Optimization Of Spare Parts Storage Strategy And Platform Development Based On Data-driven

In recent years,with the increasing demand for production,the number of industrial equipment continues to develop in the direction of complexity and precision in its structure and function,which has brought enormous challenges to the management of spare parts in enterprises.Traditional spare parts management methods rely too much on the experience of technical personnel and are highly subjective,resulting in high cost and low efficiency of spare parts management.Therefore,this article aims at three aspects of spare parts demand prediction,spare parts classification,and spare parts management platform in spare parts management.Through statistical analysis of enterprise equipment basic information and operational data,it summarizes the classification and evaluation indicators of spare parts,and establishes a life model of equipment components.At the same time,around the goal of spare parts management,it uses knowledge such as management,mathematical statistics,and machine learning to establish mathematical models for spare parts classification and prediction,A data driven spare parts management platform is designed to achieve standardized and intelligent management of spare parts data information.The main research contents are as follows:1.Aiming at the problems of high dispersion and low data volume of infrequent spare parts demand data,a reliability based method for determining infrequent spare parts reserve quota was proposed.Firstly,based on the similarity theory,the failure data of the same equipment under similar operating conditions are analyzed uniformly,and a life distribution model of components is established.Then,based on the life distribution of spare parts,the Monte Carlo method is used to simulate the process of replacing spare parts within a specified time,determine the demand for spare parts,and achieve infrequent spare parts demand prediction.2.Aiming at the problems of low prediction accuracy of traditional time series prediction models for unstable spare parts demand sequences,a prediction model for unstable spare parts demand based on a set empirical mode decomposition(EEMD LSTM)neural network is established.The set empirical mode decomposition method is used to decompose the unstable spare parts demand sequence into several relatively stable demand sequences,and the long and short term memory model is used to predict it,Finally,the data is reconstructed to achieve the prediction of unstable demand,and the prediction accuracy has been significantly improved compared to traditional time series models.3.Establish a classification index evaluation system for spare parts based on their own characteristics such as price,quantity,procurement cycle,and importance.The analytic hierarchy process and entropy weight method are used to calculate the subjective and objective weights of spare parts classification indicators,and the idea of game theory is combined to solve the conflict between subjective and objective weights to achieve combination weighting.Finally,the hierarchical clustering method is used to classify spare parts,making the classification results more reasonable.4.Based on the above research methods,this paper designs a data driven spare parts management platform,which uses the B/S framework,uses JAVA language development,and uses My SQL database for data storage.It achieves functions such as basic information management of spare parts,equipment fault data management,inventory information management,and statistical analysis,and improves the level of data-based and standardized management of spare parts information for enterprises.