420
Increasing Sorting Efficiency in Ore Pre-Concentration by
Applying Smart Combination of X-Ray and Optical Sensors in
AI Models
Jacek Kolacz, Jakub Progorowicz, Marek Dudzik
Comex Polska LTD
ABSTRACT: A new sensor-based sorting system has been developed to achieve efficient separation at different
mineral content levels of the feed material. The system employs high selectivity XRT detectors operating
together with other sensors like RGB and SWIR (Short-Wave-Infra-Red) cameras, installed in the same sorting
unit. Sorting decision is done by AI based models employing convolutional neural networks. It makes it possible
to detect minor inclusions of minerals and metal ores like copper or gold ore. The paper describes industrial
sorting application examples for iron, lead, zinc, gold and copper ores, showing significant advantage compared
to currently used sorting systems.
INTRODUCTION
Processing of minerals and especially metal ores contain-
ing copper and gold, is mainly based on complex expensive
separation processes. For metal ores, it often refers to several
processing stages including crushing, grinding, flotation,
leaching, etc. There is a high operating cost to achieve effi-
cient mineral separation (Wills, 2006). The main reason is
related to very low metal concentration, so the main mate-
rial stream from the processing plant consists of fine ground
waste particles and a relatively small amount of the metal
concentrate fraction. The input material has to be ground
to fine sizes to liberate different minerals, separated in flo-
tations circuits, and often treated by different chemicals
(leaching). The water involved in the process must be recy-
cled, which generates an additional cost. The waste materi-
als often contain toxic elements and compounds, which are
problematic for simple landfilling. All these factors gener-
ate a significant cost of processing and particularly the cost
related to environmental issues has an increasing tendency.
Pre-concentration of metal ores by sorting, has a huge
potential in reducing the processing cost and gaining
important environmental benefits. Application of sensor-
based sorting is very attractive when compared to other sep-
aration technologies. It provides the following advantages,
when compared to the traditional systems:
• Capacity flexibility with a possibility to apply in vari-
ous production lines
• Small space requirement
• No use of water (it is completely dry technology)
• Less complicated pre-processing system (only simple
screening is needed in front of sorting)
• Less need for fine grinding and further mineral pro-
cessing (only applied for unsorted material)
• No requirements of chemicals and media as used in
DMS systems
• Lower investment cost (Kolacz, 2018)
• Very low operating cost in average about 0.1 USD/t
based on current applications (Kolacz, 2018)
Increasing Sorting Efficiency in Ore Pre-Concentration by
Applying Smart Combination of X-Ray and Optical Sensors in
AI Models
Jacek Kolacz, Jakub Progorowicz, Marek Dudzik
Comex Polska LTD
ABSTRACT: A new sensor-based sorting system has been developed to achieve efficient separation at different
mineral content levels of the feed material. The system employs high selectivity XRT detectors operating
together with other sensors like RGB and SWIR (Short-Wave-Infra-Red) cameras, installed in the same sorting
unit. Sorting decision is done by AI based models employing convolutional neural networks. It makes it possible
to detect minor inclusions of minerals and metal ores like copper or gold ore. The paper describes industrial
sorting application examples for iron, lead, zinc, gold and copper ores, showing significant advantage compared
to currently used sorting systems.
INTRODUCTION
Processing of minerals and especially metal ores contain-
ing copper and gold, is mainly based on complex expensive
separation processes. For metal ores, it often refers to several
processing stages including crushing, grinding, flotation,
leaching, etc. There is a high operating cost to achieve effi-
cient mineral separation (Wills, 2006). The main reason is
related to very low metal concentration, so the main mate-
rial stream from the processing plant consists of fine ground
waste particles and a relatively small amount of the metal
concentrate fraction. The input material has to be ground
to fine sizes to liberate different minerals, separated in flo-
tations circuits, and often treated by different chemicals
(leaching). The water involved in the process must be recy-
cled, which generates an additional cost. The waste materi-
als often contain toxic elements and compounds, which are
problematic for simple landfilling. All these factors gener-
ate a significant cost of processing and particularly the cost
related to environmental issues has an increasing tendency.
Pre-concentration of metal ores by sorting, has a huge
potential in reducing the processing cost and gaining
important environmental benefits. Application of sensor-
based sorting is very attractive when compared to other sep-
aration technologies. It provides the following advantages,
when compared to the traditional systems:
• Capacity flexibility with a possibility to apply in vari-
ous production lines
• Small space requirement
• No use of water (it is completely dry technology)
• Less complicated pre-processing system (only simple
screening is needed in front of sorting)
• Less need for fine grinding and further mineral pro-
cessing (only applied for unsorted material)
• No requirements of chemicals and media as used in
DMS systems
• Lower investment cost (Kolacz, 2018)
• Very low operating cost in average about 0.1 USD/t
based on current applications (Kolacz, 2018)
