Flow path is one of the most important parts of steam turbine, also the part where faults frequently occur, which directly affects the safety and efficiency of the unit's undergoing and furthermore affects the costs. Therefore, the monitoring and diagnosing of the flow path faults are very important. However, various faults classifications and coupling between them makes it hard to timely and correctly diagnose faults in conventional means, while neural networks' strong ability in non-linear performance has a potential advantage in diagnosing flow path faults. This paper, on basis of simulation technology, utilizing both traditional thermal means and neural network, tends to a new means of diagnosing faults. Also through simulating experiments, the feasibility of the new means of application in fault diagnosis of flow path is discussed.
|