Predictive control of yankee dryer chemistry and creped product quality

Abstract

A system and method is provided for proactive process intervention in manufacturing creped products via a chemical feed stage and a Yankee dryer stage. The method includes generating signals from a plurality of online sensors, corresponding to directly measured variables for respective process components such as, e.g., pH, conductivity, and Yankee blade vibration. Models are developed including retrievable information relating combinations of certain directly measured variables to respective quality characteristics of the creped product. The method further includes indirectly determining quality characteristics (e.g., softness, bulk) for the creped product, substantially in real time, based on, e.g., signals corresponding to directly measured variables, and optionally a predicted natural coating potential. An output feedback signal is automatically generated corresponding to a detected intervention event based on the indirectly determined one or more quality characteristics and respective predetermined targets. The feedback signal may automatically regulate chemistry feed characteristics, substantially in real time.