Zhejiang Hengyang Thermal Power Co., Ltd., located in Shendang Town, is the only public cogeneration enterprise in Haiyan, serving more than 130 daily heat supply users. During the National Day holiday, many enterprises maintained their production pace, and Hengyang Thermal Power relied on its self-developed AI models and automated production lines to reduce manpower input while keeping heat supply guarantees intact, fully ensuring the production heat demand of more than 20 enterprises in Haiyan.

During the National Day holiday, a reporter walked into the Hengyang Thermal Power plant area, where the boiler room, steam turbine room and other production areas were roaring, continuously supplying heat outward through pipelines. In the enterprise's central control room, staff stared closely at real-time monitoring data on the large screen, occasionally tapping the keyboard to issue instructions and ensure the stable operation of the heat network. "Although it is a holiday, we still have more than 20 enterprise users in production and need us to supply heat continuously 24 hours a day, so we also have to stay at our posts," said Ji Jianxin, shift supervisor of Hengyang Thermal Power's production and operation department. He said that normally there are more than 130 users served, and because some users need it, production must not stop even during the long holiday.

The reporter learned that the nearly 100-mu Hengyang Thermal Power plant area normally has about 110 employees, but during the National Day period, only more than 30 employees per day are needed at the plant to meet production, operation, overhaul and maintenance work requirements. Such a contrast benefits from the enterprise's highly automated production line transformation and the application of self-developed AI models. At present, the enterprise's chemical water system workshop has already achieved unattended operation.

Jiang Yunfeng, a water treatment laboratory technician at Hengyang Thermal Power, told the reporter that the enterprise has installed high-definition camera equipment below the purified water tank, operating continuously 24 hours a day and collecting images once every 3 minutes. Then, through AI recognition algorithms, classification and judgment are automatically completed. After multiple rounds of iterative training, the model's recognition accuracy has stabilized above 98%. "After being empowered by AI, the chemical water system workshop has greatly reduced chemical consumption, while also saving on-duty operation and maintenance personnel, and the overall power consumption of the equipment has also dropped significantly, saving the enterprise about 1.8 million yuan in costs each year," Jiang Yunfeng said.

It is worth mentioning that Hengyang Thermal Power's AI technology has achieved a breakthrough from "self-use" to "output." At present, a thermal power enterprise in Ningbo has also adopted some AI application modules from Hengyang Thermal Power. Later, Hengyang Thermal Power also plans to further expand AI application scenarios and is currently developing two brand-new intelligent modules: AI inspection and AI defect elimination.

Ji Jianxin, who has nearly 30 years of experience in the thermal power industry, joined Hengyang Thermal Power 12 years ago and has always stayed rooted in the production frontline. Witnessing the plant area's technological iteration, he felt deeply: "In the past, boiler combustion and feedwater regulation all relied entirely on manual operation, and I spent most of the day guarding the boiler room. Now I sit in the central control room watching the large screen, and operation regulation can be completed through the automated system, which is worry-free and labor-saving."

It is understood that thanks to the enterprise's application and practice of emerging technologies such as AI scenarios and 5G networking in recent years, Hengyang Thermal Power's operating benefits have continued to improve. From January to August this year, the enterprise's operating revenue reached 290 million yuan, a year-on-year increase of about 40%.