2 Replies Latest reply on Jul 27, 2015 7:44 AM by Neil Faragher

    Is my thermal solution converged properly..?

    Neil Faragher

      Hey All,

       

      I am doing some thermal modelling in a solid assembly.  I have tried several mesh sizes and have compared the values I get for each solution as the mesh size decreases.  See results in table below.  My question is: do I have a properly converged solution if the temperature is not changing significantly but the heat flux is? 

       

      One thing to note is that the heat flux is sampled in a hot vertex (where 3 sides of a box make a point) so not sure if this causes some kind of singularity error?  I have taken measurements from another part of the model (nearby but not a geometrical discontinuity) and the heat flux changes for this location are -10.2% and +16.3% respectively - this more closely follows the pattern shown in the temp % change column but is obviously still a relatively high percentage.

       

      MeshMax Temp (C)Change %Heat Flux (W/m2)
      Change %
      A82.39-1.45e6-
      B83.090.8%1.52e64.6%
      C84.051.2%2.4e658%

       

      A second question is whether the increased % change in temp in row 3 vs row 2 signals that the model is not converged and that I should continue reducing mesh size? 

       

      Thanks for replies in advance,

      Neil