Research Of Fault Diagnosis System For Hydroelectric Generating Sets Based On MAS

With the development of artificial intelligence technology, malfunction diagnosisfor large-scale revolving machine has become special domain for complicate research,which covers a lot of scientific and professional fields. The paper embarks from thepresent situation of intelligent diagnosis technology, taking malfunction diagnosis forHydroelectric Generating Sets as the research object, taking the MAS theory as thefoundation and proposes special diagnosis for MAS. In view of the limitation of presentfault diagnosis system for Hydroelectric Sets, and the singleness of method, the paperbrings the new Hydroelectric Sets fault diagnosis system with the base of MAS. Thesystem designs System Agent is the key point of MAS diagnosis system, which canmake the first step of fault diagnosis, decompose the fault task as well as evaluate thediagnosis result. Communication Agent is responsible for the regulating the system andmanaging all Agent communication. Communication Agent is maintaining an onlyfunction database. Each newly agent which wants to join in the system needs to hand theother service to Communication Agent. Diagnosis Agent is used for carrying on theprecise fault diagnosis, of which the input sample fault indication weight processingmethod is proved to be good way that distinguishes the degree of fault. The paper hasstudied the malfunction diagnosis which is about the fault of Agent water and electricityunit. The fuzzy algorithm is used for analyzing the diagnosis task in System Agent. Theneural network is used for exact analyzing the fault type in Diagnosis Agent. Finally, thepaper take the simulated example about the fault which will occurs to electricalmachinery through the reality calculated. The simulation result indicated the faultdecomposition as well as malfunction diagnosis for Agent can complete the systemanalysis, which conforms to the precision requirement of system diagnosis.