Jump to content

Simon Križnik

Members
  • Content Count

    1090
  • Joined

  • Last visited

  • Days Won

    99

Everything posted by Simon Križnik

  1. No problem Solid map applied on complex shapes can have issues with node associations between source and destination geom (like the 2nd picture with skewed mesh). Defining along geom might help. Another option is to duplicate and project elements (tool>project>to surface) from source geom to dest geom and pick those elements as elems to drag and elems to match, respectively. To model the difference in geometry through thickness I would trim the solid with sweep lines in the thickness direction using 2 lines forming the "L" shape (at the tip of the arrow in the 1st picture). Basically, we want inner edges to be imprinted on outer surfaces to facilitate mesh transition. The solid should then be split using the trim with plane/surf panel into 3 longitudinal parts. Then I would 2D mesh the side and bottom faces, which would be used in solid map panel to extrude all in one go (and delete the unwanted elements) or it can be done in three sequential steps through the thickness.
  2. 1. The idea is to use the 2D mesh to impose a mesh pattern on hexa mesh. If you want a specific mesh pattern in the through-thickness direction you can also 2D mesh the sides and use nodes or elements as guides (elems to drag/match, along geom). Note that imposing mesh patterns in 2 directions requires well thought out meshing plan as it needs to be 2 directionally mappable. Maybe you will find videos on solid map helpful . 2. surf geom in the case of provided screenshot can be set to none, because it is already defined by elems to drag. Select the surface opposite to the 2D mesh as dest geom. Please hit the like button on useful tips
  3. Sure, first create 2D elements with 2D Automesh panel (F12) on the top surfaces of solids. Then use these 2D elements to create the 3D hexa mesh in two different ways: -if the component has constant thickness you can use 3D>elem offset>solid layers panel -for better mesh control in the thickness direction use the 3D>solid map>general panel and select 2D elements to drag After Hexa mesh is created delete the 2D elements. Is there a reason to split the solid in such a pattern? Otherwise, only washers should be split around holes.
  4. Hi Aslan, it looks like solid map went wrong. The outline and temporary nodes are automatically created around distorted elements. You should delete the mesh and rebuild. If the component does not have complex thickness variation it can easily be meshed: first in 2D and then extruded to 3D hexa mesh. Also, considering only the displayed geometry splitting the part seems excessive, which also causes bad meshing around holes.
  5. Hi Marta, I suggest using .sta files instead of .sty files, which is a legacy format. The procedure to carry over stress and strain state of elements is to create STATE engine cards in the 1. step: /STATE/BRICK/STRAIN/FULL /STATE/BRICK/STRES/FULL if there are shells also create additional cards (search Radioss help with keyword: STATE) /STATE/DT (define start time and frequency to write state files and declare components to be included) Run the simulation which creates model_name.sta file, which is then imported into the original model. It will overwrite the original mesh with deformed and there should also be Tables entry in the model browser (stress & strain information for each element that can be edited). Attached are the modified files and also some tips about STATE files. inista.rar StateFile_1200.docx
  6. Hi Santosh Maharana, The rigid elements used for Interface nodes of flexbody should be RBE2 type (not RBODY). 1D>rigids> to create rigid spiders set nodes multiple nodes for dependent, and calculate node for independent. Also, make sure to set the elem types to RBE2. This tutorial might help you: MV-2010: Flexbody Generation using Flex Prep and Optistruct ERROR # 176 is because SPC constraints were applied on dependant nodes of rigid elements, which is not allowed. My guess for INFORMATION # 741 is because the rotational d.o.f. were constrained on solid elements, which do not have rotational DOF. INFORMATION # 742 & 743 explain this is unnecessary and they were automatically corrected.
  7. Hi, You can access the animation output file in the analysis folder (where you keep the model files) and open model_name.h3d file.
  8. Hi Marta, Mr. George P Johnson is inactive since May, which is unfortunate because his Radioss expertise is greatly appreciated. The files can be found on hard disk with HyperWorks installation: <install_directory>/demos/hwsolvers/radioss/14_Truck_with_FXB/VPG_Flexible_body/Model_EIG/TRUCK_EIG_* The model is from a Radioss tutorial RD-E: 1400 Truck with Flexible Body. TRUCK_EIG_0000.rad.zip TRUCK_EIG_0001.rad.zip You're welcome.
  9. Hi, you can try the following: -recreate geometry from FE: Geom>surfaces>from FE and if there are solids combine the surfaces with Geom>solids>Bounding surfaces -if you have CAD geometry file: import geometry and then associate the nodes to geometry by Geometry>Edit>Nodes>Associate
  10. Hi luckyluke, please look into the attached attribute compatibility chart. DRESP1 - Responses and Attributes.pdf
  11. Hi Bakr Amine, in addition to Premanand's suggestion I would also recommend the following Youtube channels for practical demonstrations: Altair University Altair India Student Contest ELEATION By Apoorv Bapat
  12. Hi Daniel, I do not think that Boundary conditions can model such behavior directly, as they do not allow to be updated during the simulation. My suggestions would be to model end stops as: -type 4 springs attached between cylinder nodes and dummy nodes with fixed BCs and non-linear stiffnes curve to model end stop constraint. -contact interface defined between 2 rigid planes with fixed BCs perpendicular to x axis at the limit distance as master surfaces and cylinder nodes as slave.
  13. Hi, I think the issue is because kinetic relaxation option is turned on (/KEREL), which is effective for settling under gravity load (see dummy positioning example), but is not appropriate for imposed displacements. In the model you have shared there are no incompatible kinematic conditions. However, there are some minor issues with the set up of simulation: -tied interface are referring to default type2 interface definition, which is not given -there are a total of 33 nodes for which no master surface is found during starter -the mass scaling is turned on but no timestep is defined (1e-2 will give mass error below 2%) -there are no boundary conditions (constraint) applied to Acetabulum, however the simulation still gives some displacements (in the order of 5e-3)
  14. This method will slightly change the element proportions near projected edge nodes (compared to the base mesh). If the gap is too big it might lead to linear 1D elements at those edges not in alignment with the rest. But I hope it will Hyperwork
  15. Hi Nagarjuna Gundamsetty, I would suggest Motionsolve, which is a Multibody Dynamics solver. Please refer to the tutorial MV-2021: Simulating an Automotive Door Closure Event. You may also find useful the presentation from 2nd European HyperWorks Technology Conference: Door slam simulation for durability analysis with Multibody Dynamics.
  16. Hi Burner2k, after projecting the original mesh to surface, you could project those edge nodes to line in order to close the gap.
  17. Hi, RBE2 and RBE3 are rigid links in FEM used for load transfer. These elements in fact will not deform as they are considered to be infinitely rigid (diamond under atmospheric pressure for example even though it's not infinitely rigid) and the sole purpose of these elements is to transfer the loads between the grid points or nodes. RBE2 transfers the loads in such a way that all the slave (dependent) nodes have zero relative deformation after the load application. Whereas RBE3 transfers the load using weighted average of slave nodes(independent nodes) position. Perhaps you will understand even better: https://iberisa.wordpress.com/2015/10/13/rbe2-vs-rbe3-on-femap-with-nx-nastran/
  18. Hi LOKESH KATARI, you are probably referring to Flexible body generation using the CMS modeling technique, integrated with multibody analysis if the model is set up in OptiStruct for multibody dynamics. This method uses linear superposition of modes computed using the Component Mode Synthesis (CMS) method. Several variants of CMS exist, each attempting to accurately capture the deformations and boundary conditions with a minimum number of modes. A CMS-based flexible body contributes degrees of freedom equal to the number of CMS modes in addition to the six degrees of freedom associated with the rigid body motion. This body is able to handle small or linear deformations only. If you are interested in theoretical background refer to the attached file. Non-Linear Finite Element (NLFE) body: This method uses the “Absolute Nodal Coordinate Formulation” to obtain a fully non-linear finite element representation of the flexible component. As the name suggests, this body is defined with respect to the global frame and does not have a local part reference frame like the linear flexible body. Each flexible component can be made up of several finite elements that represent flexibility in the component. Similar to traditional finite elements, this flexibility is determined by the geometric and material data specified for the elements. The NLFE body allows you to model geometric non-linearity (large deformations) as well as material non-linearity (hyper-elastic materials like rubber). Current support for the NLFE body is limited to BEAM and CABLE elements only. These elements are useful in modeling long, slender structures. Since this representation is fully non-linear, no reduction analysis (like CMS) is required to create this body – the body can be created and modified entirely within MotionView/Optistruct without the need for any FE solver based pre-processing. For more information on the NLFE body, refer to the Body: Flexible modeling component. Theory.pdf
  19. Hi, from Radioss help: Maybe you could model bonding with springs?
  20. Hi, there are two possible reasons: -force/moments were applied on temporary nodes -moments were applied to solid element nodes, which do not have rotational degrees of freedom please follow the links to resolve the issue or share the model
  21. Hi Stefano, revolute joints can be modeled with compliance in all DOF (equivalent to bushing). This can be done with Allow compliance option at the time of joint creation. The motion can be given to such joint, but it must be defined using i and j markers of the compliant joint instead of directly on a joint.
  22. Hi Amasker, maybe you can use the SB allowable inter-laminar shear stress (shear stress in the bonding material) in the PCOMP, PCOMPG or PCOMPP cards. The interlaminar shear stress is probably different for matrix and the cohesive bond- so it might not be appropriate in case you have more than one successive fabric ply. Optimization can also be done in Radioss, but it is not as user-friendly as Optistruct (no GUI) and it has some limitation. Please refer to the Design Optimization chapter and tutorial RD-E: 5101 Thickness Optimization for B-Pillar in Radioss help.
  23. Hi, I am glad you have solved the issue. Care to share the method used to manage it?
  24. Hi Sundaram, The problem is with the penta elements. PENTA6 is somehow causing the problem and update penta element type to PENTA6DG. Please update the elements (2D panel>Element Types) and try running the file. If errors arise you should first search the forum (at the top right). You could also do a google search using syntax: keywords site:forum.altair.com
×
×
  • Create New...