Ship Generator Sets Based On Ann Control System Fdi Research

Along with the development of modern industry and technology, the tonnage of shipping is raising, ships propelled by electric power is rushing all the time and the capacity of generator on ship is becoming bigger and bigger. The request of the generator control system on ship becomes more after electric power propelling is applied. On ships propelled by electric power, the electricity produced by the generator provides not only the load on the ship but the propeller. All above make the factor of impacting the normal function and the potential of fault become more. Once the fault happened, it must be find quickly and exactly and eliminate as soon as possible in order to make the system run normally. Otherwise it will make graveness aftereffect, even casualty of people.The development of control theory, computer technology and artificial intelligence makes fault detection develop. Artificial Neural Networks is an embranchment of artificial intelligence, the use of it on fault detection and isolation develops quickly. Because of the collateral distributed disposal, associational memory, learning ability, non-linear mapped character of Neural Networks, it can distinguish from complex information and classify it. We can use it to classify the changes of state caused by the fault of generator control system on ships.This task chooses generator control system on ships as the research object. At first, find the faults of sensor, actor and controller such as short circuit, short circuit and plus, emulate them with SIMULINK. Secondly, use BP Network, CMAC network, Elman network detect and isolate the fault of sensor, actor and controller of the generator control system on ships, meanwhile, modify the BP network. At last, carry out the process of fault detection and isolation use MATLAB. There is an example in the paper, including the analysis of it. The result shows that all the networks above can detect and isolate the fault in real time, meanwhile, it proves the feasibility of using networks in this field.This project is supported by: Project of Shanghai Education committee: "Research on neural control method and its application in marine auto engine room." (05FZ09).Project of Shanghai Maritime University: " method and system research of motor fault detect based on data fusion" (XL0105)...