Thermal transient analysis with ANSYS

[MaxTemp2]

Dear all,

I am trying to calculate a simple thermal transient problem in ANSYS Workbench 11.0 to see how it works to make later more complicated calculations. But I have a problem:

My model:
-steel plate (vary simple geometry)
-starting temperature: uniform 300C
-boundary conditions: 150C and 10000W/m2C at one surface, perfectly insulated at other 5 surfaces
-time: 100sec with automatic time-steps

The results show, that the temperature can go above 300C, what is impossible in real.

I tried to solve the problem with refined mesh, more substeps to apply the load, calculating a steady-state solution for the transient case´s initial condition, but neither of them worked.

Can somebody explain me the problem? Or simply run the problem in ANSYS Workbench 11.0 and tell me his/her results?
The input files can be found under the following link:

http://www.math.bme.hu/~reisst/ansys.html

OR the one at the end of the post.
I am using "ANSYS Academic Teaching Advanced" licence, maybe that is the problem. (I do not have any other licences)

Thank you for your help
MaxTemp2

 

[sailoday28]

Your boundary condition at the non insulated face seems kind of odd to me. I would have expected something like one or the other.

 

[MaxTemp2]

Dear sailoday28,

Sorry, but i do not understand what you wrote. Can you specify it, please? (For a convection boundary you need to specify film coeff. and ambient temp. also)

 

[sailoday28]

Try solving a simple one dimensional transient problem.
Such as at time = 0, temp distrib is known.
At boundary x=l, perfect insulation, dt/dx=0
At x=0, Temp is Ti OR
At x=0, flux is Q/A But note the OR. Two conditions are not specified at x=0

 

[IRstuff]

Is the 150ºC a "boundary" condition, or the ambient air temperature?

Is the 10kW/m^2-ºC supposed to be input heat or somehow representing forced convection out? It seems absurdly high for a convection coefficient.

TTFN

FAQ731-376

 

[MaxTemp2]

I think I understand what you mean sailoday. I dont know how to do a 1D analysis in Workbench yet, but I will try to put the load conditions (convection) at a different time, lets say at 10 sec. (Maybe tomorrow, now I should finish working )

IRstuff: The 150C is the ambient temperature (bulk temperature of water) and 10kW/m^2C is the heat trans. coeff. The actual problem would be: "cold" water (150C) is injected into a reactor pressure vessel which is at normal operation temperature and pressure. 10kW/m^2C is an estimated value, of course it depends also on the massflux of water, temperature difference etc., but it is a good approximation.

 

[IRstuff]

Is this for school? Student posting is not allowed.

TTFN

FAQ731-376

 

[MaxTemp2]

No. I am working for an institute and thermal-structural analysis of RPV is my theme.

 

[MaxTemp2]

Applying the loads in small time-steps does not work either.
Everytime I decrease (increase) the ambient temperature, there are parts on the plate, where the temperature increases (decreases).

 

[Onno]

Dear MaxTemp


IRStuff makes a very good remark. What is the 10000 W/m^2 heat source exactly? How big is the plate where it dumps it heat?

I would consider the following: when you do transient simulations you'll end up with the steady state system situation after some simulation time. (consider the input heat load 100000 kW/m2 constant!!)

In contrast with transient problems, steady states are (more) easy to calculate. Try estimating it!! on the basis of Input heat = output heat.
So, the first question you ask yourself after a transient simulation leading to steady state: is the steady state comparable with the estimated one? Then you'll get some feeling for the values simulated.

Greetings

Onno

 

[MaxTemp2]

Dear all,

I have found the problem. If somebody is interested: look at "Thermal Analysis Guide" in the documentation, 3.4.3.2/"Number of substeps per load step, or the time step size". Here you will find a formula, which MUST be used, to get good results in transient thermal analysis.

MaxTemp2

Ps.: To my defence: while reading hundreds of pages of documentation, someone can overlook these things at the first time.

 

[IRstuff]

Well, a good lesson learned... ;-)

TTFN

FAQ731-376