A Study Of Fault Diagnosis In Hydraulic Turbine Govering System Based On Rough Set Theory

A new study is made to fault diagnosis of hydraulic turbine computer-controlled system created for the Guangxi Xijin hydraulic power plant project authored by Wuhan Sichuang control Technology Company.Hydraulic actiyator system is one of the important systems for operating quality and security of electrical power plant. Hydraulic actiyator technology has been improve to more high level - both adjusting and monitoring is integrated to form a field network control system which breaks the traditional control mode and in which industrial computer control technology, especially high reliability PLC, digital-networklized field bus technology and computer monitoring are used. Wuhan Sichuang Control Technology Company has made a great deal of contributions to this system. Based on above, a networklized hydraulic turbine computer governor system is introduced and it integrated subsidiary relay PID control solution, PLC field control station, IPC monitoring station and RS-485 field bus. Then study is made to its fault diagnosis.Governing system plays important role in the electrical generator and electrical system, and to rapidly find and eliminate its fault is significant for operating security of power systems. But on one hand the fault symptoms are complicated and various, using the expert system is difficult to comprehensively recognize the fault characteristics and position the fault components, let alone the more complex fault. On the other hand, in many conditions, the fault mechanism is ambiguous and diverse, the blindness is unavoidable in characteristics pickup, thus system status description is implicit. Therefore how to select the minimum character sets in guaranteeing the incurrence of system status approval is important, namely to reduce the character, lower work loading and reduce the uncertain factors in the conditions of guaranteeing diagnostics accuracy. And in fact these are the properly research contents of rough sets. This theory is applied to create decision tables accounting for various fault in this paper, Then to realize the searching of auto reduction of decision tables, and the critical signals or not is recognized. After applied this theory the same decision will be based on the minimum conditions, methods of how to create faultsorting regularity is proposed, and diagnostic regularity is derived directly from the fault sample sets. Thus the purpose to diagnose fault is achieved in inadequate alarming information.Based on the theory research both hardware and software is designed, and field bus is the core of the module structure and parameter decision. And Object-Oriented Programming(OOP) and VC++ are used to write the friendly human-machine interfaces software of the system monitoring and diagnosis. Preliminary experience showed that the above solution and designation is feasible. It would be applied to the engineering after further perfected.