Modelling temperature as a stress field

[RedFrame]

Hi, I am new to Abaqus I would appreciate if the experts here could give me some simple advices.

Basically I am trying to simulate the stress rate when the model is cooled from 100C to 23C.The cooling rate is 2C/min.

Since the model is cooled uniformly , I do not need any heat transfer analysis. My first approach was to set predefined field in initial step to 100 and in Step 1(static,general) I set the predefined field to 23. This however doesn't seem to be correct as from my understanding abaqus only takes the property of the material at 100C and 23C respectively and for my case the elastic modulus and CTE is temperature dependent.

Can anyone here guide me to the proper approach?

Thanks!!

 

[rstupplebeen]

When defining the material for both elastic and expansion there is a "Use temperature-dependent data" check. See attached image. I hope this helps.

Rob Stupplebeen

 

[corus]

Abaqus interpolates between temperature defined properties, and for values specified outside of the temperature range, uses a constant value.
I can't see how time comes into this unless you have strain rate data, and to be honest 2C/min doesn't sound like it would cause a problem. Of course if you have uniform temperatures then you won't get any stresses unless you have restraints that oppose the expansion/contraction.

Hoping to say Tata

 

[RedFrame]

Many thanks for the replies!!!

My problem here is because the elastic modulus and CTE is temperature dependent.. let's say its not linear.. but Abaqus only deals with temperature A and temperature B... meaning that anything between A and B is ignored.. This is verified when I entered all kinds of crazy values in between 100C and 23C.. and the results are the same with the model that only have values of 100C and 23C.

Therefore, i am thinking that setting predefined field is not the correct way?

 

[rstupplebeen]

Sounds like a bug to me. Can you create a simplified case and state your version number. I could run it on my end on a different version and see if that clears it up. I hope this helps.

Rob Stupplebeen

 

[RedFrame]

To be honest I don't think its a bug because the main problem here is the Predefined field?
I am not sure if you set the the predefined field of initial step at 100C and then set the predefined field as 23C in step 1, will this represents thermal cooling? If that is the case is there anyway to simulate cooling rate?
This is because I don't want instantaneous cooling(point A to point B in 1sec) as the elastic modulus and CTE is temperature dependent, in fact it is not linear and I personally think that cooling rate plays some role in it?

 

[rstupplebeen]

RedFrame,
God call. Step 1 100C step 2 23C.

Rob Stupplebeen

 

[RedFrame]

rstupplebeen many thanks for assisting me.. but abaqus still only consider point A to point B although its step 1 to step 2 instead of initial step to step 1.

Are there any other ways to set the temperature instead of setting it in predefined field?

This forum is superb I shall contribute bk my knowledge once Im an abaqus expert

 

[corus]

You may have the time step set to 1 so that the temperatures jump to the final value.

To set the temperatures, you could run a separate thermal analysis and just fix the temperatures to vary by some amplitude over the time interval. Then read the odb file in for your stress analysis

Hoping to say Tata

 

[RedFrame]

Hi corus,
Sorry but how do i run a thermal analysis? I only have the Elastic modules , CTE , starting and final temperature.

If i'm not wrong Heat transfer analysis will require me to specify the density, specific capacity and conductivity?

Thanks!!

 

[corus]

Just put the other properties as 1. If you fix all the temperatures then material properties are irrelevant.

For a thermal analysis choose heat transfer, transient as the step. You'll also have to change your element type too.

Hoping to say Tata

 

[RedFrame]

Hi Corus, your idea really do make a lot of senses but forgive me as I do not understand how to do you "fix the temperatures to vary by some amplitude over the time interval"?

Let's say its 100C to 23C with -2C/min cooling rate, what will the amplitude be and where can I set this in the step?

Again I can't stress how greatful am I for all your help.

 

[corus]

Use the interaction module for the amplitude and fix the temperatures in the load module.

Hoping to say Tata

 

[RedFrame]

Hi corus,
In your post are you using heat trasfer analysis or static general? This is because in the heat-trasfer analysis there is no temperature to set in the load module (only Surface heat flux , body heat flux and concentrated heat flux)?
As for the interaction module, all of them are for surface contact whereelse I just need to perform uniform cooling in the entire model?

I see that you have replied in another post of mine. If I am not wrong doesn't time dependent Elastic modulus affect the strain rate therefore it is consider as strain rate dependent material properties?

Many thanks!!!

 

[corus]

Create the fixed temperatures in the load module under Create Boundary Condition. There the temperatures can be assigned an amplitude of values against time.

In the other post (which I can't find now) I'm sure you said you had temperature dependent properties, and not time dependent properties. If somehow you have a time dependent elastic modulus then at each discrete time step you will have constant temperatures and therefore a constnat/uniform elastic modulus. How will this give you any stresses?

Hoping to say Tata