Wollastonite (CaSiO₃), as a metasilicate mineral with a chain structure, is widely used in ceramics, coatings, metallurgical protective slag and other fields due to its high aspect ratio, low thermal expansion coefficient and excellent insulation performance. China is the world's largest wollastonite producer, and Jiangxi Province contributes more than 60% of the country's production, especially the quality of ore in Xinyu and Shangrao. However, there are generally complex associated problems in Jiangxi wollastonite mines: light blue secondary ore (containing trace iron and manganese elements), black miscellaneous stones (magnetite, hematite) and silica (quartz), calcite symbiosis, which makes ore purification more difficult. Traditional mineral processing technology is difficult to balance efficiency and precision, and the introduction of artificial intelligence sorting technology provides a breakthrough solution to this problem.

1. Associated characteristics and sorting difficulties of Jiangxi wollastonite ore

1.1 Complexity of ore composition

The Jiangxi wollastonite ore bodies are mostly distributed in layers or veins, and the main impurities include:

Silica (quartz): It belongs to the same silicate mineral as wollastonite, with similar density (2.65 g/cm³), but with significant hardness difference (Quartz Mohs hardness 7, wollastonite 4.5-5);

Calcite (CaCO3): It has high calcium content and is easy to form symbiotic agglomerates with wollastonite. Traditional flotation method requires strong acid to adjust pH value;

Light blue secondary ore: It is usually iron-containing silicate (such as epidote), with iron content of 0.5%-2%, which affects the whiteness of the finished product (needs to be >90%);

Black waste stone: It is mainly magnetite (Fe3O4), with significant magnetism but uneven particle size.

1.2 Limitations of traditional sorting technology

Color sorters rely on color difference: they can only remove black debris and dark blue minerals, but are ineffective for light-colored silica and calcite;

Flotation pollution is large: oleic acid, water glass and other reagents are required, and the cost of wastewater treatment accounts for more than 30% of the total cost of mineral processing;

Low efficiency of magnetic separation: the sorting rate of weakly magnetic minerals (such as hematite) is less than 50%.

2. Technical architecture and innovation of MINGDER AI artificial intelligence sorting machine

In response to the bottleneck of traditional processes, MINGDER intelligent sorting machine achieves high-precision sorting through multi-dimensional feature recognition + AI algorithm optimization:

Multispectral imaging technology: combined with visible sensing, synchronously analyzes the color, texture, luster,gloss, shape and other characteristics of the ore;

Deep learning model training: Based on a large number of ore sample libraries (including wollastonite, silica, calcite and different chromaticity secondary ores), a dynamic recognition model is constructed to adapt to ore diversity;

Intelligent adjustment of sorting parameters: According to the differences in ore batches, the sorting threshold is automatically optimized to ensure that the rejection rate of light blue secondary ores (such as low-iron silicates) is greater than 95%, and the silica/calcite sorting accuracy is more than 98%.

Technical Advantage Comparison Table

III. Application Case Analysis of a Mine in Xinyu, Jiangxi

Project Background: A wollastonite mine in Xinyu, with an original ore grade of CaO 45%, SiO2 48%, containing light blue secondary ore (about 8%), black waste stones (3%) and silica-calcite symbiosis (12%), with the goal of purifying to CaO＞48% and SiO₂＜46%.

Traditional process bottlenecks:

Flotation requires a large amount of chemical reagents, which is costly and has great environmental pressure;

Color sorters can only remove black impurities, and the residual light blue secondary ore leads to insufficient whiteness of the finished product.

Implementation of AI sorting scheme:

Pre-treatment stage: ore is crushed to 10-50mm, and dust is removed by dry screening;

Core links of AI sorting:

The sorting machine is set to "high purity mode" to simultaneously identify silica, calcite, light blue secondary ore, and black waste stones;

Adopt high-pressure air spray device to accurately remove impurities with a response speed of 0.1 seconds;

Back-end optimization: After sorting, the ore is magnetically separated to remove iron and finally enters the grinding system.

Economic benefits:

Sorting efficiency increased by 40%, power consumption reduced by 25%;

The grade of wollastonite concentrate increased to CaO 49.2%, SiO₂44.5%, and the total amount of impurities was <2%;

Annual processing of 100,000 tons of ore, with an additional profit of more than 12 million yuan.

IV. Technology promotion and industry transformation

Applicability to other types of minerals:

Quartz sand purification: sorting feldspar, mica and other gangue minerals;

Deep processing of calcium carbonate: removing dolomite and wollastonite impurities;

Strategic minerals: efficient separation of spodumene and feldspar.

V. Industry enlightenment and future prospects

The practice of Jiangxi wollastonite mine shows that AI sorting technology can significantly break through the bottleneck of traditional mineral processing, not only in the treatment of associated minerals with similar colors and complex compositions. It can also help mining companies break through the transformation bottleneck under the "dual carbon" goal in cost reduction, efficiency improvement and green production. At the same time, the combination of AI sorting and comprehensive utilization of tailings can better promote "zero waste" of resources.