Webinar: Block Model 101: How to create different types of Block Models in K-MINE

Welcome to K-MINE's webinar on geological modeling and block model creation. This session is led by Tatiana Nistenko, K-MINE's Chief Geologist, who walks through the complete workflow of creating different types of block models in K-MINE - from database setup to resource estimation.

K-MINE currently has 12 modules for open-pit and underground mining, all of which can function as standalone tools or together as an integrated platform. The functionality covered in this webinar is part of the Geology module.

When creating a block model as part of exploration data analysis, the approach always depends on the mineral type and genetic type of the deposit. There are different interpolation and extrapolation methods, different ways of contouring (cut-off grade, sample block, selective mining unit), and different methods of mineral resource and reserve estimation, such as the polygonal method or block estimation.

Classification frameworks also vary - from the United Nations Framework Classification to CRIRSCO-aligned systems and national classifications of different countries. Performing geological modeling efficiently and correctly requires significant experience.

If the software you are using has a superficial understanding of these geological processes, you will end up with inaccurate and unrealistic geological interpretation. Geological models are based on the interpretation of core sampling data, which represents a very small portion of the ore body. These models determine the next steps in project development, which is why they are always evolving - their interpretation is never set in stone.

Incorrect geological interpretation leads to potential estimation errors that can be far more significant than those associated with grade estimation alone.

The digital database is the foundation for the entire geological, mathematical, and block modeling process. It also serves as the basis for technical and economic calculations, reserve and resource estimation, and follow-up procedures that support grade control systems at mining operations.

K-MINE uses a relational database model implemented through solutions such as Microsoft Access, MS SQL, and others. Many clients store their geological databases on a PostgreSQL server.

The digital geological database consists of several core tables:

The collar table stores drill hole coordinates. Required fields include drill hole number, longitude, latitude, and elevation (Z coordinate). Users can choose X or Y for longitude depending on the coordinate system. K-MINE allows selection of either Cartesian or geodetic coordinate systems before creating the project. Additional fields such as drill hole depth, drilling date, drill hole type, special permit name, and profile number can also be added.

The survey (inclinometry) table stores drill hole deflection data used to calculate drill hole path coordinates and sample interval positions. Required fields are drill hole ID, depth (distance from collar to measurement point), slope angle, and azimuth.

All geological information is stored in additional tables like assay, lithology, spectral analysis, composite, and others. The assay table typically includes component grades by sampling interval, along with additional fields such as level, ore body number, drill hole size, core recovery, and drift.

Rock characteristics are stored as codes for geological intervals in the lithology field. Rock codes are maintained in a separate reference table. When drill holes from surface or underground workings are composite, assay lines are created as individual records.

K-MINE creates independent tables when geological characteristics have interval lengths different from assay interval lengths. The drill hole number serves as the primary key, with depth-from and depth-to as the corresponding fields. Theoretically, there can be as many additional tables as needed - tectonics, seismology, geophysics - whenever additional geological information needs to be recorded.

Users can input data directly or upload it from Excel files. Each table must contain a header row with field names to map the import correctly. Data can also be imported from point objects into the database, including individual surface samples (chip samples, tailings sampling) or mine working entry points into the collar table.

Once the import process is complete, source data needs to be analyzed and verified to eliminate errors. Visualization can help identify errors in primary geological materials - for example, displaying exploration workings and comparing them with the corresponding geological map or section.

Users can set the coordinate grid and scale, rotate images if needed, and vectorize objects from downloaded images. Map sections from Google Maps can be downloaded, and K-MINE objects can be saved to KML format for viewing in Google Earth Pro.

Common data entry errors include missing assay results, gaps between sample intervals (non-adjacent or overlapping intervals), and total drill hole depth being less than the deepest assay interval.

Statistical analysis helps determine the type of mineralization distribution (normal or log-normal) and data set heterogeneity, which may result from several mineralization generations or variable component content. Polygenic (heterogeneous) data sets should be processed and analyzed individually.

When samples are analyzed for multiple components or by different analytical methods, correlation analysis with regression equations is essential. The coefficient of variation of grades or thickness can also be determined.

K-MINE can automatically merge sampling and lithology tables into a cross-assay table when from-to lithology intervals do not match sampling intervals. This step is critical for cases where calculations need to be performed directly in the database tables - for example, identifying grade intervals by mineral quality requirements or specifying interval assignment to a particular mineral grade based on technological properties.

Triggers allow users to create combined conditions and recalculate parameters by intervals. For complex ores, grades can be recalculated using conversion factors and brought to the conditional grade of the main component.

Compositing is the process of combining samples to ensure comparability of influence from samples with different lengths on statistical estimates within a domain. In K-MINE, compositing is performed automatically. The composite table is stored separately from the assay table, allowing updates and recalculations when a new composite length is selected, or creation of multiple tables for different conditions at any time.

K-MINE offers four methods for block model creation:

First, you can create an empty cubic block model without property values, or import one from files in CSV or BLK format.

Second, you can create a seam block model based on lithological description of mine workings in the database. In this model, blocks have a defined size along the X and Y axes, while the Z-axis size varies depending on the recalculation of rock interval thickness in the database.

Third, you can create a cubic block model for seam deposits using lithological variety codes from the database. All blocks have a defined size along three axes (for example, 10 x 10 x 1 meter).

Fourth, you can use implicit modeling with the radial basis function (RBF) method. This algorithm uses rock codes or other parameters for RBF interpolation, allowing users to create triangulated surfaces of lithological boundaries.

For explicit modeling, an empty cubic block model is created in the form of a parallelepiped, and triangulation surfaces are used to determine block attributes and define geological contact boundaries.

To create a new block model, specify the file name (ODFB format), input block model dimensions and rotation angles. In the block sizes field, set the minimum block dimensions along each axis (for example, 5 x 5 x 5 meters).

In the recently updated K-MINE version, users can change dimensions and axis rotation angles when creating a block model by selecting the rotation axis or vertex.

K-MINE uses the octree subdivision method for sub-blocking, where parent blocks are divided into regular fractions (1/2, 1/4, 1/8, etc.). The block model is constructed from the largest to the smallest block, starting from the origin point. For example, the maximum block size might be 320 meters, then 160, 80, 40, 20, 10, and finally 5 meters.

Users can sub-block within a solid or across the entire model to any required block size. A block model can contain completely different sub-block sizes in different zones. To reduce the block size below the specified minimum, create a new block model based on the existing one.

The system creates a default set of block model properties: block center coordinates, block sizes along each axis, color, and unit weight. Additional properties (property name, field type, and value) can be added as needed.

Block models can be imported from BLK or CSV files. When importing, users preview the data, set constraints for integer and fractional parts, and configure correspondence between properties of the imported and new block models.

When creating a new block model based on an existing one - for example, to change the minimum block size or add properties - property values can be recalculated using weighted average by volume, weighted average by weight, or maximum weight methods.

Block models are stored as separate ODFB files, with graphical representation in the layer editor of the current ODF file. Multiple block models can be stored in a single ODF file, each with its own ODFB file.

Block filtering uses conditions such as equal, less than, greater than, between, and contains to visualize specific subsets of blocks.

For seam deposits, a seam block model is created within a tectonic block solid. The process involves registering the chronological sequence of seams, indexing individual alternations when splitting seams, and adding alternation intervals of enclosing rocks to the reference table. Each interlayer must have its own unique code.

A wireframe of the overburden surface, oxidation zones, and other limiting surfaces is constructed. For deposits with complex tectonic structures, limiting boundaries must be created for each tectonic block.

When creating a seam block model, users specify block sizes along the X and Y axes and set a minimum block size for the Z axis. The Z-axis block size varies depending on seam intersection thickness. Symmetry sections can be specified for each calculation to ensure smooth construction.

If the deposit has tectonic dislocations, the marker surface should be created separately for each tectonic block, with interpolation performed within each tectonic block solid.

A cubic seam block model follows the same principles but with fixed block sizes along all three axes. When blocks contain multiple lithological varieties thinner than the block height, volume percentages are calculated for each variety.

Implicit modeling is the automated creation of surfaces directly from the database using radial basis functions (RBF). This method creates triangulated surfaces of geological boundaries without manual digitization.

Implicit modeling allows rapid generation of ore body models from the same data set. It also supports secondary data sources beyond drill hole intersections, including geophysics, spectral analysis, and mine face scans. Complex geological surfaces that previously required slow manual digitization can now be created semi-automatically.

To create a solid using implicit modeling, select a rock code from the reference table, specify the lithological field and construction step, and set parameters such as smoothing radius, number of iterations, and the degree of surface variability in peripheral zones.

The main advantage of implicit modeling is speed - you can perform hundreds of construction variations by changing parameters. For example, adjusting the smoothing radius changes the ore body shape. A smaller radius may divide the body into separate smaller deposits, while a larger radius produces a continuous body of greater thickness.

However, implicit modeling is not a universal solution. At early exploration stages during pre-feasibility studies, when the drill hole grid is widely spaced, explicit modeling (geological interpretation and contouring of ore bodies in cross sections) may be more appropriate. Both implicit and explicit methods can produce useful models, and both can be misused in inexperienced hands.

For explicit modeling, ore body boundaries are outlined in plan or cross-section views. Solids can be built automatically based on geological contact boundaries from cross-sections and plans, with the system accounting for actual mine working intersections in 3D space.

K-MINE includes tools for solid and wireframe editing: intersecting, splitting, merging, extruding solids from polylines at specified distances and angles, and handling ore body bifurcation (where a body splits into separate structures). Wireframe validation tools allow adding or deleting triangles and vertices to clarify geological body shapes.

One of the most common questions in mineral resource and reserve reports is how the cut-off grade was selected, or why there is a significant difference between calculated reserves of proven and probable categories compared to enterprise documents.

For deposits with simple structures and clear geological boundaries, modeling involves straightforward construction of ore body solids. Different ore properties (technological types, grades, and other varieties) are modeled as separate domains.

When strict geological boundaries are absent, mineralization must be outlined using condition parameters. Key parameters include:

The minimum grade in the assay sample (not the cut-off grade for block estimation) controls ore body thickness. Other parameters include maximum thickness of waste rock interlayers and off-grade ore, minimum ore body thickness, minimum commercial content in the calculated block, and minimum content in the drill hole.

The minimum metrogram (the product of cut-off grade multiplied by minimum ore body thickness) ensures that short, high-grade intervals are properly considered.

In the polygonal method, cut-off grade refers to a sample in a drill hole or mine working. In block modeling, it refers to block contents, which frequently differ from sample values. Samples and elementary blocks have different sizes, leading to different parameters of statistical distribution, including dispersion and asymmetry.

Within a single solid, average contents of samples and blocks may be nearly equal. However, when dividing these groups into ore and waste by the same boundary content value, there is a systematic difference in ore grade estimation. This difference depends on the natural features of the deposit, the geostatistical model parameters, cut-off grade level, exploration grid density, and conditions for selecting the mineralized zone.

Condition parameters like minimum content per edge intersection can only be applied when preliminary ore body solids are constructed. Standard calculation parameters (minimum ore body thickness, maximum waste interlayer thickness) are selected considering geological features and the mining method (open-pit or underground).

In the polygonal method for underground mining, the minimum commercial content divides reserves by balance sheet attribution. Based on this parameter, entire ore volumes can be excluded from balanced reserves and transferred to the non-commercial class.

K-MINE focuses on solving specific challenges for mining operations rather than implementing theoretical solutions. On a practical level, this means increasing the quality of ore preparation and reducing time spent on geological data audits.

Most successful projects are achieved through enhanced automation of data processing, analysis, and estimation. K-MINE's tools increase the reliability of reserve estimation at every step of the mining process and reduce risks in the decision-making process.

How is K-MINE different from other geological modeling software? K-MINE can automatically create the geological database structure, which simplifies the process and significantly reduces setup time. The software is more flexible, customizable, and affordable. K-MINE allows creating block models with a larger number of blocks and more detail. There is also a specialized block model for seam (reservoir) deposits, and the block model supports property lists that simplify and speed up processing of text data.

What interpolation algorithm is used for implicit modeling? K-MINE uses the radial basis function (RBF) method for implicit modeling.

Can K-MINE be used for economic assessments and reporting standards? K-MINE can create custom report templates for every enterprise and reporting standard. The software includes extensive tools and functions for generating individual report formats as needed.

Does K-MINE support a mine hierarchy or tree-type structure? Yes, K-MINE has a flexible structure for working with data, storing information in a tree of layers within the layer editor.

What attributes are supported in the block model? Any indicators can be recorded in block model attributes, including block coordinates, block sizes, rock density, weight, grades, rock types, and other custom properties.