MAGMA · Projects

Projects

Four industry contours in three modes: industrial operation, a deployed commercial pilot, and work with an international client. Details of each project are disclosed to the extent agreed with the client.

Russia · two regions

Forestry: cutting area control and timber turnover

Two commercial deployments

Two logging enterprises. The systems run in daily production mode.

Forest fund accounting is built on the enterprises' own data: compartment boundaries from cartographic materials, forest inventory descriptions, active logging declarations and actual shipments. Every compartment on the map is coloured by stage — developed, restored, in progress, under-cut — and linked to the declaration under which harvesting proceeds. A separate loop tracks hauling and the remainder at the upper yard.

2,876
compartments with actual boundaries on the map
23,248
inventory stands from 12 description volumes
10
working sections: map, development, shipments, control, reporting
Accounting unit
Compartment, cutting area, inventory stand, assortment lot
What we calculate
Allocated, harvested and hauled volume; stage of development per compartment; under-cutting; upper yard remainder
Reporting
Timber accounting and transactions in the national traceability system
Roadmap
Integration with the statutory contour via electronic signature, ingestion of measurement services, hauling routes on the map

Client names are not disclosed. The figures refer to the volume of data loaded into the system.

Russia · ore mining

Tungsten and rare metals

In commercial operation

An operating mining enterprise. The contour runs in production mode.

Small volumes at high unit value: the cost of an accounting error here is higher than in bulk industries. The contour covers plan versus fact by face and block, accounting for loss and dilution, grade control in the ore mass and an end-to-end metal balance across stages through to concentrate. The forecasting layer estimates concentrate output under the current regime and the plan-to-fact deviation by face.

Accounting unit
Mining block, face, ore mass and concentrate lot
Sources
Sampling and test reports, weight control, processing data, survey measurements
What we calculate
Grade and recovery by stage, loss and dilution, end-to-end metal balance
Forecast
Concentrate output, plan-to-fact deviation, anomalies in concentrate shipment

The client name is not disclosed.

Russia · nuclear industry

In-situ leaching of uranium

Deployed commercial pilot

The pilot is deployed and running on production data.

Technologically the most demanding vertical: control is exercised through the regime of solutions in the productive horizon, and the result appears with a delay in the grade of recovery wells. The platform ingests flow, pressure and grade per well, tracks cell depletion, calculates specific reagent consumption and the L:S ratio, forecasts the recovery profile and assesses the risk of solution migration beyond the design contour from observation well data.

Accounting unit
Leaching cell; injection, recovery and observation wells
Sources
Flow meters and process node sensors, assay reports for the pregnant solution, monitoring contour
What we calculate
Cell depletion, remaining reserve, recovery factor, specific reagent consumption, L:S ratio
Forecast
Recovery profile over the planning horizon, probability of a solution excursion
Special conditions
Full deployment inside the client contour, local language models, immutable audit log

Separately, a pre-pilot dialogue is under way with representatives of the Rosatom State Corporation: at the stage of exchanging materials and agreeing a possible pilot format. Trademark and software registration are in progress.

China · ore mining

Major leaching and ore mining operator

Client

MAGMA's international track. Scope is disclosed by agreement with the client.

A major Chinese operator running both leaching and conventional ore mining. The international contour uses the same platform: source model, volume and weight balance, forecasting layer and language access to data. The top layer differs — instead of Russian statutory systems, the client's reporting forms and local regulator requirements are configured.

Track
Leaching and ore mining
What transfers
Source model, volume and weight balance, ML recovery forecast, language interface
What is configured
Reporting forms, interface localisation, data residency requirements
Scientific basis
Joint work with Henan Polytechnic University on leaching control methods

The client name is not disclosed.

What this adds up to

Three commercial contours across three technologies is a test of portability in practice, not on a slide. The source model built for in-situ leaching works on an ore block and on a cutting area once the domain layer is replaced.

Proven across three technologies

Solution mining, ore mining and timber harvesting — three fundamentally different physical processes on one calculation core.

Running in production

Not demo rigs but daily operation, with accountability for the figure in the report.

Forecast trained on real data

The ML layer is back-tested on closed periods of the specific asset before it enters the production contour.

Want the same contour on your asset?

We start with a survey: the material movement chain, metering points, instruments and reporting requirements. The output is an accounting scheme and a pilot outline.