An Intelligent Fault Diagnosis Method Of Marine Seawater Cooling System Based On SOM Neural Network

Abstract

To solve marine seawater cooling system's faults better, the fault pattern recognition model of marine seawater cooling system is established. Firstly, the structure and typical faults of seawater cooling system are analyzed, and fault modes are divided. Then the LMS learning rules are selected as the learning algorithm of SOM neural network, and the fault sample set of marine seawater cooling system is constructed by using the relevant state parameters collected from the real ship to train the SOM neural network. The training results show that the model has satisfactory clustering effect. Finally, the fault identification model is verified by the real ship test data, and the results show that the model can accurately diagnose the fault mode of the marine seawater cooling system.