Research Of Fault Diagnosis Method For Ship Ballast Water System

At present,fault diagnosis and trend forecasting have become an important research hotspot in the field of ship,research and explore more practical and scientific diagnostic methods for ship equipment fault diagnosis,which will effectively improve the safety of navigation and operation of ships.Ships ballast water system is an important auxiliary system for adjusting the floating state of a ship,playing an important role in maintaining the balance of the hull,maintaining the proper steady-state height,and reducing the vibration of the hull.Ballast water system itself more equipment,space span,the system is not easy to troubleshoot hidden trouble.According to the system working environment and work process,this paper summarizes the characteristics of the system fault,the system of common fault types and fault phenomena were analyzed.In order to design a suitable fault diagnosis method for ballast water system,this paper separately studies the fault diagnosis method based on neural network and D-S evidence theory fusion method,analyzes the essence and characteristics of fault diagnosis of both and designs the specific fault diagnostic process.By applying the diagnostic method based on neural network to fault diagnosis of ballast water system,it is found that the single neural network can not fully meet the needs of system diagnosis,and there is a certain degree of fault identification error.The combination of multiple neural networks and D-S evidence theory can improve the recognition error caused by inaccurate input data and improve the accuracy of fault diagnosis.The diagnostic output of neural network can be used as the basic credibility distribution function of evidence theory to make up for the subjective defects.The uncertainty theory of evidence reasoning ability can solve the output of a single neural network instability problem,but also reduces the complexity of the network.Therefore,this paper studies the fault diagnosis method based on BP neural network and D-S evidence fusion,and analyzes the diagnostic process and characteristics of the method to meet the needs of fault diagnosis of ballast water system.Applying the above fault diagnosis method,this paper designs a ballast water system diagnosis model for specific conditions and analyzes the specific application details in the model.Finally,the data of the monitoring points of the system operation are collected and input into the system fault diagnosis type.Through the MATLAB simulation calculation,the better diagnosis results are obtained and the effectiveness of the fault diagnosis method is verified.