Intelligent trend of high speed wire rod mill – labelling robot

1. Implementation background

 

1.1 Background

The five-six high-speed line of the wire business department of Qingdao Special Steel Co., Ltd. was completed and put into operation in May 2021, actively responding to the trend of intelligent, diversified and sustainable development of the industry. The five-six high-speed line has become a key intelligent demonstration project of CITIC Group. The group invested a lot of money and technology, combined with Qingdao’s special steel rolling process, a lot of analysis and research, established an intelligent control centre integrating data collection, statistics, quantification and analysis, and cancelled many traditional workstations and replaced them with robots, such as automatic labelling robots.

The traditional manual tagging process of the high-speed wire rod mill:

After weighing and printing the tag in the weighing room, the tag and the hook are manually removed, and after the tag is threaded, then enter the PF line area, and manually drill into the inside of the wire coil to hang the tag.

There are the following disadvantages

a. Personnel safety risk:

Personnel frequently enter and exit the equipment in the PF material conveying line area; risk of material falling collision, scalding risk, lifting object risk, and data management confusion risk.

b. Reduce labour intensity and labour force:

Personnel repeat a single simple task, repetitive labour is too high, and job tasks are boring.

 

1.2 Program formulation

According to the existing production conditions and scenarios, an intelligent manufacturing plan for this position is designed, and an intelligent listing robot system is introduced. The system integrates the equipment information of the PF conveyor line of the L1 system and the material information of the L2 information system and guides the industrial robots by the AI vision system to replace the traditional manual card playing, card picking, and listing processes.

 

intelligent wire rod mill, labelling robot, AI in steel industry

 

2. Equipment principle and function

 

The main components of the automatic labelling robot system:

Industrial six-axis robot – torso and arm,

AI vision system – eyes,

automatic printing system,

hook system,

Seventh-axis driving system – legs,

Central control system – brain,

Safety and other auxiliary systems (glueing devices).

The following explains the key components of the high-line factory labelling robot.

 

2.1 Industrial six-axis robot – torso arm

The ABB industrial six-axis joint robot is used as the actuator of the tagging system; the multi-angle anthropomorphic state realizes the functions of automatic hook picking and tagging and automatic tagging; the repositioning accuracy reaches ±0.025 mm, and the protection level is IP 67.

 

2.2 AI Vision System – Eye Nervous System

 

2.2.1 Trinocular camera + line laser system:

Monocular colour lens for 2D calculation; binocular depth lens + line laser for 3D point cloud calculation; 2D and 3D spatial overlap conversion auxiliary calculation.

 

2.2.2 Eye-in-hand 3D hand-eye calibration:

AX = XB equation derivation, realize three-dimensional coordinate conversion, hand-eye combination, vision-guided robot movement.

 

2.2.3 Combined application of 2D vision and 3D vision:

2D uses a convolutional neural network and deep learning technology to realize target detection and image segmentation; 100,000 times of deep learning training, data enhancement, residual training algorithm, and continuous learning, computer-independent analysis and decision-making, much higher than traditional image processing algorithms, Compatible with various environmental interference to the greatest extent, such as no fear of direct sunlight, obstacle interference, etc.

The 3D vision part adopts advanced binocular + line laser technology and uses 3D point cloud technology to perform 3D restoration and target 3D position calculation. Use computer image technology to find the sub-pixel position of the line laser, use filter tracking technology to track the motion of the line laser, and use multi-view geometry technology to perform high-precision 3D reconstruction of the binocular line laser.

 

2.3 Automatic printing system

2.3.1 It is composed of an industrial signage printer, signage transmission and positioning and OCR (optical character recognition function) printing quality inspection device.

2.3.2 System flow: Signage template data is automatically printed, including barcode and two-dimensional code printing functions. Automatic cutting, conveying correction and positioning function, transfer the printed signage to the designated position, and wait for the robot to pick it up.

2.3.3 Print detection technology: Use OCR text recognition technology to capture pictures of printed content, extract text for comparative analysis, and monitor print quality.

 

2.4 Hook system

 

2. 4.1 System components:

The hook-making machine system includes five parts: a straightening mechanism, feeding mechanism, diameter changing mechanism, pitch changing mechanism and cutting mechanism.

 

2. 4.2 Features:

Using advanced microcomputer processing technology, through the linkage of multiple sets of servo-mechanisms, and with the linkage of robots, the functions of online hook bending, hook cutting, and hook removal are realized.

 

2.5 Central Control System – Brain

The central control system is responsible for the collection and analysis of equipment information, compatible with Modbus TCP/IP, Simes S7, RPC, Robot SDK and other protocol methods, and realizes comprehensive collection of information and logs from PLC systems, robot systems, vision systems, etc., and a unified analytics platform.

In addition, the system also has functions such as failover and alarm. It is responsible for task definition and arrangement and cooperates with actual business logic to realize system task execution and instruction transmission. It has a unique backup mechanism and multi-channel sharing to ensure the continued stability of the system.

 

3. Technology application and actual effect display

 

3.1 Coil-shaped segmentation

Functional description: Detect whether the stop position of the coil is accurate, whether the centre of the circle is within the safe range, and effectively prevent the C-disc type from being slanted and unbalanced.

Algorithm name: Image segmentation technology with the deep residual network.

Algorithm features and advantages: Segment the coil shape under complex background, and make corresponding judgments and analyses on the coil shape.

Result processing: The judgment result is transmitted to the robot system. When the coil is tilted, the robot is notified not to perform the listing action and the central control system alarms.

 

3.2 Angle positioning of packing line

Function description: Detect the position and angle of the tying line.

Algorithm name: Image Segmentation Technology for Networks.

Algorithm features and advantages: Segment the information of steel strips under complex backgrounds, and use machine learning algorithms to complete the combined calculation of angles.

Result processing: transmit the calculated angle to the robot system, guide the robot to adjust its running posture, and send an alarm to the central control system when the packing line is not detected or the angle deviation is too large.

 

3.3 Wire packing line division

Function description: Measure whether there is a packing line in the picture, and calculate the angular position of the packing line in the picture.

Algorithm name: Image segmentation technology with the deep residual network.

Algorithm features and advantages: Segment the image information of the baling line in a complex coil background and a single feature point environment.

Abnormal handling measures: When the binding line is not detected, the central control system will give an alarm.

 

3.4 Detection of signs

Function description: After the listing action is completed, take a photo to identify whether the sign is hanging normally

Algorithm name: Image Segmentation Technology with Deep Residual Network

Algorithm features and advantages: Detect the presence or absence of signs and the location information of signs in complex backgrounds.

Result processing: When it is detected that there is no sign on the screen, the central control system will give an alarm.

 

3.5 Three-dimensional target positioning of listing points

Function description: Detect the suitable position of the hangable sign inside the coil.

Algorithm name: image segmentation technology of deep residual network; multi-view geometry technology; coordinate system one algorithm; point cloud processing technology.

Algorithm features and advantages: The 3D reconstruction of the scene uses computer vision geometry, image motion tracking, image filtering and other methods, through tracking the laser data, each takes a binocular 90 frames/S image and uses multi-view geometric technology to calculate the scene. High-precision 3D reconstruction; the search for listing points is based on the results of 3D reconstruction, using computer graphics, deep learning segmentation network, point cloud processing and other technologies, collision simulation detection and other technologies to calculate the best listing point.

Result processing: Calculate the three-dimensional coordinate points (X, Y, Z, RX, RY, RZ), transmit them to the robot system, and guide the path.

 

4. Conclusion

The fifth-sixth high-line labelling machine replaces the traditional manual positions, avoiding the mixed steel accident caused by the subjective factors of the employees, reducing the total number of 8 people in the two lines and four shifts, and the direct economic benefit is about 540,000 yuan per year for each line. Provide strong support for the company to implement the policy of cost reduction and efficiency increase, safe production and clean production, which greatly promotes the company’s automation and intelligent development.

Leave a Reply

Your email address will not be published. Required fields are marked *

With machine vision as the core, DBM can provide intelligent equipment R&D, manufacturing, and transformation services such as unmanned factories in the metallurgical industry and other industries, machine vision, and robot application design.

Product category

Customer testimonials

Tell us your needs and a technician will contact you immediately.

Friendship Link

We are committed to becoming a “leader in visual applications”, advocating the core values of “integrity, cooperation, innovation, expertise, and excellence”, taking the road of “technology shaping the brand, innovation leading the industry”, and attaching importance to the investment in scientific research and technology. The research and development of new products and the construction of the team make unremitting efforts for the goal of scientific and technological development.

We have been supplied to world’s top steel manufacturer Arcelormittal, TATA Steel, EZZ steel etc. We do OEM for Concast and Danieli for a long time

Metallurgical Intelligence
Change the way of manufacturing

Save labor, improve working environment, improve efficiency, reduce error rate and save cost!

Металлургический интеллект
Изменить способ производства

Сэкономьте труд, улучшите рабочую среду, повысьте эффективность, уменьшите количество ошибок и снизьте затраты!