Intellectual Analysis System of Big Industrial Data for Quality Management of Polymer Films

Abstract

The article is devoted to application of methods and technologies of the intellectual analysis of the large production data for the large-tonnage production of a polymeric film being characterized complex structure, existence of systemic linkages, semi-structured information, power consumption and a large number of the controlled and accumulated industrial information. The novelty consists in application of methods and data mining technologies in production of polymeric films. The article describes the application of neural networks for big data analysis. Application of neural networks allowed to carry out the forecast of quality of production that in turn allowed to reduce flaws. Also, this article describes the application of other data mining algorithms, which enabled to identify the causes of the decreasing quality of products and to find the values of controls in the system, which provide the specified quality and output as recommendations to the operator. Thus, the application of the developed intellectual system for analysis of large industrial data provided to reduce defects and increase productivity. Approbation of the developed system took place at plants of Russia and Germany and confirmed its working efficiency for management and change-over of productions of a polymeric film on various types of product.