Abstract: In order to reduce the frequencies of faults caused by, for instance, misalignment, breaking and skipping of chains in operation and shorten the response time for trouble-shooting, for reducing, hence, the accidents that may occur due to failure to take timely remedial measures, a scraper conveyor fault detection system is applied at Matou Plant. With the use of the integrated explosionproof intelligent AI camera, an image identification model is development for the system. The system is integrated with the plant′s existing blocking system under centralized control, and is capable of performing monitoring, alerting and equipment shutdown functions in a real-time manner. Practice shows that with the use of the system, the number of faults and the shutdown periods of the 3113^# refuse conveyor are both greatly reduced. Statistical data recorded in March, 2024 showed that in comparison with the data recorded in December 2023 before the use of the system, the time for fault clearance, the expense of power for idle running and the costs for replacement of parts of conveyors had seen sharp declines by 2.5 h, 12,000 yuan and 14,400 yuan, respectively. Successful application of the system well demonstrates its effectiveness in reduction of management and operating costs and fault-induced risks, as well as provides valuable experience for the detection of faults of other equipment.