4.0 MicroMODEL METHODOLOGY OF USE

Subheadings for this chapter are as follows:

Generally, modeling of a deposit proceeds in the order of MicroMODEL's first six main modules. The following sections describe the methodology of use for each module. Although MicroMODEL is a flexible block modeling system, it is suggested that new users strictly adhere to the sequence and methodology described in this section, until they become more familiar with the entire system.

4.1 Data Entry Methodology

The first module, Data Entry, enables the user to define the location, size, and number of original labels (assay types) of the model. This information is referred to as the Project Information. Once the Project Information has been defined, the user can enter, graphically display, statistically analyze, and manipulate sampled drill hole information.

4.1.1 Project Information

The Project Information File must be created before proceeding with any other function. This file contains the basic information MicroMODEL will need to perform all tasks within the system.

The model's location and size should be specified such that the number of blocks are minimized, yet the blocks are small enough to provide accurate description of geological and structural geometries. Computation time is at least proportional to the size of the local grid, and geometric accuracy is increased as the blocks become smaller. Careful selection of the local grid size and block dimensions will maximize representativity and minimize program execution time. Generally, gentle trending, consistent deposits can be modeled with larger blocks, while highly variable deposits require smaller blocks.

Labels can be entered with the drill hole data, created by manipulation, or added via the Data Entry or Compositing modules. At the start of a project, the number of sample labels should be equal to the maximum number of different assay types to be entered in the sample drill hole data (e.g. gold, silver, etc.). MicroMODEL will expect a label value (including unsampled and 0.0) for each sample label in the Project Information for each drill hole interval.

4.1.2 Drill Hole Data

Drill holes are specified by their name, collar information, sampled interval data, and optional downhole survey information. The user either enter drill hole data as separate text files containing collar information, survey information, assay information, and (optional) geology information; or, the user may create a single standard MicroMODEL input file having the correct structure (see Volume II, Section 1.0).

Drill hole data entry allows the user to start a new database or append to an existing MicroMODEL database. Starting a new database deletes any previously existing drill holes in the MicroMODEL database and starts the database with the drill holes that are entered during that session. Appending an existing database enables the user to add drill holes without disturbing the previously existing MicroMODEL drill hole database. A warning is given prior to deleting any drill holes to prevent accidental deletion.

Each drill hole interval must have an whole number rock code corresponding to the geology which was logged with the drill hole. If a rock model is going to be comprised of only one rock type (a completely homogeneous model) then any whole number rock code between 1 and 9999 can be assigned to the intervals.

During compositing, if the rock code assignments are made from the rock model, MicroMODEL will assign all composites outside the model limits a rock code of 9999. Therefore, it is best not to use a rock code of 9999 so that composite intervals outside the model can be clearly identified.

The drill hole intervals also contain grade or assay data for each label. If the label value is 0.0 or unsampled (-999.99) it still must be specified. The label values can be used as input into manipulation functions which allow the user to define new labels or modify existing labels (e.g. changing measurement units, or creating additive products of labels).

4.1.3 Manipulation Uses

Manipulations can be performed on sampled and composited drill hole databases, and grade models. The manipulator programs can be used in various applications such as developing a net value for each drill hole interval or determining and flagging which blocks represent ore.

Money labels can often be created by multiplying unit value times grade. Next the money labels for each coproduct can be combined to determine a gross value for each drill hole interval or block. Finally, mining, milling, and other costs can be subtracted to obtain a net value. Any of the statistical and plotting programs can be used to aid the engineer in his evaluation of the manipulated data.

4.1.4 Digitizing

It is often more efficient to enter certain kinds of data by digitizing polygons. Two examples are Surface Data and Rock Code Data. Digitizing data is fairly simple, but if a few precautions are not taken, serious errors can result.

When prompted to place the map on the digitizer table, orient the global grid at a slight angle to the table edge. If the global grid happens to exactly coincide with the digitizer table grid, conversion and accuracy errors may result.

The instruction to locate a rectangle on the map which is parallel with the global grid is explicit. Adjacent corners must have a common coordinate.

After digitizing the four corners and supplying the coordinates of the lower left (Southwest) and upper right (Northeast) corners, Scale Factors will be calculated and displayed. Scale Factors with more than 1% error (i.e. factors >1.01 or <0.99) should be thoroughly questioned and resolved. It is possible that the reason is map stretch or inaccurate digitizing.

Once acceptable scale factors have been obtained, a couple of digitizing techniques become important. Digitizing in a clockwise direction is preferred. It is very important that the starting point of the polygon is clearly established (and also marked). This will help prevent the same area from being digitized twice in the same polygon. If an error is made, press the appropriate key to cancel the current polygon.

Certain shapes are not compatible with the digitizer routines. One example is a "figure 8". Crossing the polygon boundary will result in errors. It is necessary in such cases to digitize the area as two separate polygons.

In closing the polygon, it should be noted that the point associated with the keypad signal to close the polygon will not be recorded. It is strictly a signal device. Do not attempt to "close" the polygon by redigitizing the starting point. This has the potential for creating a "figure 8".

Upon the conclusion of a digitizing session, it is recommended to plot the digitized polygons to check for errors.

4.2 Surface Modeling Methodology

The topographic surface which separates rock from air must be modeled as a two-dimensional grid to provide for correct construction of the rock model and evaluation of pit volumes. This can be accomplished by using Inverse Distance to a Power (IDP) or kriging. A triangulated Irregular Network (TIN) model can also be constructed. The drawback of the TIN model is that you must be able to generate a set of triangles that completely covers the model area in order to avoid having unestimated cells. This means you must have at least some data points beyond the limits of your model.

Because the topography surface is used throughout the modeling process it is important that all cells within the surface model are estimated. If unestimated cells are found using a digit map of the topography, it is recommended that the surface be remodeled. If remodeling is impractical then the grid editor should be used to assign approximate elevations to the unestimated cell(s) in the topographic grid.

4.2.1 Surface Data Preparation

Regardless of the modeling method (IDP, kriging, or triangulation) used, surface data must be prepared prior to presorting and modeling. The Surface Data Preparation program enables the user to control the type of data that will be used as input for a particular presorting and/or modeling run. The input data can be comprised of digitized elevations, drill hole collar elevations, come from an X-Y-Z data file, or a combination.

4.2.2 Surface Data Variogram Analysis

MicroMODEL offers a submodule which enables the user to geostatistically analyze the surface data. The variogram is useful in determining search distances and anisotropies, and is required as input if kriging is to be used. The variogram program will process the prepared surface data only.

Since surface data tends to be geostatistically "robust", the user can generally start with a lag distance of:

Lag Distance = (Longest edge of the model) / 25

for the experimental variogram.

4.2.3 Surface Data Presort

IDP or kriging modeling method are normally used to model topo. Prior to estimating elevations of the topography grid cells, the prepared surface data must be presorted. The presort establishes a temporary data file which contains the input data to be used for each cell estimation. The presorted data is then used with any combination of modeling parameters.

A circular search is generally used for isotropic surface data during the presort. Data that falls within the user specified search radius can be used to estimate the value at the center of the circle (cell center) in accordance with other presort parameters. See Figure 4.1 for an illustration of a circular search for data points.

                        FIGURE 4.1
                CIRCULAR SEARCH (SECTOR)
             Maximum Search Radius = 250 Feet
              Maximum Number of Points = 12

For anisotropic data, the user can specify a search ellipse. Data that falls within the user specified search ellipse will be used to estimate the value at the center of the ellipse in accordance with other presort parameters. See Figure 4.2 for an illustration of an elliptical search for data.

                     FIGURE 4.2
      ANISOTROPIC (ELLIPTICAL) SEARCH (OCTANT)
         Maximum Search Radius = 250 Feet
      Maximum Number of Points per Octant = 3

The presort program has two methods of classifying data points that are within the search circle or ellipse. The first is closest point search, which selects a given number of points closest to the cell center regardless of their location relative to the cell center. Closest point search is shown in Figure 4.1. The second search method is sector search, which selects a given number of points closest to the cell center from each of the eight sectors (octants) within the search circle or ellipse. Sector search is illustrated in Figure 4.2. Generally, sector search gives better results as it tends to decluster data.

4.2.4 Surface Modeling

Once the appropriate surface data has been presorted, the grid is modeled according to user specified parameters. IDP, kriging, and triangulation modeling methods are all run from the same presort file. Modeling with the kriging option will require the input of a variogram model. Only cells that found data points during the presort process will be estimated. Unestimated cells should either be remodeled with a larger search radius or directly assigned elevations with the Grid Editing program.

For anisotropic data, a weighting ellipse can be used during modeling to weight different points according to their position in the weighting ellipse. Points that are located upon the same elliptical "shell" will receive equal weight in the modeling calculations, if weights depend only upon distances (IDP). In Figure 4.3 points A, B, and D will receive equal IDP weighting. Point C will receive less weighting than points A, B, and D since it is on a larger elliptical shell. Point E will receive the greatest weighting since it is located upon the smallest elliptical shell.

                     FIGURE 4.2
                TWO DIMENSIONAL ELLIPSE OF ANISOTROPY
                SEARCH RADIUS = 100 m
                PRIMARY AXIS LENGTH RATIO: 100      
                SECONDARY AXIS LENGTH RATIO: 50
                ROTATION ANGLE: 60 Degrees

4.3 Rock Modeling Methodology

The rock model is a three-dimensional representation of the deposit geology where each whole number rock code represents a particular geological, alteration, or mineralization zone. A very simple rock model may have rock code 1 representing mineralized, and rock code 2 representing unmineralized.

MicroMODEL offers three methods to build a rock model which can be used in any combination:

Prior to building a rock model with any of the three methods, the user is prompted for a background rock code which is assigned to all blocks that fall completely or partially below the topography. The rock model starts out with blocks that are either air (0) or background code, and is then updated by one of the three rock modeling methods.

During compositing, if the rock code assignments are made from the rock model, MicroMODEL will assign all drill hole intervals outside the model a rock code of 9999. Therefore, it is best to not use rock code 9999 during modeling so that composite intervals outside the model can be clearly identified.

4.3.1 Rock Modeling from Drill Hole Data Methodology

The easiest, but not necessarily the most accurate, method to create a rock model is using nearest neighbor assignments to the blocks from the drill hole rock codes. With this method, the program searches in three dimensions for the nearest sampled or composited drill hole interval and assigns the block center the nearest interval's rock code. Any blocks that do not have a nearest neighbor drill hole interval within a user specified search sphere or ellipsoid will remain at the background rock code.

If a structural trend is present in the deposit, an anisotropic ellipsoid can be applied to the nearest neighbor search. In this case, drill hole intervals are classified by their relative distance on the elliptical shell. In Figure 4.3, the rock code present at point "E" would be assigned to the block whose centroid falls on the center of the ellipsoid.

The ellipsoid is specified in the same manner as presort ellipsoids during grade modeling. Further discussion of ellipsoid definition can be found in Volume I, Section 3.8.

Rock modeling from drill hole data is generally used when little or no information is available on the deposit geology. Since the rock model codes are automatically assigned from the drill hole interval rock codes, the program develops a rough and often ragged rock model. This is especially true if the logging of the drill holes and assignment of rock codes to the various units has not been done consistently. Using plan view and cross-sectional polygons usually produces better results as the user can input his geologic interpretation into the model and generate more reasonable rock models.

4.3.2 Rock Modeling from Plan View Polygons Methodology

Rock modeling from plan view polygons is used when geologic data is available in plan. The user digitizes polygons in plan view and each polygon is assigned a rock code, a starting (lower) elevation, and a stopping (upper) elevation. The polygons form a volume in which all rock model blocks whose centroids are contained within the volume are assigned the corresponding rock code.

Assignments to the rock model are made in the same order as the plan view polygons are digitized. A rock model block that has been contained in multiple polygons will have the rock code of the most recent polygon which contained it. Figure 4.4 illustrates a typical polygon assignment to the rock model.

                            FIGURE 4.4
                ROCK POLYGON ASSIGNMENT IN PLAN VIEW

As long as the user does not restore (initialize) the rock model during plan view rock modeling, this method can be used to modify a rock model which was previously created using drill hole data, plan view polygons, or cross-sectional polygons.

4.3.3 Rock Modeling from Cross-Sectional Polygons Methodology

Rock modeling from cross-sectional polygons is used when geologic data is available in cross section. The user digitizes polygons in cross section, and each polygon is assigned a rock code and an influence distance for which the polygon applies on both sides of the section line. The polygons form a volume in which all rock model blocks whose centroids are contained within the volume are assigned the corresponding rock code.

Assignments to the rock model are made in the same order as the cross-sectional polygons are digitized. A rock model block that has been contained in multiple polygons will have the rock code of the most recent polygon which contained it. Figure 4.4 illustrates a typical polygon assignment to the rock model.

To simplify creation and verification of the rock model, the cross sections should be oriented along either the rows or the columns of the model. If possible, cross sections should be on consistent spacing such that a block cannot be influenced by more than one cross section. If the current set of cross sections are not oriented in this manner, plotting a set of sections square to the model and transferring the geologic zones to the new cross sections is recommended.

As long as the user does not restore (initialize) the rock model during plan view rock modeling, this method could be used to modify a rock model which was previously created using drill hole data, plan view polygons, or cross-sectional polygons.

4.4 Compositing Methodology

Composited drill holes use the same conventions and methodology as sampled drill holes. Where sampled drill holes can have multiple intervals in the same bench, bench composited drill holes usually only have one interval per bench. This generally reduces the size of the drill hole database and makes the data processing less cumbersome. MicroMODEL has three types of compositing:

Drill hole compositing prorates the sample drill hole interval assay values into composite intervals equal to the model's bench height down the hole. This method does not take into account the dip of the drill hole or the composite interval relative to the bench location. See Volume II, Compositing, Calculation of Composite Values and Figure 4.1 for details.

Mixed compositing is an attempt to bench composite all drill hole intervals. Drill holes that cannot be bench composited due to shallow drill hole dip are then composited downhole (drill hole compositing). Bench compositing prorates the sample drill hole interval assay values into composite intervals such that the "from" elevation of the composite corresponds to the top (crest) elevation of the bench and the "to" elevation of the composite corresponds to the bottom (toe) elevation of the bench.

Rock type compositing calculates composites using assay values only from within a consecutive string of intervals with identical rock codes. Composites are created in equal lengths that are as close as possible to a target length specified by the user. The user must also specify a minimum length for a composite. If a consecutive interval has a total length that is less than the minimum composite length, then no composite is created for that particular consecutive interval. Each composite is given a rock code that is the same as the code for the consecutive string of identical codes.

Rock type compositing should be used in situations where definite discontinuities exist at rock boundaries. A classic example would be a large quartz vein, where the vein carries grade, but the surrounding host rock is barren.

Composite rock codes for both bench and drill hole composites can be extracted from either the rock model or the most prominent sample rock code within the composite. Generally, the composite rock codes are extracted from the rock model since the rock model represents an interpretation of the deposit geology. This method also insures the spatial integrity between composite interval data and corresponding blocks in the grade model during grade modeling with specific rock types.

Some project drill holes may be in a composited format prior to using MicroMODEL. In this case, the drill holes can be entered directly in the Compositing Module.

Throughout the remainder of MicroMODEL, when drill hole data is required as input, the user will have the option to use sampled or composited drill hole data.

4.5 Grade Modeling Methodology

The purpose of the Grade Modeling Module is to estimate grade values into a three-dimensional grid of blocks. MicroMODEL offers three methods to create grade models:

MicroMODEL enables the user to store a grade model created by each of the three modeling methods for the same label simultaneously provided adequate disk space is available. This feature enables the user to run correlation analyses between different modeling methods for the same label. The user can also specify different grid types for the same label during pit evaluation to investigate the estimated reserves using different modeling methods.

In addition to the three modeling methods, MicroMODEL also automatically creates an Estimation Precision (Error of Estimation) grid during kriging of the current label. If requested, MicroMODEL will produce an Estimation Precision grid with IDP modeling runs, provided the user can define a variogram model. The error of estimation grid contains, for each block, the variance of the estimation error associated with the configuration of the data around the block. The square root of the variance is the kriging error standard deviation expressed in the same units as the grade. This can eventually be used to build confidence intervals on the estimated block grade.

The Estimation Precision grids can be displayed and analyzed in the same manner as any other grade model within MicroMODEL, but it is recommended to display standard deviation grids. Note that this standard deviation grid must be calculated using the grid manipulation program.

During any of the grade modeling programs, MicroMODEL will process the current label only. To model a different label, the user is required to change the current label in the File Management Module. This prevents the user from accidentally using an incorrect type of data to create a grade model.

It is not necessary to model each label present in the system. Only the labels (grade types) that are of interest need to be modeled.

4.5.1 Polygonal Grade Modeling Methodology

Polygonal grade modeling is the simplest method for creating a grade model within MicroMODEL. The drill hole database must be bench composited prior to using the Polygonal Grade Modeling program. This program makes nearest neighbor block assignments on a bench by bench basis, based upon the corresponding bench composites. Any blocks that do not have a bench composite within a user specified search radius remain unestimated (-999.99).

It should be noted that the Polygonal Grade Modeling program offered within MicroMODEL is not a "true" polygonal analysis. Rather than calculating polygonal areas for each drill hole, the program assigns grade values to the block model based upon the nearest neighbor composite interval from the block centroid. This difference is illustrated in Figure 4.5. Generally, the "true" polygon and Polygonal Grade Modeling methods produce similar, although not identical, results.

                       FIGURE 4.5
     TRUE POLYGON RESERVES VS. POLYGONAL GRADE MODELING

For convenience, MicroMODEL offers a program which generates polygon maps for user specified benches. This program displays the actual area of influence of the polygon and the corresponding grade. These maps could then be used for "true" polygonal grade calculations.

4.5.2 Variogram Analysis Methodology

MicroMODEL offers a submodule which enables the user to geostatistically analyze the drill hole assay data. The variogram is useful in determining search distances and anisotropies, and is required as input if kriging modeling is to be used. As with all modeling processes, the variogram program will process drill hole data from the current label only.

Variogram calculation and interpretation requires some knowledge of basic geostatistics. The model of variograms used in the grade modeling routines are theoretical analytical curves fitted to the experimental variogram calculated.

4.5.3 Point Validation Methodology

The Point Validation Presort and Point Validation Modeling programs use the identical parameters required by the Grade Modeling Presort and Grade Modeling program. This enables the user to determine the best set of parameters for a given data set. This is done by cross validating known data values vs. the estimated values of the same points calculated without using the known value. Good presorting and modeling parameters are indicated by a small average difference between the known and the estimated values, a minimal error variance, and a positive linear correlation relationship. Once the "best" set of parameters are discovered, the user can be confident that grid modeling of the same data with the same parameters will produce the "best" results.

In MicroMODEL, point validation is a two step process. The first step is the presort which determines which sampled or composited drill hole intervals are valid to estimate a particular location. The second step, point validation modeling, actually estimates the grade at a particular location based upon the presorted data.

Dividing point validation into two distinct tasks is advantageous because once an adequate presort has been performed, the user can make multiple modeling runs based on different weighting ellipsoids, minimum points, modeling methods, and variogram models without repetitive runs of the presort program.

The results of the most recent point validation run can be displayed as a scatter diagram with the Correlation Analysis program. This diagram plots "known" drill hole values vs. the corresponding "estimated" value. A 1:1 linear relationship with a high correlation coefficient indicates good modeling parameters.

4.5.4 Point Validation Presort Methodology

Prior to point validation modeling, which estimates grade values at the drill hole interval locations, the sampled or composited drill hole assay data must be presorted. The presort establishes a temporary data file which contains the input data to be used for each point validation. The presorted data is then point validated with any combination of modeling parameters.

A spherical search is generally used for isotropic assay data during the presort. Data that falls within the user specified search radius can be used to estimate the value at the center of the sphere in accordance with other presort parameters. See Figure 4.1 for a two-dimensional illustration of a spherical search for data points.

For anisotropic data, the user can specify an ellipsoid of search. Data that falls within the user specified search ellipsoid will be used to estimate the value at the center of the ellipsoid (point validated point) in accordance with other presort parameters. See Figure 4.6 for a two-dimensional illustration of an elliptical search for data. Specification of ellipsoids is discussed in more detail in Volume I, Section 3.8.

                       FIGURE 4.6
             ELLIPSOIDAL PRESORT ELLIPSE

                CLOSEST POINT SEARCH
                SEARCH RADIUS = 200 FEET
                PRIMARY AXIS LENGTH = 200 
                SECONDARY AXIS LENGTH = 100
                ROTATION ANGLE = 120 DEGREES

The presort program has two methods of classifying data points that are within the search sphere or ellipsoid. The first is closest point search; which simply gathers the closest group of points to the center, regardless of their orientations relative to the cell center. Closest point search is shown two-dimensionally in Figure 4.1. The second search method is sector search which gathers a given number of data points from 6 sectors within the search sphere or ellipsoid. Each sector can be visualized as a square-based pyramid whose apex lies at the point being presorted.

In addition to the above parameters, the Point Validation Presort program has the facility to exclude data points which are too close to the point being estimated. The purpose of this feature is to discount drill hole intervals lying immediately above and below the estimated point. This is done by specifying a minimum search radius. Any point lying within this minimum search radius will not be included in the point validation estimation. The minimum search radius is used in the same manner as the maximum search radius, applying the same search ellipsoid. The Point Validation Presort program determines which cross validated intervals will be estimated according to the corresponding rock codes. This feature enables the user to control the point validation modeling with rock codes from the drill hole data, preventing incorrect estimation of grade from intervals with undesired drill hole data. For example, a deposit may have 5 rock codes:

If all of the rock codes were modeled together (which MicroMODEL will allow), ore grades might be projected into intervals of waste, and the ore intervals would be diluted by surrounding waste drill hole intervals. To prevent this, the point validation runs could be made separately:

  1. Estimate #1 with #1
  2. Estimate #2 with #2
  3. Estimate #3 with #3
  4. Estimate #4 with #4
  5. Estimate #5 with #5

Perhaps statistically, the sulfide ore, #4, and the mineralized pipe, #5, can be modeled together. In this case, a point validation run for rock types 4 and 5 combined could be made by jointly estimating #4 and #5 with #4 and #5.

4.5.5 Point Validation Modeling Methodology

Once the appropriate drill hole assay data has been presorted, the data is point validated according to user specified modeling parameters. Both IDP and kriging point validation modeling methods can be run from the same presort file. Modeling with the kriging option will require the input of a variogram model. Only drill hole intervals that found data points during the presort process will be estimated.

For anisotropic data, a weighting ellipsoid can be used during point validation modeling to weight different points according to their relative position in the weighting ellipsoid. Points that are located upon the same elliptical "shell" will receive equal weight in the modeling calculations, if weights depend only upon distances (IDP). In Figure 4.3 points A, B, and D will receive equal IDP weighting. Point C will receive less weighting than points A, B, and D since it is on a larger elliptical shell. Point E will receive the greatest weighting since it is located upon the smallest elliptical shell.

The user may desire that a minimum number of data points be within the presort parameters before a point validation estimation is made. This is done by specifying a minimum number of points during point validation modeling.

Error of estimation values will automatically be calculated for point validation runs made with kriging. Error of estimation values can be calculated during IDP point validation, if the user supplies a variogram model.

4.5.6 Grade Modeling Preview

In the same manner as Point Validation Modeling, MicroMODEL's grade modeling process involves two tasks. The first task is the presort, which determines which sampled or composited drill hole intervals are valid to estimate a particular block location. The second step, block modeling, actually estimates the grade at a particular block location based upon the presorted data.

4.5.7 Grade Modeling Presort Methodology

Prior to grade modeling, the sampled or composited drill hole assay data must be presorted. The presort establishes a temporary data file which contains the input data to be used for each block estimation. The presorted data is then used to estimated within a given set of grade modeling parameters.

A spherical search is generally used for isotropic assay data during the presort. Data that falls within the user specified search radius can be used to estimate the value at the center of the sphere (block centroid) in accordance with other presort parameters. See Figure 4.1 for a two-dimensional illustration of a spherical search for data points.

For anisotropic data, the user can specify an ellipsoid of search. Data that falls within the user specified search ellipsoid will be used to estimate the value at the center of the ellipsoid (block centroid) in accordance with other presort parameters. See Figure 4.6 for a two-dimensional illustration of an elliptical search for data. Specification of ellipsoids is discussed in more detail in Volume I, Section 3.8.

The presort program has two methods of classifying data points that are within the search sphere or ellipsoid. The first is the closest point search, which selects the closest group of points to the block center, regardless of their orientations relative to the block center. Closest point search is shown two-dimensionally in Figure 4.1. The second search method is sector search which gathers a given number of data points from 6 sectors within the search sphere or ellipsoid. Each sector can be visualized as a square-based pyramid whose apex lies at the point being presorted.

The Grade Modeling Presort program determines which blocks in the grade model will be estimated, according to the corresponding rock codes in the rock model. This feature enables the user to control the modeling with the previously built rock model, preventing incorrect estimation of grade with undesired drill hole data. For example, a deposit may have 5 rock codes:

If all of the rock codes were modeled together (which MicroMODEL will allow), ore grades would be projected into blocks of waste, and the ore blocks would be diluted by surrounding waste drill hole intervals. To prevent this, the grade model could be built in 5 passes:

  1. Estimate #1 with #1
  2. Estimate #2 with #2
  3. Estimate #3 with #3
  4. Estimate #4 with #4
  5. Estimate #5 with #5

Perhaps statistically, the sulfide ore, #4, and the mineralized pipe, #5, can be modeled together. In this case, 4 grade modeling presort and corresponding modeling passes would be necessary:

  1. Estimate #1 with #1
  2. Estimate #2 with #2
  3. Estimate #3 with #3
  4. Estimate #4 and #5 with #4 and #5

4.5.8 Grade Modeling Methodology

Once the appropriate drill hole assay data has been presorted, the data is used to estimate grade values at the block centers according to user specified modeling parameters. Both IDP and kriging grade modeling methods can be run from the same presort file. Modeling with the kriging option requires the input of a variogram model. Only blocks that found data points during the presort process can be estimated. For anisotropic data, a weighting ellipsoid can be used during grade modeling to weight different data points according to their relative position in the weighting ellipsoid. Points that are located upon the same elliptical "shell" will receive equal weight in the modeling calculations, if weights depend only upon distances (IDP). In Figure 4.3 points A, B, and D will receive equal IDP weighting. Point C will receive less weighting than points A, B, and D since it is on a larger elliptical shell. Point E will receive the greatest weighting since it is located upon the smallest elliptical shell.

The user may desire a minimum number of data points be within the presort parameters before a block estimation is made. This is done by specifying a minimum number of points during grade modeling.

An error of estimation block model will automatically be calculated for grade modeling runs made with kriging. An error of estimation grid can also be created during IDP grade modeling, if the user supplies a variogram model.

4.5.9 Grade Model Manipulation

The user may need to process an input grid(s) to define another grid. This can be accomplished with the grid manipulator. For example, an ounces per ton grid can be converted to a milliounces per ton grid.

4.6 Pit Generation and Reserves Evaluation Methodology

The Pit Generation Module enables the user to quickly evaluate reserves within a particular volume. Once the user has built satisfactory topography, rock, and grade models, the programs in this module expedite the pit design process by performing the repetitive computations associated with manual design methods.

4.6.1 Open Pit Design (OPD)

Briefly, to use MicroMODEL's reserve evaluation system, the user digitizes a pit base(s) which is either evaluated directly, or expanded to the surface at user defined slope angles. The defined volume(s) is then evaluated by bench, rock type, and cutoff grade, and several types of reserve reports are generated.

OPD is a very flexible system, enabling the user to perform several tasks with a minimal number of programs. OPD does not alter the original (as modeled) surface and grade models. The rock model is only modified by OPD when specifically requested by the user and then in such a manner that the original rock model can be easily recovered.

To account for mined and unmined blocks within the pit model, OPD uses the rock model and the following rock codes:

With this system, previously mined blocks that are "re-mined" during subsequent intervals are ignored by OPD. This insures that material is not tabulated more than once for a particular design.

Another advantage to this system is the rock model is never altered from its original state. The sign of the rock code within a particular block indicates to the system whether the block is mined or unmined.

By running the Rock Model Initialization program, which takes the absolute value of all blocks within the rock model, the user can return to the original, unmined rock model at any time. This system also prevents MicroMODEL from using a "Mined or Unmined" matrix, thus saving disk space.

Each time a pit increment is evaluated, OPD makes an exact copy of the current rock model (permanent) to a temporary rock model. OPD then actually mines-out the temporary copy of the rock model, changing the block rock codes to negative where applicable. After the evaluation is complete, the programs generate reserve statements. At this point, two rock models actually exist within the database, the first is the pre-increment rock model (permanent), the second is the post-increment rock model (temporary) which has been mined using the increment.

If the user is satisfied with the results of the increment, and wants all future increments to not count the material mined-out by the current increment, then the rock model is "Updated" which replaces the pre-increment (permanent) rock model with the post-increment (temporary) rock model.

If the user is unsatisfied with the results of the increment, and does not want the increment removed from the rock model, then the post-increment (temporary) rock model is NOT updated. Since the post-increment (temporary) rock model has not been updated, analysis of the next increment will begin with the pre-increment (permanent) rock model being used as an image for the next temporary rock model.

4.6.2 Development of an Ultimate Pit

MicroMODEL's Open Pit Design (OPD) can be used in a true incremental pit design process. With OPD, an ultimate pit can be developed incrementally such that each marginal zone of ore can be evaluated against its own marginal waste stripping. A suggested procedure for this method is presented in flow chart form (Figure 4.8) and discussed here.

                      FIGURE 4.7
               OPD PIT DESIGN ALGORITHM
  1. Starting with verified surface, rock, and grade models, the user initializes the rock model which serves two purposes. First each block in the rock model is made positive so it is available to be mined. Second, each rock type within the rock model is inventoried for subsequent use by OPD.
  2. The pit generation parameters such as cutoff grades, type of grade models, rock type densities and slope angles are entered. Base expansion will almost always be used.
  3. One pit base (increment) is entered (digitized). For true incremental analysis, pit bases which isolate an ore zone on a particular bench, must be evaluated independently, from the top of the deposit downward. Plotting the grade model block values on bench maps is generally the easiest method of identifying pit bases.
  4. Evaluate the pit, using the previously digitized pit base (increment).
  5. Decide if the previously digitized pit base is acceptable. If the increment is acceptable, go to step F. If the increment is unacceptable, go to step C.
  6. Update the rock model with the previously digitized increment. This flags the mined-out blocks with negative rock codes so that future increments will not accumulate the material contained by this increment.
  7. If another increment is required, go to step C.

At the completion of each Pit Evaluation run, the user can plot the expanded pit. This program plots the polygon(s) that were generated for each pit base during the expansion. The program does not merge multiple base polygons together, nor does it show previously mined areas.

MicroMODEL can produce a surface contour map of the deposit which interfaces the modeled topography with the current permanent rock model. Since this program produces a plot relative to the mined-out rock model, the contours inside the pit are generally ragged, however, this is a true representation of the pit design relative to the mined-out rock model. This plot is usually a good starting point for the addition of haul ramps and manual smoothing of the pit design.

4.6.3 Use of an Inverted Pit

During any stage of the modeling process, the permanent rock model can be inverted. This process reverses the sign of all rock codes within the model. Negative rock blocks become available for mining, and positive rock blocks are made unavailable. This process serves two purposes, summing multiple increments and phasing within an ultimate pit design.

Generally, during incremental development of an ultimate pit, the user will have several reserve statements from multiple increments. To sum the increments together for a final reserve statement would be quite time consuming. By inverting the rock model, the increments can be easily aggregated. Inverting the rock model reverses all rock code signs within the permanent rock model, which makes all material contained within the current pit volume eligible for mining. By digitizing a simple base which will encompass the pit volume and expanding the base upward, only the positive rock code blocks contained within the current pit volume will be accumulated by OPD. The resulting reserve statement will be a sum of the current mining within the rock model.

The inverted pit should not be updated since updating would bring the rock model to a state of all negative rock codes. To mine more increments, the rock model should be re-inverted which will return it to the negative rock codes inside the pit configuration.

The second use of an inverted rock model is during phasing or scheduling within a previously developed ultimate pit. By working within an inverted rock model, the user cannot mine material outside the ultimate pit volume since the material is unavailable. This procedure is analogous to developing an ultimate pit where satisfactory increments (phases) are removed from the permanent rock model by updating the rock model with the previous increment.

4.6.4 Floating Cone Design

The Money Matrix and Floating Cone Option in the OPD Module allows the user to automatically design open pit shells. Although floating cone algorithms are not truly optimal, they can provide results that closely mimic true 3-D optimizers under most conditions. In general, deposits consisting of many "pods" of mineralization may require a true 3-D optimizers. An interface to the Whittle 3-D optimizer software is available, if the user feels this is necessary.

In order to perform floating cone optimizations, the user must first create a 3-D file of net block values. This file, called a money matrix, is created with the Create Money Matrix Program.

After running a floating cone analysis, the user can display the "cone pits" as contour maps, plan view cell plots, or as 3-D perspectives. Cone shapes can also be displayed as topographic profiles in section. Reserve calculations can be performed on material contained between any two cone shells.

4.7 File Management Methodology

The File Management Module enables the user to make global changes to the MicroMODEL database, which are not necessarily associated with a particular module. It also allows the user to create a combined plot consisting of one or more individual plot files plus optional annotation. In addition, drill hole class limits can be invoked.

4.7.1 Current Label Methodology

The statistical and modeling programs access the current label (sample, composite, or grade). This program enables the user to change the pointers to the current labels. All printed output in MicroMODEL lists the current labels. This prevents misinterpretation of statistical results.

4.7.2 Grid Editor Methodology

The grid editing program allows the user to display and make edits to the surface, rock, and grade models on a block by block basis. These edits may be necessary because the user wants to replace unestimated grid values caused by insufficient modeling data or adjust known anomalous grid values. The grid editing program is not intended to replace carefully selected modeling parameters and sound modeling practices.

4.7.3 ASCII Grid Manipulation Methodology

The output and input of ASCII External Grid programs enable the user to interact with other software packages. Use of these programs requires a thorough understanding of both MicroMODEL's and the external package's grid structure.

4.7.4 Create Combined Plot Methodology

Output from separate plotting programs in MicroMODEL can be combined into a single plot. Annotation can be added through the use of simple annotation files, which are created with any standard text editor. For example, drillhole collar locations can be displayed along with topographic contours on a single plot.

4.7.5 Drillhole Class Methodology

MicroMODEL offers the ability to limit which drillholes are used in most statistical, display, or modeling programs. The user may enter the names of Drillhole Classes (subsets of the drillhole data), indicate which class each drillhole in the data set belongs to, and then instruct MicroMODEL to include any or all of these classes in calculations.

For example, the user might want to separate the data set into categories of diamond core, reverse circulation, and trenches. Or, the user may want to categorize the drilling by which company drilled which holes.

4.8 Grade Thickness Methodology

The grade thickness module can be used to calculate and display grade-thickness values for a given sample interval label above a specified cutoff. The calculation can be made based on anywhere from one to twenty rock types. This module is a good tool for investigating general mineralization trends within a deposit.

The user calculates a set of grade-thickness data points, which can be displayed in plan view. The user may display drillhole name, northing, easting, grade, thickness, or grade-thickness at each calculated data location.

Additionally, the user may treat the grade-thickness points like a set of 2-D data, The data can be used to model a 2-D grid, similar to a topography grid. This grid can be displayed in a plan view cell plot, or contoured.

4.9 Special Tools Methodology

This module contains an assortment of special tools that enable the user to move grid files from one MicroMODEL setup to another. In addition, there are programs for adjusting the rock model for a new topographic surface, calculating reserves based on a special cutoff matrix, programs to help enter rock interval information, and a program to calculate waste dump volumes.

4.10 PolyMap Interface Methodology

This module contains a set of programs that allow the user to interface MicroMODEL with the PolyMap graphical editing and display program. The user can create a slope template to be used for open pit design. A display file can be created that shows color coded blocks on each level, depicting ore and waste, or showing differences in rock type. Cone surface contours can be exported directly to PolyMap so that they may be used as a background drawing for pit designs. The user can create appropriate files so that PolyMap can be used to design waste dumps. Finally, the user can define a set of grade files, cutoffs, and ore class information that can be used to interactively calculate reserves for a given PolyMap pit design.

4.11 3D Display Methodology

This module contains a set of programs that allow the user to generate three dimensional displays of MicroMODEL data files.