Research On Several Key Technologies Of Health Assessment And Prediction For Ship Equipment

With the rapid increase of the integration level,complexity and intelligence degree in modern ship equipment,traditional fault diagnosis and maintenance support technologies are difficult to meet new requirements.As a new concept to change the traditional maintenance support method,prognostics and health management(PHM),is one of the new topics of ship equipment maintenance and management.At present,the PHM has some limitation in real ship application,due to several problems,such as the monitoring difficulties of radar,lack of fault samples,ship machinery and equipment are greatly affected by environmental conditions.This dissertation is focused on the key aspects of condition monitoring,evaluation and prediction,and maintenance decision in the field of PHM.The overall structure,hierarchical structure and information flow of PHM are studied.Key technologies are proposed according to the characteristics of three types of typical ship equipments.The main work and results are as follows:1.Taking the shipborne tracking,telemetering and command radar as the object,the health condition monitoring and evaluation method is studied.The condition monitoring system of tracking,telemetering and command radar is designed,including online monitoring information and offline test information.On this basis,the radar health assessment method is studied,the health assessment index system is established,and the radar health assessment method based on fuzzy comprehensive evaluation is proposed.In order to further eliminate the fuzzy boundary problem,an FCCS-SVR evaluation model is proposed by combining the fuzzy comprehensive evaluation and SVR model,and using the CS algorithm to optimize the SVR parameters.According to the performance comparison analysis of PSO-SVR,GA-SVR,BA-SVR,CS-SVR and FCCS-SVR models,the case simulation results testify that FCCS-SVR has higher accuracy.2.Taking the ship power equipment as the object,the fault pattern recognition method under small sample conditions is studied.A CS-LSSVM fault identification model is proposed.The improved ICS-LSSVM model is proposed for the problem that the CS algorithm is easy to fall into the local optimum.The simulation comparison and case analysis results of this model and various intelligent optimization models show that the recognition accuracy of this model is higher.In addition,the fault identification method based on HMM model is studied.The case analysis shows that it can effectively improve the fault pattern recognition ability by using HMM to convert the slowly changing signal characteristics into a large log likelihood probability.3.The prediction method of power equipment fault state and defect state is studied.A CS-SVR-HMM state prediction model is proposed.The simulation case study shows that the model has high fault pattern recognition ability and state prediction accuracy.According to the time-delay theory,a defect state recognition method based on HSMM is proposed.The simulation results show that this model has a good predictive effect on the defect state and remaining life.4.The maintenance strategy,which considers environmental factors,is studied for deck machinery and equipment.The definition of environmental factors and the estimation method of parameters are given.The evolution rules of equipment degradation that considers the influence of internal and external factors are analyzed.A comprehensive objective dynamic decision model based on the highest availability and the lowest maintenance cost rate is established.The Weibull life distribution model is adopted.Through the case study,the influence of different environmental factors on the decay evolution of equipment and the change of prevention and maintenance time interval are discussed,and the validity of the decision model is verified.In the end,the dissertation is summarized.It further puts forward the prospects for future research.