Research On Collaborative Supply Strategy Of Spare Parts For Intelligent Ship Based On Uncertainties

Spare parts are the important material basis for equipment maintenance and support,which will not only affects the integrity of equipment and the reliability of the system,but also affects the operation efficiency and maintenance cost during the equipment life cycle reasonable and effective spare parts storage.Especially with the improvement of modern equipment intelligence and system function integration,the category and quantity of equipment composition increase greatly,which poses a severe challenge to the storage and supply mode of spare parts.Condition monitoring based on state perception improves the prediction of equipment health trend and spare parts demand to a certain extent,but due to the inevitable existence of various uncertainties such as operating environment,mission conditions and parameter measurement,still can cause the problem and so on paroxysmal breakdown,uncertain spare parts consumption and partial emergency demand.Therefore,this research chooses the high-tech intelligent ship as the research object,based on the spare parts category decision,the demand forecast and the inventory coordination,constructs the intelligent ship’s uncertainty analysis frame and the solution in the spare parts coordinated supply,it provides a theoretical support for the decision-making of equipment support and scheduling management of intelligent ships,and has obtained the following main research results:(1)Based on G-DCNN data processing method proposed,the traditional multi-attribute spare parts classification problem is transformed into image recognition and classification problem.First of all,the G-DCNN classification model is based on an explanatory hierarchical classification structure which is constructed according to a variety of attributes considered by different decision-makers and their causal relationship.Secondly,the construction rules of hierarchical diagram are studied by the AlexNet neural network because of its powerful recognition ability,which greatly improves the computing power and speed of the model.(2)In view of the problems and challenges encountered in the development of spare parts demand prediction,the multi-layer nonlinear network structure is constructed to extract,learn and excavate the high-dimensional complex input data based on Deep Learning theory,whose purpose is to realize the accurate prediction of the running process state and spare parts demand.In this paper,faced with the key equipment with advanced state monitoring system,the large amount of time series data obtained by the perceptive system in operation are made full use.And the prediction models of AEPCA-LSTM and GAPLS-LSTM are constructed step by step from two aspects.The variable correlation analysis and selection of high latitude parameters in state monitoring system are taken into account in the traditional PCA pure data-driven model.Due to the important characteristics of long-term memory neural network to selectively retain long-term memory,the demand prediction model of spare parts is constructed by fitting the health curve on the basis of operation trend estimation.Finally the replacement demand point is obtained,which thus realizes the important support for preventive maintenance of equipment.(3)In view of the problem of collaborative supply of spare parts,a state-based spare parts co-inventory strategy is proposed.The dynamic coordination and cost checks and balances between demand forecast and physical inventory under optimal strategy are verified by dynamic study of the benefit of state-based strategic decision-making through Markov Decision-making Process.For the Markov chain characteristics shown in the process of equipment degradation,MDP considers the impact of decision-making and decision-making on the system,through the state of the equipment in the process of degradation and the state of the existing inventory constitutes the state space.When the equipment degrades within a certain period of time and the state transfer occurs,the decision maker can dynamically determine the order time and order quantity of spare parts based on the state change.There is also an emergency procedures set up to ensure the reliability of supply,in order to reduce fixed inventory at the inventory point and reduce its inventory holding costs.The paper verifies the cost optimization advantage of the proposed collaborative strategy through a large number of numerical experiments,and studies the specific impact of object monitoring level and sample size and supply capacity on the implementation of the strategy.Compared with the theory of safe operation of traditional data statistics,this study is based on advanced data processing technology and operation management concept.The spare parts supply guarantee strategy for the uncertainties in the process of intelligent operation and maintenance of ships is observed by using time-series "big data" of the long-cycle and high-reliability equipment condition monitoring system.An intelligent operation and maintenance system based on state perception-demand prediction-dynamic inventory-collaborative supply has been constructed,which achieves the autonomy and optimization of maintenance under the premise of ensuring the safe and reliable operation of ships.