This Tips & Tricks provides helpful suggestions on how to refine interpolation parameters and improve the accuracy of your 3D block model (3DBM).

The Pintrpq.dat procedure in MineSight Compass uses MineSight Torque drillthrough data to interpolate values in the block model. Several different interpolation methods can be performed with this procedure: kriging techniques, including ordinary kriging and indicatrix kriging, the gradient technique, among others. You can create multiple block models or multiple model items for any parameter in order to compare and verify the results of different interpolations.

Limitation of the interpolation area

The ability to limit the area of your interpolation implies that the pintrpq.dat procedure will run faster, making your operations more time efficient. If you have a particularly large model that needs a lot of memory, with multiple areas of interest, for example 2 ore bodies within a model, interpolation may be limited to a certain range of the PCF-based model, as shown in Figure 1.

How to handle clustered data

When modeling deposits, it is important to consider clustered data. One effect of clustered data on the interpolation process is that, in a high-content area of the model, the aggregated data is biased because it increases the true mean value.

One option in the interpolation procedure in MSCompass is the ability to split the search ellipsoid into “special selection” segments of octant, quadrant, split octant and split quadrant.

Each of the threads looks for composites individually and prevents the procedure from using too many composites in a given direction. It is possible to further limit the interpolation model, establishing that, for example, if there are a certain number of empty octants/quadrants, the block will not be interpolated.

how to treat the outliers

MineSight also has the ability to treat values considered as outliers from the composite data to ensure that any bias on the interpolation results is reduced.

A value outside the cutoff can be specified along with a maximum search distance from a block centroid. This can be applied for values higher or higher than that cut-off grade.

E.g.: if a value greater than this cutoff appears within the original search, the maximum distance considered will be the 3D search distance specified in the field below the grade, Figure 3.

Construction of variographic maps

In MSCompass it is also possible to build variographic maps based on imported data. The routine Contour Variogram Data allows from the directional variograms made by MSCompass to generate a variographic map, using the MineSight surface file 13 for plotting.

Block Limiting and Geologic Matching Options

The option Block Limiting allows limiting the blocks used in interpolation calculations only to those with specific codes Figure 5. This is useful when you have different interpolation parameters for different geological domains.

The option Code Matching or Geologic Matching limits the composites used in the interpolation to those that have exactly the same code as the model blocks, Figure 5. This method is commonly used for hard geological boundaries and prevents a composite from a different domain from influencing the block being calculated.

Validation of the search ellipsoid

Another useful option in the MineSight interpolation procedure is the ability to generate an ellipsoid to validate the estimates. This option only interpolates a block and generates a search ellipse geometry object, which can be imported into MS3D. This object also includes a table of the composites that the block used in its interpolation and polylines that connect the centroid of the block to the selected composites, Figure 6.

The ellipsoid will allow you to check if the search parameters are satisfactory and if there are any changes that need to be made before the actual interpolation of the entire model.

MineSight Compass's interpolation procedures provide many options for modelers to refine and add detail to the final model. MineSight is a powerful tool for anyone dealing with model tweens, allowing them to create more accurate models that are both auditable and detailed.

If you have questions about the topics in this bulletin, or any other task, please contact our MineSight/Prominas Technical Support (support@prominas.net). And if you still do not have this tool and are interested in obtaining more information, contact us via email. client-relations@prominas.net.