Gradient Output in Stress Analysis

In the analysis of large FE models, in nonlinear analysis, and when layered materials are used, the amount of gradient data that is generated and that is stored can become very large. This poses two problems: More memory is used and more disk space is used. In this case, the amount of gradient data can be reduced:

  • In the material definition, specify a failure criterion and set the option filter max_of_element. For each element, gradient data (of any kind) is stored only for the integration point with the maximum failure index.

  • In the material definition, specify a failure criterion and set the option fi_min v. Gradient data is stored only for integration points having a failure index greater than or equal to v.

  • In the case definition, use the gradients to control at which load increments gradient data is written to the database.

In the following example, the amount of gradient data written to the database is drastically reduced:

material 1 type orthotropic
  # T300/1034-C graphite/epoxy, units in [MPa]
  e1 146.86e+03
  e2 11.376e+03
  e3 11.376e+03
  nu12 0.3
  nu13 0.3
  nu23 0.3
  g1 6.186e+03
  g2 6.186e+03
  g3 6.186e+03

  failure hashin
    t1  1730.0
    t2    66.5
    c1  1379.0
    c2   268.2
    s12  133.763
    s23  101.4
    filter max_of_element # take only the integration point with maximum fi
    fi_min 0.5            # and only if fi >= 0.5
  end
end

case 1
  ebc            1
  analysis       nonlinear
  gradients      -1 # only at the end of the nonlinear analysis
end