Conveyor belt,as a important media of ore transport.It's safety and efficiency influence directly the stability of  production line. The foreign object in the conveyor belt will ruin the belt even to destroied the belt and productional device of the ore.Facing the problem,traditional way is to take manual sorting way,but there are problems such as low efficiency, high misjudgment rate, and significant safety hazards.

With the breakthrough of industrial intelligence technology, the foreign object removal system based on AI visual recognition and robot collaboration is reshaping the operation mode of material sorting in mining tunnels.

一、The principles of foreign removel system

The foreign object removal system consists of a three-tier architecture consisting of a perception layer, a decision-making layer, and an execution layer, forming a closed-loop control.The perception layer(optical system）consists of a collection module and a recognition module.

Decision-making layer (early warning system): the edge computing unit analyzes the information of foreign objects in real time, completes the classification and coordinate positioning of foreign objects within 2ms, and dynamically plans the capture path in combination with the material flow rate.

The independently developed trajectory prediction algorithm can compensate for coordinate offset caused by conveyor belt vibration (error<± 3mm). According to the size and weight of the foreign object, dynamically select the grasping mode (automatic grasping, warning shutdown+remote operation, manual intervention).

Execution layer (robotic arm system): The modular design of the six axis collaborative robot is equipped with an electromagnetic vacuum dual-mode end effector, with iron foreign objects using magnetic suction devices and non-metal foreign objects switching to vacuum suction cups.

The average response time of the execution unit is less than 0.8 seconds, the single grasping cycle is less than 2.5 seconds, and the foreign object is transferred to the recovery device without stopping the machine. In warning mode, the conveyor belt is controlled to stop by IO signal, and the sound and light alarm prompts manual intervention.

二、Typical application scenarios of foreign removel system

1. Coal power industry (coal transportation system)

Pain points: Metal objects such as residual detonators and broken anchor rods can easily cause the crusher to jam, and wooden debris can cause abnormal friction coefficient of the belt

Solution: Deploy a dual station robot system in front of the coal crusher, equipped with a detection module that can identify metal parts in the coal flow. According to data from a 2 × 1000MW power plant, there are 2175 hazardous materials per year, and if the system intercepts the equipment, the failure rate can be reduced by 73%.

2. Alumina plant (ore pretreatment)

Challenge: Steel filter mesh fragments (size 5-30cm) mixed in red mud can easily cause damage to high-pressure diaphragm pumps

Innovative application: Deploy foreign object sorting robots at the outlet of the vibrating screen, combined with detection technology, to capture non-magnetic stainless steel fragments. The annual production of 800000 tons of alumina project can reduce the cost of spare parts replacement by more than 2 million yuan per year.

3. Cement industry (raw material transportation)

Special requirement: It is necessary to distinguish between normal agglomerated clinker (allowed to pass through) and foreign objects

Technological breakthrough: Developing a classification model based on the recognition of agglomerated clinker and foreign object characteristics, and determining agglomerate properties through material surface features. The false catch rate can be reduced from 15% to 1.8%.

4. Iron ore, copper ore, lead-zinc ore (roadway transportation)

Composite scenario: Metal and non-metal foreign objects mixed together, iron parts easily disturbed by ore, and work in high dust environments

Engineering practice: Installing a dust cover on the robotic arm and integrating an air curtain cleaning module into the visual system. Isolate and protect the equipment, achieve a foreign object recognition rate of ≥ 92%, intercept over 4000 iron pieces annually, and reduce belt tearing accidents by 70%.

In the future,the foreign removel system will promote the transformation of mining sorting from "passive disposal" to "active prevention and control".