I am using Multiphysics 2012 and performing a steady state heat transfer analysis. The model consists of a single part with heat source loads and convection loads. Since the input heat flux loading varies over the heated region, I have modeled it as pixels of discrete sizes and heat flux values. All of the water cooling is assumed to have the same convection coefficient and ambient temperature (the flow is not being simulated), and is applied to several faces (32 in total) due to the contours of the cooling channels. Since there are twenty unique values of heat flux, placed on 35 pixel surfaces, I would like to be able to determine the value of the total heat removal rate by the convection load over all the surfaces it is applied to, in order to quickly check the simulated heat flow rate through the model against the intended heat flow rate (to catch errors in computing and/or transferring the loading map).
Does a command (or set of commands) exist in the results environment to quickly and simply provide the heat added or removed by a convection load? When I select a convection loading and choose "Inquire Loads and Constraints" from the right-click menu, it only supplies the coefficient, ambient temperature, and part and element numbers for the load, not the amount of heat it adds/removes.
Using the inquire command for the heat rate through the involved faces will not work because the geometry of the model and the manner in which heat rate through a face is calculated provide heat rates that do not approximate reality. In fact, if the heat rate through a face is summed over all of the extrenal element faces on this model, the result is approximately 500 W (22% of the known applied load and much greater than zero); an impossibility if the system is at steady state. The best single element I could find had a heat rate sum over all of its faces on the order of 10^-5 W, but the majority ranged from 10^-2 farther from applied loads up to 10^2 at regions of applied loading. These results occur whether the element definition is set to "Heat Flow Calculation > Projected at Centroid" or "Heat Flow Calculation > Linear Based on BC".
Solved! Go to Solution.
Can you elaborate on your statement "the manner in which heat rate through a face is calculated provide heat rates that do not approximate reality"? In what way is the heat going through a face different than h*A*(T-Tenvironment)?
Granted, the heat rate through the face is an approximation of an approximation (the analysis itself), but with the heat flow rate calculation method set to "Linear based on B/C", the two values should match pretty closely. (But there may be some issues with 2D elements, just in case that is what you are using.)
Summing the heat leaving the model through convection should be no more difficult than it was to apply the convection load in the FEA Editor:
- Display the "Heat rate of face" result.
- Set the selection method to surface.
- Select the multiple surfaces where convection is applied. (If may even be possible to use a "Filter" in the browser to select all of the faces that have a convection load, but it has been awhile since I used the filter, so I do not know the steps.)
- Inquire on the results.
- Change the drop-down in the bottom left corner (summary?) to Sum.
16 years experience with Simulation Mechanical
You are right, that statement makes little sense. What I was trying to get at is just what you addressed in your second paragraph; the accuracy was insufficient with the settings I was using.
Based on your post, it seems I was on a misguided search for a mesh-independent readout of the heat removal rate provided by a convection load; a function that evidently does not exist. I had observed that temperatures at points of interest (and presumably throughout the model) were nearly independent of mesh, so I was not particularly concerned with mesh settings, but heat rates through faces are heavily dependent on the mesh according to the calculation methods described in the documentation (and observations I made late yesterday). The absence of the function I sought is probably for the best, as it forces a verification of mesh quality before heat rates through faces can be used to verify loadings.
Thank you very much for your help. This has cleared up my confusion and set me in the right direction.