The National Energy Administration recently issued the Notice on Carrying out Pilot Work for Technology Upgrading and Application of Coal Mine Intelligence, proposing to iteratively improve the operational level of coal mine intelligent construction. Among the key areas, the intelligent coal preparation system became one of the five priorities. Since the release of the Guiding Opinions on Accelerating the Development of Coal Mine Intelligence by eight ministries and commissions in early 2020, the construction of intelligent coal mines in China has been continuously accelerating and improving in quality.

As an indispensable part of the coal processing and utilization process, intelligent coal preparation has attracted significant attention and achieved certain progress.

To date, the National Energy Administration and the National Mine Safety Administration have announced the first batch of 66 national intelligent demonstration coal mines (including 52 coal preparation plants). Among these, there are 45 intermediate-level intelligent coal preparation plants and 7 primary-level intelligent coal preparation plants.

“Promoting intelligent coal preparation is an inevitable choice for the safe, green, and clean development of the coal industry, and it represents a technological revolution,” said Ma Jian, President of the China Coal Processing and Utilization Association.

“China’s coal washing technology and equipment have reached internationally advanced levels, but the journey of intelligent construction is still long. Currently, the intelligent construction of coal preparation in China is still in its infancy,” said Zhou Juanhua, Deputy Director of the Coal Industry Intelligent Coal Preparation Engineering Research Center.

Intelligent Technology Has Yet to Achieve Deep Empowerment

According to Zhou Juanhua, research on intelligent coal preparation began to increase significantly in 2019, covering aspects such as heavy medium, flotation, dry separation, coal preparation plants, design, equipment, detection, and application.

Currently, research on intelligent coal preparation in universities focuses on the practical needs of industry development, without delving into basic theoretical research or the evaluation system for coal preparation effectiveness.

“Due to bottlenecks in basic research, process parameter settings rely on empirical judgment. There is a lack of research on coal preparation process control models driven by the fusion of mechanisms and data. Intelligent technology has not yet achieved deep empowerment,” Zhou said.

Zhou noted that alongside the lack of basic research, the intelligent construction of coal preparation also faces challenges on the technical level.

On one hand, with the rapid development of intelligent coal preparation plants, numerous technology companies have entered the market, generally implementing projects through technology transplantation, lacking adaptability analysis for the production scenarios of coal preparation plants. The depth of application for transplanted technology is insufficient. Equipment condition detection sensors and data platforms are typical examples of transplanted technology. While this technology can support coal preparation plants in achieving data trend analysis and data anomaly alerts, it rarely involves advanced functions such as equipment fault prediction and predictive maintenance.

On the other hand, core intelligent technologies for key process stages and key equipment instrumentation still require breakthroughs. The absence or insufficient precision of key process detection instruments for coal preparation is a bottleneck in the intelligent construction of coal preparation plants.

“We need to increase R&D investment and strive to break through core technological bottlenecks,” Ma Jian stated. He emphasized the need to pay close attention to core technologies such as high-precision coal quality analysis and large coal preparation models.

How can high-precision coal quality analysis be achieved? Ma Jian stated that it is necessary to achieve the localization and large-scale production of intelligent sorting equipment, improving the reliability of domestically produced intelligent sorting equipment. For example, vigorously develop online detection and equipment control instruments, meters, and computer software management systems with high precision and good reliability; develop stable and reliable detection technologies, and develop control models for processes such as gravity separation, flotation, coarse slime separation, and slime water treatment, thereby continuously improving control accuracy.

“We must focus on building a full-chain, high-precision coal quality analysis system from raw coal to product. At the same time, accelerate the construction of a high-level comprehensive management platform covering production, safety, equipment, and operations, promoting the real-time aggregation, unified scheduling, and collaborative optimization of multi-factor data, achieving closed-loop management of ‘perception-analysis-decision-execution,’ to make production scheduling more efficient and resource allocation more precise,” Ma said.

How should large coal preparation models be built? Kuang Yali, a professor at the School of Chemical Engineering and Technology, China University of Mining and Technology, stated that massive amounts of high-quality data and corpora are key factors for promoting the deployment of large models. Currently, the construction of large models in the industry is generally in the stage of language model application. There is still a long way to go to achieve feedback control and intelligent decision-making.

“In the future, we should solidly promote the overall data governance of enterprises, establish massive databases, and achieve the integration of large models and small models (mechanism models),” Kuang said. “Making good use of small models is our key to establishing massive databases.”

“By learning massive amounts of production data, dynamically optimizing separation parameters, predicting equipment performance degradation, and accurately matching market demand, large models will propel coal preparation technology from ‘experience-dependent’ to a new stage of ‘data-driven’ intelligence,” Ma said.

In response to technical challenges such as high-precision analysis and large coal preparation models, relevant enterprises have already begun research and development efforts.

“The core of the economic benefits of a coal preparation plant lies in maximizing the clean coal yield, and high-quality online sampling, detection, and analysis technology is a prerequisite for achieving intelligent coal preparation,” said Bao Liangqing, Chairman of Liaoning Dongfang Measurement & Control Group.

According to Song Qingfeng, R&D Director of Intelligent Instruments at Liaoning Dongfang Measurement & Control Group, the group uses technologies such as neutron activation, X-ray fluorescence, and ultrasound to achieve precise online detection and analysis of coal ash content, sulfur content, and slurry concentration, particle size, and flow rate data. The heavy medium separation project that Liaoning Dongfang Measurement & Control Group built at the Shaqu Coal Preparation Plant of Huajin Coking Coal Company, part of Shanxi Coking Coal Group, uses a neutron activation high-precision coal quality analyzer as the core to help stabilize production.

“We have built a multi-modal large model coal preparation expert system called ‘Coal Xiaohui.’ Using a localized large model, we integrate expert knowledge graphs, management and control platform data, analysis and decision-making small models, and event linkage interfaces to create an AI expert knowledge system, forming an intelligent human-machine interaction model,” said Chen Xiaoxia, Deputy Director of the Washing and Processing Center of Shandong Energy Zaozhuang Mining Group.

“We are building a research and development system for large-scale dry separation equipment, focusing on solving the technical challenges of dry separation, developing highly reliable equipment, pursuing more precise separation results, and laying a solid data foundation for intelligent decision-making,” said Yang Ruifeng, Director of the Intelligent Coal Preparation Research Institute and Deputy Dean of the Coal Preparation Engineering Institute at China Coal Tianjin Design and Engineering Co., Ltd. “Next, we will pay more attention to integration and openness, optimize relevant parameters in real-time, deepen the application of intelligent video, and build a closed-loop management system of ‘perception-analysis-decision-execution’.”

Management and Technology Need to Progress in Sync

“Some people think that intelligence is like a universal key that can solve all problems that arise in an enterprise. This idea is unrealistic,” Zhou Juanhua said. “In the process of implementing intelligent technology in coal preparation plants, many on-site management issues have already been exposed.”

Coal preparation plants with a relatively stable source of raw coal often no longer focus on improving their technical management level and lack lean management. Some enterprises control the number of employees, leading to insufficient coal preparation technical inspection personnel, making it difficult to achieve comprehensive process technology management. Some enterprises’ coal preparation plants are secondary units of mines without independent financial and property rights, resulting in relatively rough daily production management, and basic equipment cannot meet the needs of intelligent production management.

“Due to the lack of stable, advanced production equipment, enterprises have to invest substantial funds in upgrading and renovating production equipment, pipelines, non-standard components, and basic automation electrical equipment during the intelligent upgrade and transformation process of coal preparation plants. The funds actually invested in intelligent technology transformation are extremely limited,” Zhou said.

“We need to lay a solid foundation and fully utilize intelligent technology on the basis of lean management,” Yang Ruifeng said. “Intelligent coal preparation is a profound systemic upgrade. Its essence is data-driven, process-centric, value-oriented, and with safety and environmental protection as the bottom line. This means we not only need to catch up in technology but also undergo profound changes in management concepts, organizational methods, and talent teams.”

Zhou Juanhua admitted that currently, some managers have insufficient understanding of intelligence, affecting decisions related to intelligent coal preparation. Some managers believe that their coal preparation plants have already reached an intelligent level, leading to low-level repeated construction and investment. Others are taking a wait-and-see attitude towards intelligent coal preparation.

“As a production entity, a coal preparation plant can only identify problems in a timely manner, find the causes of problems, formulate improvement plans, and achieve intelligent decision-making if it can achieve real-time analysis of production results,” Zhou said. “The imperfect evaluation system for separation effectiveness greatly restricts intelligent decision-making in coal preparation.”

Ma Jian said: “Currently, there are still issues such as non-uniform standards and imperfect evaluation systems in the intelligent construction of coal mining and preparation in China, which constrain the overall advancement of intelligent mining and preparation.”

The Guidelines for the Construction of Coal Mine Intelligent Standardization Systems released by the National Energy Administration last year proposed that by 2025, a well-structured, clearly layered, clearly categorized, scientifically open coal mine intelligent standardization system should be initially established to meet the basic needs of coal mine intelligent construction. By 2030, the coal mine intelligent standardization system should be basically perfected, with a relatively complete series of standards formed in areas such as intelligent coal mine design, shaft construction, production, management, operation and maintenance, and evaluation.

In the Guidelines for the Construction of Coal Mine Intelligent Standardization Systems, standards for intelligent washing systems and equipment mainly include intelligent production control, intelligent coal quality detection, intelligent production support, intelligent production processes, intelligent washing and screening equipment, and intelligent storage, loading, and transportation.

As the pace of coal mine intelligent construction accelerates, the work of setting standards for intelligent coal preparation has also been put on the agenda, with some provinces beginning active exploration.

In July this year, the Shanxi Provincial Energy Administration, the Shanxi Provincial Emergency Management Department, and the Shanxi Mine Safety Supervision Bureau of the National Mine Safety Administration issued the Measures for the Evaluation and Management of Coal Mine Intelligent Construction, specifying the relevant conditions for applying for the evaluation of intelligent coal preparation plants. The measures will come into effect on September 1.

“Recently, we initiated preparations for the 2026 intelligent coal preparation standard-setting work. We will formulate technical specifications for intelligent coarse slime separation construction and technical specifications for intelligent coal preparation management and decision-making construction, adding technical specifications for intelligent collaborative control of the entire separation process,” said Wang Ranfeng, Director of the Department of Coal Mine Electromechanical Engineering, College of Mining Engineering, Taiyuan University of Technology.

Kuang Yali said that after three years of refinement from 2022 to 2025, China University of Mining and Technology submitted draft group standards for technical specifications for data interfaces in intelligent coal preparation plant construction, technical specifications for standardization of control data in intelligent coal preparation plant construction, and technical specifications for standardization of management data in intelligent coal preparation plant construction.

“These group standards come from extensive engineering practice and are currently supported by relatively mature software, enabling direct implementation,” Kuang said. This year, China University of Mining and Technology has also initiated two industry standards: the Guidelines for Data Governance in Coal Preparation Plants and the Technical Guidelines for Standardization of Design Data in Coal Preparation Plants.

Currently, the first batch of national intelligent coal preparation plants has set benchmarks for the intelligent construction of coal preparation plants. These demonstration projects cover processes such as heavy medium separation, flotation, and dry separation, and will form replicable and scalable construction models, driving the overall upgrade of the coal preparation industry through pilot projects.

Full-Process Intelligent Collaboration Is a Key Development Direction

“Rather than answering the question ‘What does an intelligent coal preparation plant look like?’, I would first like to talk about what an intelligent coal preparation plant is not,” Zhou Juanhua said, listing several points.

The goal of building an intelligent coal preparation plant is not simply to reduce or eliminate personnel, but to improve efficiency. Intelligent construction requires technological accumulation and continuous improvement. In the initial stage, it may lead to a reduction in on-site operators; in the long term, it will lead to an increase in high-quality, highly skilled professionals.

Algorithm models are not a panacea. Regarding model errors, the general consensus in the industry is: model error in the mechanical industry is 1%, in the electronics industry is 10%, in the chemical industry is 30%, and in the metallurgical industry is 50% to 100%. Coal preparation belongs to the metallurgical industry. The parameters actually used by coal preparation enterprises are often not calculated by models but derived through practical exploration. Therefore, the algorithm models of coal preparation enterprises are highly tailored to each plant.

Intelligent construction cannot replace basic management. Process and management issues in coal preparation should be addressed before intelligent construction. There is a saying in the coal preparation community: do not automate backward processes, do not digitize backward management, and do not pursue intelligence without a foundation of digitalization and networking. This reflects the importance of basic management.

An intelligent coal preparation plant cannot be built in a single attempt. Intelligent construction requires technological accumulation and continuous improvement. Data modeling, industrial app development, and other digital methods support continuous improvement in enterprises, requiring long-term persistence and ongoing refinement throughout the enterprise lifecycle.

In May last year, the National Energy Administration issued the Notice on Further Accelerating the Intelligent Construction of Coal Mines to Promote High-Quality Coal Development, specifying three key construction elements for intelligent coal preparation plants: high-precision online coal ash content detection, intelligent sorting control, and full-process intelligent monitoring, decision-making, and control.

In April this year, the Expert Committee on Mine Intelligent Construction of the National Mine Safety Administration compiled and released the Blue Book on Coal Mine Intelligence Development (2025), proposing full-process intelligent collaboration in coal preparation plants.

“In the future, full-process intelligent collaboration in coal preparation plants is a key development direction, and general artificial intelligence has great potential,” Wang Ranfeng said. He suggested constructing a full-process data standardization and collaborative optimization system, breaking through bottlenecks in intelligent perception and adaptive control technology, overcoming core technologies for process dynamic control and self-optimization, enhancing intelligent decision-making and full-process closed-loop optimization capabilities, and promoting the leap from single-point intelligence to full-process self-optimization in coal preparation plants.

“Intelligent coal preparation construction is a systematic project,” said Zhang Jianming, Vice Chairman of the China Coal Society. In the next stage, we should focus on overall planning and systematically promote intelligent coal preparation construction.

“In promoting intelligent coal preparation construction, enterprises are the main players and should fully play their main role,” Ma Jian said. For coal enterprises, it is necessary to proactively increase investment in funds, technology, and talent, establish a collaborative innovation mechanism involving industry, academia, research, and application, and jointly tackle technical challenges with research institutions and universities. At the same time, focus on cultivating intelligent talent by creating a team of versatile professionals who understand both coal mining and preparation processes and are proficient in information technology through school-enterprise cooperation and internal training.

“The operation of an intelligent coal preparation plant requires a team with specialized skills and knowledge. Currently, the talent pool in the coal industry is relatively insufficient, especially lacking professionals with both intelligent technology and management experience,” Zhou Juanhua said. This leads to talent shortages in some coal preparation plants during the intelligent transformation process, often resulting in intelligent systems “not being used.”

Ma Jian said that industry associations will continue to play a bridging role, further intensifying efforts in cultivating coal preparation professionals and updating the knowledge of engineering and technical personnel. They will leverage the National Open University to add coal preparation majors and encourage outstanding front-line technical workers to participate in learning and training programs.

(Source: China Coal News)