19 Replies Latest reply on Nov 19, 2014 10:30 AM by Mike Pogue

    Transient thermal results different with different time increments

    V. S

      I am trying to conduct a transient thermal analysis and the results I am getting seem to be quite different as I change the time increment. They also do not agree with my first principles understanding.


      I am modeling a solid cylinder (20mm in diameter and 80 mm in length) before modeling a larger system. It is a part file with only one feature (Boss-Extrude).


      Material and model properties: 1) Total Mass of 44.5 grams. 2) Thermal conductivity in X and Y (radially) = 1 W/(m*K). 3) Thermal conductivity in Z (longitudinally) = 40 W/(m*K). 4) Specific Heat = 800 J/(kg*K)

      Thermal Loads = (see attached snapshot) If we integrate that curve, I should be adding 4518.27 Joules into the system.

      Initial Temperature = 25C defined for Boss-Extrude

      Mesh parameters were the default. Standard mesh. 2.59mm with 0.129mm tolerance.


      I probed several points around the cylindrical and one along the center axis at the top face. The resulting graphs of 0.2 second time increment and 0.1 second time increment are attached. The total test time was 20 seconds.


      By first principles, my guess would be that since I am putting 4518.27 Joules into the system, I would see a steady state temperature increase of about 100C. Q = m*c*dT --> dT = 4518.27/(800*0.0445) ~= 123K or 123C. So from the intial temperature of 25C, I'd imagine I'd see the final temp to be on the order of ~150C. Why am I seeing temperatures on the order of 10^4 C? Are my first principles and/or simulation understanding incorrect?


      Thank you in advance for any help.