The system combines X-Ray Transmission (XRT) sorting with Tomra’s Obtain and Contain software platforms, which use artificial intelligence to improve particle classification and separation.
Soma expands process optimisation in Colombia
Soma Gold Corp. is active in gold exploration, development and production in Colombia, with operations concentrated in the Antioquia region. At El Bagre, the company processes gold-bearing material from its own mines as well as third-party sources. This operating model requires a processing setup that can handle material with varying grades and characteristics while maintaining stable plant performance.
A central objective at the site is to reduce the amount of non-valuable material entering downstream stages such as crushing and grinding. Early removal of waste is intended to improve feed consistency, reduce wear and energy demand, and make better use of plant capacity.
Sensor-based sorting for earlier waste rejection
The technology selected for El Bagre is based on dual-energy XRT, which distinguishes particles according to atomic density and enables separation of mineralized and non-valuable material. Tomra has combined this sensing approach with artificial intelligence tools designed to improve classification accuracy under demanding operating conditions.
Obtain uses deep learning to classify individual particles at high throughput, including under conditions of high belt occupancy. Contain is intended for more complex inclusion-type material and is designed to identify valuable mineralization embedded in host rock that may be less visible to conventional sorting approaches.
According to Soma, the objective is not only to upgrade mineralized material, but also to improve the performance of the overall plant by reducing unnecessary load in the downstream circuit.
Test work supports investment decision
Before committing to installation, Soma completed a benchmark test programme at Tomra’s test centre using representative material from its Colombian operations. The programme covered different size fractions and mineralized material types in order to assess system performance under variable feed conditions.
The results indicated a clear upgrade of the mineralized material stream and consistent sorting performance across different material classes. The tests also examined challenging scenarios, including smaller particle sizes and dense material flow, where the system maintained accurate detection and separation.
Tomra said the work showed that sensor-based sorting can separate lower-value material from mineralized streams and support a more controlled processing strategy.
Performance under demanding plant conditions
One of the factors highlighted during testing was the sorter’s ability to maintain precision across both coarse and fine particle ranges and under high belt loading. The system also incorporates Tomra’s TS100 ejection module, which the company says can reduce compressed air consumption by as much as 70% while maintaining reliable separation performance. This has implications for both operating efficiency and running costs.
Integration into the El Bagre flowsheet
Following the test campaign, Soma placed an order for a COM Tertiary XRT 1200 sorter for direct integration into the main El Bagre processing flowsheet. The sorter is intended to improve material handling, support more stable plant operation and increase the value of the processed stream by rejecting waste earlier in the circuit.
The project is being developed with Dismet, Tomra’s local integration partner, which is supporting the design and implementation of the installation for the requirements of the El Bagre operation.
Material value and process control
For Soma, the planned installation is part of a broader effort to improve both process control and mineralized material utilization. By reducing the variability and volume of feed entering the plant, sensor-based sorting is expected to support more efficient throughput management while increasing the concentration of valuable material delivered to downstream processing.
With the El Bagre project moving into implementation, Tomra’s XRT and AI-based sorting technologies are set to become part of Soma’s operating strategy in Colombia. The company is using the system to improve upgrading performance, reduce unnecessary processing load and strengthen plant efficiency at a site handling variable gold-bearing material streams.
