1. Generalities#

1.1. background#

The objective is only to test the non-regression of future evolutions of Code_Aster and HOMARD. Even if the test cases are realistic from a physical point of view to represent real studies well, one should not attach importance to the value of the results.

In particular, there is no need to derive/to establish anything about the performance of the error indicator/mesh adaptation pair.

These test cases validate the functioning of the two macrocommands MACR_INFO_MAIL and MACR_ADAP_MAIL which control the entire process.

More specifically, the functionalities tested are as follows:

  • reads and writes meshes and fields in MED format. These are the IMPR_RESU, LIRE_CHAMP, and LIRE_MAILLAGE commands with the MED keyword as the format,

  • writing the data file for HOMARD,

  • launch of the procedure managing the execution of HOMARD. It’s command EXEC_LOGICIEL; it calls a script with a variable number of arguments,

  • control of the entire process by python: macr_adap_mail_ops.py.

The process is a prima facie insensitive to the modeling envisaged. The important points that cause different treatments in the exchange of data between HOMARD and Code_Aster are the types of elements, the management of the adaptation and the updating of fields on the new mesh.

Non-regression is tested on the value of the displacement, stress, or temperature field in a free node. The test takes place for several resolutions, those with the meshes from the first and from at least one other adaptation. In fact, the data transmissions and the management of HOMARD are not the same for the first and subsequent adaptations. It is therefore necessary to test at least two passages.

1.2. Calculation method used for the reference solution#

These test cases are non-regression test cases. The reference solution is the one obtained with a Code_Aster calculation.