit what surfaces interact with the radiated heat. The process is started in the default enclosure ’1′, and the first step is to define the surfaces that will be included in any radiation study. The Define Surfaces button opens the dialog.
Once inside, you can identify which surfaces from which parts will be defined. In my case I chose the ‘all’ option to pick up all surfaces from each part. Then the emissivity settings of each can be defined.
Once saved, the surface definitions are available to include in the enclosure. Two other options are toggled on or off here: Surface Shadowing and Included Enclosures participation. We’ll talk more about surface shadowing and emissivity in an upcoming article. The ‘Enclosure participates in calculation’ option is very important, and dictates whether this particular enclosure will be considered in the simulation or not. This feature allows you to define multiple enclosures, and select only the ones you want in a particular analysis.
Applied temperatures (aka Controlled Temperature) dictates what temperature will be applied to selected parts, surfaces, or nodes. In my case, 200°C was applied to one part. The Stiffness setting directs the analysis how well the temperature will persist in the part.
Initial Temperature (aka Part Temperature) indicates what temperature the part will enter the analysis at. This has no stiffness as it is only an initial setting, which is the fundamental difference between the two temperature features.
With no other features applied, the analysis calculations will show temperatures between the minimum and maximum of the applied and part temperatures.
Remember to run through your analysis settings. The first stop is the multipliers tab, where various features are turned on and off in the scenario. In my case the radiation multiplier was set to 1.
Body to body radiation triggers the analysis into non-linear calculations, and as a result only two solvers are available: Sparse and PBiCGStab. Both use multiple processor cores. I went with the Sparse solver, and BCSLIB-EXT type.
If non-linear operations will be performed, this tab needs to be addressed. This is one of the most important setting groups that I want to point out. Here, various settings of how the non-linear iterations will be conducted are available. In my study, the default setting would not converge.
First step is to select the ‘Stop when corrective norm <E1 (case 1)’ option. This is the basic best option to use to stop iterations once a criteria is met. Under this the maximum number of iterations can be set. Anything over 200 will flag a query from Simulation, indicating that this is unorthodox (and that something else might be needed). My tolerances were relaxed a bit in the process of the study, but the big deal is the Relaxation Parameter.
I’ll follow this up with a quick discussion about this option, but for now understand that this is a factor of restriction of the magnitude in the iterations. If you are having trouble getting a good convergence, this setting may need to be lowered. I used 0.03, and was then able to get good convergence.
As always, I am amazed at what Autodesk Simulation Multiphysics can do, and just how much I have to learn. So far I have not found the limit to it’s capabilities. While it is often quite frustrating when trying something new, a little understanding of certain settings goes a long way. I hope that I have been able to shed a little light into the Simulation Steady State Heat transfer.
The information in this article was collected from many sources, including two classes at Autodesk University. The two most important factors that I found in this process was learning body to body radiation enclosures, and the relaxation setting. Once I understood these better, everything else fell into place.
One last thing I’d like to mention is that the results of this analysis can now be applied in other studies. We’ll see some of that in the near future.