Data Analysis And Research Based On Industrial Internet Platform Of Bearing Production Line

In the context of the rapid development of digital and intelligent industry,the low level of informatization between bearing assembly production lines,discrete value of data,and poor data application efficiency will limit the production capacity innovation and optimization decision-making of bearing line production.Therefore,it is particularly important to establish an industrial data platform for the bearing production line,design its processing and calculation architecture based on its data characteristics,and analyze its application methods and feasible solution algorithms for production line optimization.This article starts with the batch processing and stream processing of bearing industry data.Besides that,this article also analyses how to solve the optimization of balance of production line,analyzes the data characteristics and application perspectives of bearing production industry data.The structure which can extract,transform and load data was designed for data cleaning and solution,the main contents are as follows:Different from the medical data or e-commerce data industries that have formed a unified development and modeling standard,bearing production industry data has not yet established unified procedures in terms of data naming,type,and coding methods.This article designs a data standard set for some application parameter indicators and targets The realization of the difference function is designed to realize the three layers of ODS,DWD and DWS as the data warehouse,which realizes the completion of data cleaning and deduplication and the calculation of target parameter indicators in the data batch processing process.In terms of the application of analysis data to optimization of the bearing production line,it started from the perspectives of crowded production line queues and unbalanced job unit planning,and deduced its mathematical models according to different scenarios.Aiming at the problem that the frequently-used single-objective optimization model cannot solve the optimal value of multiple parameters,a multiobjective optimization mathematical model is designed for it.This mathematical model can effectively summarize the optimization requirements of the multi-objective optimization in the optimization solution.In view of the general convergence performance of the traditional optimal solution search algorithm and the tendency to fall into local optimality,the SFLA-PSO algorithm and the improved genetic algorithm are proposed for the two problems in this context.Compared with ordinary search algorithms,it has better search speed and convergence performance.Finally,for the batch processing calculation of the data warehouse and the stream processing calculation of the optimal configuration of the capacity of the bearing buffer,according to the applicable multi-enterprise multi-production line operation scenario,the data visualization interface is designed and implemented with partial simulation data testing and verification.