Research On Identification Of Abnormal Motion Causes Of Ship Type Buoy Based On BP Neural Network

The ship type buoy is the navigation aid mark used to indicate the dangerous area on the Yangtze River route.The accuracy of its position and the normal function are the key to ensure the safety of the navigation channel.However,buoys often change due to interference from external factors,resulting in abnormal function or deviation from accurate position.Because there are few studies on transposition identification at sea and at home and abroad,there is no effective method to determine the cause of the buoy transposition,and maintenance personnel cannot formulate targeted preventive measures to reduce the occurrence of buoy transposition.Therefore,researching a set of methods for intelligent identification of the reasons for ship-type buoys can promote the development of buoy research in China.It can also effectively reduce the probability of floating reversal and has great significance for improving shipping safety.This article takes the ship type buoy of the Yangtze River Channel as the research object,and uses the acceleration of the buoy’s transposition as the judgment basis of the reason of the change.The data acquisition device was designed for the special working environment of the buoy and the buoy transposition experiment was implemented.Through the theory and experiment,we study the characteristics of the transient acceleration and propose a method of characteristic acceleration selection.Using the principle of artificial neural network,training and testing neural network in MATLAB environment.The simulation results show that the BP neural network identified in this paper can accurately identify the corresponding cause of the change based on the change of motion,and can be used to identify the cause of the buoy’s change in the future.The specific research work in this paper is as follows:(1)This paper analyzes and discusses the characteristics and application of BP neural network principle in recognition and fault diagnosis.Combined with the current situation in buoy transposition studies at home and abroad,this paper proposes a method of using BP neural network to study the reason identification of buoy transposition.The BP neural network was established and trained in the MATLAB environment to identify the causes of the buoys.A good recognition rate has been obtained,which shows that BP neural network can be used to classify and identify the type of buoy transposition.(2)The real scene of the buoy transposition was analyzed,and the buoy transposition simulation experiment was designed and implemented.The experimental data was used to analyze the characteristics of the acceleration of the buoy when the buoy changed,and a method for selecting the acceleration value to represent the characteristics of the buoy’s different movements was proposed.The method was used to obtain the input samples of the buoy intelligent recognition neural network.Experiments show that the method is effective and can be used to select training samples for neural network buoy transposition identification.(3)The special working environment of the buoy is analyzed,and an acceleration acquisition device suitable for working in this environment is designed for the working environment and equipment conditions.The acquisition device is used to collect the buoy acceleration and provide data support for the research.It also provides a template and reference for the development of the buoy shipborne equipment in the future.(4)Theoretical analysis and calculation of the force of the buoy in the water,and the influence of different external factors on the buoy acceleration,proves the feasibility of the simulation experiment,and provides a reference for future buoy experiments.(5)The accuracy of neural network recognition based on conjugate gradient descent method and the accuracy of neural network based on L-M method are compared.It is shown that L-M method is more suitable for identifying the cause of buoy transposition.