rserrano
Computer
- Nov 23, 2008
- 7
Hi, I need to simulate the firing of a ceramic component.
The mechanical model of the component during firing is viscoelastic. When there is a constant stress, the strain of the material increases with time. The stresses will be induced by the thermal profiles calculated.
For my problem, the thermal model will not receive information from the structural model.
I have thought of several ways to couple them, unfortunately all of them are wrong:
1) I calculate the final thermal profiles, then do the structural part using it. If some part of the piece is heated too fast but at the end the temperature becomes uniform, no stresses will be reported at all.
2) Each substep of the viscoelastic part will receive the stresses from the last substep and the profile of temperatures correspondent. Obviously wrong, the stresses will increase arithmetically and the more substeps, the higher and worst the stresses.
3) Each substep of the viscoelastic part will receive the displacements from the last substep. I think that it will have the same behaviour as option 2). Besides, the strain rate does not behave linearly (in both senses of the word, the model is Norton's law).
Any Ideas? what is the right way to do this?
The mechanical model of the component during firing is viscoelastic. When there is a constant stress, the strain of the material increases with time. The stresses will be induced by the thermal profiles calculated.
For my problem, the thermal model will not receive information from the structural model.
I have thought of several ways to couple them, unfortunately all of them are wrong:
1) I calculate the final thermal profiles, then do the structural part using it. If some part of the piece is heated too fast but at the end the temperature becomes uniform, no stresses will be reported at all.
2) Each substep of the viscoelastic part will receive the stresses from the last substep and the profile of temperatures correspondent. Obviously wrong, the stresses will increase arithmetically and the more substeps, the higher and worst the stresses.
3) Each substep of the viscoelastic part will receive the displacements from the last substep. I think that it will have the same behaviour as option 2). Besides, the strain rate does not behave linearly (in both senses of the word, the model is Norton's law).
Any Ideas? what is the right way to do this?