XANSYS Message: 18528 [Go back to message list]
[bookmark on del.icio.us]
No rating yet

Rate item: 10 (Good) 5 0 -5 -10 (Bad)

---

Subject: Re: Steady State to Transient Thermal, works
Author: Carlos Shultz
Date: 2000-11-30 16:33:00

Chris,
when you use the LDREAD command the temperatures are read in as BF loads
which are structural loads. you need to insert the following code to change
those to ICs for use in the thermal transient analysis. with this
additional code the procedure that you wrote,under the heading: TED HARRIS
SUGGESTION (DOESN'T WORK), will work.

Note, I didn't have time to vectorize this with implied do loops which is
the way to go if you work with large models. Also, I believe that the *get
of the bf temp is not yet documented.

*get,numnod,node,,count
*do,i,1,numnod
nodnum=ndnext(-1)
*get,nodtemp,node,nodnum,ntemp
ic,nodnum,temp,nodtemp
nsel,u,node,,nodnum
*enddo

good luck,
Carlos

Carlos Shultz
PADT
Phoenix Analysis and Design Technologies
www.padtinc.com
480-813-4884
480-813-4807 Fax

-----Original Message-----
Sent: Thursday, November 30, 2000 12:19 PM
To: xansys

JASON HUSBAND AND TED HARRIS,

> I am not convinced that your suggestion of turning on time
integration
> and restarting the steady state analysis will work. If you can do
this
> please post the deck.

HERE IS THE INPUT DECK. IT IS VM116 PROBLEM. IT DOES STEADY-STATE
SOLUTION AND RESTARTS THE SOLUTION WITH TIME INTEGRATION EFFECTS ON.
HENCE A USER DOES NOT NEED TO RERUN THE INITIAL STEADY-STATE ANALYSIS.

YOU SHOULD ASK YOUR ANSYS SUPPORT DISTRIBUTOR IF THIS STILL DOESN'T
CONVINCE YOU.

> I believe that Mr. Harris's strategy is sound and could be
> made to work. We regularly perform tricks such as this to map

I CAN ALSO PROVE THAT TED HARRIS'S SUGGESTION DOES NOT WORK.
STRUCTURAL THERMAL LOADS DO NOT EQUAL INITIAL TEMPERATURE D.O.F.
BOUNDARY CONDITIONS. I MODIFIED VM116 BELOW (SECOND INPUT DECK) TO
PROVE THIS AS WELL.

*******************SS TO TR RESTART*******************************

/PREP7
/TITLE, VM116, HEAT CONDUCTING PLATE WITH SUDDEN COOLING
C*** PRINCIPLES OF HEAT TRANSFER, KREITH, 2ND. PRINTING, PAGE 161,
EX. 4-11
ANTYPE,STATIC
ET,1,LINK34,,3 ! CONVECTION LINK, USE (TI-TJ) FOR HF EVALUATION
ET,2,LINK32 ! HEAT CONDUCTION BAR
R,1,1 ! UNIT AREA ASSUMED
MP,KXX,1,2 ! CONDUCTIVITY, DENSITY AND SPECIFIC HEAT
MP,DENS,1,800 ! INPUT USED BY CONDUCTION ELEMENTS
MP,C,1,0.833
MP,HF,1,2,.02 ! TEMPERATURE DEPENDENT HF (USED FOR ELEM 1)
N,1 ! NODES 1 AND 2 DEFINE THE CONVECTION LINK
N,2 ! (ARBIRARY LENGTH)
N,10,(8/12) ! CONDUCTION LENGTH IN FT.
FILL
E,1,2 ! ELEMENT 1 IS CONVECTION LINK
TYPE,2 ! ELEMENTS 2 THROUGH 9 ARE CONDUCTION BARS
E,2,3
EGEN,8,1,-1
FINISH
/SOLU
D,2,TEMP,500 ! INITIAL SURFACE TEMPERATURES
D,10,TEMP,100
OUTPR,,LAST
OUTRES,,ALL
TIME,0.001 ! INITIAL STEADY-STATE CONDITION
SOLVE
FINI
/POST1
SET,LAST
PLNS,TEMP
FINI
/SOLU
ANTYPE,STATIC,REST
TIMINT,ON ! TURN ON TIME INTEGRATION ON FOR
TIME,7 ! TRANSIENT OVER 7 HRS
DDELE,2,TEMP ! DELETE NODAL TEMPERATURE
D,1,TEMP,100 ! ENVIRONMENT TEMPERATURE IS DECREASED
KBC,1 ! SUDDENLY
AUTOTS,ON
NSUBST,20
SOLVE
FINISH

/POST26
NSOL,2,2,TEMP
PRVAR,2 ! PRINT TEMPERATURE HISTORY AT NODE 2
/AXLAB,Y,TEMP
/GRID,1
PLVAR,2
FINISH

/POST1
ETABLE,TI,SMISC,2 ! NODAL TEMPERATURES FOR CONDUCTION ELEMENTS
ETABLE,TJ,SMISC,3
PLLS,TI,TJ ! DISPLAY TEMPERATURE VARIATION ACROSS PLATE (AT 7
HRS)
PRNSOL,TEMP
*GET,TN2,NODE,2,TEMP
*DIM,LABEL,CHAR,1,2
*DIM,VALUE,,1,3
LABEL(1,1) = 'T,F(AT X'
LABEL(1,2) = '=0.0 in)'
*VFILL,VALUE(1,1),DATA,285
*VFILL,VALUE(1,2),DATA,TN2
*VFILL,VALUE(1,3),DATA,ABS(TN2/285)
/COM
/OUT,VM116,VRT
/COM,------------------- VM116 RESULTS COMPARISON -------------
/COM,
/COM, | TARGET | ANSYS | RATIO
/COM,
*VWRITE,LABEL(1,1),LABEL(1,2),VALUE(1,1),VALUE(1,2),VALUE(1,3)
(1X,A8,A8,' ',F10.0,' ',F10.0,' ',1F5.2)
/COM,-
/OUT
FINISH
*LIST,VM116,VRT

***************TED HARRIS SUGGESTION (DOESN'T WORK)*********

/PREP7
/TITLE, VM116, HEAT CONDUCTING PLATE WITH SUDDEN COOLING
C*** PRINCIPLES OF HEAT TRANSFER, KREITH, 2ND. PRINTING, PAGE 161,
EX. 4-11
ANTYPE,STATIC
ET,1,LINK34,,3 ! CONVECTION LINK, USE (TI-TJ) FOR HF EVALUATION
ET,2,LINK32 ! HEAT CONDUCTION BAR
R,1,1 ! UNIT AREA ASSUMED
MP,KXX,1,2 ! CONDUCTIVITY, DENSITY AND SPECIFIC HEAT
MP,DENS,1,800 ! INPUT USED BY CONDUCTION ELEMENTS
MP,C,1,0.833
MP,HF,1,2,.02 ! TEMPERATURE DEPENDENT HF (USED FOR ELEM 1)
N,1 ! NODES 1 AND 2 DEFINE THE CONVECTION LINK
N,2 ! (ARBIRARY LENGTH)
N,10,(8/12) ! CONDUCTION LENGTH IN FT.
FILL
E,1,2 ! ELEMENT 1 IS CONVECTION LINK
TYPE,2 ! ELEMENTS 2 THROUGH 9 ARE CONDUCTION BARS
E,2,3
EGEN,8,1,-1
FINISH
/SOLU
D,2,TEMP,500 ! INITIAL SURFACE TEMPERATURES
D,10,TEMP,100
OUTPR,,LAST
OUTRES,,ALL
TIME,0.001 ! INITIAL STEADY-STATE CONDITION
SOLVE
FINI
/POST1
SET,LAST
PLNS,TEMP
FINISH

/PREP7
ET,1,8
ET,2,8
FINI

/SOLU
LDREAD,TEMP,,, , ,,rth,
FINI

/PREP7
ET,1,LINK34,,3 ! CONVECTION LINK, USE (TI-TJ) FOR HF EVALUATION
ET,2,LINK32 ! HEAT CONDUCTION BAR
FINI

/SOLU
ANTYPE,TRANS
TIME,7 ! TRANSIENT OVER 7 HRS
DDELE,2,TEMP ! DELETE NODAL TEMPERATURE
D,1,TEMP,100 ! ENVIRONMENT TEMPERATURE IS DECREASED
KBC,1 ! SUDDENLY
AUTOTS,ON
NSUBST,20
SOLVE
FINISH

/POST26
NSOL,2,2,TEMP
PRVAR,2 ! PRINT TEMPERATURE HISTORY AT NODE 2
/AXLAB,Y,TEMP
/GRID,1
PLVAR,2
FINISH

/POST1
ETABLE,TI,SMISC,2 ! NODAL TEMPERATURES FOR CONDUCTION ELEMENTS
ETABLE,TJ,SMISC,3
PLLS,TI,TJ ! DISPLAY TEMPERATURE VARIATION ACROSS PLATE (AT 7
HRS)
PRNSOL,TEMP
*GET,TN2,NODE,2,TEMP
*DIM,LABEL,CHAR,1,2
*DIM,VALUE,,1,3
LABEL(1,1) = 'T,F(AT X'
LABEL(1,2) = '=0.0 in)'
*VFILL,VALUE(1,1),DATA,285
*VFILL,VALUE(1,2),DATA,TN2
*VFILL,VALUE(1,3),DATA,ABS(TN2/285)
/COM
/OUT,VM116,VRT
/COM,------------------- VM116 RESULTS COMPARISON -------------
/COM,
/COM, | TARGET | ANSYS | RATIO
/COM,
*VWRITE,LABEL(1,1),LABEL(1,2),VALUE(1,1),VALUE(1,2),VALUE(1,3)
(1X,A8,A8,' ',F10.0,' ',F10.0,' ',1F5.2)
/COM,-
/OUT
FINISH
*LIST,VM116,VRT

--- In xansys , Jason Husband wrote:
> Mr. Corbes
> I definately have performed transient analysis before so please
allow me
> to comment. I believe that Mr. Harris's strategy is sound and could
be
> made to work. We regularly perform tricks such as this to map
results
> sequentially back and forth between thermal and structural analyses.

> I believe the original problem of Mr. Singer was that he had
results
> from a static run and then wanted to continue with transient
analysis.
> For reasons unknown, he is reluctant to re-run this 1 step but is
> willing to run X number of steps in a transient (probably a few
percent
> increase in CPU time). Several Xansys members then offered possible
> solutions to help him.

> I am not convinced that your suggestion of turning on time
integration
> and restarting the steady state analysis will work. If you can do
this
> please post the deck.

> In general, for XANSYS as a whole, some of the suggestions are good
and
> some are not as good. Hopefully we can keep this group as a
discussion
> forum where people can offer solutions, good or not as good, and
not
> recieve a hostile reply.

> Best Regards
> Jason Husband
> QuEST

> Steve Cobes wrote:

> > TED HARRIS,

> > YOUR KIDDING RIGHT???!!!??? YOUR TELLING ME THAT STRUCTURAL
> > TEMPERATURE LOADING IS THE SAME AS THERMAL D.O.F.
SOLUTION???!!!???
> > THAT IS RIDICULOUS!!! YOU OBVIOUSLY HAVE NEVER DONE THERMAL
> > TRANSIENTS BEFORE.

> > EDMUND SINGER,

> > AS EVERYONE ELSE HAS RECOMMENDED, RESTART SOLUTION AS STEADY-
STATE
> > AND TURN ON TIME INTEGRATION EFFECTS. THIS DOES NOT REQUIRE RE-
> > SOLVING OF STEADY-STATE SOLUTION. DO NOT DO WHAT TED HARRIS
> > RECOMMENDS, SINCE IT IS OBVIOUSLY MORE STEPS AND IS WRONG.

> > --- In xansys , "Ted Harris" wrote:

> >> Edmund,
> >> There may be another way to do it, but I quickly tried using the
> >> LDREAD command to read the temperatures from the steady-state

> > analysis

> >> into the transient thermal analysis. I had to temporarily
change

> > the

> >> element type in /prep7 to a structural element, read in the
temps,

> > and

> >> then change back to a thermal element type. There may be a
better

> > way

> >> to do it, but hopefully this will work for you. Here is a
snippet

> > of my

> >> .log file:

> >> !steady-state run has been completed
> >> /SOLU
> >> !*
> >> ANTYPE,4
> >> !*
> >> TRNOPT,FULL
> >> LUMPM,0
> >> !*
> >> FINISH
> >> /prep7
> >> et,1,42
> >> FINISH
> >> /SOLU
> >> LDREAD,TEMP,,, , ,therm,rth,
> >> FINISH
> >> /prep7
> >> et,1,55
> >> FINISH
> >> /SOLU

> >> Ted Harris
> >> Phoenix Analysis & Design Technologies
> >> Gilbert, Arizona
> >> 480-813-4884
> >> 480-813-4807 fax
> >> www.padtinc.com
> >> www.anspak.com

> >>> -----Original Message-----
> >>> From: edmund_singer@u... [mailto:edmund_singer@u...]
> >>> Sent: Tuesday, November 28, 2000 11:48 AM
> >>> To: xansys
> >>> Subject: [xansys] Steady State to Transient Thermal

> >>> I am trying to run a transient analysis with starting

> > temperatures

> >>> that I have obtained from a previous Steady State analysis. I
am
> >>> unable to restart the analysis since the analysis type
changes.

> > Is

> >>> there a slick way to start with the steady state temperatures,

> > and

> >>> then run the transient, or must I store them in a table or

> > impress

> >>> them directly?

> >>> Thanks

> >>> Edmund Singer
> >>> Senior Engineer
> >>> Applied Mechanics
> >>> United Defense L.P.

---

Posts possibly associated with message #18528		Author	Date	Score
18528	Re: Steady State to Transient Thermal, works	Carlos Shultz	2000/11/30
18534	Re: Steady State to Transient Thermal, works	chankumar8@	2000/11/30
18571	Re: Steady State to Transient Thermal, works	Carlos Shultz	2000/12/01