شماره ركورد كنفرانس :
3208
عنوان مقاله :
Robust Fault Detection on Boiler-turbine Unit Actuators Using Dynamic Neural Networks
پديدآورندگان :
Daneshnia, Arash Department of Electrical Engineering - Amirkabir University of Technology , Menhaj, Mohammad Bagher Department of Electrical Engineering - Amirkabir University of Technology , Barazandeh, Farshad Department of Mechanical Engineering - Amirkabir University of Technology , Kazemi, Ali Department of Mechanical Engineering - Amirkabir University of Technology
كليدواژه :
model error modeling , neural network , actuator , boiler-turbine
عنوان كنفرانس :
چهارمين كنفرانس بين المللي كنترل، ابزار دقيق و اتوماسيون
چكيده لاتين :
Due to the important role of the boiler-turbine
units in industries, it is important to diagnose different types of
faults in boiler-turbine units. Actuators as the main part of the
system can be affected by different types of faults. In this paper
fault detection of boiler-turbine actuators is studied. In order to
detect the fault, a dynamic neural network with an internal
feedback is applied. After generating the residuals, the decision
making step has to be followed. In order to design a proper
threshold which is sensitive to different types of faults and
insensitive to noise, the robust threshold is designed using the
model error modeling method. The results show the effectiveness
of this approach for designing the threshold. As a practical case,
the dynamic model of the boiler-turbine unit presented by Bell
and Astrom is considered.
