Contact us
AI big model opens a new era of steelmaking
This year, the work report of the municipal government clearly puts forward, will vigorously develop and apply the steel vertical field of artificial intelligence model, steel enterprises key process CNC rate of more than 85%.
As an important steel enterprise in Handan, Hebei Yongyang Special Steel Group Co., Ltd (hereinafter referred to as Yongyang Special Steel), with the concept of ‘innovation-driven, high-quality development’ and the goal of ‘standardisation of business, production intelligence and industrial intelligence’, has constructed With the concept of ‘innovation-driven, high-quality development’ and the goal of ‘business standardisation, production intelligence and industrial intelligence’, the company has constructed an integrated system of data collection, storage, governance, analysis and application, carried out feasibility tests of key technologies and pilot applications of key scenarios, and actively built a large model of artificial intelligence in the steel industry.
The picture shows the digital intelligent control centre of Hebei Yongyang Special Steel Group Co. Handan News Media Centre reporter Gong Yuqiang photo
Yongyang special steel ‘special’ where? A few days ago, the reporter went into the front-line workshop, feel ‘AI + converter wisdom steelmaking model’ and digital intelligence to the enterprise brought great changes.
‘One key steelmaking’ farewell experience steelmaking
‘Before using the converter ‘one-key steelmaking’, the gun operator needs to pay full attention to the flame at the mouth of the converter, and adjust the position and flow of the oxygen gun according to the slagging situation, and control the timing and weight of the auxiliary materials, which not only tests the operator's knowledge, experience, attention, and responsiveness, but also easily causes vision fatigue.’ He Chunlai is China Unicom sent to Yongyang Special Steel artificial intelligence model experts, and what he said ‘one key steelmaking’ is ‘AI + converter wisdom steelmaking model’.
What is ‘AI + converter wisdom steelmaking model’? He Chunlai introduced that the big model mainly applies the metallurgical mechanism deep fusion AI multimodal fusion big model technology, adopts ‘big model decision-making + deep learning’ and metallurgical mechanism computing architecture, developed the main raw material model of the converter, the secondary charging model, the optimisation of slagging model, the dynamic oxygen lance model, smelting state model, slag state calculation model, flame analysis model, flue gas analysis model, the slag state calculation model, the main raw material model of the converter. model, flame analysis model, flue gas analysis model, sub-lance model, end-point prediction model, alloying model and self-learning model.
Look at the steelmaking application scenarios: dynamic monitoring of the converter steelmaking production process, production guidance, one-key steelmaking and intelligent steelmaking can be achieved, which can reach the adaptive calculation of raw material fluctuations, the whole dynamic calculation adjustment, optimisation of slag making and de-P, the prediction of the intervention of sputtering back to dryness, increase the metal yield, enhance the three-hit rate of each steel type, shorten the smelting cycle, reduce the use of sub-lance, and, through self-awareness, self-learning and self-optimization, achieve cost reduction and efficiency increase, improve quality, and improve the quality of steel. Through self-awareness, self-learning and self-optimisation, it achieves the purpose of cost reduction and efficiency enhancement, quality improvement and production enhancement. Steel material consumption can be reduced by 2-3 kg per tonne, and the cost of tonne of steel can be reduced by more than 5 yuan.
See the operation of the application scene: through the machine vision, auditory feedback, intelligent steelmaking system automatically control the oxygen gun, silo and other charging equipment, the operator only needs to be on the side of the monitoring, greatly reducing the operation action, converter smelting operation has become more ‘leisurely’.
Through the large model of learning, Yongyang Special Steel converter steelmaking will soon say goodbye to the era of relying on artificial experience in steelmaking. ‘The traditional converter steelmaking is very dependent on the experience of the operator, the furnace and how to operate the situation to rely on manual judgement, which requires years of experience accumulated by the operator, and after the shift, everyone's judgement is not the same, which can easily lead to the smelting control is not stable, affecting the quality of steel.’
He Chunlai said that through the learning and thinking of the big model, the artificial experience can be transformed into a stable operation standard, and every step of the big model of converter steelmaking can provide corresponding guidance, which greatly liberates the manpower, and further improves the stability of the converter steelmaking, and solves the problems of low hit rate of converter steelmaking.
At present, ‘AI + converter intelligence steelmaking model’ has been in Yongyang Special Steel training for six months. ‘The application of ‘one key steelmaking’ is in a leading position in the iron and steel industry in Hebei Province, attracting many peers to come to learn.’ He Chunlai, ‘In the future, we will continue to optimise the big model, through continuous learning and accumulation, so that the “AI brain” is more intelligent.’
Digital control to make manufacturing into ‘intelligent manufacturing’
Enterprise production, supply and marketing, the whole process overview, this is the Yongyang Special Steel from 2023 has been building the ‘intelligent brain’ - digital intelligent control centre.
Push open the door of Yongyang Special Steel control centre, only to see the room surrounded by screens, from the procurement of raw materials, to the steelmaking process, and then to the sale of products, enterprise production, supply and marketing data in real time jumped on the big screen. In the middle of the room, two staff members are closely monitoring the data to ensure the normal operation of all aspects of the enterprise.
‘Yongyang Special Steel was founded, has been committed to enterprise information technology and digital, through the local information technology, business and financial integration and digital transformation of three stages. The current digital intelligent control centre is a typical example of our digital transformation.’ Yongyang Special Steel information section chief Xi Caihong said.
According to the introduction, the digital intelligent control centre includes production, supply and marketing model, energy model, equipment model. Among them, the production, supply and marketing model integrates the existing system of Yongyang Special Steel to achieve the enterprise's full data carrying capacity, and provide data extraction, cleaning, transformation, standards, service provision and other capabilities. Take the production execution system as an example, to achieve the integration of production, supply and marketing, quality consistency, control integration, process synergy, refined cost, digital support, real-time and accurate raw material inventory, raw material quality inspection, efficiency improvement, support for procurement and settlement, and achieve the closed loop of procurement business.
‘The application of the production, supply and marketing of large models can effectively reduce the intensity of personnel work, improve the efficiency of data collection and costing efficiency, which in turn improves the efficiency of data flow and process standardisation, so that the quality of the quality traceability and quality analysis means to be enhanced.’ Xi Caihong introduced.
Iron and steel industry is energy-intensive, improve energy utilisation is a key factor in the sustainable development of steel enterprises.
Yongyang Special Steel energy model to establish a large energy flow network scheduling model, the main production process model describes the use of energy and recycling, submedia model describes the energy generation, conversion and transmission and distribution of energy, energy flow network integration model will be energy use, recycling, conversion, transmission and distribution of the three links in the energy information associated with the formation of a variety of energy medium between the regulation. After system analysis and processing, the information needed for the energy balance and its prediction model is obtained, and the balance prediction results are presented in data and graphical form. Through the dynamic monitoring and management of energy production, transmission and distribution and consumption, the energy balance is optimised to achieve a 2% reduction in the integrated energy consumption of tonnes of steel, forming a 100,000 tonnes of standard coal energy-saving capacity.
Xi Caihong said that the application of artificial intelligence model in the steel enterprise, is reshaping the development pattern of the steel industry, promote procurement, production, sales and other intelligent transformation of the whole chain, and gradually achieve cost reduction and efficiency and sustainable development, promote the steel industry from manufacturing to ‘smart’ leapfrog development.
Now, with the rise of DeepSeek in China, Yongyang Special Steel has also DeepSeek access to enterprise production management.
‘Now, we have completed the construction of the enterprise knowledge base in DeepSeek.’ He Chunlai introduced that the knowledge base contains local production protocols, and staff can conveniently check production and management standards with just one command, greatly reducing the retrieval process and time. In addition, real-time data on production is also being built in DeepSeek. ‘When the time comes, workers will be able to search the data and reports generated in production in their own language, avoiding the difficult-to-learn and understand code search, and they can also dialogue with DeepSeek, greatly improving production efficiency and allowing AI to empower steel companies with greater energy.’
Pathfinding for intelligent upgrading of industrial enterprises
Since last year, Handan City adheres to the combination of government guidance and market-led, and actively promotes the innovative application of artificial intelligence big model in the industrial field, so far, the city has 31 enterprises in the production optimisation, quality control, equipment maintenance, safety supervision, energy management and other areas of the application of artificial intelligence big model, the application of 66 application scenarios (including sub-application scenarios 238).
‘Yongyang Special Steel is one of the first batch of advanced-grade intelligent factories in Handan City, and the enterprise takes data-driven as the core, comprehensively reshapes the production business process, and creates an innovative model of digital transformation in the iron and steel industry, which provides the city and even the whole province's iron and steel enterprises with a highly valuable reference and example.’ Qiu Huimin, Secretary of the Party Group and Director of the Municipal Bureau of Industry and Information Technology, commented.
Qiu Huimin said that the next step, Handan City will promote the application of artificial intelligence in four aspects of the big model - deepen the ‘wisdom to digital transformation’, 500 enterprises to provide digital transformation diagnostic services, to diagnosis to promote the turn, to point to bring the surface, to build a solid foundation for the application of artificial intelligence innovation. Create benchmarking demonstrations, create 100 application scenarios in the industrial field, and lead the innovation of large model applications with benchmarking demonstrations. Strengthen the linkage between government and enterprises, and hold monthly AI application presentations.
Handan News Media Centre Reporter Su Sheng