Track: Case Studies

Abstract

Pigment dispersion process is the most crucial process in printing ink manufacturing. A multinational printing ink manufacturing company produces ink through semi-finished or intermediate products prior to finished ink production. Pigment concentrated is an intermediate with the most complex manufacturing process in the company because it involves a pigment dispersion process. During the 2019-2021 period, the Right First-Time rate of pigment concentrated was below than 70%. Along with technological developments, the use of Data Mining methods in the industrial world has been widely studied. The objective of this study is to classify printing ink quality based on the pigment dispersion process by bead mill with CART using historical production parameter and quality data from a multinational printing ink manufacturer. This research was conducted using the QM-CRISP-DM framework by Schafer et al. (2018). The data was obtained from the production logbook and the company's ERP database from 2019 to 2021 period. The model was made using 18 independent variables and particle fineness status as target variable with (0) as defected and (1) as not defected. This study shows that the fineness of printing ink is highly affected by temperature of Pass 1 during the production. The decision rule to reduce defect in pigment concentrated production process is also proposed. The use of the over-sampling method on training data and the Entropy Criteria has the highest accuracy.