Promoting High-Quality Development in the Energy Industry Through Advanced AI Applications
-
The real economy is the foundation of national economic strength, the source of wealth creation, and a crucial pillar of national prosperity. Coal, as the primary energy source, serves as the 'ballast' and 'stabilizer' for national economic development, ensuring energy security. Building a modern coal industry is a key component of establishing a modern industrial system and strengthening the real economy.
As a leading energy enterprise, Shandong Energy Group adheres to the mission of 'creating green momentum and leading energy transformation,' practicing the core values of 'safety, innovation, green, responsibility, and excellence.' Focusing on the vision of 'building a clean energy supplier and a world-class enterprise,' the group is deeply advancing the energy technology revolution, striving to create a modern diversified energy supply system. In the field of intelligent mine construction, the group has implemented the national 'Guidelines for Accelerating the Intelligent Development of Coal Mines,' investing over 10 billion yuan in joint innovation efforts. It has established a number of national intelligent demonstration coal mines, with intelligent mining output exceeding 80%. The group has also created innovation platforms such as the 5G+ Smart Mining Joint Laboratory and the Intelligent Mining Test Center, collaborating to develop the world's first high-reliability mine-specific 5G network system, pioneering a new path for intelligent mine construction.
Shandong Energy Group's Practice in Mining AI Large Models
In early 2022, Shandong Energy Group and Huawei established a joint innovation center, leveraging their respective industrial and technological advantages to form a comprehensive strategic partnership. Deeply integrating industrial scenarios with digital technology, they focused on the goal of 'fewer or no people' in coal mine intelligence, launching 20 joint innovation projects to explore digital technology applications in the coal industry. One key project was the joint innovation of mining AI large models.
The group released the mining industry's first AI large model, using computing power as the foundation and large models as tools, driven by scenarios to achieve deep integration, making AI a new labor force for Shandong Energy Group. The computing power center, built with Ascend-based AI computing power and a hardware-software integrated approach, provided the necessary computing resources for mining large model development. The development of mining large models is not a one-time effort; the first version achieved higher comprehensive benefits than smaller models. In 2023, frontline workers were deeply involved in the application of large models, identifying 21 scenarios in actual production. After six months of optimization, the group released the mining industry's first AI large model in July.
Empowering Intelligent Coal Mine Applications
Shandong Energy Group's rich application scenarios and digital conditions provided diverse data for mining large model training, while the models, in turn, empowered intelligent coal mine scenarios. Meanwhile, leveraging large model capabilities, the group internalized AI technology, creating the Yunding AI service platform to achieve closed-loop management from model training and inference to business applications. This platform serves nine industries, including mining, chemicals, steel, and power, helping enterprises enhance safety, reduce costs, and improve efficiency.
AI as the Inevitable Choice for Fewer or No People in the Coal Industry
Looking back, the coal industry's workforce has decreased from 5.25 million in late 2012 to about 2.6 million today, with a significant reduction in underground workers, thanks to the achievements of intelligent construction. Looking ahead, like other industries, the traditional coal sector is undergoing a transformation in production methods supported by new-generation information technology. As a high-risk industry, reducing or eliminating human presence in coal mines is an inevitable trend to protect miners' lives. Further breakthroughs in coal mine intelligent construction require new technologies like AI, with large models serving as the tool to integrate AI technology with industry scenarios.
This year, the rise of general large models like ChatGPT has injected new momentum into AI industry development. China's economic structure and industrial characteristics, rooted in the real economy, dictate that AI must follow a path with Chinese characteristics—focusing on 'practical application' to penetrate all sectors of the real economy, creating a landscape of large-scale development across industries and deep, scenario-specific refinement within individual industries.
Currently, AI adoption in the coal industry remains low due to three main reasons: high barriers to entry as an emerging technology requiring specialized talent, equipment, and environments that are difficult to deeply integrate with coal industry R&D and engineering; complex scenarios with significant variations in production conditions across different regions and companies, making intelligent applications costly and hard to replicate; and insufficient integration where the industry's large scale and diverse scenarios become challenges for intelligent transformation. Only through efficient integration of fragmented scenarios and capabilities, accumulating industry data, experience, and expertise for widespread circulation and sharing can AI technology accelerate its fusion with the industry, turning fragmentation disadvantages into scale advantages. Shandong Energy Group has developed an effective methodology combining its scenario advantages, equipment strengths with AI technology to accelerate large-scale AI deployment in underground mining.
Promoting Large-Scale AI Implementation and Innovation in Energy Industry Through "Four Transformations"
Engineering Transformation: Integrating AI with mining scenarios to drive industry development. AI must transition from theory to practice, from research innovation to engineering applications, moving from data centers to production sites. Only by combining with coal mine safety management and production scenarios to form value loops can it gain industry vitality. Based on the Pangu Mining Model, Shandong Energy Group initially focused on CV applications, developing over dozens of scenario algorithms covering belt transportation, vertical shaft lifting, and auxiliary transport within six months. Next steps include combining the model with mining and mechanical theories for predictive equipment maintenance and intelligent rock burst prevention.
Industrialization Transformation: Achieving "one-mine-one-policy" high-efficiency, low-cost self-learning capabilities for localized customization at scale. With 350 million tons annual capacity across varied geological conditions, Shandong Energy Group centralized dispersed mining expertise onto its platform through its AI computing center and Yunding Model, improving AI accuracy by 10% and training efficiency by 85%. Its subsidiary Yunding Technology has established electronic seal projects at 76 mines, forming preliminary AI industrialization capabilities.
Standardization Transformation: Providing continuous massive data flows to sustain AI vitality. During digital transformation, Shandong Energy processes 200 million data entries daily. By standardizing data architecture and specifications, it integrates OT, IT, and CT systems to enable unified multi-source data collection, enriching the mining model's capabilities.
Modularization Transformation: Addressing scenario fragmentation through industry collaboration. Avoiding outdated approaches, Shandong Energy built an integrated AI service platform using Ascend, MindSpore and large model technologies. It modularizes general and industry-specific algorithms while enabling cross-industry model sharing through streamlined workflows from data labeling to model deployment.
The industry generally regards artificial intelligence as a general-purpose technology similar to the steam engine. The invention of the steam engine did not immediately trigger the Industrial Revolution; rather, it was the development of a series of machine tools powered by steam engines that enabled its widespread use in textiles, mining, metallurgy, papermaking, transportation, and other fields, ultimately leading to the transformative Industrial Revolution of the 18th century. Similarly, the application of artificial intelligence across various sectors in China will bring revolutionary changes, much like the steam engine did. We must adhere to the principles of "engineering, industrialization, standardization, and modularization" to integrate AI with China's industrial characteristics and deepen its application in coal mining scenarios. By pursuing a development path for AI with Chinese characteristics, we will significantly advance the coal industry toward safety, intelligence, and high-quality growth, contributing industry wisdom and strength to the construction of a modern industrial system.