In order to solve effectively the issue of on-line detection and fault diagnosis of power transformer in FengMan hydropower plant, a kind integrated project of combining embedded network with BP neural network and fuzzy expert system is proposed. Embedded network is embedded in detecting set of power transformer, it takes charge of collecting and detecting information of dissolved gas in transformer oil. In diagnosis center, diagnosis system access collected information, exact dissolved gas concentration values are obtained by multi-sensor consistency information fusion and fusion computing that integrate arithmetic average with batch estimation algorithms, BP neural network takes charge of exact identification composing element and concentration of gas, fuzzy expert system takes charge of fuzzy inference. Diagnosis system reports fault causes and processing methods and steps. Tests on experimental device show that the proposed method is effective and can be used for network on-line detection and fault diagnosis of power transformer.
[1]
Zhang Jing,et al.
Nerual network expert system and their application to forecasting water invasion of colliery
,
1995
.
[2]
Ren C. Luo,et al.
Dynamic multi-sensor data fusion system for intelligent robots
,
1988,
IEEE J. Robotics Autom..
[3]
Thomas Parisini,et al.
Keynote paper: Fault diagnosis and neural networks: A power plant application
,
1995
.
[4]
Zhao Hai,et al.
D-S evidence theory applied to fault diagnosis of generator based on embedded sensors
,
2000,
Proceedings of the Third International Conference on Information Fusion.
[5]
Erik Blasch,et al.
Multilevel feature-based fuzzy fusion for target recognition
,
2000,
SPIE Defense + Commercial Sensing.