Prediction of offset ink film thickness using machine learning

Abstract

The wet film thickness of printing ink in offset printing process is an important parameter which is related with the print quality and cost. In this paper, the prediction of ink film thickness in offset printing based on machine learning has been proposed. For measurement and prediction of wet ink film thickness in offset process, an experimental inking model is designed and used to show the effect of various factors like machine speed, run time, ink color, ambient temperature, and relative humidity on it. The machine learning based prediction models can provide very close approximation of the accurate information about the control parameters so that the print technician could make a good setting work during real-time production process. With the help of this, the technician can make a good decision in the setting of control parameters for ink thickness based on this predictor. The results show that the prediction models can provide about 95 % accuracy in predicting the ink film thickness. Thus, the prediction system not only can help technicians and greatly improve their production efficiency, but also can save the cost of production.