Fixed Defects

Issues		Short description of the source problem resolved in this version
Data Import/Export context	FX-31125	The Flux internal function of the Import/Export context for identifying the mesh element where a given set of coordinates is situated was affected by some limitations impacting its usage when evaluating data collections on moving mechanical sets in Skew module. Consequently, data collections could be wrong in very specific configurations. Flux 2023.1 version fixes the issue.
Flux e-Machine Toolbox	FX-27444	Flux e-Machine Toolbox versions before 2023.0 was developped with Java 8 and NetBeans Platform 12.5 which was causing warning messages due to deprecated libraries. Since version 2023.1, Flux e-Machine Toolbox moved to Java 11 and do not longer use NetBeans Platform. Warning messages do not longer prompt in batch mode.
Mesh Import/Export	FX-31313	For a mesh import coming from SimLab, due to the conformity of the ids of the instances of geometric entities between SimLab and Flux and when these ids were of the order of a million (e.g. edge id = 2,200,142) then a significant slowdown occurred during the mesh import of the project. This had an impact on the use of the dynamic tables of the import algorithms because they were sized with the maximum ids. This has been fixed in version 2023.1.
Parallel Computing	FX-30762	A bug was preventing any project using parameters in a mechanical set to be run using parametric distribution. This issue is now fixed.
Parallel Computing	FX-30830	In previous versions, using the batch solving feature of the Supervisor was not compatible with the parametric distribution, as solving was launched sequentially. In this new version, the parametric distribution can be used with batch solving transparently.
Physics	FX-27839	In Flux 3D only, an issue in the management of the orientation of B vector values impacted the accuracy of both Bertotti and LS iron losses on regions associated with a moving mechanical set (typically, the rotor of an electrical machine). The issue has been fixed in Flux 2023.1 version; nevertheless, in cases where the magnetic flux density presents large time variations in the area close to an airgap, it is strongly advised to use 2nd order mesh elements to get accurate iron losses computations.
Physics	FX-30760	Under certain circumstances, the static initialization of Transient Magnetic applications through a file led to a bug that ultimately resulted in the corruption of the Flux 2D project. This issue has been solved in Flux 2023.1 thanks to improvements to the file reading and interpretation processes.
Physics	FX-26010	The predefined sensor labelled as "Inductance (Magnetic flux divided by the current)" is no longer available in standard mode: it has been moved down to "Advanced mode" only. In fact, this sensor provides as a result the magnetic flux in the considered coil divided by the current flowing inside, without subtracting the contributions of the other coils which are coupled to the considered coil by their mutual inductance. In some particular cases (e.g., only one coil in the project or specific current values supplying the coils), the results provided by this sensor have a physical meaning, but in many other situations results are confusing and lead to misunderstanding. Users working in "standard mode" and interested in computing the ratio between the magnetic flux and the current flowing inside a coil are invited to define a "Magnetic flux" sensor and then the appropriate formula for an I/O parameter. Moreover, the computation of inductance matrices (apparent and incremental) is available in post-processing via the "Computation" menu, for the moment for 2D Transient Magnetic application only, being very time consuming for 3D projects.
Physics	FX-32017	The "New B(H) 2D Curve (Material)" command, when applied to any of the nonlinear magnets, displayed the transversal and the module curves in addition to the B(H) curve in the magnetization direction, since the "Transversal relative permeability" is one of the required characteristics to describe such magnet models. As these transversal and (especially) module curves are meaningless from a physical point-of-view and can be confusing for the user, they are no longer displayed for all nonlinear magnets, i.e. for "Nonlinear magnet described by Hc and Br module", "Nonlinear magnet described by Hc and Br module + knee adjustment" and "Nonlinear magnet described by a spline" models.
Physics	FX-27336	1) The creation and the displaying of the B(H) 2D curve of a material which had a spline B(H) property with 103 points was not possible and the following message was displayed: Message: MRC2BE => Arrays TABH and TABB out of bounds (1) Message: MRC2BE => Failed Message: MRC1BE => Failed … This first bug has been corrected. 2) A project has a region of type "Solid conductor region described by surface impedance", assigned to a material with a spline B(H) property with a B(H) curve which has an inflection point at the end of the Rayleigh zone. With this project, the solving process was stopped, and the following message was written in the Flux3D.report file: Z1CZS2 => No convergence of NR algorithme Z1CZS2 => Failed MRCCZH => Failed ... This second bug has been corrected. Notice that the results obtained with a project containing a region of type "Solid conductor region described by surface impedance", assigned to a material with a nonlinear B(H) property will give different results than in the previous versions, i.e., until version 2022.3. Bugs that led to slightly wrong results have been corrected. The results are better in the new version 2023.
Physics	FX-30983	In the previous version, while editing a sensor some irrelevant prompts were being displayed in the output. In this new version, these prompts are disabled and thus improving a bit the editing time of a sensor.
Physics	FX-31910	Spatial quantities of subtype “constant tabulated” or “for multiphysics” that were created prior to all the regions were not automatically taken into account or updated, thus resulting in an error when trying to solve the scenario. This issue has been fixed in 2023.1 Flux version.
Postprocessing	FX-13950	In Flux projects containing periodicities (both even/cyclic or odd/anticyclic), Laplace force arrows (dFLapV quantity) were incorrectly reconstructed in the periodical parts when the user displayed the full device. This problem has been resolved in Flux 2023.1.
Solving	FX-31569	In the previous version, using the torque ripple as an objective could lead to an error in the free-shape optimization. In this new version, the torque ripple can be use for all optimization.
Solving	FX-30947	In Steady State AC Magnetic application coupled with Transient Thermal, when defining "first order approximation" for both scalar and nodal vector variables, Flux mistakenly stopped the solving process at the end of the first step. The issue has been fixed and now Flux works as expected.
Solving	FX-30594	In version 2022.3 in 2D axisymmetric applications the 1st order mesh (which was not the default setting) did not produce optimized results in some cases. In Flux 2023.1 to avoid any confusion it has been decided to disable the usage of this 1st order mesh for 2D axisymmetric applications: the 2nd order mesh is - from now on - the only available, except if you have selected in Flux supervisor the "Advanced mode" which is intended for expert users. A message is displayed to the user when he runs the "check physics" command and the solving process is stopped in case he sets the 1st order mesh.
Solving	FX-30807	A clearer error message is now provided when a user does not have Optistruct installed and tries to run a free-shape optimization with remeshing. Many other error messages were also improved for free-shape optimization with remeshing.
Solving	FX-30940	In previous free-shape optimization version, saving a project in a different directory before computing an optimization could lead to blocking errors. In this version, the saving issues are resolved.
Solving	FX-30924	In previous versions, the units for volume constraints were confusing. No help was provided to the user in the graphical interface, and the values requested from users were in m3. For this version, the values requested from users for volume constraints are in mm3 and this is explicitly written into the graphical interface. The retrocompatibily of older projects created with versions previous to Flux 2023.1 has been ensured. This means that Flux 2023.1 will automatically convert values stored in m3 to mm3. Similar changes have also been made to the move limit and the sensibilities filter optimization options. Previously, values were requested in m. Now values are requested in mm.
Solving	FX-27898	In the case of the MT3TOM and MT3TSCRHJ formulations with real movement, the recommended solver and automatic solver have been replaced by MUMPS instead of PETSc, as we cannot currently solve this type of project with PETSc.