In-cycle Deterministic and Stochastic Dynamics of Extrusion Blow Molding

Abstract

Abstract Optimization of the thickness distribution in extrusion blow molded parts is crucial to minimize cycle times and resin usage. Closed-loop parison programming can be cycle-to-cycle or in-cycle. Corrective action occurs between cycles for the case of cycle-to-cycle control. However, this control scheme does not correct for high frequency disturbances that enter the system, such as periodic flowrate changes, noise from the gap position transducer, uneven melting and compression of the polymer and uneven inflation. Corrective action occurs during the cycle for in-cycle control. This allows for regulation of the high frequency disturbances. Deterministic in-cycle model identification was performed by obtaining the impulse responses of the bottle thickness distribution to pulses in the die gap input during the extrusion cycle. The response was found to be time variant along the individual cycle. Stochastic identification was also performed in order to ascertain a noise model for the system. A first order ARMA noise model was found to adequately describe the process. Control simulations were performed to determine the feasibility of in-cycle control. A minimum variance controller with a constraint for minimizing of manipulated variable movement was found to adequately regulate both a deterministic and a stochastic disturbance input during the extrusion cycle.