Thermal transient analysis in ANSYS Workbench 2

[MaxTemp2]

Dear all,

I have a simple geometry in ANSYS Workbench 11.0 created to study, if it can handle transient thermal analysis. The goal would be later to build a more detailed geometry and solve a coupled structural - (transient)thermal analysis.

After one week trying out everything my conclusion is, that something is wrong with my version of Workbench or my Institutes licence is not the proper one, because I get nonrealistic results.

It would be very kind, if somebody looked at the attached project and find my mistake. Or could somebody please recalculate the results (open the attached file, SimplePlate.zip in Workbench 11.0, click the solve button, wait 2-3 minutes) and tell me them? (maximum and minimum temperatures at each time-step are enough)
And please also write what kind of licence did you use.

Thanks for your time and help
MaxTemp2

Ps.: it is not a homework or something like that, I only want to figure out why my Workbench produces 304°C at certain time-steps.

 

[Transient1]

ANSYS lets you try to solve the problem but you are not getting the correct solution because the boundary conditions are not correct for a transient problem. Check out the theory manual for some background reading.

You have an initial condition, a convection heat load (a boundary condition), but no time dependent boundary conditions (i.e. heat load or fixed temperature conditions). You'll note that if you solved this as a steady state problem or just look at the Min/Max temperature at the 1 second mark you will see the answer you expect 300 C on the block because it starts at 300 C and your removing heat with a 300 C reference temperature so there would be no heat flow. The answer gets screwy because the problem set up doesn't make sense after the initial condition passes.

Basically, the correct way to set up this problem would be to select the block and apply a temperature condition to it (300 C). Then, delete the applied insulated boundary conditions. If a surface does not have a specified heat load or other boundary condition ANSYS treats the surface as a perfectly insulated boundary.

If you make a slice plain down the center of the block you will see a temperature distribution and see the outside surfaces are maintained at 300 C.

 

[MaxTemp2]

Thank you for your answer Transient1.
I tried what you said (300C on the block, after time-dependent convection boundary on the upper surface), but the max. temperature is going above 300 anyway. I think I am still doing something wrong.
Could you please send your solution to the problem? (I have read the whole help, but this is the first time i am working with Workbench...)

 

[MaxTemp2]

Dear Transient1 (and all others),

after reading your answer again 10-20 times and experimenting, I think, I would like to do another calculation, than that you were thinking, but I have problems anyway. I will try to explain my problem more detailed, I hope you can help:
- geometry: for simplicity let assume a simple plate (brick) of steel
- initial condition: is known from steady-state analysis, lets say again for simplicity 300°C everywhere.
- now at t=0, water is starting to flow over a surface of the brick at decreasing temperature (from 300°C to 200°C), while the remainder surfaces are perfectly insulated (later i would like to model symmetry, and I think this is the only way in thermal analysis). So I specify the heat transfer coeff. and bulk temperature for water (and for completeness the perfectly insulated boundary conditions).
There is no heat source in the model. I would like to know only, how the brick is cooling down with time (because of heat conduction inside the brick and heat transfer between brick and water).
The problem: the temperature is rising above 300°C significantly, and this is not realistic.

I attached this problem made with ANSYS Workbench 11.0. Please take a look at it. Any comments, solutions are welcome.

Thank you for advance
MaxTemp2

Ps.: I am new to ANSYS Workbench, and have not found any tutorials specifically for Wokbench (not for Classic) and transient thermal analysis.

 

[Transient1]

MaxTemp2,

Your right that the my suggestion didn't solve the problem. I only quickly looked at the first load step. I've looked the problem over again and the issue comes from first principles. Try to solve the problem by hand, you have a steel brick (4m x 10 m), 300 degrees celsius to start and you have a condition on one face of -40,000,000 Watts.
( In your problem you have 10,000 W/m^2-C, 40 m^2 area and have a 100 degree difference. So that means that convection condition is removing 40,000,000 Watts. ) Pull out a thermal text book. Can you solve that problem? If you can't solve this by hand, you cannot solve it using ANSYS.


There are tutorials in the help file for ANSYS workbench, just type in transient thermal. Additionally, you can go to

http://www1.ansys.com/customer

and get into the workbench forums which have numerous tutorials loaded. Also, check out

http://www.padtinc.com/epubs/focus/default.asp

You should understand that workbench is just a Graphical User Interface(GUI) for setting up problems the same way ANSYS classic does. So understanding something about how classic works will help you with workbench. For example, try setting up a simple beam bending problem with one side fixed and force on the end. You can select the solution file then go to tools>>output ANSYS input file. It's this input file workbench generates before issuing a solve.

 

[MaxTemp2]

Transient1,

I have calculated the results also by hand, and the minimum temperature is in good accordance (the temperature decrease on the convection surface is -0,6°C/s) with those from ANSYS. My problem is, that the maximum temperature is rising steadily above 300°C, which is nonsense, since there is no heat source in the problem.

Thx for the homepage, I will check it out.

 

[MaxTemp2]

Sorry to write again, I forget to ask you: do you have Workbench 11.0? If yes, could you please run the problem and tell me results?
I just tried out ANSYS Classic through "Open analysis in ANSYS". The values are OK (everything is the same; Workbench says maximum temp. after 10 sec is 307°C, ANSYS Classic 298,01°C). Now I am really confused. It seems my Workbench does not calculate well...

 

[Transient1]

Can you run through your calculation?

The difference between workbench and classic, I believe is due to the way results are presented. One is node values, the other is element averages. If you look up in the manuals for postprocessing you can probably find a definative value.

 

[MaxTemp2]

My calculations looks OK.


Due to the help, in ANSYS Classic the results are for the nodes. (See: Release 11.0 Documentation for ANSYS/4.3 - Types of data available for postprocessing - "Primary data consist of the degree-of-freedom solution calculated at each node ... Primary data in thermal analysis is temperature").
I have not found yet what kind of results Workbench calculates primarily. But if the values at the nodes are all 300°C or below it, than you cannot get values significantly over 300°C, whatever happens.

Transferring the results from ANSYS Classic back to Workbench does not work either: Workbench messes up all the data, and again, I see values above 300°C.

Anyway, thx for your comments. Now I am looking forward to the weekend. Next week I will look at the problem again, and if I dont find a solution, I will have to use Classic instead of Workbench.

Any further help is greatly appreciated.

 

[Transient1]

Well, this problem has confused me. If you look at the reaction heat loads (in Watts) at the surfaces it seems that you do not have a balance of forces. This would explain how you can get a temperature above 300 degrees celsius. So something is screwy with boundary conditions...but I've tried a few different methods of simulating your cooling problem and have come up with similar results.

By the way, I am using version 11.

Have a good weekend.

 

[89062517817]

Dear all,
I have a problem with my transient analysis using Birth and Death element's behavier. I use ansys commands and APDL programming language. I use DO Looping to perform my analysis. I want to kill elements which exceed strain criteria after finished one looping . In every looping i change loading condition and time at the end of load step.
Ansys performs all looping . But Ansys don't change load step number. Ansys performs every looping as load step number (1) and its substeps . After finished analysis ,i have only one load step number(1)and its sunsteps of the last looping. I want to be like that 1st looping is load step(1), 2nd looping is load step(2), 3rd looping is load step(3), and so on .

If i use ( ANTYPE, ,REST,,, )command in 2nd /SOLU routine, i don't get answer and Ansys give me warnings as follow;
[ *** WARNING ***
Both solid model and finite element model boundary conditions have been applied to this model. As solid loads are transferred to the nodes or elements, they can overwrite directly applied loads.
*** WARNING ***
Transient dynamic analysis has no mass.
*** WARNING ***
The requested EPLOT command cannot be executed.
Either elements do not exist or the selected elements
are not supported by PowerGraphics.
To correct this situation, try selecting different
elements,etc, or turning off PowerGraphics. ]

I attached my commands text file. Please take a look at it. Any comments, solutions are welcome.

Ps; Now in my text file, i closed !ANTYPE, ,REST,,, command.

With best regards
Aung Myo

 

[cbrn]

Hi 8906...etc...

please don't "hi-jack" other's threads for questions that don't have any relation. Open a new thread instead.

BTW, your question has been covered somewhere in this forum, I only don't remember where but a Search will certainly find it out.

Regards

 

[suganth]

Oh I have the same problem as Maxtemp 2 was having. Temperatures keep going above the initial value when it is supposed to cool, for me in a different case. I am trying on a hot body brought in contact with a cold one and want to observe the cooling profile of the hot body. I saw MaxTemp2's other thread explaining that we need to play with the time step size to get the right answer. Maxtemp; Do you mean getting the time step size to suit L square / alpha??
That didnt really work for me...I ve tried various other values for the time step size. None of them seem to work. do you have any other comments, were you able to solve your problem after all??