Research On Production-Logistics Modeling And Spatial-Temporal Analysis Prediction Method Based On Digital Twin

According to the characteristics of multiple models and variable batches of products in the intelligent manufacturing workshop of spacecraft,production logistics needs to have real-time monitoring,logistics tracking,online detection,intelligent decision-making,and other functions.Under the support of information technology,a large amount of data in the intelligent manufacturing workshop has been collected but not effectively utilized,and the production decision-making process lags due to insufficient predictability,leading to the loose integration of information space and physical space in the production logistics process,which is difficult to meet the requirements of dynamic real-time mapping.Aiming at the above problems,the main research conten ts of this paper include:(1)Based on the assumption and mathematical description of the production logistics problem,the production logistics model and operation process based on the digital twin five-dimensional model are proposed.Aiming at the proble m of information space perception lag to physical space in digital twin workshop,a production logistics state prediction method is proposed to improve the real-time information interaction.(2)In order to solve the problem that it is difficult to track t he interlaced and mixed logistics caused by multi type product processing in the flexible manufacturing workshop,the workshop logistics supply process is studied,and the influencing factors of spatial temporal are analyzed.In spatial,the spatial heterogeneity and autocorrelation of workshop layout affects the material storage and transportation process;In terms of temporal,due to the disturbance in the manufacturing process,the loss of Overall Equipment Effectiveness and other reasons,local logistics is blocked,and the smooth flow of materials in the whole workshop is affected by the transfer effect.Finally,based on the spatial-temporal correlation analysis of logistics status,a network modeling method of production logistics diagram is proposed,which combines workshop layout and workpiece processing path.(3)Aiming at the problem that it is difficult to obtain effective information from massive spatio-temporal data in digital twin workshop,a production logistics state prediction method based on spatio-temporal characteristics is proposed.This method takes the historical data recorded in the information system as the basis for forecasting the material flow in each buffer zone in the workshop,uses the Graph Convolutional Network(GCN)and Gated Recurrent Unit(GRU)to obtain the spatial and temporal characteristics of the production logistics state,and finally applies it to the logistics state prediction task.Finally,for the purpose of solving the practical engineering problems in th e intelligent manufacturing workshop of a spacecraft in an aerospace institute,a digital twin production logistics prototype system was built,and the method for predicting the logistics status of the production workshop based on the spatial-temporal characteristics proposed in this paper was numerically verified using the processed workshop layout and production data.The method showed better prediction performance than the baseline method in different prediction ranges.The prediction accuracy is above 96%.Therefore,this method is helpful for the real-time mapping of information space and physical space in the digital twin and can provide sufficient time for subsequent planning and adjustment.