• Announcements

    • Rahul Ponginan

      Please click here for a short but important announcement   03/26/17

      Dear Users Our Commercial and Academic users around the world can use these same forums here as before i.e. the Altair Support Forum , Commercial users from India with solver queries can go to the Solver Forum for India Commercial Users , Academic Users from India and AOC India Participants are requested to go to the Forum for India Academic Users and AOC India Participants , We will be tending to all queries in all the forums promptly as before, thank you for your understanding. 

Q.Nguyen-Dai

Members
  • Content count

    840
  • Joined

  • Last visited

  • Days Won

    13

Q.Nguyen-Dai last won the day on April 14

Q.Nguyen-Dai had the most liked content!

3 Followers

About Q.Nguyen-Dai

  • Rank
    Super User

Profile Information

  • Gender
    Male
  • Country
    Belgium
  • Interests
    FEA, Scripting Perl/Tcl/PHP/Bash, Linux Sysadmin, Network
  • Are you University user?
    No
  1. Have a look at .... Hooke's law of Elasticity Theory
  2. Generally, you can not setup a Cluster with "normal" computers A Cluster needs some specific hardward & software. Cluster installation is the job of IT Professional, not for everybody Myself I can't do that.
  3. You need MPI Solver when you have a Cluster. You NOT need MPI solver when you have only ONE computing box with multiple cores.
  4. Maybeyour simulation does not correctly present real behavior of the part? Check BC, material, loading,... and meshing.
  5. Just checked your model. For Solid mapping, you have to split your complex geometry in several "mappable" solids. Have a look at some tutorials of Hypermesh and TRY firstly on your geometry.
  6. Ahhh, I understand now. You got script from someone and you don't know how to customize. If you want to modify structure of config file, you have to adapt TCL too. So if "FORCE" card is modified, go to "includes\create_loadcases.tcl" and look for the line proc ::CreateLoadcase::ReadConfigFile { ConfigFile } { Within this proc, you have to modify the code for "FORCE" card: ... } elseif { $Card == "FORCE" } { set NodeName [ string trim [ lindex $Data 1 ] ] set Factor [ string trim [ lindex $Data 2 ] ] set ForceData [ string trim [ lindex $Data 3 ] ] set LoadGroup [ string trim [ lindex $Data 4 ] ] set FX [ string trim [ lindex $ForceData 0 ] ] set FY [ string trim [ lindex $ForceData 1 ] ] set FZ [ string trim [ lindex $ForceData 2 ] ] lappend FORCE($Index) "{ $NodeName } { $Factor $FX $FY $FZ 1 } { $LoadGroup }" } elseif { $Card == "MOMENT" } { ... At the end of this script, you have to insert added data into list to return. And so on... Good luck!
  7. But without neither your TCL script nor your TXT files, how could I help you?
  8. Just write a small proc to extract ID-system from TXT file. Why use such complicated format in your TXT file?
  9. Already created by my script within the set "E_T"
  10. I know nothing about BC & Loading of your analysis. So I can't say more about the results. But in reality, you can not tie the pin to 2 parts. In general, we tie pin to one part (external) and make slide contact to another part (internal). Numerically, you can make both slide contact to 2 parts, but you have to add some BC to pin to keep them out of rigid modes.
  11. As you do all with solid meshing, contact should be defined between ONE part and the pin. Another part will be "tied" to the pin. This method needs some attention about meshing technique: contact zones should be correctly splitted on all parts. Why you make all in TET mesh? This structure is rather simple so it's better to work with HEX mesh, right?
  12. Could you share some pictures to better understant your question?
  13. Here's the TCL script, based on tinh's ideas: *createmark elements 1 "displayed" *findedges1 elements 1 1 0 0 30 hm_createmark elems 1 "by comp name" "^edges" *findmark elems 1 257 1 elements 0 2 *entitysetcreate "E_T" elements 2 *clearmark elems 1 *clearmark elems 2 *clearmark sets 1 *createmark comps 1 "displayed" foreach i [hm_getmark comps 1] { set cname [hm_entityinfo name components $i] if {$cname == "^edges"} { continue } set set_name "E_$cname" hm_createmark elems 1 "by set name" "E_T" hm_createmark elems 2 "by comp id" $i *markintersection elems 1 elems 2 eval *entitysetcreate "$set_name" elems 1 } *clearmark elems 1 *clearmark elems 2 My script will also create one element Set for each component.
  14. If you can share a simple model (with some component only) so I can do the test?
  15. There are several methods to do that. I can imagine a simple method: Make a node set for each component Look for intersection (nodes) between sets From intersection nodes, look for elements which share these nodes HTH,