알테어 하이퍼웍스(Altair HyperWorks)는 모든 산업에서 이용되는 가장 포괄적인 오픈 아키텍쳐 CAE 시뮬레이션 플랫폼 입니다. 최고의 기술력을 바탕으로 설계, 성능 최적화, 중량 최적화 및 혁신적인 제품 제작에 도움을 줍니다.
알테어의 목표는 언제나 고객이 원하는 최고의 기술을 제공하는 것이며, 이번 하이퍼웍스 2017 릴리즈 역시 예외는 아닙니다.
이번 하이퍼웍스 2017 패키지는 다음과 같은 부분에서 다양한 새롭고 향상된 기능들을 제공합니다.
- 효율적인 모델 메싱과 관리
- 구조 최적화 및 설계
- 복합재 구조 해석
- 멀티피직스 및 계산 성능
- 고주파, 저주파 전자기장 해석
자세한 내용은 릴리즈 노트를 참조하시기 바랍니다.
… Release note for HyperWorks_2017_ReleaseNotes
하이퍼웍스 2017 공식 업데이트의 간략한 내용은 아래와 같습니다.
- Expanded support of faces and edges selection to additional panels.
- Quick window selection and deselection is supported for face and edge selection.
- Improved selection for contact definition via the Entity Editor.
- Improved performance and memory usage in Cross-Reference, Include File Isolate and Advance Delete functionalities.
- Added “Displayed” selection for CSV file export from ID Manager dialog.
- Import and export tabs fail to remember the correct tab location when you close and reopen the session.
- Selection of elements and surfaces occasionally selects elements and surfaces that are not visible.
- Application crashes while moving morph handles.
- Picking elements by domain sometimes stops working.
- Updating an independent node location for RBE3 creates a temporary node at its original location.
- Changing the set type option resets all the previously selected entities in the set.
- New preference option is added to reverse the background of a non-transparent legend when capturing an image or video.
- Image capture is enhanced to properly capture high-resolution images.
- Ability to set precision and format of displayed values is provided in the tensor panel.
- New report template parameters for parametrizing contour datatype and data component is provided.
- Advanced query is enhanced to work with result value for min/max/extreme layer contours.
- Ability to save and use a custom legend for FLD via preference file is added.
- Up and down arrows are provided in the vector and tensor panels to quickly scale the size of the arrows.
- Size of the circular border of a note is optimized to fit more closely around the text.
- By face selection method does not work when normal of elements are flipped.
- Result math operator BCNodeToElem shows N/A on mid-side nodes for 2nd order elements with corner bound values.
- Application crashes when animating an adaptive model with solid elements.
- Modifying averaging method drop down resets resolved-in system to Elemental system.
- Drop in performance when expanding all folders in the Include browser in Linux.
- Video tracking is not restored when session file is re-opened.
- Contour Legend appears vertically compressed on an Ultra-High Definition display in Linux.
- Saved/reloaded views result in the model
- Issue related to capturing image or printing graphics with category legends.
- Capturing video of graphics area flips transparency of legend background.
- Note is not displayed in graphics when created from the browser on some h3d files.
- Note shows N/A instead of actual values on a top/bottom based contour with corner data enabled and show mid-side node results option unchecked.
- HyperView holds results files loaded using advanced result-math template in memory even after clearing the session.
- Application crashes when working on Nastran and OptiStruct files and using the browser selector to pick a material or property in the graphics.
- Certain time increments of an Abaqus ODB not animating due to floating point precision issues.
- Default layer selection for top/bottom contour option is not as expected.
- Legend does not restore correctly when loading an H3D file created from a model with an N/A contour plot.
- Advanced query shows incorrect information when querying part-bound Min/Max/Extreme Layer contour.
- Nodal quick query does not work in some cases where the pool names are long.
- Toolbar disappears momentarily in HyperView Player when dragging and dropping an H3D file.
- FLD zone contour legend colors are used for Distance contour legend when using default button in edit legend dialog.
- Variation information not displayed in contour legend when set to zero.
- Loadcase names are not exported to h3d file after re-naming the derived loadcases.
The major items for the 2017 release are:
- Puck Failure Criteria
- Inertia Relief for models with greater than 6 rigid body modes.
- User-defined material (beta)
- Friction table support for Nonlinear Analysis
- Material Coordinate System for Solid Elements
- Complex Eigenvalue Analysis for Acoustics
- Power output for Heat Transfer (SPCFORCE)
- 64-bit OptiStruct support (Beta)
In addition to the OptiStruct 2017 Release Notes, this document also contains a list of all features released after OptiStruct 14.0. The version in which a particular feature was released can be determined by looking at the corresponding version number in parentheses.
Equivalence Two Grid Sets (2017)
All degrees of freedom of grid points of two different grid sets of a model can now be equivalenced. The pairs of equivalenced grid points are not required to be adjacent to each other. The GMATCH Bulk Data Entry can be used to achieve such equivalence.
Thickness Output Support for PCOMPP (2017)
Thickness output via the THICKNESS I/O Options entry is now supported for composite elements referencing PCOMPP Bulk Data entries.
64-bit OptiStruct (2017)
The 64-bit OptiStruct run can be activated using the –i64 run option. This feature avoids the 32-bit integer overflow in any OptiStruct module (Beta).
Duplicate Element IDs Allowed (14.0.230)
Duplicate element IDs between scalar elements (CELAS/CMASS/CDAMP) and other type of elements are now allowed. SYSSETTING,ELEMENTID is added for this purpose.
Missing Mid-side Nodes Allowed for 2nd Order Solid Elements (14.0.230)
SYSSETTING,ADDMIDSIDE is added for this purpose.
Bar/Beam Stress Output is Available for Transient Analysis (14.0.230)
Bar/Beam stress output is available for Transient Analysis.
Temperature Field Input with Temperature Gradient – TEMPP1 (14.0.230)
Temperature field input for plate elements that allows specification of a temperature gradient through the shell thickness.
H3D Element Erosion (14.0.230)
Supported for Sub-modeling (SUBMODEL) and Model Change (MODCHG).
OP4 Reader (14.0.230)
ASSIGN,OP4DMIG is now available to read reduced matrices in the .op4 file.
Neuber Stress and Strain Output on Surface of Solid elements (14.0.230)
The NEUBER option can now be used in conjunction with the SURF option on the STRESS and STRAIN entries.
Improved Stress Calculation for 2nd Order Shell Elements (14.0.220)
PARAM,CURVSHL2,YES will activate this calculation.
New Threshold Option (PTHRESH) has been Added (14.0.220)
This new option takes PTHRESH*maximum response value as threshold. This is available for ESE, STRESS and THERMAL output request.
Threshold Options for Normal Mode Analysis and Modal Frequency Response/Modal Transient Analysis (14.0.220)
Threshold option (TOP, RTOP, RTHRESH, and PTHRESH) has been added to Normal mode analysis and Modal Frequency Response/Modal Transient Analysis with MODAL output option.
Corner Strain is now Available for Modal Frequency Response/Modal Transient Analysis (14.0.220)
Corner Strain is now available for Modal Frequency Response/Modal Transient Analysis with MODAL output option.
Restart Functionality Added for Zooming (Global-Local) (14.0.220)
The results from global subcase can be saved and retrieved for local subcase job.
Fatigue Analysis can be Included in Zooming (Global-Local) Setup (14.0.220)
Fatigue Analysis can be included in Zooming (global-local) setup.
Globalization of any ID Group, GRIDs will be Available for Parts and Instance Feature (14.0.220)
With new I/O options entry IDGLOBAL is added. This card allows to disable treating specific group of cards as part-localized. For example, IDGLOBAL,GRID will treat all GRIDS as existing in a global part, and they can be referenced on other cards only by numeric ID, instead of qualified ID.
Centrifugal Force (RFORCE) can be Applied on Specific Element Sets (14.0.210)
The IDRF field on the RFORCE bulk data entry can be used to apply centrifugal force to a set of elements in a model. This allows the definition of different rotational forces to different parts of the model.
RADIOSS is used across all industries worldwide to improve the crashworthiness, safety, and manufacturability of structural designs. For over 25 years, RADIOSS has established itself as a leader and an industry standard for automotive crash and impact analysis.
In the RADIOSS 2017 release, physics modeling capability has been expanded to include additional advanced material and failure laws. Improved simulation driven innovation was enabled by improvements to solver robustness and accuracy and strong ties to optimization. Computation performance has been improved such that 16300+ cores could be used to simulate a 10 million element full vehicle model. RADIOSS has become even easier to use with the addition of new automated ways to handle poor solid element quality and contact initial intersections. Ease of use was improved with new composite modeling input that better matches the manufacturing process.
- Improved Finite Volume Mesh airbag robustness, efficiency, and modeling capabilities
- Improvements to cohesive materials and failure models
- Arruda-Boyce hyperelastic material model
- Stability improvements for existing hyper-visco-elastic materials
- General visco-elastic Maxwell model that can be added to many material models
- Simplified Composite modeling with new /PLY, /STACK, and /DRAPE entities
- Composites with multiple integration points per ply
- Composite ply-xfem shell with delamination prediction capabilities
- Hot-forming using new Thermal options for contacts
- Metal forming springback improved usability due rigid body motion removal
- Drawbeads contact parallelization and MUMPS solver for springback
- Automatic TYPE24 Contact segment offset to easily resolve initial penetrations and intersections
- Global time step method which calculates the stable time step based on the full model allowing higher steps and thus shorter run times
- Nodal Time Step compatibility with Advanced Mass Scaling
- Automatic methods to deal with poor solid element quality including negative volumes
- Extended element type and formulation support for Solid to SPH modeling
- Topology Optimization for RADIOSS Optimization
- RADIOSS Optimization with composite entities
- Improved usability with WARNING and ERROR message simplification and summary
Notable Resolved Issues
- Resolved a crash in AcuPrep with models that contained large numbers of simple boundary conditions and surface outputs.
- Resolved a problem in AcuSolve when trying to restart from a simulation that contained a moving mesh and a solid element set.
- Resolved a problem in AcuSolve when using ELEMENT_OUTPUT in a simulation that contained a moving mesh and a solid element set.
- Corrected a problem with the satisfy_boundary_condition option of the NODAL_INITIAL_CONDITION command such that it does not apply to all variables in the simulation. Instead, it is applied only to the variable referenced in the current NODAL_INITIAL_CONDITION command.
- Resolved a problem with compression heating that led to incorrect results.
- Fixed a bug associated with specifying anisotropic thermal conductivity through UDF.
- Fixed a segmentation fault in AcuPrep when a SIMPLE_BOUNDARY_CONDITION command was applied to an element set having medium=none.
- Fixed a number of issues associated with special characters in the PATH variable on Windows platforms.
- Resolved an issue with AcuProj when projecting results from a mesh that contained prism elements.
- Resolved a segmentation fault in AcuPrep when encountering improperly formatted input.
- Fixed a bug that impacted the behavior of ALE free surface mesh displacement when using periodic boundary conditions.
- Fixed a bug that caused spurious pressure and velocity fields on guide surfaces when using match_mesh_velocity.
- Fixed a bug in AcuTrans that led to invalid Ensight files.
- Fixed a bug in AcuSolve that prevented the user specified Jacobian from being included in the left hand side matrix when applying a surface heat flux element boundary condition through UDF.
- Fixed a bug in AcuSolve that caused incorrect results when specifying a nodal boundary condition for omega using a UDF
In version 2017, MotionSolve brings new capabilities, added functionalities to the existing capabilities and improvements in performance.
2D Curve to Curve Contact
The 2017 release of MotionSolve enables you to define rigid body contact between two curves that are each defined in two dimensions. Simulating rigid body contact in two dimensions between two curves is advantageous when:
- It is known a-priori that the contact occurs only within the plane in which the two curves are defined i.e. there are no out-of-plane contact forces that are expected.
- The curves between which the contact is to be calculated are smooth and represent the curvature of the 3D geometry well
The advantages of simulating rigid body contact using 2D curves over 3D tessellated geometry are many:
- More accurate results: Results are more accurate since there is lesser error due to discretization. Discretization of a 3D geometry into triangles occurs over 3 dimensions as opposed to 2 in curves
- Easily improve accuracy: Accuracy of the results can be improved quicker by introducing more points in the curve within MotionView.
Improving the accuracy of results when using 3D contact requires you to re-import the graphic and re-mesh or go outside the MotionView environment into a CAD tool to re-mesh. Re-meshing a 3D geometry will certainly take more time than refining a 2D curve.
- Faster simulation times: For geometries where contact occurs over a curve, using a curve-curve contact is a lot faster than using a 3D tessellated geometry since the solver has to work harder to determine contact for a 3D tessellated geometry which increases the overall simulation time
For more details on how curve-curve contact can be defined and solved in MotionView and MotionSolve, please refer to the respective user and reference manuals.
Faster Solution Times for Linear Solver
Parallelization has been enabled for certain operations within the default linear solver in MotionSolve. For models that take a large time to simulate, you should expect to see an improvement in linear solution time. The overall speedup in simulation time will be model dependent and will be influenced by pre/post processing times, time spent in user subroutine calls along with the time spent in the linear solution.
NLFE Validation Manual
Included in the documentation accompanying MotionSolve 2017 is a validation manual for the Non-Linear Finite Element capability in MotionSolve. The goal of this validation manual is to compare simulation results from models that include NLFE elements to known, analytical results.
The validation manual is available as part of the user’s guide for MotionSolve. Also included in the HyperWorks installation are models that are used in the validation manual. Please refer to the user’s guide for more information and the location of these models.
New Behavior for Kinematic Analysis
Previously, the solver time step for models running kinematic analyses was determined by the output step size. If your model contained event sensors, the only way to increase the accuracy of the sensors was to reduce the output step size which would result in larger output files (mrf, h3d etc.). Simulations like these thus demanded a lot of disk space, if the simulation end time was large, or if the model was simulated repeatedly.
With the current release, this behavior has changed. If your kinematic model contains event sensors, you may now specify a solver time step that is different from the output step size by using the parameter h_max. This lets you control the size of your output files independent from the accuracy of the sensors in your model.
Failure in Quasi-static analysis
With the previous release, a failure in analysis was observed for a class of models that contained massless bodies and ran quasi-static analysis. This was identified as a bug and has been fixed with the current release.
FEKO offers several frequency and time domain electromagnetic (EM) solvers under a single license. Hybridisation of these methods enables the efficient solution of a broad spectrum of EM problems including analyses of antennas, microstrip circuits, RF components and biomedical systems, placement of antennas on electrically large structures, calculating scattering effects and performing electromagnetic compatibility (EMC) studies.
The FEKO 2017 release is packed with features and improvements to create a platform for simulation-driven innovation. Since the 14.0 release, features have been made available to users in quarterly updates allowing users to take advantage of the extended capabilities as soon as they are ready. The most notable extensions in this release are:
Computational Performance Improvements
- The finite difference time domain (FDTD) solver supports OpenMP and MPI parallelisation allowing users to take full advantage of machines with multiple cores and multiple computation nodes in cluster environments.
- The ray launching geometrical optics (RL-GO) solver has been improved considerably in terms of speed and resource efficiency (memory reduction). Innovative algorithms select the most suitable ray distribution and automatically determine when enough ray interactions have been taken into account.
Extensions to the EM Solver Expands the Portfolio of Physics that can be Solved
- 3D anisotropic materials can be modelled using the FEM and FDTD solvers. 3D anisotropic media make it possible to solve circulators and other interesting devices.
- Improved multilevel fast multipole method (MLFMM) stabilisation allows large problems with intricate detail, that traditionally could have prevented the MLFMM to converge, to be solved.
Improved Model Creation and Mesh Preparation
- Extensions to the loft operator make the creation of transitions fast and easy. The new loft extensions also provide the user with more control over the loft surface creation.
- A new mesh engine generates improved meshes that often consist of fewer mesh elements that directly relate to a reduction of resources and faster simulation.
Model Interrogation, Validation and Reporting Improvements Make Simulation-Driven Innovation Easy
- The Cartesian surface graph allows users to visualise more data on a single graph, making it easier to identify patterns and make design decisions for further simulation. (14.0.430)
- Improvements to the windscreen visualisation allow users to see the layers and their relationship with regards to the active elements. This ensures correct model setup.
- Graphs have been extended with options to add text and shapes that allow improved reporting from within the FEKO interface.
The items mentioned above are only a few of the highlights of the last year’s development and the sections that follow provide a more detailed list of the changes. In addition to the FEKO 2017 Release Notes, this document also contains a list of all features released as updates to FEKO 14.0. The version in which a particular feature was released can be determined from the corresponding version number in parentheses.
Approach Specification Defaults
For each approach specification step, the presentation of the methods has changed. A subset of all available methods is shown, but all methods are available from an expanded list. This change makes the interface consistent with HyperStudy’s suggested best practices.
Input Variable Constraints
Define constraints that are functions of only the input variables. These special constraints can be enforced without any solver analysis to avoid running simulations that are known to fail. This can save run time and increase efficiency.
D-Optimal Design of Experiments
This is a new space filling DOE. The scheme distributes points in a space for use with a Fit. This is an excellent choice for sampling data for use in least squares regression modeling.
The Ordination tab is a general post-processing tab. The tab contains Principal Component Analysis bi-plots. These plots are used to identify relationships between variables and responses, especially in multi-dimensional problems.
Enhanced Fit Diagnostics
The Fit diagnostics table is enhanced with spark lines and color coding. The visual features aid in quickly assessing the Fit’s predictive quality.
- Modified Lattice Sequence sampling is available in a Stochastic approach
- String filtering tools added to the vector source and readsim expression builders
- Added a quick start example with discrete and categorical variables
- The HyperStudy editor supports drag and drop files
- New Fit approach report type: HyperStudy Fit (*pyfit)
- Improved support for 4k resolution including new visual icons
- Workbench connection upgraded to work with version 17.1
- Resolved an issue with ARSM and “Ignore failed analysis”
- Miscellaneous font and graphical improvements
- The DSS post-processing module is removed.