Research On Fault Diagnosis System For Maritime Ship

Maritime Ship undertakes many important tasks of specific waters such as ships navigation safety,water pollution prevention and control,cruise law enforcement and so on.It is of vital importance to guarantee Maritime Ship voyage safety and completing tasks successfully by implementation of real-time and reliable fault diagnosis system for Maritime Ship equipment.For the purpose of solving the problems existing in the process of Maritime Ship equipment fault diagnosis,the system framework of Maritime Ship Fault Diagnosis System is designed,missing data imputation technology and real-time fault diagnosis technology based on improved Radical Basis Function Neural Network are researched,and the main functional modules of Maritime Ship Fault Diagnosis System are implemented.The requirement analysis of the Maritime Ship Fault Diagnosis System is conducted,the framework of the system is designed,the key technologies of the system are described and the main business logic processes of the system are introduced.An improved algorithm based on K-Nearest-Neighbor is presented to impute the missing data from the Maritime Ship equipment state data acquisition process.To overcome the shortage of traditional Euclidean distance in representation of correlation degree in traditional K-Nearest-Neighbor algorithm,an improved Gray Correlation Degree is used to represent the correlation degree between data objects.The experiment results show that the proposed algorithm has an advantage in imputation accuracy over the traditional K-Nearest-Neighbor missing data imputation algorithm.For the Maritime Safety Ships equipment fault diagnosis problems with lack of applicability and accuracy during ships sailing,a radical basis function neural network method for fault diagnosis is designed.An Improved Artificial Bee Colony algorithm combining opposite learning initialization strategy and auto-adapted search strategy is designed and used in parameter optimization of Radical Basis Function Neural Network for constructing a better performed classifier.The results show that the accuracy and usability of Maritime Ship fault diagnosis process can be improved by the Improved Artificial Bee Colony-Radical Basis Function Neural Network framework,and the real-time requirement of Maritime Ship equipment fault diagnosis can be satisfied,too.The application background of the Maritime Ship Fault Diagnosis System is introduced,the system implementation of Maritime Ship equipment real-time fault diagnosis is given based on the key technologies above.The implementation of the key technologies is applied in the Maritime Ship Fault Diagnosis System,and good running performance is achieved.