Solver Problem. Zero pivot when processing D.O.F 11 of 49 nodes

[Drigo]

Hello,
I have a question regards a thermal-electric analysis. I am trying to model how the electric potential will behave in a conductive material. I apply electrical potential at the corners of a material then I had to use a subroutine to use the electrical conductivity interaction GAPELECTR. Though, my program gives me the following warining:

Solver Problem. Zero pivot when processing D.O.F 11 of 49 nodes. The nodes have been identified in node set WarnNodeSolvProbZeroPiv_11_1_1_2_1



Any suggestions on how to fix this? Also, the nodes that are shown vary with different runs..... :S

 

[VaderstadErik]

D.O.F. 11 is the thermal degree of freedom. zero pivots indicate overconstraints in your model. So, start looking at your thermal loads or boundary conditions.

 

[Drigo]

I dont have any thermal loads (I am just working with the electrical parameters). My only two boundary conditions are electric potientials of 5e-6 and -5e-6 on the top and bottom surfaces. It does work with 5e-6 and -5e-6 but when I change it to 5e-3 and -5e-3 (which is my real values) and get the zerp pivot error again. Any clue?

 

[xerf]

I am not familiar with this type of analysis, but I would say something is wrong with the boundary conditions. I would check in the documentation what are the default boundary conditions when you do not prescribe anything for the other boundaries.

For example, for steady state heat transfer the default boundary condition is zero flux, which means "insulated" b.c.. In this case (heat transfer), if you prescribe a certain flux for a portion of boundary and nothing else for any other boundary region, the analysis will abort with zero pivots messages.