Learning Forum

[Chad.Martin]

Im running into an issue using Ansys Thermal-Electric for Bus Bar Design. I think maybe the software is limited and not built for this.

It seems that Ansys Thermal-Electric can only run a single line of bus bar run with one input (amperage applied) and one output(voltage grounded 0V). Unless the model is symetric about the inputs and outpus you can not have multiple equal inputs and equal outputs behave correctly. Ansys Thermal Electric does not allow you to specify exact values of current inputs and outputs at the same time without throwing an error. From what I have seen after solving is that the current will get pulled from the closest output or similarly pulled from the closest input and not distribute among the outputs evenly as needed.

I need help finding a way to force the inputs to be a certain amperage supply and force the outputs to draw a known amperage value. If i am not using the correct Ansys tool for this please let me know what i should be using.

I have multiple equal inputs(amperage input known and constant) and i have multiple outputs(amperage output value known and constant).

[wrbulat]

Hi Chad,

I know you said each of your multiple input/output currents are "constant" but I just want to make sure... is the current AC or DC? If AC, you'd want, in general, to perform a harmonic response electromagnetic analysis to account for transient phenomena (skin & proximity effects, self and mutual inductance). This analysis type is not natively exposed in Mechanical but some relatively simple APDL command objects can be used to set this up in a Magnetostatic analysis system (details to follow if you're interested - let me know). 

If DC, and your setup is in a steady state thermoelectric analysis system or (preferred in recent releases) a steady state coupled fields analysis system, then using the electric BC's such as those you described can be problematic. It sounds like you aren't constraining a VOLT DOF in your model. This is like performing a steady state thermal analysis without constraining the TEMP DOF anywhere, and hoping applied thermal fluxes are in perfect balance. Or like performing a static structural analysis with BCs allowing rigid body motion in some direction and hoping forces acting in opposite directions are in perfect balance. The system matrix in all these cases will very likely be ill-conditioned if not downright singular (unsolveable). 

That being said, you could try turning off pivot checking (insert a command object under the analysis branch of the tree containing this: PIVCHECK,OFF). This might get you past the errors you're seeing.

If that doesn't help and you're APDL-savvy, you might try coding up a command object that adds independent current source circuit elements, each coupled to one of the faces where you want current added/removed:

Mind you, I just now thought of this and have never tried it before, so I'm not 100% certain it'd work. My hunch is that it will work (I do have a lot of experience using these element types). This should allow you to fix VOLT DOFs in the model.

Best,

Bill