Research On Real-time Cyber Physical System Of Intelligent Workshop Based On Edge-cloud Collaborative Computing

The Cyber-Physical System(CPS)in the intelligent workshop is an important carrier in the field of intelligent manufacturing.It is a multi-dimensional and complex production system that integrates perception,calculation,and control.The workshop CPS obtains the dynamic and static information of the software and hardware resources of the workshop through the perception function of the physical layer,and uses the powerful computing power of the industrial cloud platform to calculate and analyze a large amount of industrial data in combination with related application models,and returns the analysis results to the workshop through the control module Issue corresponding action instructions to realize independent decision-making and control in workshop environment monitoring,product process optimization,product quality inspection,etc.,to improve production efficiency and product quality.However,the workshop CPS architecture based on the cloud platform faces new demands and needs to sense,transmit,and calculate more and more industrial production data,which makes the system control delay and difficult to control in real time.However,intelligent precise control requires rapid calculation and analysis of large amounts of data,and feedback results drive corresponding operations,which puts forward requirements for the real-time nature of the production system.In order to solve this problem,this paper introduces edge computing technology into the traditional workshop CPS system architecture based on industrial cloud platform,proposes the workshop CPS system architecture of edge-cloud collaborative computing,and starts from the two perspectives of real-time task scheduling and real-time task data processing.The main research content and innovation points are as follows:(1)Propose a workshop CPS system architecture based on edge-cloud collaborative computing.Based on the workshop CPS architecture controlled by the traditional industrial cloud platform,the edge computing technology is introduced,and the scattered software and hardware manufacturing resources of each workshop are abstracted into network nodes,and they are connected as a whole through the workshop network to establish the four sides of edge-cloud collaborative computing.The CPS system architecture model of the multi-level workshop,and abstractly defines the industrial big data analysis problems such as workshop production environment monitoring,product process optimization,and product quality inspection as CPS computing tasks,which are regarded as an application.(2)Establish a real-time processing model of workshop CPS task data based on edge-cloud collaborative computing.First,according to the workshop CPS system architecture of edge-cloud collaborative computing,mathematically model the CPS computing tasks and establish a real-time task scheduling model,including task calculation time cost model and task data transmission time cost model;secondly,the task execution time cost optimization goal is proposed Function,the objective function is solved by an improved ant colony algorithm to realize real-time processing of computing tasks;finally,experiments are carried out on the Cloudsim cloud simulation platform.The results show that the model has obvious time performance advantages as the number of tasks increases,which is far stronger than the traditional workshop CPS architecture,And the performance of the model can be further improved by increasing the number of edge computing nodes.(3)Establish a real-time processing model of industrial data based on edge-cloud collaborative computing.With the application of tool wear monitoring in the workshop as the background,a real-time processing model of industrial data is proposed for the CPS calculation tasks that need to be scheduled to the cloud platform for execution.First,establish a data preprocessing model at the edge,perform data preprocessing close to the data source,build a tool wear monitoring application model based on residual convolutional neural network in the cloud,and receive preprocessed task data at the edge for calculation;second,propose task data The weighted transmission time is the model evaluation index,and the calculation accuracy of the model is taken into account to optimize the task data transmission time.Finally,the experimental results show that the data preprocessing can extract the detailed characteristics of the signal data and compress a large amount of transmission data.The performance of the cloud application model is also better.Other comparison models and the overall model greatly reduce the task data transmission time with a slight loss of accuracy,thereby shortening the task execution time and obtaining the effect of real-time task processing.