| |  | Gut zu wissen: Hilfreiche Tipps und Tricks aus der Praxis prägnant, und auf den Punkt gebracht für Ansys | | |  | Funktionale Sicherheit: ein Kinderspiel?, ein Fachartikel
|
Autor
|
Thema: Coupling restart problem (4277 mal gelesen)
|
serg1976 Mitglied mathphysicist - researcher

 Beiträge: 50 Registriert: 03.05.2007
|
erstellt am: 13. Nov. 2008 13:31 <-- editieren / zitieren --> Unities abgeben:         
After the restarting thermal->structural- the output only for the thermal analysys, what is wrong? Code:
/NOPR ! Suppress printing of UNDO process /PMACRO ! Echo following commands to log FINISH ! Make sure we are at BEGIN level /CLEAR,NOSTART ! Clear model since no SAVE found /input,menust,tmp,'',,,,,,,,,,,,,,,,1 /GRA,POWER /GST,ON /PLO,INFO,3 /GRO,CURL,ON /CPLANE,1 /REPLOT,RESIZE WPSTYLE,,,,,,,,0 /prep7 et,1,solid87 mp,kxx,1,3 ! Conductivity BLOCK, 0,0.1, 0, 0.3, 0, 0.5 ESIZE,0.05 VMESH,ALL tunif,100 ASEL,S,AREA,,1 NSLA,,1 ! SELECT TARGET NODES ON BIGGER CYLINDER d,all,temp,500 ! Set outer wall temperature allsel ASEL,S,AREA,,2,6 NSLA,,1 ! SELECT TARGET NODES ON BIGGER CYLINDER sf,all,conv,2,100 allsel physics,write,thermal ! Write the thermal physics file physics,clear ! Clear all bc's and options et,1,solid92 mp,ex,1,30e6 ! Define structural steel properties mp,alpx,1,.65e-5 mp,nuxy,1,.3 ASEL,S,AREA,,1 NSLA,,1 ! SELECT TARGET NODES ON BIGGER CYLINDER d,all,all ASEL,S,AREA,,2 NSLA,,1 ! SELECT TARGET NODES ON BIGGER CYLINDER cp,NEXT,ux,all *GET,NC,NODE,,NUM,MIN ! GET LOWEST NODE NUMBER (MASTER) allsel d,NC,ux,0.1 ! physics,write,struct ! Write the thermal physics file physics,clear ! Clear all bc's and options fini ! /SOLU physics,read,thermal antype,trans,new OUTPR,,LAST OUTRES,,ALL TIME,10 DELTIM,1,1,2 NSUBST,20 SOLVE FINI ! /SOLU physics,read,struct antype,stat,rest ldread,temp,,,,,,rth SOLCONTROL,0 NLGEOM,ON AUTOTS,ON OUTPR,,LAST OUTRES,,ALL TIME,20 DELTIM,1,1,2 NSUBST,20 SOLVE
Eine Antwort auf diesen Beitrag verfassen (mit Zitat/Zitat des Beitrags) IP |
ife Mitglied Berechnungsdienstleister FEM
   
 Beiträge: 1397 Registriert: 29.10.2002 IFE Deutschland Simulation ANSYS Workbench MAPDL Multiphysics CFX
|
erstellt am: 13. Nov. 2008 14:51 <-- editieren / zitieren --> Unities abgeben:          Nur für serg1976
|
serg1976 Mitglied mathphysicist - researcher

 Beiträge: 50 Registriert: 03.05.2007
|
erstellt am: 13. Nov. 2008 15:20 <-- editieren / zitieren --> Unities abgeben:         
The idea to solve in cycle thermal structural, thus in any case it need: antype,stat,rest !,and than: antype,trans,rest !,and than: antype,stat,rest Possible you know how to transfer the mesh into another stage of the solution like forces in ldread-comand? Eine Antwort auf diesen Beitrag verfassen (mit Zitat/Zitat des Beitrags) IP |
ife Mitglied Berechnungsdienstleister FEM
   
 Beiträge: 1397 Registriert: 29.10.2002 IFE Deutschland Simulation ANSYS Workbench MAPDL Multiphysics CFX
|
erstellt am: 13. Nov. 2008 16:28 <-- editieren / zitieren --> Unities abgeben:          Nur für serg1976
if you intent to cycle, construct a *DO *ENDDO loop containing conditional *IF statements: which makes the execution initially pass over 1) the script-snippet containing antype,stat,new and subsequently makes the execution pass over 2) the script-snippet containing antype,stat,rest cycle.txt (see attachment, please verify syntax) is provided as a template only, for the structural part. ------------------ Gruesse, Frank Exius IFE Deutschland www.ife-ansys.de Mo-Fr 9:00-18:00 Uhr durchgaengig Eine Antwort auf diesen Beitrag verfassen (mit Zitat/Zitat des Beitrags) IP |
MGebhardt Mitglied Dipl.-Ing. Masch., WiMi

 Beiträge: 95 Registriert: 15.09.2008 2 * Intel Core 2 Duo 2,66 GHz, 2* 4 GB RAM, OpenSUSE 11.0 Ansys 11.0 Academic Research
|
erstellt am: 13. Nov. 2008 18:08 <-- editieren / zitieren --> Unities abgeben:          Nur für serg1976
Hello, I recently had some problems with restarting, as ANTYPE,,REST overwrites variables (including the loop counter) to their very first value. That drove me mad. I needed to write them into a file and get them back, after the restart. Haven't checked your code yet, just an initial guss. /EDIT: Just checked your problem: my guess has nothing to do with your problem. Your problem really is none. Results are stored in different files (.rst for structural and .rth for thermal). If you would like to have a look the temperature distribution, you can do that in the GUI with plot results -> nodal solution -> (scroll down) -> body temperature. As far as I can remember: If you would like to have the full thermal solution at hand you have to issue the appropriate PHYSICS command PHYSICS,read,thermal and then ->read results (sorry I'm a GUI user when it comes to postprocessing) I can't check it right now, so please try by yourself and give me a feedback. I think of a workaround if it does not work. EDIT/ Kind regards Moritz ------------------ Moritz Oliver Gebhardt [Diese Nachricht wurde von MGebhardt am 13. Nov. 2008 editiert.] Eine Antwort auf diesen Beitrag verfassen (mit Zitat/Zitat des Beitrags) IP |
serg1976 Mitglied mathphysicist - researcher

 Beiträge: 50 Registriert: 03.05.2007
|
erstellt am: 17. Nov. 2008 14:07 <-- editieren / zitieren --> Unities abgeben:         
Thank you for the interests  Thus the problem is simplier, it doesn't create *.rst at all!, but it's calculating structural part. (I suppose it's not possible: antype,trans,new -> antype,stat,rest). So how to create *.rst file (for my input-code)? Eine Antwort auf diesen Beitrag verfassen (mit Zitat/Zitat des Beitrags) IP |
MGebhardt Mitglied Dipl.-Ing. Masch., WiMi

 Beiträge: 95 Registriert: 15.09.2008 2 * Intel Core 2 Duo 2,66 GHz, 2* 4 GB RAM, OpenSUSE 11.0 Ansys 11.0 Academic Research
|
erstellt am: 17. Nov. 2008 15:49 <-- editieren / zitieren --> Unities abgeben:          Nur für serg1976
Ups, should have read the other comments, starting with the first from IFE (I promise, I'll never answer posts shortly before going home again ) Another try to convince you from ANTYPE,STAT,REST is wrong in the way you use it: Both Physics domains are solved independently from each other. Just start the first structural run with ANTYPE,STATIC,NEW Later, when you call the next cycle you must use, ANTYPE,STATIC,REST but not the first time you call it. This is independent from the thermal-model, which you did correctly start initially with ANTYPE,TRANS,NEW. Just change this (as all the other posts have already suggested) and it should work. Some hints: -I did not understand, why you set a time for the structural analysis -If you encover problems with your time counter when cycling through a *do-loop, just tell and also see my first post (Might only occur under LINUX). Maybe its easier to start with issuing the solution procedures one by one, till you are statisfied. Then get the whole thing condensed in a loop. -you can issue all /SOLU arguments before PHYSICS,WRITE. This keeps you from executing them in every cycle. I'm slightly stunned that your ANSYS does not stop with a huge ERROR-Message. Kind Regards Moritz ------------------ Moritz Oliver Gebhardt Eine Antwort auf diesen Beitrag verfassen (mit Zitat/Zitat des Beitrags) IP |
serg1976 Mitglied mathphysicist - researcher

 Beiträge: 50 Registriert: 03.05.2007
|
erstellt am: 24. Nov. 2008 13:17 <-- editieren / zitieren --> Unities abgeben:          Nur für serg1976
OK, the idea to enforce ANSYS to solve thermal-structural task at the single database (avoid additional load transfer actions), as I understand it is possible only through the restart, thus how to do so? Eine Antwort auf diesen Beitrag verfassen (mit Zitat/Zitat des Beitrags) IP |
| Anzeige.:
Anzeige: (Infos zum Werbeplatz >>)
 |