O modeling ============== Characteristics of modeling ----------------------------------- We use a 3D modeling in non-linear thermal with hyper-reduction of the model and finite element correction (option CORR_COMPLET). Characteristics of the mesh ---------------------------- The mesh contains 216 elements of type HEXA8. Tested sizes and results ------------------------------ With CORR_COMPLET ~~~~~~~~~~~~~~~~~~ The value of the temperatures obtained with a hyper-reduced model (with correction) is tested by compared to those obtained in the complete model: .. csv-table:: "**Location**", "**Instant**", "**Component**", "Temperature", "**Precision**" "Knot in :math:`({1.5,1.5,3})` "," :math:`t=1s` "," TEMP "," ", "29.9070478306"," 0.007%" "Knot in :math:`({1.5,1.5,3})` "," :math:`t=4s` "," TEMP "," ", "139.467784038"," 0.002%" "Knot in :math:`({1.5,1.5,3})` "," :math:`t=7s` "," TEMP "," ", "567.145887211"," 0.0004%" "Knot in :math:`({1.5,1.5,3})` "," :math:`t=10s` "," "," TEMP ", "999.500384048"," 0.0006%" "Knot in :math:`(3,0.5,0.5)` "," :math:`t=1s` "," TEMP "," ", "22.6152896917"," 0.45%" "Knot in :math:`(3,0.5,0.5)` "," :math:`t=4s` "," "," TEMP ", "90.756408026"," 0.05%" "Knot in :math:`(3,0.5,0.5)` "," :math:`t=7s` "," TEMP "," 380.605752326", "0.007%" "Knot in :math:`(3,0.5,0.5)` "," :math:`t=10s` "," TEMP ", "931.460613403"," 0.015%" Without CORR_COMPLET ~~~~~~~~~~~~~~~~~~ The value of the temperatures obtained with a hyper-reduced model (without correction) is tested by compared to those obtained in the complete model: .. csv-table:: "**Location**", "**Instant**", "**Component**", "Temperature", "**Precision**" "Knot in :math:`({1.5,1.5,3})` "," :math:`t=1s` "," TEMP "," ", "29.9070478306"," 0.007%" "Knot in :math:`({1.5,1.5,3})` "," :math:`t=4s` "," TEMP "," ", "139.467784038"," 0.0025%" "Knot in :math:`({1.5,1.5,3})` "," :math:`t=7s` "," TEMP "," ", "567.145887211"," 0.0009%" "Knot in :math:`({1.5,1.5,3})` "," :math:`t=10s` "," "," TEMP ", "999.500384048"," 0.00075%" notes --------- We can see that the calculation with CORR_COMPLET is a bit better on the evaluation of temperatures.