Fixed Defects
| Issues | Short description of the source problem resolved in this version | |
|---|---|---|
| Environment | FX-17947 | FluxSupervisor was using Sigar library to get memory information which seems to crash.
Sigar library is no more used and we are using our custom algorithms to get memory information (tested and validated on Centos 8.2 ) |
| Environment | FX-17111 | The script used to launch a file in a web browser was deprecated for newest versions of linux desktop environments.
An update of this script has been performed (tested and validated on centos 8.2 using Gnome Classic with X11 display server) |
| Environment | FX-12844 | The sketcher background options "grey" (available on Supervisor options) is now taking into account as the sketcher background of the modeler 3D |
| Environment | FX-11472 | With Flux2020.1, if you run:
newProject() exit() -->a directory "1531985695567" (or similar) was not clean in user tmp (pex: C:\Users\toto\AppData\Local\Temp on Windows or /tmp on Linux)
Now the directory appears after the new project and is deleted when the application is exited. |
| Environment | FX-11315 | With Flux 2020.1 if a project was saved into the "temp" directory the Flux project was automatically deleted when the project was closed.
On Windows this "temp" directory is the one pointed by the %TEMP% environment variable. On Unix this is usually "/tmp" or "/var/tmp"
In Flux 2021, this problem has been fixed, and a project saved in the "temp" directory is no longer deleted on closing. |
| Examples | FX-17782 | The PDF document of the 3D example "induction heating with shower and hardening" accessible via the supervisor didn't match with this example.
The right PDF document associated to this example has been implemented in Flux 2021. |
| Macros | FX-13263 | The macro HalbachMagnetization2D.PFM was stored in the bad directory "Macros_Flux2D_Postproc ", now it is stored in the right directory "Macros_Flux2D_Physics".
The macro HalbachMagnetization3D.PFM was stored in the bad directory "Macros_Flux3D_Postproc ", now it is stored in the right directory "Macros_Flux3D_Physics". |
| Mesh | FX-16763 | The button "Mesh Face" is now masked in the application of type "Magneto-Static 3D - Integral Equation Methods" since face meshes are not supported. |
| Mesh Import/Export | FX-17950 | Defect on mesh import devices: the solving process is impossible when a sliding surface between two mechanical sets is defined by more than one line (2D) or face (3D). The 2D lines (or 3D faces) were detected as not belonging to a single curve (or surface). Now with Flux 2021, this problem has been fixed and the solving process can be done. |
| Mesh Import/Export | FX-17040 | In Flux 2020.1 the "Check mesh" failed after importing a specific .bdf mesh from SimLab.
The "check mesh" command just after a mesh import returned an error because some information useful to Flux was missing. This problem is now fixed in Flux 2021 and a treatment is performed at the end of the mesh import process to fill these missing information related to the mesh. |
| Parallel Computing | FX-16985 | When a project uses parametric distribution it is split into multiple parallel computation jobs.
In Flux 2020.1 when a computation was failing, the corresponding job was sometimes reported as successful in the distribution manager. In Flux 2021 when such a computation error occurs the corresponding job is correcty reported with an "Error" status. |
| Solving | FX-14265 | Fix corrupt project when some certain parametric step are not possible with a mechanical set in movement
It happens that during the resolution of a project with geometric parameters and with mechanical sets, in translation and with a compressible, some calculation steps cannot be resolved (collision of mechanical assemblies for example). the project saved at the end of the resolution was corrupted due to a bad management of the error on these problematic calculation steps. This defect is now fixed in Flux 2021 and now the Flux project is correctly saved. |
| Solving | FX-16825 | A limitation on solving and post-processing computations has been removed: if the jacobian of an element was negative, these computations were forbidden.
Case of null jacobian is not allowed because it comes from a poor quality of the meshing element. |
| Solving | FX-16640 | A bug has been reported while using "Direct solver" and the out of core strategy on Linux. The error was related to the folder specified to store temporary files, which was not found by Flux.
This issue is now solved. |
| Solving | FX-12653 | The correction of interpolation of nodal variables (scalar or vector variables) on the sliding surface was not correctly done when:
- The mesh has a second order after the solving of the project - The option "Approximation order for scalar variables" is set to "First order" or the option "Approximation order for vector variables" is set to "First order", or both options are set to "First order" in the "Formulation model" tab of the dialog box of the defined magnetic application. In these cases, the bug was leading to wrong results. The bug is now corrected. |
| Solving | FX-16596 | A bug has been reported when computing the inductance matrix on projects that uses parametric distributed scenario.
This issue is now fixed. |
| Solving | FX-15051 | A limitation, previously set to one hundred, on the number of circuit-coupled conductors in Flux3D projects has been removed. |
| Solving | FX-16767 | With Altair Flux 2021, you can solve and have good results for a Steady State AC Magnetic coupled with Transient Thermal project with the magnetic vector potential formulations and moving mechanical sets. For this type of project, movement is now well taken into account in order to obtain the right results. |
| Solving | FX-16765 | With Altair Flux 2021, you can solve a Steady State AC Magnetic coupled with Transient Thermal project with the magnetic vector potential formulations without error messages during the solving. |