Altair Manufacturing Solver 2022.2 Release Notes

General

Altair Manufacturing Solver is a state-of-the-art solver suite for manufacturing applications built on a parallel, modular, and extensible framework that is suitable for simulations of manufacturing processes. This release contains solutions for the following modules.
  • Additive Manufacturing
  • Injection Molding (3D Solution)
  • Injection Molding (Shell Solution)
  • Metal Casting
  • Welding
  • Polymer Material Data Analytics (PMDA)
  • Molding Toolkit

Highlights

Highlights of this release include:
  • Altair Compose-based molding toolkit with calculator for injection molding analysis
  • Superfast shell-based computations of sink mark index and weld lines in injection molded parts

Additive Manufacturing (3DP for SLM)

New Features

Layer lumping technology to improve the accuracy of results and performance of the solver
The new layer lumping is a proprietary technology of Altair which estimates correctly the effect of layer-by-layer deposition when they are activated together. This allows for use of coarser meshes while maintaining good accuracy.
Recoater distance and accident prediction result
One of the main problems in 3D printing occurs when unsupported or partially supported regions move upwards. Due to the very thin layers used in metal 3D printing, displacements of a few tenths of a millimeter can lead to a collision between the recoater and the part. This result shows the minimum distance of the recoater with any node of the voxel. In the case that the model deformed too much in the Z direction and would have collided with the recoater, the solver stops with an error statement. Lumping can affect the accuracy of this result and it is strongly recommended to use a small layer to capture defects. (AMSLVR-417)
Failure index result
The contraction occurring every time a new layer is printed creates residual stresses that can lead to an increase of computed stresses beyond the elastic limit and potential failure of some regions or even the complete parts. This result shows the ratio between the effective stress and the yield stress in an element. When the numerical value of this index is above 1.0, it means it has exceeded the yield stress and will lead to part failure. The stresses are usually highest during the printing stage, and some portions of these stresses are released after springback. (AMSLVR-420)
Layer shifting result
Layer shifting affects the dimensional accuracy of the printed part and one of the causes of this shifting is the contraction of the immediately printed layer. After printing each layer, the subsequent melting and cooling produces contraction of that newly deposited layer. This leads to a deviation of the shape of the new layer from the intended printed shape. With no feedback control, the printer's tool head misses this deviation resulting in a small step and shifted layer. This new result shows the displacement occurring when a new layer is printed and is important for both surface finish and dimensional tolerance. The numerical lumping technique can affect the accuracy of this result and it is recommended to use small layers to capture this defect. (AMSLVR-419)

Metal Casting

New Features

Specify a symmetry plane
You can now specify a symmetry plane so only one-half of the part needs to be simulated. This greatly reduces the computational cost. (AMSLVR-479)

Enhancements

Improved convergence in sub-stepping during thermomechanical simulation
If in a certain time step during the thermomechanical simulation the solver does not converge with the current time step, this time step is divided in half until the solver converges. During this operation some internal variables were not being correctly reinitialized, leading to some difficulties in convergence. These have been corrected so the solver runs more smoothly during sub-stepping. (AMSLVR-494)

Resolved Issues

Filling issues when runner goes below the piston position
At the beginning of a filling simulation with a piston shot, the solver fills the runner system with the selected amount of material. When part of the runner was lower than the piston cavity in the gravity direction, the runner was filled in an unrealistic way at the start of the simulation. This was because the solver was filling from bottom to top in the gravity direction. Now, the solver ensures that the initial volume is set only on that part of the runner that is directly connected to the piston cavity, instead of only following the gravity direction. (AMSLVR-490)
Dependence of temperature results on number of threads
There was a minor error in the parallelization of the linear solver used in the mold thermal analysis. As a consequence of this, temperature results had small variations depending on the number of threads used for the analysis. This issue is now resolved. (AMSLVR-524)
Unfilled nodes in pouring simulation with crucible
In tilt pouring with a crucible, there are nodes that are not filled at the end of the simulation. The solver now extrapolates the filling time from the filled nodes to overcome this issue. (AMSLVR-504)

Injection Molding (3D)

New Features

Thermal residual stress in warpage computation
In this release, thermal residual stress that occurs within the mold from the instance of gate freeze until the part is ejected is also considered in the warpage calculations. This stress is included as initial stress while performing the OptiStruct analysis. This increases the accuracy of the predicted deformations due to part warpage. (AMSLVR-538)
Temperature dependent properties in warpage computation
Temperature-dependent material properties are now used for computing thermal residual stress in each timestep. This is computed using Tait's two-phase equation of state model and increases the accuracy of the predicted residual stress. (AMSLVR-538)
Fiber-fiber interaction in core and wall layers
Injection-molded parts often show three distinct layers due to the way fibers are oriented. There are two highly aligned wall layers on either side and a somewhat randomly oriented core layer in the middle. Higher shear rates near the wall lead to fiber orientation being more aligned with the flow, and very low shear rates in the core result in a random orientation and also a greater influence of incoming material orientation. Both these regions, however, show the effect of fiber–fiber interaction but with a differing degree of interaction. In this release, the solver is enhanced to account for this difference in the degree of interaction by introducing two different fiber-fiber interaction coefficients for the wall and core layers. The solver assumes 40% of the thickness is the core layer with 30% thick wall layers on either side. This data can be changed in the input deck. The solver also assumes a greater interaction (0.5) in the core than the wall (0.1). This data can also be changed in the input files. (AMSLVR-553)

Enhancements

Diagnostic statistics - additional results in CSV file
Additional results for each timestep are added in the filling stage CSV output file. These include maximum temperature, maximum velocity, maximum strain rate, maximum viscosity, and v/p control at the injection. (AMSLVR-539)

Resolved Issues

Temperature oscillations in models that include the mold
While solving the models including mold, there will no longer be numerical oscillations in temperature during the early stages of filling simulation. (AMSLVR-507)
Incorrect OptiStruct solver location
For warpage or FSI calculations, if the specified OptiStruct solver location in the input files is incorrect, the solver now issues a warning message in the output file. Even though the OptiStruct analysis cannot be performed, the molding solver exports the FEM input deck for stress analysis. This helps you manually run the OptiStruct analysis without having to rerun the molding simulation. (AMSLVR-512)

Injection Molding (Shell)

New Features

Sink mark indicator
A new and very fast module to compute a sink mark indicator is added in this release. The module assumes the shrinkage ratio in the thickness direction is a constant to predict sink mark depth and this shrinkage ratio varies depending on the polymer. Since the sink mark depth computation is not via a detailed 3D analysis, the results are displayed as an indicator using the computed sink mark depth. This result is classified into four stages: critical, visible, barely visible, and not visible. This result is useful for redesigning the part, especially the local thickness in the regions with significant sink marks. (AMSLVR-527)
Weld line prediction
A fast module for weld line prediction is added to fill pattern results. Weld lines occur when two melt fronts meet during the filling process. These are usually found around holes or obstructions and cause visible flaws and locally weak areas in the injection molded parts. You can move the weld line positions by adjusting the gate positions. (AMSLVR-528)

Polymer Material Data Analytics (PMDA)

Enhancements

Additional material models
PMDA now supports the power-law model to describe the polymer viscosity. In addition, Arrhenius and exponential temperature dependence functions are now supported.

Resolved Issues

Plots not closed during report generation
Some of the plots created in matplotlib were not closed and this was causing an issue in SimLab. This is now resolved by closing the plots after generating the images.

Molding Toolkit

The Molding Toolkit is a collection of utilities and libraries aimed at those performing injection molding and other simulations with polymers. This collection is programmed in Open Matrix Language (OML) used by Altair Compose The libraries can be installed either as a whole collection or the specific individual utility, or utilities, desired can be installed individually using the OML Library Manager in the Compose user interface. The specific application calculators can also be launched via Windows Explorer by double-clicking the <.OMD> file, as this is a "distributable" application that summons a lite-client of Compose to run the user interface.

The features and utilities that are made available in this release are:
  • Fiber volume fraction calculator
  • Polymer material library
  • Virtual injection unit
  • Virtual clamp unit
  • Viscosity pressure dependence (WLF D3 calculator)
  • Viscosity fitter

New Features

Fiber volume fraction calculator
This utility computes the fiber volume fraction needed in Multiscale Designer to create the unit cell model and develop the material model for short fiber-reinforced polymers. Material vendors typically provide the fiber content as a mass fraction; however, the simulations need them as a volume fraction. This utility helps to compute the fiber and matrix volume fractions. This feature has options to vary the inputs depending on what information is available and hence is useful for all levels of users.
Polymer material library
This utility is used to obtain the temperature-dependent elastic properties of unfilled polymers and produce either a user table file or a file to be included with OptiStruct that contains the MATT1 and TABLEM1 cards for elastic modulus, Poisson's ratio, density, and coefficient of linear thermal expansion. The data is fit by taking partial derivatives of the specific volume function to get the bulk moduli and thermal expansion rates and fitting the Poisson's Ratio to a TANH based function. The resulting parameters can then be saved for table creation and reuse.
Virtual injection unit
This application is designed to assist the performance of benchmark simulations of injection molding. The polymer melt at the molding press is compressible but in simulating the molding process the melt is treated as incompressible in the filling stage. Hence, a straight calculation of the screw positions taken from the process sheet or controller of the molding machine cannot be used to determine the transfer positions. This calculator helps to address this issue by determining the process conditions that can be correctly used in the simulation. In addition, it also takes into account the machine's hydraulic intensification ratio and computes the plastic pressures from the reported hydraulic pressure. This application helps to determine:
  • Material density for the analysis
  • Melt temperature for the analysis
  • Maximum ejection temperature
  • Transfer position
  • Hold pressure
Virtual clamp unit
This application takes information about a molding press and the applied hydraulic force from the mold cavity during the pack/hold phase, including the center of the hydraulic load with respect to the center of the platen, and computes the added stretch in the tie bars where the load has exceeded the applied force from the machine's clamp. It also computes the added force and added stress which could be used for fatigue calculations. Units are available in the application and changing will invoke a translation of the existing inputs and recalculation of the outputs.
Viscosity pressure dependence (WLF D3 calculator)
This utility computes an estimate of the D3 pressure dependence parameter on shear viscosity based on the transition temperature parameters from the specific volume parameters. For Tait, these are b5 and b6. The inputs are designed around a familiar expression of the Cross model of viscosity used with the Williams-Landel-Ferry (WLF) equation. It refits the viscosity so that the transition temperature is set to b5 as opposed to the D2 value in the input. In this way, the b6 value from Tait becomes the D3 parameter. The tool then refits the D3 in the refit viscosity back to the system of the original coefficients and provides the D3 estimate relative to the original inputs.
Viscosity fritter
The Viscosity Fitter performs a sophisticated fitting of corrected data from a capillary rheometer. All temperature sets must have the same number of shear rate and viscosity pairs (for example, shear rate and viscosity can be expressed as n x m matrices with as many rows as there are temperatures in the data). The data is assumed to have industry standard corrections for wall slip, shear rate, and entrance losses.
The fitter attempts to fit the Cross model first, then the Modified Cross model, and finally the Carreau-Yasuda model. Once the best fit is established for the constituent equation, the temperature dependence model is fit using the zero-shear-rate viscosities at the specified temperatures from the original data. The viscosity fitter can determine the transition temperature for the polymer from a file with specific volumes versus temperatures and pressures, or you can enter the transition temperature value manually. The following Time-Temperature-Superposition (TTSP) models are run together and then ranked based on the lowest residual error:
  • Williams-Landel-Ferry (WLF)
  • Exponential (Beta-Delta-T)
  • Arrhenius via Activation Energy (Q/RT)
  • Arrhenius via Reference Temperature (Tb/T)
  • Masuko-Magill (MM)
  • Normalized-Williams-Landel-Ferry (NWLF)
Reports are generated by each of the constituent fitters, and a final report is generated after the TTSP models are fit and ranked.