Strategies for automated statistical process control in food manufacturing

This research addressed a number of important issues involved in implementation of automated statistical process control for food manufacturing. Based on models developed for twin-screw food extrusion, various methods for determining control limits were analyzed for their applicability to correlated quality variables. Existing techniques to identify unknown or varying time delays in process dynamics were compared. The merits and limitations of each technique were given on the basis of computer simulations and experimental data from the food extrusion process. To facilitate implementation of fuzzy logic control, the equivalence between fuzzy logic and conventional controls were studied theoretically and experimentally. A tuning procedure was developed and experimentally tested for fuzzy logic control. An automated real-time statistical process control system was successfully designed and implemented. A machine vision system was developed to sample and measure quality attributes of an extruded product. Process adjustment actions were determined and implemented on-line based on automatic control techniques.