Utilization of deep learning technology Establishment of a comprehensive monitoring system for combustion disorders
[일렉트릭파워 배상훈 기자]Korea East-West Power (President Park Il-jun) has completed a system that can predict abnormal symptoms of power plant facilities based on artificial intelligence deep learning technology.
The East-West Power Digital Technology Convergence Institute announced on February 2 that it has built a’combustion disorder comprehensive monitoring system’ that can detect clinker formation early.
The clinker is the melting of coal ash and solidifying into a lump during the combustion of thermal power plants. It sticks to the inside of the boiler or the steam tube, lowering the power generation efficiency, and may cause equipment failure if it falls.
The system that East-West Power has built this time is not a method of observation through CCTV. The sensor detects clinker generation in advance and allows the driver to take action through an alarm.
This system uses sensors installed in existing power plants and data accumulated in the sensors without installing additional equipment.
East-West Power analyzed data and tracked sensors that showed significant changes before the clinker fell. After evaluation by an in-house boiler expert, final sensors required for system development were selected.
In addition, a deep learning model was created that trained the normal state in which clinker was not generated. In particular, when a certain criterion is exceeded, a clinker generation alarm is generated.
This system is currently being applied to East-West Power Dangjin Power Plant Units 1-4. It is planned to expand to all units within the Dangjin Power Plant by the end of this year.
An official from East-West Power said, “We will continue to do our best to contribute to stable power plant operation by utilizing big data and artificial intelligence, which are core technologies of Industry 4.0.”
Meanwhile, East-West Power has been promoting the development of a facility health monitoring system that can monitor the health of major facilities in power plants in real time using artificial intelligence technology since last year. The service is scheduled to start on February 15th.
Copyright © Electric Power Unauthorized reproduction and redistribution prohibited