Figure 1 illustrates the two primary production models that are typically used in the steel industry: the long process of ironmaking and steelmaking procedure (LPISP) and the short process of steelmaking procedure (SPSP). The burden materials utilized in the LPISP are very different from those used in the SPSP. For instance, metallurgical coke, sinter, and pellets are used in the LPISP, while the SPSP uses scrap steel. In addition, the smelting furnace used in the steelmaking process differs depending on whether it is a converter or electric furnace (LPISP vs. SPSP).

Since 2015, China’s steel industry has begun to conduct small-scale intelligent transformation pilots to eliminate data islands, promote data sharing between processes, improve the quality control and traceability efficiency of the product production process, improve equipment fault diagnosis capabilities, improve energy efficiency, reduce logistics costs, reduce nonessential post staff, reduce the R&D cycle of new products, and reduce management. Table 1 shows the typical intelligent manufacturing projects used in China’s steel industry in recent years, which are mostly dispersed in the mine, rolling, and steelmaking and rolling processes, with the majority of them being carried out in the rolling procedure.

On the one hand, the physical changes in materials during the rolling process can be easily defined by control models and detection methods; and the level of automation, informatization, and modeling is higher than in the smelting operation. On the other hand, as the final step in the steel production process, the problems that occurred during the previous smelting and continuous casting procedures must be corrected and eliminated from the rolling link in a timely manner. Otherwise, the product quality and brand of the entire steel production line will suffer, causing significant harm to the enterprise. Decision-makers at steel businesses, logically, are also focused on intelligent manufacturing in the rolling process.

China developed the “Guidelines for the Construction of the National Intelligent Manufacturing Standard System”; researched the fundamental similarities, key technologies, key industry standards, and specifications of intelligent manufacturing; built a standard test verification platform; carried out technical specifications and standards throughout the entire process of test verification; and actively promoted all manufacturing-related fields. First, we discussed some fundamental rules and principles. Standard fundamentals include term definition, reference model, metadata, object ID registration, and parsing. Standards for management evaluation include the evaluation index system, measurement method, and implementation guide. Standards for information security included architecture, security, management, and evaluation. They also include adaptability to the environment, equipment dependability, and other criteria for quality. Second, we then discussed important technological norms and requirements., including equipment and product standards for industrial robots, industrial software, intelligent IoT devices, additive manufacturing, human–computer interface, and other technologies. We also discussed standards for network design, connectivity, and interoperability; convergence of fieldbus and industrial Ethernet; industrial sensor networks; industrial wireless systems; and industrial gateway communication protocols and interfaces. Standards for “smart factories” include network collaborative production, intelligent detection, intelligent logistics, and accurate supply-chain management. Standards for the industrial big data and cloud include cloud services, data analysis, and quality. Industrial processes use energy efficiency analysis software standards and service-oriented manufacturing standards, such as personalized customization and remote operation and maintenance services. Third, we discussed important industry rules and standards. Digital workshop integration applications in typical discrete industries and smart factory integration applications in process industries each represent one of ten important areas of industry standards and requirements.

The steel industry, as a typical process industry, has many production procedures, a large number of pieces of equipment, a large amount of energy consumption, a large number of operators, and a large number of complex reactions, including chemical reactions during the smelting procedure and physical changes during the rolling procedure. The interweaving of various influencing elements provides a huge barrier to intelligent manufacturing of the whole steel production process, further complicating the quality, energy, and product management procedures. Specifically, in the steel industry, there are quality data islands in various manufacturing processes, process adjustment is complex and difficult to achieve, the product development cycle is long, and product quality control cannot be accurate to the process. Furthermore, quality control and traceability are difficult to implement, and the smelting process is largely reliant on manual experience. Intelligent manufacturing, as a new generation of information and communication technology and advanced manufacturing technology, can quickly respond to internal and external changes in enterprises and establish a new model of personalized customization and resource optimization configuration centered on customer needs by improving the manufacturing industry’s intelligence level.

Therefore, intelligent manufacturing has resulted in a significant transformation in the steel industry’s organizational structure, namely a transition from a production organization to a service-oriented manufacturing organization. As a result, the steel industry’s goals for intelligent manufacturing in China may be stated as follows. Manufacturing enterprise operations will shift from responsive to customized, with optimization shifting from local to global. While management is focused on backward and in-the-moment pre-event transformation to achieve flexible production, equipment control has shifted from automation to intelligence. The following four factors are the primary goals for the completion of intelligent manufacturing in China’s steel industry.

First and foremost, boost productivity. On the one hand, breaking down data silos within a single process and between multiple processes involving ironmaking, steelmaking, continuous casting, and steel rolling, as well as leading the substance, energy, and information flow, could be efficiently utilized to reduce defective product rates and eliminate production efficiency losses caused by data islands and information gaps. Robots, on the other hand, are being used in large numbers to replace humans in repetitive, harsh environments and labor-intensive positions, gradually merging the main control position, improving personnel efficiency and intelligence level, and real-time online monitoring and fault prediction maintenance of key equipment to reduce equipment failure rate and unplanned downtime.

Second, boost product quality. The production process status monitoring system will cover many dimensions such as raw materials, equipment, process parameters, and laboratory data, effectively avoiding the influence of abnormal working circumstances on the quality of the steel manufacturing process. Then, by raising the intelligence level of the production line, the equipment control accuracy and equipment failure online monitoring and early warning system may effectively avoid the occurrence of recurring quality mishaps. Finally, a quality monitoring and control system was implemented across the ironmaking and steelmaking processes to eliminate worker interference on product quality and increase the precision and dependability of quality control. In addition, product quality optimization was implemented; and digitization, networking, and intelligence of product quality control methods were improved through the traceability of the whole steel product life cycle.

Third, cut production expenses. An energy medium optimization system and an automatic batching system are built to achieve the automatic entry of molten iron and scrap into the furnace, minimize material preparation response time, optimize smelting rhythm, and reduce energy consumption. Then, to shorten the smelting time and enhance smelting efficiency, information technologies such as automated detection of molten steel components, intelligent blowing control, and automatic component analysis were used. Finally, lowering manufacturing costs via the use of robots and intelligent maintenance decreases the rate of faulty goods and increases the rate of worker productivity.

Fourth, reduce the development cycle of new products. A virtual simulation system for the entire ironmaking and steelmaking process is constructed, and the number of experiments for new product development is reduced by replacing the traditional experimental trial-and-error method with a digital process simulation system, thereby shortening the R&D cycle and improving product development efficiency.

Above all, following the aforementioned goals for intelligent transformation, the typical steel industry’s product research and development cycle is typically reduced by about 30%, the rate of product defects is decreased by 10–20%, the energy used to produce a ton of steel is decreased by 10%, the production cost is decreased by 10%, and the production efficiency is increased by 10–25%. From the standpoint of financial investment revenue, intelligent manufacturing is one of the key avenues for the development of the steel industry in the future. The investment in intelligent manufacturing of steel production lines is expected to pay for itself in 3 to 5 years.

As illustrated in Figure 2, the levels of the framework for intelligent manufacturing in China’s steel industry were split into the device layer (DL), network layer (NL), big data layer (BDL), and application layer (AL).

DL: The equipment parameters produced in the primary system, secondary system’s control model parameters, and tertiary system’s quality control parameters should be gathered in a timely and accurate manner during the overall ironmaking and steelmaking processes; particularly those affecting production efficiency, cost, and quality, such as key raw material parameters, process parameters, and quality process parameters. Equipment condition monitoring, temperature detection, defect detection, composition online inspection, shape detection, and other critical process parameters should all be detected online. Thus, the DL might provide a vast amount of raw data, laying a robust data basis for future big data applications and intelligent manufacturing.

NL: Due to the steel process’s big-scale, multi-equipment, and long process features, the physical distance between each piece of equipment is rather long, and huge data in diverse formats must be transmitted in real time across a significant number of networks. Not only this might help to enhance the manufacturing process’s reaction time, but it also could help to support BDL and AL applications. Therefore, data transport is mostly accomplished using optical fiber transmission, industrial Wi-Fi, 5G CPE, and a 5G industrial gateway. Data capture and analysis are dependent on device access, protocol analysis, data processing, data integration, and edge data processing.

BDL: Similar to the function of the data middle platform, the main objective of the BDL is to supply data sources and various application components for the application layer, as with the preceding data provision, data acquisition, and preprocessing. Thus, the BDL for steel production processes includes industrial data modeling and analysis, as well as application and microservices components. The former primarily consists of mechanism, rolling mechanism, machine learning, and visualization, whereas the latter provides the algorithm, development tools, primary model, and operating knowledge. Furthermore, the goals of big data analytics in the steel industry are mostly focused on industrial data cleansing, heterogeneous data processing, data management, data analytics, and data visualization.

AL: The primary goal of the AL is to realize real-time process diagnosis and optimization of the steel production process using a large number of data such as procedure, equipment, quality, logistics, order, and scheduling. In order to effectively improve production efficiency and the quality of the steel product, the rate of defective products and production costs should be reduced. Normally, a typical framework for intelligent manufacturing includes systems such as the visualization of production status system, equipment status monitoring and fault diagnosis system, digital process simulation–optimization system, online monitoring and control of quality in the entire ironmaking and steelmaking system, intelligent logistics tracking system, order digitization and intelligent scheduling system, intelligent decision-making system, energy management and optimization system, and intelligent decision-making system.