Model-based fault monitoring of a plastic film extrusion process

Abstract

This study reports the modelling of a plastic film extrusion process and the development and implementation of a model-based fault monitoring system. The model is mainly derived from the first-principles of chemical and mechanical engineering in addition to empirical knowledge related to the behaviour of polymer and designed to provide a safe offline platform for developing monitoring and control systems. The fault monitoring system constructs a residual via parity relations, and a multi-objective optimisation problem must therefore be solved in order that the residual can be sensitive to faults but insensitive to disturbances and modelling errors. In this study, a genetic algorithm is exploited for solving this multi-objective optimisation problem, and the resulting fault monitoring system is applied to the model. Simulation results demonstrate that various types of faults can be detected and diagnosed successfully.