Ship’s Cooling Water System Fault Diagnosis On Radial Basis Function Neural Network

The ship’s cooling water system is a very important ship power piping systems to ensure normal and stable operation of the marine power plant. Through mandatory continuous cooling water circulation, the heat generated in the running process of marine diesel engine which isn’t translated into mechanical energy has been taken away, and thus preventing the diesel engine components damaged due to too large heat load and the lubricating oil burnt failure. Therefore, the ship’s cooling water system working properly has a direct impact on reliability, economy and system equipment life of the diesel power plant. Fault diagnosis of marine main engine cooling water system mostly relies on the experience of engineers to detect and judge in previous time, but in that situation it is difficult to promptly exclude once there is malfunction in the system. Therefore, research on how to help the engine department judge the nature of fault quickly and then exclude fault has a special significance to ensure the normal navigation of the ship.At first, Dalian Maritime University training ship MV "YUKUN" are taken as an example, and three subsystems of its central cooling system, sea water cooling system, low temperature fresh water cooling system and high temperature fresh water cooling system and their typical failures are detailed described. Secondly, a variety of fault diagnosis methods commonly used is considered and neural network method for fault diagnosis is selected; then the artificial neural network is introduced, and further the radial basis function (RBF) neural network is selected to build fault diagnosis model. Furthermore, in accordance with historical operating data of MV "YUKUN", respectively aiming at sea water cooling system, low temperature fresh water cooling system and high temperature fresh water cooling system, several sets of data are selected to establish sample sets of faults and fault symptoms, and fault diagnosis program is built using the language MATLAB and then is used to distinguish faults.Finally, ship’s cooling water system fault diagnosis module is built using the language Visual C++; by calling the module and set the module different parameter values, the training sample data can be used for training the model and test data can be conducted fault diagnosis. The results show that fault nature can be distinguished well using ship’s cooling water system fault diagnosis module. With the intelligent level improvement of ship systems, it is an inevitable trend that artificial intelligent fault diagnostic techniques are applied to modern ship engineering management. Ship’s cooling water system fault diagnosis module built using the language Visual C++in the thesis can set up structure parameters easily and has a good portability, so it has a certain practical significance.