Research On Fault Diagonisis System Of Ship Generator Rotor

Vessel generator is the core component of the ship power system, which influences the whole ship running , once breaking down, it must be repaired as soon as possible and make it operated normally, otherwise it may cause huge economic losses, even more endanger the staff's lives who in the ship. Therefore, it has vital significance to research fault diagnosis of ship generator.By the analysis and research of the Ship generator's fault diagnosis technology , according to the current situation and development trend of this technology ,the paper for one of the common faults of ship Generator, that diagnosis technology for rotor eccentricity fault to depth Discussion and research. in this paper ,Ant colony algorithm for the neural network is to be as fault diagnosis technology, and use stator current detection methods, which constitute the neural network fault diagnosis system which based on ant colony ship generator rotor eccentricity .finally achieved the Online Diagnosis and of ship generator's rotor eccentricity .In the specific research process, it firstly introduces the structure, working mechanism and common faults of the ship generator, in which Analyzed the ship generator's Features when the rotor eccentricity fault occurred . at the same time , with the spectrum diagrams of the ship generator in the rotor eccentricity fault, comes the corresponding feature frequency and amplitude ,that also can reveal the inner relation between characteristic frequency and the rotor eccentricity fault.Secondly, Through analysis ,comparison and research the ,for the shortcomings of BP neural network in fault diagnosis, and Taking into account of the robustness, global optimization, easy to combine with other methods of the ACO optimization algorithm ,this paper which based on the traditional neural network fault diagnosis, apply Ant Colony Neural Network algorithm to ship generator rotor eccentricity fault diagnosis, and build the Ant Colony Neural Network Structure used in ship generator rotor eccentricity fault diagnosis system .By submitting a large sample data to train the ant colony neural network, then making the good ant colony neural network be used on ships generator rotor eccentricity fault diagnosis, through a large number of simulation, and comparing with the result which comes by BP algorithm, we can get a better diagnostic result ,Whether in terms of training speed or accuracy . Thus provide that the neural network algorithm has some validity and better fault tolerance in the Intelligent Fault Diagnosis system of the ship generator ,and it is an effective diagnostic method. and well realized the on Line detection and diagnosis of the ship generator's rotor eccentricity fault.Finally, based on ant colony neural network technology, Mixed C++ and Matlab ,Developed a new intelligent fault diagnosis subsystem which can realize the rotor eccentricity fault better.