Precision Coal Mining: Transforming the Mining Sector with Cutting-edge AI Technologies

Yuan Liang, Academician of Chinese Academy of Engineering, President of Anhui University of Science & Technology

Coal will remain the main energy source in China for the foreseeable future due to the country's resource reserves and economic development needs. China's National Bureau of Statistics has reported steady growth in annual coal production in recent years. In 2024, China produced approximately 4.78 billion tons of coal, with this resource accounting for 53.2% of the nation's total energy consumed.

Nonstop advancements in digital and AI technology are laying the groundwork for AI-assisted precision coal mining, which is vital to China's energy security and worker safety.

In 2020, China released the Guiding Opinions on Accelerating the Development of Intelligent Coal Mines. Since then, coal mines in the country have made major strides in intelligent transformation, including 66 mines at the national level and more than 200 at the provincial level. These mines are applying cutting-edges technologies for geological transparency assurance, intelligent excavation technology and equipment, and intelligent disaster identification.

Key tech #1: Geological transparency assurance – generating high-precision geological information and enabling visibility and prediction of disaster factors

In the mining industry, ensuring the transparency of geological conditions is essential for coal resource security and intelligent mining. We need advanced technologies for high-precision geological exploration, visualization of geological information, and real-time updates and early warnings for hidden disaster-causing factors.

Several mines have improved their intelligent recovery and disaster warning capabilities for their mining faces (see Figure 1). They have achieved this by integrating different types of data (e.g., radar, in-seam wave, and borehole exploration data) from different physical fields. These data inform high-precision geological models for mining faces. In addition, these mines have implemented new process-based approaches that combine geological transparency assurance, excavation, production, and disaster prevention.

Nevertheless, unmanned mining is still hindered by the inadequate precision, reliability, and speed of conventional exploration technologies. These days, mining companies look to conduct exploration in parallel with excavation (also known as simultaneous exploration and excavation) on their fully mechanized mining faces and driving faces. There are two types of technologies for this job: conventional technology and geological transparency technology. Conventional technology lags far behind in data collection speed, data integration efficiency, and exploration costs.

• The conventional approach to simultaneous exploration and excavation is to have human workers operate drilling machines and collect samples. Single-hole drilling often takes weeks, which makes data collection slow.
• Integrating the data is another challenge for two reasons. First, there is a lack of 3Dmodels that support data integration. Second, data is siloed due to different data formats.
• Conventional technology cannot provide precise predictions for underground geological structures, so engineers have to verify their predictions through extensive drilling. As much as 30% of these drilling operations produce no results. The average annual cost of exploration at a mine may exceed CNY10 million.

There is an urgent industry need for advanced exploration technologies, especially high-resolution seismic exploration, computed tomography (CT) monitoring, 3D laser scanning, and transient electromagnetics. We also need to develop intelligent drilling machines that can work in complex geological conditions. These technologies and gear will allow us to better integrate dynamic monitoring information from multiple physical fields, and create digital twin models that encompass the geological information of entire mines and their different parts. This will give us the precise and transparent geological information needed for intelligent excavation and disaster warning.

Geological transparency technology is a combination of 5G communications, 3D geological space, and intelligent drilling technologies. It's the ideal solution for the real-time needs of simultaneous exploration and excavation, delivering huge gains. First, the technology leads to at least 30% higher data collection efficiency. This improvement relies on integrating data from different sources (e.g., geophysical exploration systems, drilling systems, and IoT sensors), and transmitting the data in real time over a 5G network. Second, 3D geological space technology is used to create underground structure models that help us precisely predict the formation of coal-bearing rocks and the distribution of hidden disasters. Precise predictions, coupled with intelligent algorithms, enable us to optimize drilling plans to minimize futile drilling operations.

Key tech #2: Intelligent excavation technologies for greater machine adaptability and efficiency in complex geological conditions

In China, excavation machines are rapidly becoming smarter with the help of technologies for intelligent sensing, fast data transmission, intelligent decision making, and automation. These machines are now in use in more than 1,800 intelligent excavation faces of coal mines across the country.

These machines are backed by a state-of-the-art technical architecture that employs 5G networking and 3D digital twins for fully mechanized mining faces. The results are impressive: Shearers can operate on their own; hydraulic support systems can move automatically with the shearers; tunneling machines can work efficiently in different coal seam conditions; and vehicles like trackless rubber-tired vehicles, monorail cranes, and mining electric locomotives can drive on their own. The efficiency of the entire working face is improved significantly (see Figure 2). However, these machines are not without problems. These include a high failure rate of underground sensors and controllers, weak adaptability of machines to complex mining conditions, and inadequate system efficiency in day-to-day use.

That's why our future R&D initiatives should focus on enhancing the machines' environmental adaptability and reliability. We will need key technologies for high-precision geological exploration, high-precision and dynamic identification of coal-bearing rocks, intelligent sensing and warning of hazard sources, multi-purpose underground robots, and integrated, intelligent systems for high-efficiency excavation, support, and transportation.

Simultaneous exploration and excavation technology offers a safety net for underground workers, enabling an at least 70% reduction in accidents. In particular, we can perform dynamic exploration based on real-time collection of data from diverse sources as well as dynamic data modeling and presentation. Then, we can create holograms of underground spaces with 3D geological transparency technology. These holograms are essential for precise prediction of geological hazards, such as coal seam faults, water-conducting fractures, and abnormal stress zones. With precise prediction and data visualization, our excavation and recovery engineers can accurately determine the spatial paths for underground work and make sure their work plans will steer clear of hazards.

Key tech #3: Intelligent disaster identification – transitioning from reactive risk response to proactive warning

AI is changing the way we prevent mining disasters: We are shifting an empirical, qualitative approach to one that is more precise and quantitative. AI models can play a significant role in disaster prediction and analysis, especially for disasters involving gases, water, and roof pressure. If we combine these predictions with geospatial information from our geographic information systems (GIS), we can obtain a transparent view of underground geological environments, and generate precise and visualized space data for disaster prevention.

Conventional approaches for disaster prevention are inadequate in two critical aspects.

First, without systematic data at hand, workers have to perform inspections underground and make decisions empirically. So there may be a high chance of post-event disaster handling (e.g., gas outbursts and water disasters), serious delays in risk warnings, and a high likelihood of accidents.

Second, conventional systems are only precise up to 30 meters, which makes them inadequate for identifying hidden disaster-causing structures like minor faults and water-conducting fissures. As a result, we cannot identify risks with precision and take prompt measures for disaster prevention.

Geological transparency technology is a game changer. It allows us to transition from reactive risk response to early prevention. The system monitors geological changes in real time, while AI algorithms use the data for risk prediction, diagnosis, and early warning. This enables sub-meter-level precision for locating hidden disaster-causing structures, which significantly reduces accidents.

In particular, we can more precisely identify hidden disaster-causing structures with advanced technologies, like seismic exploration in sync with excavation as well as ground-penetrating radar. The resulting sub-meter-level identification precision is a massive improvement from the meter-level precision of conventional approaches. Risks can now be identified and mitigated at the source.

Our intelligent disaster prevention platform (see Figure 3) delivers major benefits: real-time monitoring of multiple disaster types, coordinated disaster response, and high-precision prediction and identification of multi-hazard disasters. To develop this platform, we have integrated heterogenous data generated by different underground sensors, and created a holistic system that covers every step of disaster prevention, from sensing and analytics to early warning and decision making.

However, the development of this platform is just the first step. Further R&D is needed to address unresolved challenges. Typical dynamic disasters, for instance, are nonlinear, complex problems and involve multi-phase and multi-physics interaction. It is still unclear how these disasters emerge in the first place and how they evolve over time.

Moving forward, mining companies need to leverage geological transparency assurance systems to strengthen mining disaster prevention capabilities. Specifically, they need to:

• Develop intelligent technologies for sensing and preventing underground multi-hazard disasters in complex geological conditions (featuring physical fields like stress, fracture, seepage, and temperature fields).
• Research and develop disaster prevention machines with high reliability and self-adaptability.
• Build fusion sensing technologies that collect heterogenous data from diverse sources and encompass all underground scenarios, and design a multi-network converged data transmission method.
• Implement multi-source information tiering and holistic monitoring.
• Develop a multimodal intelligent monitoring and warning system based on the cloud.

The road ahead: From labor-intensive to technology-intensive mining

Coal security and AI-assisted precision mining are high priorities in China's national energy strategy. To achieve these targets, we need to build a modern mining system that attaches great importance to data, AI, and standardization. We need to transform the entire industry value chain with digital technology, that covers all areas of mining, from geological exploration and equipment R&D to production and control.

The goal is to make coal mining less labor-intensive and more technology-intensive. This will allow China to ensure energy security and attain its dual carbon goal of peak CO2 emissions by 2030 and carbon neutrality by 2060. To recap, we need to focus our efforts on the following areas:

(1) Developing key technologies to achieve high-precision 3D geological dynamic modeling, integrate heterogeneous data from different sources, and standardize industrial Internet protocols; and making tech breakthroughs for millimeter-level coal-bearing rock identification, multi-physical-field dynamic sensing, and intelligent collaboration of equipment clusters.

(2) Creating a national standards system covering exploration design, intelligent mining, and ecological restoration; formulating specifications for data interfaces of intelligent equipment; and building a national mine data center.

(3) Cultivating a prosperous ecosystem for joint innovation between businesses, academics, researchers, and end users; removing barriers that restrict access to the data of businesses, equipment vendors, and scientific research institutes; and building up a talent support system and encouraging all players (universities, scientific research institutes, and businesses) to foster multidisciplinary innovators.

(4) Promoting exemplary intelligent mining projects that use 5G + Industrial Internet, and using intelligent operation and maintenance modes for equipment throughout their lifecycles, in order to systematically reduce the energy consumption per ton of coal produced.