Intelligent ladle system

The ladle is a container used to hold molten steel for refining and pouring. It consists of three parts: the outer shell, the inner lining and the casting flow control mechanism. During the use of the ladle, the lining of the ladle comes into contact with molten steel and slag and goes through multiple cyclic stages. At each stage, due to changes in the molten steel level, molten steel temperature, mechanical loads, alternating temperature stresses, mechanical load stresses, and Mechanical impact, result in damage to refractory materials. In the steelmaking process, each production process is organically linked by ladles. Through dynamic monitoring, position tracking, and operation management of ladles, a reasonable ladle turnover model can be established to guide ladles to optimize selection and scheduling when participating in steelmaking production. Provide a strong guarantee for achieving efficient continuous casting.

At present, the status of ladles mainly relies on manual inspection and empirical judgment. Subjective factors have a greater impact on the judgment results. Some melting defects existing at the edges cannot be discovered in time, which poses safety risks of missed inspections and even steel leaks in ladles. In order to ensure production safety, when operators discover melting damage defects, they will take the ladle offline for repairs in advance, which is not conducive to saving ladle refractory materials. Conventional ladle judging methods cannot effectively accumulate ladle status information, which is not conducive to improving the ladle turnover rate and has low production efficiency.

The intelligent ladle system based on the Internet of Things, machine vision, cloud computing and deep learning can realize the monitoring and information feedback of the real-time status of the ladle, establish a reasonable ladle turnover model, and judge the status of the ladle well. The combination of machine vision and deep learning realizes the automatic collection of ladle lining images, and recognizes the lining images to determine the degree of melt loss of the ladle lining, providing auxiliary decision-making for production.

 

Intelligent ladle system

 

1. Introduction to intelligent ladle system

 

1.1 Intelligent ladle system architecture

 

The system collects equipment information such as ladles in real-time, including ladle position information, ladle lining image information, ladle outer wall temperature information, process parameters, steel temperature, steel type, smelting process and production information such as the actual start and end time of each stage schedule. Based on machine vision and deep learning on the computing platform, a combination of data model and mechanism model is trained, and finally, on the application platform, equipment information visualization, equipment fault diagnosis and analysis, ladle melting loss prediction and production-assisted decision-making are realized.

 

1.2 Objectives of intelligent ladle system

 

(1) Implement the ladle IoT system architecture, and the communication middleware receives the real-time status data information of the ladle.

(2) Establish interconnection with the production management system to collect and feedback ladle production schedule information and turnover and transportation plan information.

(3) Dynamically calculate the physical property parameter library, mechanism model, etc. of ladle refractory materials.

(4) Provide real-time temperature monitoring and early warning for the ladle and management of the remaining thickness of the lining of the ladle.

(5) Use machine vision to recognize the real-time collected image information of the ladle lining to determine the melting damage of the ladle and predict steel breakouts.

(6) Based on the deep learning algorithm, the auxiliary decision-making of ladle production schedule information and turnover and transportation plan is realized.

 

2. Implementation of an intelligent ladle system

 

2.1 Ladle IoT monitoring system

 

The intelligent ladle IoT monitoring system is mainly divided into a data layer, model layer, transmission layer and decision-making layer. The data layer collects and stores data, including production data and training databases. The model layer mainly includes components such as feature extraction, target positioning, target classification, training, and model output. The transport layer is used to connect the data layer and decision-making layer. The decision-making layer is the main component of the production environment, including the storage unit, data visualization unit, prediction and decision-making unit, etc.

 

2.1.1 Ladle positioning and ladle number identification system

 

The laser photoelectric switch of the ladle monitoring system detects that the ladle reaches the designated position, the scanning module obtains the ladle number, the ranging sensor obtains the precise position, the collected data is uploaded to the server, the parameters are compared with the original data, a judgment is made, and prompts and warnings are given.

 

2.1.2 Ladle temperature detection system

 

The ladle temperature detection system consists of an infrared temperature measuring camera, which scans all sides of the ladle and uploads it to the server. After the molten steel is packaged, the temperature of the ladle shell is scanned and measured again at the detection point. The time interval is recorded at the same time. The two temperature measurement data are compared. The system uses expert experience to determine whether the temperature is abnormal and issues an early warning to remind whether to perform ladle maintenance.

 

2.1.3 Ladle thickness measurement and melting loss identification system

 

At present, recording the melting loss of the ladle lining is based on manual judgment, using a camera device and manual shooting. The angle and colour of the photo are unstable and may be blocked by foreign objects, etc., which will cause the model training to jitter and be difficult to converge. At the same time, there are insufficient data samples and a lack of training samples labelled by experts.

Intelligent defect solutions based on laser thickness measurement equipment, machine vision artificial intelligence, and machine vision technology can well solve the above problems. The laser thickness gauge based on the principle of laser ranging measures the thickness of the ladle lining at important tank times and key positions of the ladle. Through data fusion and machine vision, it can realize intelligent identification and determination of melting damage defects in the ladle lining and Model training for ladles.

Combined with the ladle lining images obtained by the model, different melt loss types and damage levels are explained. image on ladle
Carry out melting loss annotation, and transfer the ladle lining image to the model for machine vision learning and training. Finally, the development and optimization of the model can be completed, and the recognition rate can be continuously improved.

 

2.2 Process and Quantitative System

 

Collect and organize data during the ladle transfer process, predict the heat loss of the ladle and the temperature of the middle ladle based on historical data and judgment, and provide information to operators to assist decision-making. Unstable pouring temperature will lead to a series of adverse consequences. Analyze and quantify the various factors that affect the pouring temperature, collect data from each stage of the ladling process, and establish a temperature prediction model based on deep learning, which will help improve the quality of the slab and increase production efficiency.

The basis for the implementation of the intelligent ladle system is the digitization, informatization and reliable cloud storage of the necessary data, real-time status, resume information, technical files and other information of the ladle, as well as the interconnection of this information on the Internet and mobile terminals. The historical data of the ladle needs to be collected from the following process stages: ladle preheating, electric arc furnace and ladle metallurgical furnace smelting, pouring, and conditioning. These data mainly include historical basic data of each ladle (including ladle thickness, ladle wall erosion rate, ladle wall temperature, whether there is a ladle cover, etc.); real-time thermal status of the ladle throughout the process; real-time position tracking of the ladle; pouring steel volume – Pouring speed: production information such as steel temperature, steel type, smelting process, and actual start and end times of each stage scheduling; physical model of the ladle, refractory structure design drawings; physical property parameters and physical and chemical indicators of various refractory materials of the ladle; supplier information; Information such as the installation position and insertion depth of each measurement sensor.

Design and optimize the system model based on the accurately collected data information as above. First, establish the relationship between input variables and expected output data, conduct an in-depth study of the heat loss of the ladle system, conduct an evaluation, determine the main factors affecting the heat loss of the ladle, and define these behavioural factors that affect the temperature of the ladle as input variables.

Machine learning requires data cleaning to create a suitable data structure, and the data information obtained through the above devices
Used for model research, functional verification and testing, they are classified into training sets, verification sets and test sets respectively.

After data processing, model training, model performance evaluation, result evaluation and error verification, the ladle prediction results are finally obtained, realizing intelligent identification of ladle status, digitization and informatization of full-cycle operation and maintenance, and visualization of production process information such as molten steel composition and temperature. , steelmaking production and ladle production schedule planning information assist decision-making, forming a ladle lining refractory monitoring database, thereby improving the stability and product quality of steelmaking production.

 

Source of article: Gao Bing, Liao Xiangwei, Chai Mingliang, Zhao Chenglin, Wang Lijuan. “Development of Intelligent Ladle System” ANGANG TECHNOLOGY

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.

Find Your Plant

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

Customer testimonials

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!