Abstract: With the rapid development of China's industrial economy，the demand for strip materials is

increasing，and the construction of strip production lines both domestically and internationally is growing year by

year. However，the complexity of large-scale rolling mill systems and the harsh working environment make

equipment fault diagnosis increasingly challenging. Traditional fault diagnosis techniques lack intelligence and realtime capability. Therefore，real-time diagnosis based on equipment operating status，combined with remote

monitoring and fault diagnosis technology that provides timely feedback through human-machine interfaces，has

become a hot topic in the field of automated production lines. To address this challenge，a large-scale rolling mill

fault diagnosis expert system was developed. This system integrates BP neural networks，Bayesian networks，and

traditional expert systems. It employs a combination of manual and active knowledge acquisition methods to build

an expert knowledge base and establishes a comprehensive reasoning mechanism. By applying this fault diagnosis

model to the production line，real-time monitoring of the operating status of the large-scale rolling mill and diagnosis and prevention of faults can be achieved. The application of this technology can simplify the fault diagnosis process of strip production lines，improving the efficiency and stability of the production line.

      Key words: large rolling mill；remote monitoring；fault diagnosis；BP neural network；Bayesian network；expert system