In order to meet the intelligent production needs of blast furnaces, Shandong Steel Group Rizhao Co., Ltd. equipped two 5100m³ blast furnaces with a set of blast furnace expert systems. To align the expert system closely with production, the project team optimized and adjusted the parameters of each model in the expert system and added systems such as the iron ladle docking system, ore leveling system, and hot blast stove optimization system based on production needs, providing more technical support for blast furnace production.
Keywords: Blast Furnace; Expert System; Optimization; Modules; Parameters
In recent years, domestic blast furnaces have developed towards large-scale operations, requiring control over a significant amount of information and many control points, making traditional blast furnace control methods difficult to achieve precise control. Therefore, achieving intelligent control through an expert system is of great significance. The two 5100m³ blast furnaces at Shandong Steel Group Rizhao Co., Ltd. were put into production in December 2017 and April 2019, respectively. To build a fine-quality base and achieve intelligent production, the company cooperated with Beijing Beike Yili Technology Co., Ltd. in March 2017 to introduce the blast furnace expert system. The installation and development of the expert system were completed before and after the commissioning of the first 5100m³ blast furnace, and it officially went live in July 2018.
1. System Application and Optimization
The blast furnace expert system includes modules such as the main interface, safety warning, material and capacity utilization, smelting process, production management, blast furnace big data, message inquiry, and system settings, as shown in Figure 1. The expert system provides detailed diagnoses and analyses of critical parameters such as economic indicators of the blast furnace, material utilization, total output and total pig iron output, thermal index, thermal load of the furnace body, and pig iron composition information, offering comprehensive data support to the blast furnace operators.
After the expert system went live, each module provided certain guidance for blast furnace production, but the parameters of each module needed continuous adjustments and optimizations based on the actual production conditions to make the expert system more aligned with practical production and provide better guidance. According to the research plan, the project team utilized the opportunity of debugging the 2# 5100m³ blast furnace expert system with developers from Beike Yili to study and continuously optimize the models of each module in the expert system. Based on production needs, new modules were added to better serve production. The team summarized the handling of actual production cases based on the problems encountered during blast furnace production and optimized the expert knowledge base for the system’s warnings and reminders. According to the operational characteristics of the Rizhao company’s blast furnace, the team studied the mechanism models of various modules, such as material and energy utilization, safety warning, economic and technical indicators, and smelting process modules, compared them with the actual operating parameters of the blast furnace, and adjusted and optimized the parameters of each model. Currently, the blast furnace expert system can predict furnace temperature trends, analyze overall thermal load trends, detect airflow, track material levels, draw Rist operation lines (as shown in Figure 2), provide safety warnings for tuyeres and furnace bodies, and analyze the trends of important parameters and operational guidance for the blast furnace.
1.1 Furnace Temperature Trend Prediction
The blast furnace expert system analyzes data from operational parameters, tuyere imaging, slag and iron discharge, raw fuel composition, and operational adjustments to predict the trend of the furnace temperature. The blast furnace foreman performs daily operations based on the expert system’s recommendations, improving the precision and intelligence of blast furnace operations, avoiding reverse operations due to human negligence, and reducing fluctuations in furnace conditions, which is beneficial for stabilizing the quality of pig iron and reducing smelting costs.
1.2 Material Level Tracking
The material level tracking module visually displays the distribution status of the ore and coke layers in the blast furnace, accurately calculating the smelting cycle. When changes occur in the ore batch, coke ratio, or distribution matrix, the module can mark the changing ore batch. Over time, the material level tracking module can accurately show the position of the changing ore batch within the blast furnace, facilitating analysis and adjustments by the blast furnace operators. During the wind recovery process, markings can accurately determine the specific locations of additional coke at the tuyeres and furnace belly, providing a basis for analyzing furnace condition trends and evaluating the timing and quantity of additional coke additions, as shown in Figure 3. The specific positions of the materials in the blast furnace are shown in Table 1.
1.3 Furnace Condition Analysis
The expert system diagnoses changes in furnace conditions based on variations in operational parameters, airflow, cooling wall temperatures, and raw fuel composition, providing operational recommendations. On April 18, 2022, at 14:21, the expert system analyzed the results: 16 segments of westward static pressure fluctuations, a decrease in gas utilization rate, and a wind reduction of 200m³/min. The blast furnace reduced wind flow from 7350m³/min to 7132m³/min based on the operational recommendations, which led to material collapse in the furnace.
Post-analysis: Due to poor pig iron discharge and excessive slag iron, a pipeline journey occurred, and reducing wind flow damaged the pipeline, preventing the situation from worsening. On July 15, 2022, at 23:02, the expert system analyzed the results: a decline in coke quality, an increase in coke ratio was requested by the blast furnace foreman after consulting the area supervisor, raising the coke ratio from 345kg/t to 355kg/t. Post-analysis: Fluctuations in the coking plant’s production caused the coke ash content to rise from 12.3% to 12.9%, and the increase in coke ratio achieved a smooth transition.
1.4 Blast Furnace Safety Warnings
The safety monitoring and warning model for the blast furnace cooling wall in the safety warning module conducts online monitoring and adjustments of parameters such as cooling wall parameters and water velocity. To provide a more intuitive display, the monitoring interface includes data on heat flow intensity and temperature, allowing operators to quickly and intuitively understand changes in airflow and cooling wall slag skin through color variations. In the furnace operation model, the team adjusted selected heat conduction parameters based on the study of the model mechanism established by Beike Yili. Simultaneously, stable slag skin operation periods during actual furnace operation were considered, and a control standard for the central furnace model was established based on big data analysis of similar blast furnaces, as shown in Table 2, to standardize the adjustments in the central furnace.
On May 3, 2022, fluctuations in furnace conditions occurred, with a reduction in airflow. Observing the operation module of the expert system, it was found that the thickness of the cooling wall slag skin in the central furnace was increasing and exceeding control standards, indicating a thickening phenomenon in the furnace wall. Immediate measures were taken to adjust the distribution matrix, develop edge airflow, increase furnace temperature, and reduce slag alkalinity for thermal acid washing. By May 6, the slag skin thickness gradually decreased and returned to normal control range, restoring normal furnace conditions.
During the design of the iron ladle system, the data transmission issue between the blast furnace and the iron ladle was not considered, resulting in the inability of the iron ladle system to operate. To address this, a data transmission scheme between the iron ladle and the blast furnace expert system was proposed from a process perspective, adding a communication and data interaction module for iron ladle scheduling in the blast furnace expert system. This module works in conjunction with the iron ladle scheduling system to upload and issue scheduling instructions, generating, uploading, sending, and sharing blast furnace iron discharge information such as iron weight, ladle number, iron batch number, start and end times of iron discharge, ensuring the normal and stable operation of the iron ladle scheduling system.
Based on the production needs of the blast furnace, a large screen display system was developed to show real-time data of iron weights at each iron discharge point, real-time vehicle scale weights, radar levels, blast furnace airflow, wind pressure, pressure difference, furnace temperature, iron flow speed, and slag flow speed, achieving mutual communication of parameters inside and outside the furnace and effectively guiding operations in front of the furnace, improving the qualification rate of iron discharge and the accuracy of ladle filling, and reducing iron quantity discrepancies.
3. Hot Blast Stove Automatic Optimization System
After the implementation of the blast furnace expert system, issues encountered during the implementation of the hot blast stove automatic optimization system were jointly studied with personnel from Beike Yili, achieving the launch of the hot blast stove automatic optimization system and implementing closed-loop control. Due to issues with flow meters in various gas branches of the hot blast stove, where the flow reached maximum capacity with the valve opened to 50%, the automatic stove operation could not commence. Adjustments were made to the flow meters based on their detection range to meet gas flow requirements, and the hot blast stove automatic operation system was put into use. During the implementation, the calculation process of the hot blast stove automatic operation was audited, changing the design concept centered on saving gas, recalibrating key parameters and calculation methods such as gas calorific value and combustion temperature, and air-fuel ratio selection. While considering gas savings, the combustion temperature during stove operation was lowered, ultimately achieving automatic operation of the hot blast stove and resolving the issue of NOx exceeding limits.
4. Blast Furnace Ore Layer Distribution Model
To control the mixing degree of different ores entering the blast furnace, a standard for controlling ore layer distribution was designed during the material discharge process, controlling pellet ore distribution at 50%-60%, lump ore distribution at 45%-55%, and coke distribution at 40%-50%. Based on these control requirements, a module for displaying the ore layer distribution rate was added to the blast furnace expert system, showing real-time data on the flow lengths and relative positions of different materials, calculating the current level of distribution, and immediately correcting any discrepancies to meet the ore flow distribution control standards.
The expert system was designed with a CS and BS dual architecture service platform, but only the CS client mode was contracted. The project team, in collaboration with Laigang Electronics and Beike Yili, tackled data transmission barriers from the perspectives of process, primary automation, and expert systems. A gateway communication machine was deployed in the blast furnace master control, with one network card connected to the primary control system and another connected to the secondary system, using dual network cards and a hardware firewall for communication with the primary control system. The entire secondary system was deployed on the group’s intranet, isolating it from the primary control system. The master control room primarily accesses the secondary expert system of the blast furnace through the CS client mode, while the BS server for the secondary system is deployed on the web application server and published externally. Company management personnel can access the secondary system via the BS mode as long as they can access the company intranet. This allows blast furnace management technicians to view the actual operation status of the blast furnace remotely, and through Rizhao company’s mobile cloud desktop, they can check at any time, facilitating better control and adjustments of the blast furnace operations, creating favorable conditions for stable furnace operation.
6. Enriching the Expert Knowledge Base and Improving the System
In the blast furnace expert system, a reasoning and diagnosis model for furnace conditions has been established, currently based on ironmaking theoretical knowledge, with nearly a hundred reasoning rules established for issues such as sliding materials, collapsing materials, airflow distribution changes, pipeline issues, suspended materials, accumulation in the furnace body, and changes in blast furnace conditions. Although the theoretical knowledge for ironmaking is the same regardless of the size of the blast furnace, the specific manifestations of different furnace conditions vary significantly. Based on the production experience since the commissioning of the 5100m³ blast furnace in Rizhao, the project team continuously summarizes typical handling methods for blast furnace production, establishing various control models. Once these models mature, they are integrated into the blast furnace expert system to standardize blast furnace operations. In 2019, the project team combined model reasoning rules with on-site process realities, adjusting operational values according to the actual situation of the blast furnace, optimizing the expert knowledge base.
To meet the operational habits of the blast furnace and the needs for data analysis, the report system was improved for more targeted and practical use; various trends were optimized to align with the operational habits of Shandong Steel Group Rizhao Company, enabling operators to observe trend changes more intuitively; and the accuracy of data was continuously improved. The blast furnace expert system requires accurate data, and since the commissioning of the blast furnace, the project team has consistently prioritized data accuracy, continuously checking and calibrating detection and measurement data, replacing substandard measuring instruments to ensure data accuracy.
(1) The models of the expert system are continuously optimized to be more aligned with actual production, improving the accuracy of model predictions and warnings, and reducing fluctuations in furnace conditions; the report system is improved; and the expert knowledge base is continuously enriched to provide expert recommendations that align with actual production, supporting blast furnace operations. Real-time online monitoring of blast furnace safety conditions is achieved, preventing major safety accidents and extending the service life of the blast furnace.
(2) The optimization of the iron discharge module ensures the stable operation of the iron ladle scheduling system, achieving mutual communication of parameters inside and outside the furnace.
(3) The optimized hot blast stove automatic optimization system has been launched, resolving the issue of NOx exceeding limits.
(4) The blast furnace ore layer distribution model has been established.
(5) Remote login to the secondary system of the blast furnace has been achieved.
The Rizhao company blast furnace expert system has significantly improved its auxiliary role in blast furnace production operations, helping operators grasp the operational status of the blast furnace, enhance operational levels, reduce smelting costs, and improve product quality, achieving safe, long-lasting, efficient, and scientific management of the blast furnace. In the next steps, the blast furnace expert system will continue to optimize in production, continually improve the data support system, enhance the accuracy and timeliness of various data, and optimize the parameters of each model to provide better data and technical support for blast furnace operations.
[1] Zhou Chuan Dian, Liu Wan Shan, Wang Xiao Liu, et al. Blast Furnace Ironmaking Production Technology Manual [M]. Beijing: Metallurgical Industry Press, 2002.
[2] Chen Jian Hua, Xu Hong Yang. “Current Status and Development Trends of Blast Furnace Expert Systems” [J]. Modern Metallurgy, 2012, 40(03): 6-10.
Source: Fu Ting Qiang (Shandong Steel Group Rizhao Co., Ltd.)
Welcome to Join the Ironmaking, Steelmaking, and Rolling Mill Managers’ Communication Group
Add 15031180999 to Join the Group
The Metallurgical Information Equipment Network provides various technical, product promotion services, technical exchanges, industry consulting, business cooperation, and article submissions. Please add the editor’s WeChat15031180999 Liu Wen Qing
Scan to Add Me on WeChat