Research and analysis of intercalated meltblown nonwovens based on machine learning

Abstract

This paper explores the relationship and change law between process parameters, structural variables and product performance, and through seeking the optimal combination of process parameters, the filtration efficiency of the product is as high as possible and the filtration resistance is as small as possible. Firstly, the correlation between structural variables and product performance is analyzed by Spearman correlation coefficient method after the relevant data are processed by factor analysis method. Secondly, linear regression, random forest regression and lightGMB regression models are used to judge the relationship between process parameters and structural variables. Finally, the relationship between structural variables and product performance can be obtained by using multidimensional scaling transform (MDS). The results showed that the product filtration efficiency was the highest when the acceptance efficiency distance was 20m and the hot air velocity was 800km/h.