Polymer extrusion production control using active recognition and adaptive control system

An active recognition and adaptive control (ARAC) system for a polymer extrusion production line is proposed. The objective of the ARAC is to emulate not only the control experience of the human expert, but also the strategies or thinking of the expert in monitoring and controlling real-time operating systems. An ARAC consists of two interactive models, namely, pattern analysis and adaptive control, both using fuzzy logic technology. The pattern probe of the pattern analysis model is used to interact with the system actively to recognize its operating state or mode. In the multilayer rule architecture of the adaptive control model, the upper-layer rule is capable of adjusting the lower-layer rules and membership functions according to system dynamic characteristics