Non Stationarity Option

This second page is reachable if Proportions and/or Trend option is checked.

Proportions

Check this option to provide a local/global proportion constraint. For a categorical variable, the proportions must be defined by categories. For a continuous variable, the proportions must be defined by classes of values.

  • Reference Variable: The proportion constraints can be defined only on one variable. You must define the reference variable. Some options depend of the type of the reference variable (categorical or continuous)
  • Acceptance Threshold: This parameter corresponds to the threshold for probability constraints. The lower the threshold is, the greater the proportion constraints are applied (and slower the simulations are). Both proportion acceptance threshold and pattern acceptance threshold have to be satisfied to simulate a node.
  • Local Support Radius: This parameter is available only for local proportion constraints. It is the ellipsoid size (distance) used to respect the local proportion constraint. The proportion of the neighbors inside the ellipsoid in the simulated grid is compared to the proportion given in the input local proportion map at the simulated node. You can use this parameter to manage the scale at which the local proportion has to be respected. Its default value is 12 times the maximum grid cell size (dX, dY or dZ).
  • Deactivation Distance: The proportion constraint is deactivated if the distance between the current simulated node and the last node in its neighbors is below the given deactivation distance. Use it to reduce noise in the results induced by the proportion constraint, but it will increase computation times. Its default value is 6 times the maximum simulated cell size (dX, dY or dZ).

  • Use Local: Check this option to activate Local Proportions.

    • Global Categorical Proportions: In the Proportion column of the table, you have to define proportion for each category (Lithotype). If the sum of the proportions is not equal to 1, the proportions will be normalized. The proportions must be greater than zero. The default value corresponds to the TI global proportion.
    • Local Categorical Proportions: If you want to use local proportions, you have to define a Proportions Macro Variable (indices must be consistent with the reference categorical variable). If the sum of the proportions is not equal to 1, the proportions will be normalized. This variable could be defined in an Auxiliary Grid. In this case, the variable will be migrated in the simulation grid using a bilinear interpolation method.

    • Global Continuous Proportions: In this case, you have to define several classes of values and the associated proportion.

      • Minimum: Included minimum class value. The current Minimum must be greater or equal to the previous Maximum and is expressed in the variable unit class.
      • Maximum: Excluded maximum class value. The current Maximum must be greater than the current Minimum and is expressed in the variable unit class.
      • Proportion: Proportions have to be defined for each category class. If the sum of the proportions is not equal to 1, the proportions will be normalized. The proportions must be greater than zero.

    • Local Continuous Proportions: As for global proportions, you have to define several classes of values and the associated proportion. Those variables could be defined on an Auxiliary Grid. In this case, the variables will be migrated in the simulation grid using a bilinear interpolation method.

      • Minimum: Included minimum class value. The current Minimum must be greater or equal to the previous Maximum and is expressed in the variable unit class.
      • Maximum: Excluded maximum class value. The current Maximum must be greater than the current Minimum and is expressed in the variable unit class.
      • Proportion: Proportions have to be defined for each category class. If the sum of the proportions is not equal to 1, the proportions will be normalized. The proportions must be greater than zero.

Trend

Choose this option to apply a trend. Trend is a continuous variable. It must be exhaustively informed on the Training Image and the Simulation Grid.

You can define as trend variables as you want.