12. Summary of results#

The four models give results where some values exceed the tolerance set initially (5%):

  • for modeling A (AXIS: QUAD4, TRIA3), the maximum difference is 15.3% and 4 values out of 16 exceed the tolerance,

  • for B modeling (PLAN: QUAD8, TRIA6), the maximum difference is 13.8% and 3 values out of 24 exceed the tolerance,

  • for C modeling (3D: HEXA8, PENTA6), the maximum difference is 15% and 5 values out of 24 exceed the tolerance,

  • for D modeling (AXIS: QUAD9), the maximum difference is 13.7% and 4 values out of 16 exceed the tolerance.

These tolerance excesses are observed for values close to 0.

The calculations were carried out in °C. Determining the difference, considering the temperatures in °F (as in the reference solution), gives a maximum difference that is very different from that obtained in °C (3% instead of 15%).

In addition, it was not possible to obtain the original reference (Kreith’s book), which was cited in the ANSYS audit manual. The method for obtaining the reference solution (graphical estimation) and its uncertainty are therefore not known.

The results are considered to be acceptable taking into account the points mentioned above.

This test made it possible to test the consideration of variable thermal conductivity within several models. The main commands tested are as follows:

  • DEFI_MATERIAU associated with the keyword THER_NL, making it possible to define the characteristics of a material whose characteristics vary according to temperature,

  • THER_NON_LINE command allowing the resolution of a non-linear thermal problem, stationary or not.