The Study Of Fault Diagnosis For Ship Fuel Oil System Based On Artificial Neural Network

Ship fuel oil system is the most important part of the ship propulsion system as the heart to people. The function of this system is to supply oil with certain pressure and appropriate viscosity to the machines such as main engine, diesel generator, auxiliary boiler which consume oil. It has the greatest effect on the normal operation of the ship. It is very usual to see the serious accidents occur on the main engine and diesel generator because of the fuel oil system fault, so it is very important to forecast and judge whether the fuel oil system is working properly, whether the fault is existed and find out the causes of the fault timely and accurately.With the development of the ship automation, the ship machinery and equipments become more and more advanced, which adds difficulty to the fault diagnosis of the ship equipments and machinery. Ship fault diagnosis is of integrated and cross technique and it’s hard for traditional ship fault diagnosis technology to meet the practical requirements of modern ship management. The application in the ship fault diagnosis of the neural network is of important direction in the marine engineering field. This paper applies neural network to fault diagnosis of ship fuel oil system and a fault diagnosis structure which by neural network mainly and fuzzy theory as a useful complement is presented.Firstly, this paper summarizes the research status and common methods of the ship marine engineering fault diagnosis technology as well as its developing direction. Secondly, gives a brief introduction to the basic theory and methods of artificial neural network, studies the network structure, regulations, designing methods of the BP neural network, also analyzes the defects of BP neural network and corresponding improved methods. Thirdly, makes a brief introduction to fuzzy theory and puts forward the fault diagnosis mode with the combination of fuzzy theory and artificial neural network. Lastly, analyzes the fuel oil system of the teaching and training ship "YUKUN" vessel, studies the fault modes and reasons of the main engine fuel oil supplying system, establishes fault diagnosis model by applying BP neural network, simulates the fault diagnosis by using the MATLAB software and optimizes the fault diagnosis model according to its defects. The simulation statistic shows that the fault diagnosis model optimized can identify the fault quickly and accurately, and this fault diagnosis method is effective. Based on above study, the fault diagnosis system is developed by using hybrid programming technology.