Improvements
Issues | Short description of the source problem resolved in this version | |
---|---|---|
Data Import/Export context | FX-32299 | In Flux 2024.1, the internal evaluation of DataSupports in IO context has been renewed to reduce the creation time and improve the stability of all processes of the IO context, especially for Skewed applications |
Data Import/Export context | FX-33374 | When exporting several DataCollection in a single H3D file, there may be incompatibilities between collection. In Flux 2024.1, the error messages in case of incompatibilities have been clarified. |
Environment | FX-34057 | In release 2024.0, the third party
'Groovy' embedded in Flux was in version 2.4.8 and was including 'Xstream' jar
in version 1.4.7. This version of 'Xstream' contains some security
vulnerabilities. In release 2024.1, we have updated 'Groovy' to version 4.0.20 which is using 'Xstream' version 1.20. It fixes the vulnerabilities detected. Warning: Groovy scripts written by our customers may no longer be compatible with the new version of Groovy embedded in Flux. |
Physics | FX-34525 | In Flux 2024, it was impossible to create
a composed non-meshed coil with many points to define the coil path (of the
order of a thousand), because the generated python command became too long in
characters for Flux python interpreter. Now, in Flux 2024.1, we have introduced the possibility of defining a composed non mesh coil by importing a csv file which contains the definition of the points on the path: each line contains (x, y, z) coordinates of a path point and its curvature radius. Note: the parameters are supported during file parsing. |
Solving | FX-30911 |
In Flux Skew projects with at least one scenario already solved, users who set the 2D initialization for solving a new scenario, unintentionally perturbed the Flux solving process whose convergence was slower. In Flux 2024.1 version an error message is displayed and this option is automatically changed to prevent the former issue. In other words, the 2D initialisation for Flux Skew resolutions is available only for the first scenario to solve. |
Solving | FX-31141 | In 3D, a second order mesh with additional
nodes may be generated by Flux just before the resolution of a scenario if
required by the formulations of a magnetic application. Consequently, the command “Show recommended configuration” may display less than optimal recommendations when used in projects with a first order mesh. To remember users about this behavior, a warning message is now displayed by Flux 3D under such conditions. Thus, in version 2024, if the “Show recommended configuration” command is executed in projects discretized with first order elements, Flux will warn the user to regenerate the mesh with second order elements. Moreover, it is no longer possible to execute this command in projects that have not been previously meshed.Consequently, a warning message remembering users to mesh the geometry before using this command was added in Flux 2D and 3D. |
Solving | FX-32420 | The "Show recommended configuration"
command that suggests optimal solving options for a scenario is no longer
available under certain circumstances. More specifically, it's no longer possible to use this command in 2D/3D Electric applications ("Electric conduction", "Electro Static", "Steady State AC Electric", "Transient Electric" and "Electrolysis") and in Flux 3D "Magneto Static - Integral Method" applications because the recommendations were found inaccurate in those cases. |
Solving | FX-32983 |
Solving: - either 3D Steady State AC Magnetic projects containing "solid conductor regions described by surface impedance conditions" - or 3D Transient Magnetic projects containing "rotating mechanical sets" and "solid conductor regions" where nodal finite elements are employed for representing the electric vector potential T was only possible with MUMPS direct solver because of a limitation on the parallel iterative solver using PETSc. This limitation has been removed and it is now possible to solve such projects with PETSc, which guarantees better performances in terms of computation time and memory resources for large Flux projects. |
Solving | FX-33692 | Until Flux 2023.1, the “Show recommended configuration” command did not suggest using parametric distribution in the case of scenarios with fewer than ten parametric steps (physical or geometrical) in all cases. This behavior changed in Flux 2024 for Transient Magnetic scenarios. The user will now obtain recommended settings that require enabling parametric distribution in scenarios with at least two parametric steps and with at least ten steps, parametric and time steps. |
Solving | FX-34371 | In Standard mode, the information on the Linear Solver displayed in the Output window after a Check Physics command or a Show recommended configuration command has been improved: in both columns (current settings and recommendations) Mumps and PETSc terms have been replaced by Direct solver and Iterative solver, respectively, with the aim to be consistent with the labels indicated in the "Solving process options" dialog box. |
Solving | FX-34434 | In Flux 2024, when running a Flux solving
process with several pilots on a scenario and where a very large number of I/O
parameters/sensors could exist, Flux could seem to freeze! The cause was the automatic creation of curves at the beginning of the resolution. In some Flux projects, around 100,000 curves could be created and initialized, blocking the resolution for a while. So, for Flux 2024.1, we have added an optional field to the scenario to create these curves automatically or not. Here is the command, which is only managed by python script (no GUI): Scenario['NameOfTheScenario'].automaticCurvesCreation='no' |
Solving | FX-34712 | From Flux 2024.1, SuperLU solver, Solver by user subroutine and External solver, available only in advanced mode, will no longer be available. |