Decision And Optimization For Condition-based Maintenance Of Equipment Based On The Defect Forecast

Condition-based maintenance(CBM),which is generated based on the condition monitoring of equipment.Through monitoring and analysis of condition parameters to judge the status of equipment,the CBM could arrange the proper maintenance time and maintenance modes.This thesis lucubrates to the decision-making and its process,which are the important parts of CBM.A method of maintenance modes-selection based on important degree and fault pattern of equipment is researched,after that an early recognition model and a prediction model of defective status are researched.At last,the thesis establishes a model of decision and optimization for CBM of equipment based on the models above.In addition,through the computer simulation research and case analysis of the models,the availability and feasibility of relevant models for decision-making process are verified.Firstly,the maintenance modes-selection is studied from two following ways: the important degree and the fault pattern features of equipment.A method for sorting the equipment important degree is researched,after that the thesis analyzes the characteristics and applicative condition of the Analytic Hierarchy Process(AHP)and Entropy Value method.Then a combination weighting approach method based on the AHP and Entropy Value method is put forward,and the Monte Carlo method is also used for keeping the objectivity of the method.After studying some recognition methods for fault pattern features of equipment,proposing an improved Cuckoo Search algorithm which is combined with the Least Square Support Vector Machine(LSSVM)method to accurately recognize the fault patterns.At last,a case analysis of ship power equipment is used for studying the method of equipment maintenance mode-confirmation based on the important degree and the result of fault patterns.Secondly,the characteristics of equipment operation and the early recognition method for defective status are analyzed and researched.Considering from the aspect of equipment condition recession,it will get a research on the Hidden Semi-Markov Model(HSMM).This thesis uses the algorithms of Forward-Backward method,Viterbi algorithm and Maximum Likelihood Estimation for solving the questions of evaluation,decoding and learning.After studying a solving method for the confirmation and selection of the recession factors based on the theory of stochastic filtering,the thesis puts forward amethod for early recognition and residual life prediction of defects with recession factors based on the HSMM.The computer simulation is used for verifying the availability and feasibility of these methods above.Finally,the problems of decision and optimization for maintenance activities and interval time are made,then giving a solving method which is based on the objectives of the least maintenance cost.Aiming at getting the optimization objectives of maintenance modes which are simplistic,one method of decision and optimization for multiple-objects maintenance after combining the highest equipment reliability and the least maintenance cost is proposed.The corresponding decision and optimization models are established,and the availability of these methods is verified through the computer simulation.