Although launched less than a year ago, China's first AI large model for the diagnosis, operation and maintenance of industrial equipment has already been widely adopted in the coal, chemical and power industries.

Developed by CHN Energy Digital Inteltech, the large model has built a comprehensive and smart knowledge base with the function of AI expert consulting. Equipment operation and maintenance personnel can diagnose faults and learn maintenance methods by doing Q&A with the large model based on the real faults the equipment encounters.

"The development of the coal industry highly depends on all kinds of mechanical equipment; thus stable, reliable and smart systems for the operation and maintenance of industrial equipment are indispensable," Guan Feng, product R&D technology manager at Digital Inteltech R&D Center, said.

During the research phase, the company identified several sore points. One of them came from the workers at the frontline of the coal industry. They said daily operation and maintenance work was difficult because there were numerous equipment models and types and the equipment mechanism and structure were complicated, which made digital transformation hard to achieve.

Another was that the mature models for equipment diagnosis abroad are mostly small ones that cannot comprehend the technical language of industrial equipment, resulting in high operation and maintenance costs and low efficiency. Domestic models were still in the early stage of technological innovation.

The large model solved the dilemma of low efficiency and high costs. It is like an "intelligent hospital" for industrial equipment, as it can give accurate diagnoses and provide targeted "prescriptions" by offering professional maintenance decisions, according to Duan Wei, deputy director of the Lijiahao Coal Mine, owned by CHN ENERGY Baotou Energy Co., Ltd.

When the large model detects something amiss in the bearing at the drive end of an operating belt conveyor, it sets off a smart alarm immediately, reminding staff to strengthen routing inspection and provide solutions in accordance with the different causes like abnormal vibration of bearings or overly high temperature of the equipment.

It is calculated that the large model can effectively reduce 10 percent of the time spent on the operation and maintenance of industrial equipment, enhance the accuracy of fault location by more than 30 percent, and save about 30 percent of human resource cost for industrial enterprises.

Gao Yanchao, vice general manager of Digital Inteltech, said, "We will continuously push forward the integration of AI technologies and industrial application, and driving the upgrade and iteration of the large model."