Research On Demand Forecasting And Classified Management Of Multiple Units Maintenance Spare Parts

In recent years,more and more vehicles and more and more overhaul models are attached to each section of bullet trains,which brings great challenges to the spare parts management of bullet trains.The traditional spare parts demand prediction method has poor generalization ability,which can only deal with some spare parts with stable demand,and has poor prediction effect when facing the spare parts with discrete and irregular demand.Secondly,the unified management strategy is difficult to adapt to the management needs of EMU spare parts with huge differences in attributes,which makes the problems of high cost and low efficiency of EMU spare parts management difficult to be effectively solved.Demand forecasting and inventory management,the author of this paper two key links of classified management,put forward the demand of the adaptive prediction method and classification management method based on multiple attributes,solved the emu spare parts demand prediction error is big,the problem of low efficiency of spare parts management,realize the diversification of inventory management strategy,provide an important guarantee for the normal maintenance of bullet train.The main research work is as follows:(1)Aiming at the problems of various types of maintenance spare parts and large differences in spare parts requirements for EMU,this paper proposes an adaptive combination model prediction method based on demand characteristics.Firstly,the discretization and variability of spare parts demand are decompressed and analyzed to summarize the adaptability of spare parts demand characteristics to the prediction model.Then,based on Boost regression decision tree,exponential smoothing method,and other prediction methods,the prediction effect of different models under different demand characteristics is explored.Finally,the combination mechanism of various models is studied by the extraction and comprehensive ability of demand characteristics of BP neural network,and the adaptive integration of different models is carried out,so as to achieve better prediction accuracy in a wider range of spare parts categories.(2)In view of the complex technology of spare parts for EMU,and the great difference in the importance,demand,and purchase cycle of spare parts,a classification management method based on spare parts multi-attribute optimization clustering is proposed in this paper.Firstly,the spare parts are divided into special spare parts and common spare parts by the hierarchical clustering method,and the adaptive classification number of spare parts is explored according to the category distance according to the distribution characteristics of the constructed hierarchical structure tree.Secondly,aiming at the local optimization problem of hierarchical clustering,the global search ability of the genetic algorithm is used to optimize the clustering classification.Finally,according to the classification results,the appropriate inventory management strategy is selected for all kinds of spare parts,and the inventory fund is rationally allocated to optimize the inventory management structure.(3)In view of the existing emu spare parts system incomplete,inefficient data management,data flow hysteresis problems,based on the study of demand forecasting and classification management method as the core,set up a set of the emu forecasting and management system of the spare parts,including data management services,the core algorithm,data analysis,and human-computer interaction interface module,The intelligent management of spare parts for EMU has been realized,which improves the efficiency of spare parts management and reduces the cost of inventory management.To sum up,starting from the analysis and summary of EMU spare parts scenarios,this paper constructs the adaptive demand prediction method and the classification management method based on multiple attributes,which realizes the theoretical research breakthrough of spare parts inventory management and the implementation of system application and provides strong support for the effective management of EMU spare parts.