Application of box behnken design to optimize some parameters for flexographic printing process

The flexographic printing process parameters are crucial elements in ensuring the appropriate print quality and efficient use of resources associated with the printing process. For this, an effective study on establishing the relationship between the multiple input factors and the key output responses are necessary. This study investigates the print quality assessment and optimization of process parameters associated with the flexographic printing process based on Design of Experiment (DoE) carried out by one of the most efficient quadratic models in Response Surface Methodology called Box Behnken Design. The experimental approach designates three factor inputs such as three kinds of anilox roller screen rulings, three grades of paper substrates with different smoothness and three levels of halftone dot percentages. The dot gain and print contrast are taken as the output responses that denote the degree of accomplishment of print quality. The flexographic print trials are executed with the printing of a grey scale image on three grades of paper substrate with three anilox roller screen rulings conditions. The halftone image is created with AM square dots for study. The Box Behnken Design (BBD) is implemented here for the optimization of process parameters with proper employment of analytical approach such as ANOVA and regression analysis. The results found the suitability of BBD model and regression analysis in selecting the parameters to get the better print in flexography. The value of Variance Inflation Factor (VIF) in regression analysis is found below 1.5 that indicates the less occurrence of multi-collinearity in the regression model. The ANOVA result for the dot gain gives R-sq value 97.37% which implies the suitability of the regression model in defining the relation between input variables with the output (dot gain). The ANOVA result for the print contrast shows R-sq value 99.77% which suggests that independent variables are capable of explaining the variance of the dependent variable (print contrast). It is also found that the roughness of paper surface, anilox roller screen rulings and dot percentage influence the dot gain and print contrast.