Big Data Aid Decision-Making System Of Fiber Optic Production

With the large-scale construction of China's communication infrastructure,the optical fiber and cable manufacturing industry has developed rapidly.At the same time,intelligent manufacturing is leading the wave of industrial 3.0 revolution.China's traditional manufacturing industry has gradually begun to upgrade and transform towards intelligent manufacturing industry.China's fiber preparation enterprises have basically realized information and intelligent management,but there are still some problems in the process of fiber preparation.How to solidify expert preparation experience;How to accurately predict the influence of technological parameter changes on future product quality,etc.This thesis describes the status and shortcomings of the traditional fiber preparation process,explores the use of machine learning to assist the decision-making of fiber preparation,and forms a visual system.Analyze the prefabricated rod data and fiber data,establish the regression decision tree model,form the fiber preparation rules according to the model,analyze the internal correlation between the prefabricated rod parameters and fiber parameters,and visualize the results to form a complete system.The main research contents of this thesis are as follows:1.Conducted in-depth understanding of prefabricated rod data and optical fiber data,conducted a series of operations including noise removal,normalization and data linking,and conducted preprocessing for feature construction and modeling of machine learning;Based on the understanding of the data and service,the information gain method is used to select the characteristics of the prefabricated rod data.A regression decision tree model is constructed with prefabricated rod data as the feature and a parameter of optical fiber as the target.The validity of the regression decision tree model is verified by experimental analysis.2.Obtain multiple fiber preparation rules according to the splitting path of the regression decision tree model;According to the information gain method,the prefabricated rod features with high importance are obtained as the influence factor of the fiber.Through the analysis of historical data,the relationship between the fiber target and the influence factor is obtained.3.Build a big data platform and a web server,visualize the fiber preparation rules and influence factors,and form a complete auxiliary decision-making system for fiber preparation of big data.For the designed fiber optic big data aided decision-making system,its advantages are summarized: multiple fiber preparation rules are obtained through the splitting path of the regression model,and the experience of curing preparation is summarized;Through the analysis of the influence factors of the fiber parameters,the influence of the prefabricated rod parameters on the fiber and their internal relations are found effectively.Build a big data platform to store and analyze massive data and improve the efficiency of the system;The complete visualization platform can directly display the results of the algorithm and provide convenient decision-making assistance for the preparation experts.