AT1 model

[Batmanabcdefg]

Is it really possible for us to use "AT1" model in "cracks"_adaptive? if the degradation part is (1-phi).

I have changed the following lines:

for residual:

• G_c alpha_eps grad_pf * phi_i_grads_pf/1.333333
• G_c/alpha_eps (- 1.0) phi_i_pf/1.333333 for tangent:
  • G_c alpha_eps phi_i_grads_pf[i] phi_i_grads_pf[j]/1.333333
    // + G_c/alpha_eps phi_i_pf[i] * phi_i_pf[j]

The results (AT1) of Miehe's tension benchmark are wrong as all the values of phase field become zero after few iterations.

Thanks in advance!

[Batmanabcdefg]

(1-phi) is the geometrical part, not the degradation part, sorry :)

[tjhei]

We have not played with AT1 so far and I am not sure what is going wrong. I think this is worth exploring.

[Batmanabcdefg]

Dear Prof. Heister, thanks for your reply/support! Yesterday, I talked to Prof. Wick about AT1 model in cracks.cc in Deal.II, AT1 should work in cracks.cc, I will find where is the error in my modfied codes and let you know in this week. .;) Actually, I want to know about the numerical performances of AT1 Model, as AT1 Model has been widely used in 2018, even for the simulation of fatigue. Kind regards

[Batmanabcdefg]

Dear Prof. Heister and Dear Prof. Wick,

I have removed the bugs in the modified codes, and now I attach the modified part for AT1 phase-field model. Tests: Tension & Shear with G_c = 0.5 (AT1 Model)

// Tangent
//     + G_c/alpha_eps * phi_i_pf[i] * phi_i_pf[j]
  + G_c * alpha_eps * phi_i_grads_pf[i] * phi_i_grads_pf[j]***3/4**
// Residual
                         - G_c/alpha_eps*phi_i_pf***3/8**
 //                                        - G_c/alpha_eps * (1.0 - pf) * phi_i_pf                        
                         + G_c * alpha_eps * grad_pf * phi_i_grads_pf***3/4**

Kind regards