Study On The Health Assessment Model Of Large Unmanned Surface Vessel Integrated Power System

In recent years,with the rapid development of the technology of unmanned surface vessel(USV),large USV driven by electric power have gradually become one of the research hotspots at home and abroad.The large USV adopts the most advanced electric propulsion mode.The integrated power system has complex structure and many equipment,and it often carries out long voyage operations.Therefore,ensuring the healthy operation of integrated power system is the key factor of large unmanned master ship.It is necessary to develop the intelligent monitoring and health management module of large USV integrated power system based on the long sailing and high load health requirements of large USV.Health assessment,as the key technology of the health management module of large USV integrated power system,has considerable research value.At present,there are shortcomings in the selection of feature parameters for health assessment based on expert experience.And the health assessment algorithm also has shortcomings such as low accuracy and long assessment time.Therefore,in view of these problems,this paper relies on the research project of intelligent monitoring and health management technology of large USV integrated power system to study the problem of health assessment.The purpose is to establish a high accuracy health assessment model for integrated power system of USV.This paper focuses on the research and design of the health assessment of integrated power system for USV:First of all,based on a large amount of research,analyzes the research status quo of USV at home and abroad and the main gap;then analyzes the research status of the health evaluating method;and then combined with the actual needs of the project,design the overall evaluation system of the integrated power system for USV health assessment.Secondly,the related parameters that affect the health state of the power system are analyzed.From the main system equipment parameters,USV sailing state parameters,environmental parameters analysis of the three aspects affect the health of a system parameter,and the preliminary selection of voltage,frequency,the probability of such nine parameters as the characteristic parameter;parameter selection due to excessive use of rough sets is proposed cumulative dynamic reduction algorithm to simplify the parameter evaluation based on factor,finally obtains the characteristic parameters of this study selected four.Finally,based on the analysis of the health assessment algorithm,proposed health assessment system using convolutional neural network,and then according to the characteristic parameters to establish convolutional neural networks with different network parameters of the health evaluation model,studied the effect of various network parameters of convolutional neural network on the accuracy of assessment,verify these parameters can reflect the health system state.The feasibility and superiority of the convolution neural network algorithm used in this paper is verified by comparison with the BP neural network and the depth belief network.