Modeling A ============== Characteristics of modeling ----------------------------------- The modeling is D_ PLAN, only the edge of the frame is represented. Two calculations are performed with different linear solvers. Characteristics of the mesh ---------------------------- .. image:: images/100000000000034E00000321FF280112D7BB7D25.png :width: 4.1335in :height: 3.8673in .. _RefImage_100000000000034E00000321FF280112D7BB7D25.png: Number of knots: 396 Number of meshes and types: 320 QUAD4 for the plate and 32 SEG2 for the frame. Tested sizes and results ------------------------------ First calculation (master-slave normal, controlled geometric update, algorithm 'PENALISATION', solver 'MULT_FRONT') .. csv-table:: "**Identification**", "**Reference type**", "**Reference value**", "**Tolerance**" ":math:`\mathit{DX}` at the point :math:`A` instant :math:`1.0` ", "'SOURCE_EXTERNE'", "2.86E-5", "5.0%" ":math:`\mathit{DX}` at point :math:`B` instant :math:`1.0` ", "'SOURCE_EXTERNE'", "2.72E-5", "5.0%" ":math:`\mathit{DX}` at the point :math:`C` instant :math:`1.0` ", "'SOURCE_EXTERNE'", "2.28E-5", "5.0%" ":math:`\mathit{DX}` at point :math:`D` instant :math:`1.0` ", "'SOURCE_EXTERNE'", "1.98E-5", "5.0%" ":math:`\mathit{DX}` at the point :math:`E` instant :math:`1.0` ", "'SOURCE_EXTERNE'", "1,5E-5", "5,0%" Second calculation (master-slave normal, controlled geometric update, algorithm 'PENALISATION', solver 'LDLT') .. csv-table:: "**Identification**", "**Reference type**", "**Reference value**", "**Tolerance**" ":math:`\mathit{DX}` at the point :math:`A` instant :math:`1.0` ", "'SOURCE_EXTERNE'", "2.86E-5", "5.0%" ":math:`\mathit{DX}` at point :math:`B` instant :math:`1.0` ", "'SOURCE_EXTERNE'", "2.72E-5", "5.0%" ":math:`\mathit{DX}` at the point :math:`C` instant :math:`1.0` ", "'SOURCE_EXTERNE'", "2.28E-5", "5.0%" ":math:`\mathit{DX}` at point :math:`D` instant :math:`1.0` ", "'SOURCE_EXTERNE'", "1.98E-5", "5.0%" ":math:`\mathit{DX}` at the point :math:`E` instant :math:`1.0` ", "'SOURCE_EXTERNE'", "1,5E-5", "5,0%" We also test the projection by considering the penultimate slave node on the right. .. csv-table:: "**Identification**", "**Reference type**", "**Reference value**", "**Tolerance**" "PROJ_Xde CONT_NOEU instant :math:`1.0` ", "'ANALYTIQUE'", "3.88E-002", "1.0E-6%" "PROJ_Yde CONT_NOEU instant :math:`1.0` ", "'ANALYTIQUE'", "0.00E+000", "1.0E-6%" notes --------- The results obtained are close to the external source to within 5% (code mean). The linear solver has no influence on the results.