| At present,China is the world’s largest copper producer,but it is still not a strong copper producer,with relatively excess low-end production capacity,and some high-end copper products,such as high,fine and sharp,are still dependent on imports,which is difficult to meet the demand for high-end copper in various industries in China.In addition,the product quality level and yield of China’s large number of small and medium-sized copper processing enterprises are generally low,which is also a prominent problem restricting the development of China’s copper processing industry,and the reason is that the product quality problems of copper processing enterprises are mainly caused by the extensive management of production technology.In recent years,in order to achieve the goal of improving quality and efficiency,domestic copper processing enterprises have been carrying out intelligent transformation under the encouragement of national and local government policies.However,the application of industrial big data has become a prominent obstacle to the intelligent transformation of copper processing enterprises.In jiangxi are studied in this paper some copper belt production enterprise production process parameters and product quality testing data and other related industrial data collection,build industrial data collected in the big data platform for data cleaning,data storage,data analysis and so on a series of work,to achieve the business part of the production process parameters of intelligent optimization,for copper belt or other copper production enterprise based on the industrial production process parameters optimization of large data made beneficial exploration,and laid a certain foundation.The main research work of this paper is as follows:(1)The research status of copper-strip process parameter optimization and big data technology application was analyzed,and the feasibility and necessity of using big data technology to realize copper-strip production process optimization were proposed;(2)Analyze and study the functions and structure of the industrial big data platform,analyze and compare the advantages and application scenarios of the mainstream architecture of the Hadoop system,draw on the advantages of each architecture,and take data collection,data storage and data analysis as the main line,and propose The lightweight industrial big data platform architecture for small and medium-sized enterprises,and on this basis,the technical details of the Hadoop-based industrial big data platform such as ETL data preprocessing and loading,Hive data warehouse,and Spark big data analysis framework;(3)Deployed and implemented the high-availability copper-strip industrial big data platform with FLUME data acquisition,Hive data warehouse data storage management and Spark data analysis framework as the core modules,and tested the stability,I/O throughput and other performance of the platform;(4)The industrial big data platform was applied to collect,preprocess and store the production process parameters and product quality data of a copper strip enterprise in Jiangxi.Based on the platform,a multiple linear regression model of process parameters was established by using Spark machine learning.Taking the reduction rate and copper content as independent variables,mathematical expressions were obtained by fitting with elongation and hardness respectively.The fitting degreesR~2were all greater than 0.9,indicating that this model has good fitting ability for multiple linear regression.Taking hardness and elongation as optimization objectives,a mathematical model of multi-objective optimization was constructed,and the non-dominated sorting genetic algorithm(NSGA-II)with elite strategy was used to solve the multi-objective optimization.The results show that the NSGA-II algorithm can approach the Pareto front very well,and the generated non-dominated solution set has good distribution characteristics. |
PDF Full Text Request