Use of Predictive Modeling to Explore Techniques to Improve Recycled Board Coating Performance

摘要

Coating paper or board is a complex operation involving many process variables that can change with each new run. The goal of process optimization involves analyzing enormous amounts of data in order to make the best possible decision. Traditional methods consist of manual regression by experienced individuals. That process is tedious and time consuming and may sometimes miss key relationships. The following study uses predictive modeling to analyze process data for optimization. A recycled paper mill wanted to know what size diameter rod to install on their new precoater in order to optimize performance. Pilot coater trials were conducted to study the impact of rod diameter, tube pressure, and base coat formulation solids on the recycled board coating system. These results were statistically analyzed to determine the impact of the study variables on brightness, smoothness, coat weight of the base-coat as well as brightness, DuPont appearance, smoothness, coat weight, and K&N % brightness drop of the top coat. It was determined that the base coat formulation solids had the most impact on DuPont appearance, brightness of both the base coat and top coat, and coat weight pick up. Rod diameter and base rod solids both had a high impact on K&N % brightness drop. Predictive modeling was used to optimize properties by using the desirability function to identify the highest DuPont appearance value possible at the lowest air knife coat weight.