2. Reference solution#

2.1. Calculation method#

The reference solution is obtained by integrating the equations of the model in Mathematica for axisymmetric modeling. To do this, it suffices to write the formal equations of the system, to apply to them the transformation rule characterizing Hooke’s law and to solve the nonlinear differential system. Users wishing to obtain more information can refer to the note HT-2C/97/016/A

Here are the results of integrating the behavior, depending on the case where the coefficient \({K}_{D}\) is chosen depending on the temperature and on \({\sigma }_{0}\) (modeling a and b):

_images/100002010000032C00000264BBCC59A999A46512.png _images/100002010000032C0000026467E2A8D3195EB6A6.png

SIGM for K_D (T, SIG)

r for K_D (T, SIG)

_images/100002010000032C00000264EA70E8DE8B783DB4.png _images/100002010000032C00000264C2994BE6692D2949.png

p for K_D (T, SIG)

D for K_D (T, SIG)

Here are the results of integrating the behavior, depending on the case where the coefficient \({K}_{D}\) is chosen depending only on the temperature (c and d modeling):

_images/100002010000032C00000264EAF1F9F38F9CDCAF.png _images/100002010000032C0000026467E2A8D3195EB6A6.png

SIGM for K_D (T)

r for K_D (T)

_images/100002010000032C00000264338DB73851BB0382.png _images/100002010000032C00000264F434C0AE1E85C502.png

p for K_D (T)

D for K_D (T)

In the graphs above, \(D\) is the damage variable corresponding to the internal variable \(\mathit{V9}\), \(r\) is the multiplicative viscoplastic work hardening variable corresponding to the internal variable \(\mathit{V8}\), and \(p\) is the cumulative viscoplastic deformation corresponding to the internal variable \(\mathit{V7}\).

We also have the following correspondence, in relation to the parameters of the VENDOCHAB keyword:

\(N\mathrm{=}{N}_{\mathit{VP}}\)

\(M\mathrm{=}{M}_{\mathit{VP}}\)

\(K\mathrm{=}{K}_{\mathit{VP}}\)

\(A\mathrm{=}{A}_{D}\)

\(R\mathrm{=}{R}_{D}\)

\(k\mathrm{=}{K}_{D}\)

2.2. Benchmark results#

The coefficient \({K}_{D}\) is chosen depending on the temperature and on \({\sigma }_{0}\). ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~

Evolution of the constraint, \({\sigma }_{0}\), as a function of time. This value is tested at various times:

Instant

Reference

20

252.76091

2000

164.261

200000

101.596

1000000

75.978499999999997

1600000

55.54209999999999998

Table 2.2.1-1 : Benchmark results for K_D (T, SIG)

Evolution of the damage variable, \(D\) as a function of time. This value is tested at various times depending on the modeling:

Instant

Reference

20

2.3168400000000001E-4

2000

2.77144E-3

200000

0.032255100000000002

1000000

0.110134

1600000

0.28131600000000001

Table 2.2.1-2 : Benchmark results for K_D (T, SIG)

Evolution of the viscoplastic isotropic work hardening variable, \(r\), as a function of time. This value is tested at various times:

Instant

Reference

20

1.6445100000000001E-3

2000

2.2312600000000001E-3

200000

2.6251500000000001E-3

1000000

2.747799999999999999E-3

1600000

2.79276E-3

Table 2.2.1-3 : Benchmark results for K_D (T, SIG)

Evolution of the viscoplastic isotropic work hardening variable, \(p\), as a function of time. This value is tested at various times:

Instant

Reference

20

1.644569999999999999E-3

2000

2.231879999999999999E-3

200000

2.6301200000000001E-3

1000000

2.7607899999999999E-3

1600000

2.8147799999999998E-3

Table 2.2.1-4 : Benchmark results for K_D (T, SIG)

2.2.1. The coefficient \({K}_{D}\) is chosen only depending on the temperature.#

Evolution of the constraint, \({\sigma }_{0}\), as a function of time. This value is tested at various times:

Instant

Reference

20

253.02000000000001

2000

164.36000000000001

200000

102.16

1000000

79.920000000000002

1600000

70.900000000000006

Table 2.2.2-1 : Benchmark results for K_D (T)

Evolution of the damage variable, \(D\) as a function of time. This value is tested at various times depending on the modeling:

Instant

Reference

20

2.32E-4

2000

2.7399E-3

200000

0.02756099999999999999

1000000

0.06626656500000000006

1600000

0.090278800000000006

Table 2.2.2-2 : Benchmark results for K_D (T)

Evolution of the viscoplastic isotropic work hardening variable, \(r\), as a function of time. This value is tested at various times:

Instant

Reference

20

1.645999999999999999E-3

2000

2.2339E-3

200000

2.6282800000000002E-3

1000000

2.7522900000000001E-3

1600000

2.7992999999999998E-3

Table 2.2.2-3 : Benchmark results for K_D (T)

Evolution of the viscoplastic isotropic work hardening variable, \(p\), as a function of time. This value is tested at various times:

Instant

Reference

20

1.6461E-3

2000

2.234499999999999999E-3

200000

2.63290000000000000001E-3

1000000

2.762699E-3

1600000

2.8137000000000001E-3

Table 2.2.2-4 : Benchmark results for K_D (T)

2.3. Uncertainty about the solution#

Accuracy of the codes

2.4. Bibliography#

  1. HT-2C/97/016/A, Dupas P., Description of the law of viscoplastic behavior coupled with the isotropic law of behavior of Chaboche introduced in Code_Aster, 1997.

  2. Aster R5.03.15 reference manual, Viscoplastic behavior with Chaboche damage.