Webinar: Planning Automation with K-MINE Scheduling Module

Anya, Business Development Manager at K-MINE, opens the webinar and introduces the topic – a detailed walkthrough of K-MINE's scheduling module for open pit mine production planning. She is joined by Eugene Chernichenko, Mining Engineer at K-MINE, and guest speaker Gabo Machado, CEO and Founder of Argelas, K-MINE's partner in Brazil.

K-MINE has been on the market for over 25 years, developing integrated mine planning and design solutions for both open pit and underground operations. Over this period, the company has built 15 mine models and worked with users transitioning from other mine planning platforms. Users who switched to K-MINE reported more precise estimation algorithms, better forecasting accuracy, reduced operating expenses, and improved production levels.

Gabo Machado presents a brief history of mine planning technology, starting with one of the earliest recorded mine plans from approximately 500 years ago – a holistic document covering vein-type metallic ore, mine design, metallurgical processes, and tool descriptions. He draws an analogy between that early integrated approach and the direction modern mine planning is heading.

In the 1950s, punch cards were state of the art. The concept of computerized mine planning originated with a mining engineer who later joined IBM, and Lane's theory was applied to cut-off grade optimization – the initial steps toward what we now know as mine planning and scheduling.

The Lerchs-Grossmann algorithm, introduced around 60 years ago, became the dominant method for pit optimization and remained so for nearly 30 years until 64-bit technology enabled the first practical 3D geological modeling and mine design applications. In 1994, Mr. Nasarenko, the founder of K-MINE, was already working on automated survey calculations and graphic modeling.

By the early 2000s, the industry had progressed to optimal contour determination, strategic and operational planning, and integration with dispatch management systems. A shift occurred from the Lerchs-Grossmann algorithm to direct block scheduling, which is naturally more suited to sequential short- and mid-term production scheduling due to the block-by-block nature of the approach.

The modern mine planning process follows a clear sequence: geological exploration feeds the geological model, which justifies further investment. From there, the workflow splits into mine design and mine scheduling. Today's scheduling tools consider all operational aspects – haulage, environmental constraints, and equipment parameters – as either inputs or constraints to the schedule.

Gabo highlights several key trends shaping the future of mine scheduling:

The mining industry is moving toward systemic, integrated management. Companies like BHP and Gold Fields treat planning as the operational cornerstone, with all other activities – topography, drilling, haulage – feeding into and receiving feedback from the plan in real time.

Mineral scarcity and the shift to a circular economy are changing which commodities receive investment. Coal and iron ore are expected to shrink in market share, while lithium, cobalt, copper, uranium, and rare earths are growing. Mining companies are rebalancing portfolios accordingly.

The digital era is transforming mine operations through digital twins, artificial intelligence, advanced analytics, and real-time sensor data. Digital twins enable two types of workflows: simulation (evaluating the impact of operational changes – such as switching explosive types or blast patterns – on the entire value chain) and prediction (making forward-looking operational decisions based on projected outcomes). Real-time sensor-driven automation, such as automatic reagent dosing in water treatment, is already reducing costs at operating mines.

The scheduling module is designed to determine the mining sequence of individual blocks within an open pit, ensuring compliance with quality, volume, and technical requirements across defined time intervals.

The block model is the primary data source. In the demo, K-MINE uses an artificial iron ore deposit model containing approximately 4 million blocks. Each block has a cubic shape of 5 x 5 x 5 meters and carries attributes including coordinates, rock code, total iron content, magnetic iron content, grade, and weight.

Rock types are classified by code. In this example, all rocks except codes 2 and 4 are classified as overburden (waste). Rock code 2 is assigned to ore type K2 and rock code 4 to ore type K3. Quality parameters – total iron and magnetic iron – are defined for each ore type.

Wireframes define the 3D extents of the mineable area. Two optimal pit boundary positions are used, both previously determined by K-MINE's pit optimizer module. The inner (smaller) boundary represents the initial pit position. The outer boundary represents the final pit position at the end of the last scheduling period. The volume between these two boundaries is what needs to be scheduled for extraction.

Users can create multiple scheduling versions, each stored in a separate project with a unique name and directory. All source data and calculation results are kept in the project folder. A tree diagram organizes the data hierarchy: block model at the top level, followed by sections, excavators, and scheduling parameters.

After loading the block model, the software calculates the number of mineable blocks and computes volume and weight by rock type within the pit boundaries.

The blocks between the two wireframes are divided into sections – geographic zones that represent autonomous mining areas within the pit. Sections are important for controlling pit wall movement, managing equipment access, and ensuring that mining progresses safely and efficiently.

To create a section, the user defines its plan-view boundary, sets upper and lower elevation limits, and confirms. Blocks are automatically assigned. If a new section overlaps an existing one, the overlapping blocks are transferred to the new section and removed from the original.

Each section contains level-by-level statistics: average total and magnetic iron, number of blocks, volume, and weight for the entire level and by rock type (overburden, K2, K3).

Each section is assigned an excavator. Excavator parameters include the start date, assigned section, and capacity by rock type. In the demo example, overburden capacity is set to 1,000 cubic meters per day, and ore (K2 and K3) capacity is set to 2,500 tons per day.

The user defines the scheduling start date, interval unit (e.g., one month), and total number of intervals (e.g., 10 months). This establishes the scheduling timeline.

With the minimum required inputs configured – block model, wireframes, sections, excavators, and time intervals – the schedule is calculated. Results are displayed in a table showing, for each period, which levels the excavator works on, start and end dates per level, volumes of rock mass mined, and average quality values (total and magnetic iron) for each ore type.

Final summary statistics include total extracted volumes, total weight of K2 and K3 ore, and overall quality parameters across the scheduling horizon.

By default, the scheduling algorithm uses a retreating stope method – mining begins at the sections of each level that are in contact with the existing open pit space. The excavator advances along the free boundary, gradually extracting blocks according to a specified stope width.

The stope width parameter controls the size of each mining increment. In the demo, changing the stope width from 30 meters to 10 meters alters the block grouping pattern – blocks are now grouped in 10-meter-wide strips – but the total mining time remains the same because excavator capacity is unchanged.

When multiple sections are created (five in the demo), each with its own excavator, the software automatically handles dependencies between sections. If blocks in one section cannot be mined until overlying blocks in an adjacent section are removed, the schedule reflects this constraint – work on dependent sections is delayed accordingly.

Sections can be modified after creation. Users can adjust section boundaries, shift upper or lower elevation limits, and re-create sections as needed.

To simulate ramp cutting or other preparatory works, the user can carve out a sub-section (e.g., a small portion of the overlying layer from an existing section) and assign it as a separate mining task. For example, excavator 1 first mines section 6 (the ramp area) in 4 days, then moves to section 1 for the main extraction.

Sub-sections can also be selected from non-uppermost levels to model scenarios where mining on a lower bench must precede work on adjacent sections.

The module supports flexible event-based scheduling. Excavators can be triggered by several event types:

Beginning of plan – the excavator starts at the beginning of the first period. On a certain day – the excavator starts on a specified date. After mining a section – the excavator starts after another section or level is fully mined. After partial completion – the excavator starts after a specified percentage (by weight) of a level is mined. For example, excavator 2 can begin after 50% of level 2.5 on section 1 has been extracted.

Users can configure the number of excavators (or mining faces) operating simultaneously on a section – up to three in the demo. When the worksite opens on an underlying level, blocks on the side are scheduled for mining, enabling multi-level extraction within a single section.

Ramp parameters are also configurable. Users specify the ramp length to the underlying level and the minimum tonnage that must become available before mining on the next level is justified. If the available tonnage is insufficient, the algorithm waits until the overlying level is fully mined before descending.

For detailed control, users can manually adjust the mining sequence on individual levels. By going to the manual settings tab, the user specifies stope width, display color, and the preferred order of block extraction. The schedule recalculates to reflect these manual overrides.

In addition to the semi-automatic mode (where the specialist controls each section), the module offers fully automated scheduling. This mode determines the optimal distribution of mining volumes across sections to meet target production indicators for each period.

The user sets target values for the scheduling interval – for example, 25,000 cubic meters of overburden (tolerance: 5,000 m3) and 36,000 tons of ore (tolerance: 6,000 t). The algorithm then recommends which sections to mine and in what volumes to meet these targets.

In the demo, the algorithm recommends mining sections 2 and 5, yielding approximately 30,000 cubic meters of overburden and 33,600 tons of ore – both within the specified tolerances.

Users can further constrain the solution by specifying volume ranges for individual sections. Adding section 3 with a range of 1,000 to 4,000 cubic meters produces a revised distribution of production volumes.

For each section and level, the user can interactively increase or reduce scheduled volumes by moving a pointer. The system dynamically recalculates dependent levels – if volumes on an overlying level change, available volumes on the underlying level adjust automatically.

Each scheduling period can have different target parameters. In the demo, the second period targets 30,000 cubic meters of overburden and 40,000 tons of ore with adjusted tolerances. The algorithm recommends sections 2 and 3, with the option to add section 1 for volume redistribution.

For sections that are not automatically included but are needed operationally, the user can manually specify the mining zone and stope configuration, then recalculate. The process continues iteratively across all periods until the complete life-of-mine schedule meets all volume and quality requirements.

Eugene Chernichenko answers questions from the audience:

Horizon elevation in quarry planning: Yes, users can specify a list of horizons with upper and lower edge elevations and inclination angles. The quarry development plan accounts for these parameters.

Equipment maintenance planning: K-MINE includes a dedicated input form for managing equipment maintenance schedules within the scheduling workflow.

Blast timing and operational parameters: Users can create reference directories containing rock drillability, explosiveness, borehole parameters, days off, blast dates, duration parameters, and other operational data. These feed directly into the schedule.

Excavator movement between sections: Excavator events allow sequential operation across multiple sections at different time periods. Travel time between sections can also be accounted for.

Stockpiles and buffering: This question from Pierre Lafleur was noted for a detailed follow-up response to be posted on K-MINE's LinkedIn page.