Altair HyperXtrude 2023 Release Notes

Altair HyperXtrude is a suite of finite element solvers for simulating the following manufacturing processes. These solutions also have interfaces in Inspire.
  • Binder Jet Sintering
  • Metal Extrusion
  • Polymer Extrusion
  • Quenching
  • Calibration
  • Metal Rolling
  • Friction Stir Welding
  • Resin Transfer Molding

Highlights

  • The coupled analysis with OptiStruct will now consider the effect of tool deflection on choke/relief in the bearing region.
  • Improvements to die gap opening thickness at the start of the bearing region
  • Improvements to heat transfer correlations used in quenching analysis including support for user-defined functions
  • New material models for polymer extrusion

All Solutions

Resolved Issues

Default BC is disabled
When BCs were missing on some external faces, the solver assumed a default BC condition. This was leading to some inconsistencies due to the assumed data for boundary conditions. In this release, the default BC option is turned off and the solver will stop execution when it finds external faces with missing BCs. (HXT-406)

Binder Jet Sintering

Enhancements

Exported STL file format improvements
To facilitate the import of created STL files in all other third-party CAD software, minor corrections were made to the exported STL file. (HXT-633)

Resolved Issues

Computing part center
The solver computes the part center and there was a minor inconsistency between the part scaling computation and when the part center was computed. This internal issue is now resolved. (GIT-57aec7fc)

Metal Extrusion

New Features

Effect of tool deflection on choke and relief
The metal extrusion solver uses Bearing Reference Surfaces (BRS) captured from the tool side of the mesh to determine the choke distribution on bearing die contacts and enables the specification of a complex choke in a natural manner. This is now extended to coupled analysis with OptiStruct. This major improvement recomputes the choke/relief angles, after the tool deflection analysis with OptiStruct, and then uses the recomputed angles in the bearing friction specification. To enable this, some improvements were made to data deck creation to preserve the behavior on the tool side 3D mesh for nodes on reference surfaces. (HXT-172)
User-defined function for quenching heat transfer specification
The quenching solver now supports UDF/DLL to enable specifying your own correlations to compute the HTC/Nusselt number based on the specific requirements for the air fan quenching. (HXT-675)
Improvements to die opening gap thickness computation
The die gap opening thickness at the start of the bearing region is determined using the Delauney Constrained Triangulation (DCT) method. This method can fail sometimes and is mesh dependent. When this approach fails, the solver resorts to a simple Ray Tracing (RT) method, even though this method is not accurate. RT is also mesh dependent and gave different results for each hole in a multi-hole die. (HXT-489)
In this release, bugs in the RT methods are fixed and it was made less mesh dependent and is no longer sensitive to meshes in symmetric models.
An improved Rotated Ray Tracing (RRT) method is also implemented as a default instead of the RT method. This is a better alternative when DCT fails.
This update also implements another RFE to predict the opening gap thickness variation. To achieve this, a CSV file is written with the initial and final gap thickness prediction for each bearing point based on the profile exit velocity. (HXT-612)

Enhancements

Material disappearance to numerical reasons in charge weld computations
In the charge weld (also referred to as billet interface tracking and transverse weld), material was disappearing from the dead zones even after the exit had 100% new billet material. This was due to the numerical smoothing of the hyperbolic governing equation and this issue is resolved now. (HXT-558)
Quenching results are written only after profile exits die
The model setup for quenching requires the entire profile length to be meshed, however, the results computed are relevant only for the extruded portion of the profile at any given time. The solver is now enhanced to write the results only for this portion and not write the initial data for the non-extruded section, which is a numerical artifact. To support this, a new parameter named DieAssemblyExitLocation is added to the solver. (HXT-640)
Improvements to strain prediction
The strain should be zero on the walls with no-slip condition with non-moving contact BC. The solver was enforcing only the values at the inlet and the numerical solution was expected to naturally capture the dead zone and zero strain on the tool contact surface. This issue is resolved now and will lead to more reasonable values of strain without hot spots and the dead zones are captured more accurately. (HXT-253, HXT-278, HXT-668)
Write unit information in the HMASCII file
To avoid errors because of the unit mismatch between extrusion analysis and tool deflection analysis, comment lines are written in the HMASCII file mentioning the units for the extrusion analysis. (HXT-571)

Resolved Issues

Inconsistent ram speed values in the data deck
The value of ram speed is specified in multiple places in the deck due to differences in the analysis types. There was an issue during multi-cycle analysis with different ram speeds and billet lengths for each cycle. The computed scaled variable step and change in billet length was leading to incorrect results. This issue is now addressed by making the ram speed consistent. (HXT-626)
Error in conform extrusion computation
For conform extrusion, there are two exits. One for the profile to exit and the other to account for the material to exit that does not go through the bearing and leaves the domain. The Lagrangian data structure creation was failing for the latter surface as it was a tetrahedral mesh. The Lagrangian data structure for nose-cone analysis is required only for the profile exit. Therefore, a check was added to avoid the creation of a Lagrangian data structure for the other exit. However, this fix would work for a regular tetrahedral mesh with a prism for the bearing/profile region. (HXT-94)
Dummy block position in multi-cycle analysis with different billet lengths
The multi-cycle analysis supports the starter billet length option and the possibility to change the billet length from cycle to cycle. This feature is often used without the dummy block and works correctly. When the dummy block (tool) is included in the analysis, this tool component has to reposition itself to be in contact with the billet-dummy block contact surface. There was an issue with this feature and the dummy block was not repositioning. This issue is resolved now. (HXT-13, HXT-163, HXT-676)

Polymer Extrusion

New Features

Masuko-Magill temperature dependence model
This model frequently fits the material data better and extends to a wider range more accurately than the WLF or Arrhenius material models. This model relates the viscosity at any given temperature to the viscosity at the glass transition temperature, two parameters A and B. More details of this model can be found in the reference paper – MASUKO, Toru, and Joseph H. MAGILL, "A comprehensive expression for the temperature dependence of liquid viscosity." Nihon Reoroji Gakkaishi 16.1 (1988): 22-26. (HXT-653)
A generalized polynomial viscosity model
A generalized viscosity model to fit more complex material behavior developed by Altair is implemented in this release. This model is more suited for rubber-like materials. The viscosity fitter in the compose toolkit can efficiently fit the viscosity data to the best-fitting material model and it includes this generalized polynomial viscosity model. (HXT-654)