Prediction of viscosity, density and solids content in inks employed in printing industry production chain combining infrared and neural models

Abstract

In this work, some characteristics of the black and whiteinks that are part of the graphic printing processes,calledrotogravure, were evaluated. In this process, the ink playsa decisive role in the quality of the material produced,therefore, its properties must be evaluated and guaranteedin order to obtain a product that works properly in theprinting process.Thus, an analytical method was devel-oped that combines infrared spectroscopy with ArtificialNeural Networks (ANN) to estimate the viscosity, densityand solids content of inks, having the advantage of provid-ing highly accurate results very quickly and with littlecomputational effort. The best models were those devel-oped for density, with average percentage errors of: 1% intraining and validation, and 2% in testof the black andwhite inks together; 1% in training and validation, and0.7% in test of theblack ink; 0.2% in training, 0.8% invalidation and 0.7% in test of white ink. The method de-veloped has the potential to be applied in printing indus-tries as an improvement for the production of high qualityrotogravure printed material

Keywords

• Inks
• Viscosity
• Density
• Solids content
• Artificial neural networks.