Configure Simulation Search Properties

The algorithms for Sequential Gaussian and Sequential Indicator simulation both use kriging at each node in the simulation grid to derive an estimate of grade. This requires variogram models to be created for each domain in the dataset.

Note: For Sequential Gaussian Simulation a set of normal scores continuity models for each domain are required for each domain. If normal score variograms do not exist then they will be inserted automatically when a conditional simulation component is inserted in the data tree.

Activity Steps

1. Display the graph.
2. Select the Vario/Search tab from the Property Panel.
3. Select the domain from the Domains list. The Variogram and Search Properties options only apply to the selected domain.

   Note: Domains that are not checked are not calculated in the conditional simulation.

4. Configure the following options as required.
   • Select Variogram – Click the ellipsis (...) to browse for the variogram.
   • Use Continuity Models – Set search range and angle using data from continuity models.
   • Dir – Manually set the search range for each direction. Only available if Use Continuity Models is unchecked in the Search Range section.
   • Angle – Manually set the search angles. Only available if Use Continuity Models is unchecked in the Search Angle section.
   • Original Data Samples field group:
     • Assign Data To Node – Whether to assign data to the node to speed up the simulation. Only available for Sequential Gaussian Simulation (SGS).

       Notes: Assigning data to nodes is designed to improve the speed of the simulation. However, there is a trade-off in the quality of the results. In particular, you should consider the difference in density of the defined node grid compared to the data spacing. In situations where the node grid is significantly denser or sparser than the data spacing, this process produces lower quality results.

       If checked, each sample is assigned to it's nearest node. A node may only have one sample assigned to it (the closest one) and any sample that is not assigned to a node is not used during the simulation. This feature significantly reduces the time needed to run the simulations. However, if the node grid spacing is wider than the sample spacing, then a significant amount of data may not be used during the simulation.

       When performing the search, the maximum number of previously simulated nodes property is consulted to select the samples. From this selection the number of assigned original samples is checked and if the number is less than the minimum number of original samples, the node is not simulated. The process also checks the octant requirements and does not generate a simulated value if those criteria are not met.

       It is recommended that you test the process with and without the Assign Data to Node property in order to understand the impact on the results for their data inputs.

     • Number Required – The minimum and maximum number of original data samples required for the node to be simulated. Points closest to the node are selected, up to the maximum number specified, or if * is entered then all data points within the search area are used. Only available if Assign Data To Node is unchecked.
     • Minimum octants informed – To enable octant searching, enter a value between 1 and 8 to define the minimum number of octants that must contain data.
     • Number required per octant – The minimum and maximum number of original data samples required for each informed octant.
   • Previously Simulated Nodes field group:

     • Note: After a node has been simulated, its value may then be used in calculating a value for another node in the simulation grid. After collecting all simulated nodes within the search ellipse, the nodes are ranked based on their distance from the node to be calculated. Each is then assigned a random number between 0 and 1 and if it falls below the probability of acceptance it is used. The process continues up to the maximum number required. By setting the acceptance probability to less than 1 you can increase the likelihood that the calculations use some nodes from slightly further away from the simulated node. A probability of acceptance of 1 is the equivalent of a 100% acceptance rate and only the closest nodes are used.

       Probability of acceptance – Set to a value between 0 and 1. See note above.
     • Maximum number to use – Maximum number of previous simulated nodes to use.
   • Domain Density field group:
     • Density – Select a density from a column in the dataset. Only available if a density data column component is present.
     • Override – Override the density data and enter a custom value. Option is automatically selected if there is not density column in the dataset.