Analysis On Fault Diagnosis For Turbine Generator Unit Based On Neural Network

With the online monitoring fault diagnosis of turbine generator unit vibration as the researched object as well as fault diagnosis of system in neural network for turbine generator unit vibration as the development object, this paper analyzes and diagnoses these data whereby obtaining the conlusion which is of significant important for thermopower plant practice.With the development of technology, power is used in all fileds, at the same time, turbine generator unit becomes larger and more complex. In this condition, safety engineering is facing more serious troubles on online monitoring fault diagnosis of turbine generator unit vibration. To begin with, this paper deals with the importance of turbine generator unit fault diagnosis, and analyzes the existing researched systems and develops trend of diagnosis in details. Accordingly, the vibration theory, vibration sources and vibration fault featurs of turbine generator units are analyzed, sorted out and formalized so as to lay a solid foundation for the establishment of fault diagnosis bank. Secondly, by analyzing the detector of turbine generator units and the theory of safety engineering, a way of solving the problems are established, using the artifical neural network on online monitoring fault diagnosis of turbine generator unit vibration, lay the foundation of preventing accidents by safety technology. By discussing the way of artifical neural network work, and then analyzes the vibration of turbine generator unit by the frequency of the vibration. Design a table which is to be used in turbine generator unit, according to the Sohre's table on symptom of vibration. Based on fuzzy clustering algorithm , this paper develops some common fault of both standard models and model center. By inputting fault signals combined with competitive learning mode of the artificial neural network, we can detect where the fault is and therefore fulfils the needs of online monitoring fault diagnosis of turbine generator unit vibration. Drawn on the ground of the author's extensive experience in power industry, this paper tries to give measurements of signals of faults and how to preven the faults, effectively. In the conclusion of this paper, it anlyzes the shortcomings of the existing systems, based on the needs of online monitoring fault diagnosis of turbine generator unit vibration, and points out a new convergence algorithm instead, thus it is a try for further research on similar topics.