Research On Filling Method Of Missing Data In Polyester Fiber Data Stream

The scope of application of polyester fiber is becoming wider and wider,the market demand is constantly increasing,and the production scale is rapidly expanding,which makes the competition between polyester fiber manufacturers increasingly fierce.With the rapid development of high-tech industry,the polyester fiber industry has also opened a new era.The launch of "Industry 4.0" makes the manufacturing industry spare no effort to combine cloud computing,Internet of Things,big data and other technologies to promote its own development,which is a rare development opportunity for polyester fiber production.Against this background,this article designs a fullprocess industrial cloud platform for polyester fiber production,and designs solutions of industrial field,cloud and client in detail.It evaluates and fills in the incomplete data and quality of the data for the entire polyester process.Laying the foundation for ensuring the quality of polyester fiber production.This paper mainly completes the following work:(1)For the entire process of polyester fiber,by defining the completeness,accuracy,and consistency of the data flow,the missing data is filled using the extreme learning machine(ELM)to solve the problem of missing data and improve the quality of the data.The experimental results show that the extreme learning machine(ELM)can effectively improve the quality of the polyester industry data.(2)Aiming at the incomplete production process data,an artificial immune inspired data filling algorithm is proposed,which can effectively improve the integrity of large-scale complex data generated in the process industry.This method first calculates the missing rate of the data stream.When the missing rate is within a certain range,it is evaluated by calculating the individual’s fitness,similarity,and expected reproduction rate.The optimal individual is selected after iteration and the incomplete data stream is replaced.When the missing rate exceeds a certain threshold,calculate the comprehensive evaluation value of the data stream in the standby antibody library,and replace the incomplete data stream with the comprehensive evaluation value.All data with missing values are stored in the antigen diagnostic area,which is convenient for data analysis and the adjustment,improvement and optimization of the polyester fiber industry site.The experimental results show that the artificial immune-inspired data filling algorithm has better filling effect than ELM method,and time complexity is better than EM algorithm.(3)Establishing an industrial cloud platform for the entire process of polyester fiber,so that the data of the entire process of polyester fiber is integrated,and the visualization and controllability of the production process is improved.According to the industrial characteristics of polyester fiber,design solutions are designed from three aspects: industrial site,cloud computing platform,and client.First,arrange appropriate wireless sensors to form WSN in the industrial site,and then select Hadoop components according to the actual computing needs.Analysis of on-site hardware and Hadoop components.Then,the effectiveness of the proposed scheme is verified by verification experiments based on the performance prediction of the polyester fiber cloud platform.