Temperature decreases instead of rising !!

[76185]

Hi guy,
I would be really happy, if you could help me to solve my problem !!

Here is it:

I study residual stresses induced in orthogonal cutting.
The simulation is "dynamic-explicit-temp-disp"
The workpiece is modelled with a Johson-Cook law, conductivity, heat capacity, inelastic heat fraction etc...
And so is the tool.
The chip is built thanks to a Johnson-Cook failure criterium with element deletion.

The contacts are defined with Coulomb-friction and "heat generation"

PROBLEM: During the process, upstream of the tool's nose, the temperature of some elements decreases! But I get residual stresses under the machined surface.

If the conductivity of the material is set to a very small value (0.001 for example), there isn't any problem but no residual stresses!


It seems really that the problem comes from the conductivity...It's not a problem of units...


It makes no sense ?!?!
If the temperature decreases in some elements, where does the heat go???

Thank you for any suggestions !

 

[Eser]

Hi!

I think you may check the energy dissipated from the contact, as an output variable. You can check first of all if there is really an energy dissapation during the process.

 

[vumat721]

Try turning off your inelastic heat fraction and see if you still have issues. I suspect it's a problem with your Johnson Cook coefficients, probably a bad specific heat. Alternatively, turn off heat generation on the contact pair.

In other words, you should try eliminating one of your heat sources to see if that helps you isolate the problem.

The reason that the elements are getting colder is probably because one area is drawing heat away from them erroneously in order to maintain energy balance. By reducing the conductivity, you are reducing the ability of the bad zone to pull heat away from those far away elements... But you still have a bug; the low conductivity is just masking it.

 

[76185]

Thank you for your replys Eser and Vumat721.

I already tried to turn alternatively off the heat generation and the inelastic heat fraction but it doesn't change anything to my problem.

And I still have the problem.

Any other suggestion?

 

[vumat721]

Do you have any other heat sources? For example, through surface loads or boundary conditions? Are you initial temperatures specified properly?

 

[76185]

I specified an initial temperature of 293K with a predefined field.
But there is the problem without too.

 

[vumat721]

Your initial temperature sounds right. Did you specify Absolute Zero and Boltzmann's contant in consistent units with the rest of your model?

For you:
Absolute Zero = 0 K
Stefan Boltzmann = 5.67 × 10–8 kg s^-3 K^-4

From Abaqus manual:

*Physical Constants

ABSOLUTE ZERO
Set this parameter equal to the absolute zero on the temperature scale chosen. For example, if the analysis uses temperature in degrees Celsius, set ABSOLUTE ZERO=–273.15.

STEFAN BOLTZMANN
Set this parameter equal to the Stefan Boltzmann constant. For example, STEFAN BOLTZMANN=5.669 × 10–8 Joule per sec m2 Kelvin4 in SI units.

 

[76185]

I set the absolute zero temperature, but not the boltzmann-konstant. I am going to try that.
Really thank you for your help Vumat721.

 

[76185]

It didn't change anything with the Boltzman-constant.
Here are some files which could help you to understand better the problem:

1: The temperature in an element on the surface of the machined surface:

http://files.engineering.com/getfil...d49-4538-9b9c-5c65585f3a41&file=Decrease2.png

2: The domain in which the temperature decrease (in dak-blue):

http://files.engineering.com/getfil...7db-4a0b-a2e6-beea41d235c2&file=Decrease1.png

Other suggestion ?

 

[vumat721]

This definitely looks like an issue with your material model. Are you able to send your *Material section from your input deck?