Home > News>Full-Process Intelligent Agent Collaborative Steelmaking System for BOF Converters

Full-Process Intelligent Agent Collaborative Steelmaking System for BOF Converters

Oct. 11, 2026

Full-Process Intelligent Agent Collaborative Steelmaking System for BOF Converters

As the steel industry moves toward digitalization, intelligent manufacturing, and low-carbon production, traditional converter steelmaking methods that heavily depend on operator experience are facing increasing requirements for process stability, production efficiency, quality consistency, and cost control.

The full-process intelligent agent collaborative steelmaking system for BOF converters integrates industrial big data, metallurgical process models, artificial intelligence, machine learning, Industrial IoT, digital twin technology, and multimodal sensing technologies to connect key processes including converter blowing, slag making, bottom stirring, alloying, and tapping.

The system creates a closed-loop control process of “sensing → analysis → decision-making → execution → feedback”, enabling more standardized, adaptive, and data-driven converter steelmaking.

As a professional electric arc furnace, converter, and steelmaking equipment manufacturer, Sanrui Electric Furnace provides EAF, LF/LRF, converter-related equipment, and intelligent steelmaking solutions for steel producers seeking to upgrade their production processes.

1. Background of Intelligent Converter Steelmaking

The basic oxygen converter is one of the key processes in modern steel production, connecting hot metal pretreatment with secondary refining and continuous casting.

Traditional converter operations rely heavily on operator experience. Differences between operators and shifts can result in variations in blowing control, endpoint accuracy, slag conditions, and molten steel quality.

One-click steelmaking systems based on sublance measurement and static thermodynamic models have been adopted by many steel plants. However, these models generally depend on relatively stable raw material conditions. Variations in hot metal temperature, composition, scrap ratio, charging conditions, or auxiliary materials can reduce model accuracy and may increase the risk of abnormal conditions such as slag collapse, slopping, or reversion of slag conditions.

A new generation of intelligent converter steelmaking systems therefore needs to move beyond fixed-model control toward multimodal sensing, dynamic analysis, intelligent decision-making, and coordinated autonomous control.

By combining off-gas analysis, acoustic slag monitoring, flame monitoring, historical production data, and metallurgical models, the system can provide a more comprehensive understanding of converter conditions and support dynamic process optimization.

2. Full-Process Intelligent Agent Collaborative Control

The intelligent steelmaking system divides key converter operations into specialized intelligent control agents and connects them through a unified data platform.

The system can integrate 10 core intelligent agents and 5 supporting functional modules.

The core intelligent agents include:

Intelligent Blowing Control Agent
Slag-Making Material Agent
Alloying Optimization Agent
Bottom Stirring Maintenance Agent
Dynamic Slag Condition Analysis Agent
Bottom Argon Stirring Agent
Slag Splashing and Furnace Protection Agent
Tapping and Converter Tilting Agent
Parameter Error-Proofing Agent
Big Data Lance Position Agent

Supporting modules can include process operation data recording, key smelting data queries, process quality evaluation, online production cost monitoring, and comprehensive converter information visualization.

Through data exchange between different intelligent agents, the system connects the main converter production process and enables coordinated control across multiple operations.

The core control logic is:

Real-Time Sensing → Data Analysis → Intelligent Decision → Automatic Execution → Feedback → Continuous Optimization

This closed-loop architecture reduces dependence on individual operator experience and improves process standardization and consistency.

3. Multimodal Data Sensing and Fusion

The interior of a converter is characterized by extreme temperatures, intense reactions, dust, and rapidly changing process conditions. Many important metallurgical parameters cannot be measured directly using conventional instruments alone.

The intelligent steelmaking system therefore combines multiple sensing technologies to obtain information from different dimensions.

The main data sources include off-gas, acoustic signals, and flame images, providing a comprehensive perception of converter conditions.

3.1 Off-Gas Analysis

The off-gas analysis system continuously measures gas components such as hydrogen, carbon monoxide, and nitrogen in the converter exhaust gas.

Changes in gas composition over time can be analyzed to evaluate the carbon-oxygen reaction and decarburization process.

The system can identify important process stages, including:

Slag formation
Changes in decarburization
Slopping-sensitive periods
Blowing endpoint trends

When combined with other process data, off-gas analysis can improve endpoint prediction and dynamic blowing control.

3.2 Acoustic Slag Monitoring

During converter blowing, carbon-oxygen reactions and slag conditions generate characteristic acoustic signals.

The acoustic slag monitoring system collects and processes converter noise signals and extracts sound intensity and other relevant features to evaluate slag behavior.

Changes in acoustic intensity can provide information about the reaction intensity and slag foaming condition. Stronger sound signals may indicate a tendency toward slag reversion or reduced foaming, while lower sound levels can indicate more stable slag foaming conditions.

Combining acoustic data with off-gas and visual information can improve the reliability of slag condition analysis.

3.3 Flame Monitoring

The converter flame provides important visual information about the blowing process.

High-definition cameras capture flame images at the converter mouth. Image-processing technologies, including color-space conversion and edge detection, are then used to extract flame color, brightness, and morphological characteristics.

By correlating flame features with process parameters and decarburization behavior, the system can assist in evaluating the condition of the molten bath.

Multimodal data fusion combines off-gas, acoustic, and visual information to reduce the limitations of individual sensors and improve overall process perception.

4. Data-Driven Intelligent Decision-Making and Optimization

An intelligent steelmaking system must not only perceive converter conditions but also make appropriate decisions based on real-time information.

The system combines metallurgical process knowledge with machine learning, data analytics, and intelligent optimization algorithms to develop models for slag condition analysis, endpoint prediction, lance position optimization, and alloying optimization.

4.1 Intelligent Alloying Optimization

The alloying optimization agent can evaluate alloying strategies based on:

Current molten steel composition
Target steel grade
Alloy inventory
Alloy prices
Deoxidation requirements
Target composition range

Different carburization, deoxidation, and alloying options can be evaluated using intelligent optimization algorithms.

The system automatically selects a suitable alloying strategy while meeting steel grade requirements and reducing unnecessary alloy consumption and production costs.

4.2 Big Data-Based Lance Position Optimization

Oxygen lance position has a significant influence on converter blowing efficiency, decarburization, and slag behavior.

The big data lance-position agent analyzes historical heats with similar hot metal conditions, charging conditions, and process parameters.

By comparing the process data of the previous N similar heats, machine learning models can generate real-time recommendations for oxygen lance position and flow rate.

Compared with fixed lance-position curves, this approach enables more adaptive control according to actual converter conditions.

4.3 Autonomous Intelligent Agent Decision-Making

The intelligent agent architecture improves the system's ability to adapt to changing production conditions.

When hot metal temperature, composition, charging conditions, or other external factors change, the system can reassess the current process state and dynamically adjust its control strategy.

Converter control can therefore evolve from fixed-parameter operation toward dynamic optimization, autonomous decision-making, and adaptive control.

5. Converter Digital Twin and Digital Human Technology

Digital twin technology creates a virtual representation of the physical converter and its production process.

The system uses industrial animation and real-time data interfaces to build a digital converter model. Key parameters such as oxygen lance position, blowing flow rate, slag conditions, and other process data are synchronized with the virtual model.

Operators can visually monitor:

Oxygen lance trajectories
Blowing flow changes
Converter conditions
Process trends
Key metallurgical parameters
Intelligent agent operating status

This transforms conventional parameter monitoring into a more intuitive and dynamic visualization environment.

Digital Human-Assisted Operation

The system can also incorporate a digital human interface that allows operators to query process information and interact with the system using natural language.

Operators can ask about:

Current blowing status
Temperature trends
Carbon content prediction
Lance position
Slag conditions
Historical similar heats

When abnormal trends are detected, the digital human can provide multimodal alerts and assist operators in responding to potential process problems.

6. System Architecture and Implementation

The full-process intelligent converter steelmaking system adopts a layered architecture consisting of:

Data Acquisition Layer → Data Processing Layer → Control Decision Layer → Application Layer

Data Acquisition Layer

The data acquisition layer acts as the intelligent agents' “sensing interface.”

Sensors, off-gas analyzers, acoustic monitoring devices, high-definition cameras, and industrial control systems continuously collect real-time production data.

Data Processing Layer

The data processing layer performs data cleaning, abnormal-data identification, feature extraction, and model calculations.

It provides intelligent models for slag condition analysis, endpoint prediction, and other key process parameters.

Control Decision Layer

The control decision layer acts as the “execution center” of the intelligent agents.

Based on real-time analysis, the system generates control commands for oxygen lance position, blowing flow rate, material addition, bottom argon stirring, and other parameters.

PLC and industrial control systems then execute the corresponding commands.

Application Layer

The application layer provides visual interfaces for intelligent agent status, process parameters, prediction results, and warning information.

Operators can monitor system performance in real time and intervene manually when necessary.

7. Industrial Application Results

The application of a full-process intelligent converter steelmaking system can deliver benefits in quality, productivity, cost control, and process standardization.

According to industrial application data, the system can achieve:

Converter endpoint hit rate increased from approximately 85% to 95%;
Slopping occurrence reduced by approximately 60%;
Steelmaking cost reduced by approximately RMB 10–20 per ton;
Converter heat cycle shortened by approximately 1–2 minutes per heat.

Through real-time analysis of multimodal data, the system can identify converter reaction conditions more accurately and adjust blowing parameters in a timely manner.

During slag making, the system calculates auxiliary material additions and timing according to slag conditions and hot metal parameters.

During carburization, deoxidation, and alloying, intelligent optimization models can select appropriate alloying strategies and work with automatic charging systems to reduce alloy consumption and operator-related errors.

Actual performance may vary depending on steel grade, raw material conditions, equipment configuration, process control capabilities, and plant operating conditions. Final performance indicators should therefore be verified through individual industrial applications.

8. Driving Intelligent Transformation of Converter Steelmaking

Traditional converter steelmaking relies heavily on operator experience. Intelligent steelmaking systems can convert operational experience, historical production data, and metallurgical knowledge into repeatable digital process-control capabilities.

By integrating multimodal sensing, Industrial IoT, artificial intelligence, digital twins, and automation, the system can reduce unnecessary manual intervention and improve consistency between different operators and shifts.

The resulting benefits can include:

More stable molten steel quality
Lower slopping and abnormal-condition rates
Shorter steelmaking cycles
Lower alloy and auxiliary material consumption
Improved equipment utilization
Reduced operator-related errors
Greater production transparency

In the long term, intelligent converter steelmaking is more than an automation upgrade. It is an important step toward data-driven, adaptive, and autonomous steel production.

9. Sanrui Electric Furnace: Steelmaking Equipment and Intelligent Solutions

As a professional electric arc furnace and steelmaking equipment manufacturer, Sanrui Electric Furnace is committed to providing efficient, reliable, and intelligent equipment solutions for steel producers.

From Electric Arc Furnaces (EAF) and Ladle Furnaces / Ladle Refining Furnaces (LF/LRF) to converter-related steelmaking equipment and automation systems, Sanrui Electric Furnace focuses on the integration of equipment, metallurgical processes, and intelligent control.

As steel production becomes increasingly digital, green, and intelligent, modern steelmaking equipment needs to provide not only the required production capacity but also reliable automation interfaces, data acquisition capabilities, and opportunities for intelligent control upgrades.

Sanrui Electric Furnace can develop equipment configurations and customized steelmaking solutions according to customer production capacity, steel grades, raw material conditions, process requirements, and automation needs.

This enables steel producers to progressively develop more efficient, stable, and intelligent modern steelmaking production systems.

10. Conclusion

The full-process intelligent agent collaborative steelmaking system for BOF converters integrates industrial big data, metallurgical process models, artificial intelligence, multimodal sensing, digital twin technology, and industrial automation to transform traditional experience-based converter operation into data-driven intelligent control.

Through the closed-loop process of “sensing → analysis → decision-making → execution → feedback,” the system can dynamically optimize critical operations including blowing, slag making, alloying, bottom stirring, and tapping.

With continued advances in artificial intelligence, Industrial IoT, and digital twin technologies, converter steelmaking is expected to move further toward self-learning, self-optimization, autonomous decision-making, and reduced-operator operation.

For steel producers, intelligent steelmaking can improve production efficiency and quality while supporting cost reduction, energy efficiency, emissions reduction, and the development of smart steel plants.

Sanrui Electric Furnace will continue combining its steelmaking equipment manufacturing capabilities with intelligent technologies to provide advanced EAF, LF/LRF, converter-related equipment, and intelligent steelmaking solutions for steel producers worldwide.

Previous: Ferroalloy Production Methods Explained: Blast Furnace, Submerged Arc Furnace, Refining Routes & How to Choose

Next: How to Choose a Submerged Arc Furnace for Industrial Silicon Smelting?

5 Ton Electric Arc Furnace

The 5 ton electric arc furnace for steel-making is a special purpose equipment that makes ordinary steel, quality carbon steel, alloy steel and non-corrosive steel with electric arc as heat source and scrap steel (iron) as raw material.