Dile Pontarolo Stremel, Eduardo Coselli Vasco de Toledo and Rubens Maciel Filho*

Faculty of Chemical Engineering

State University of Campinas (UNICAMP)

CP 6066, CEP 13081-970

Campinas, SP, Brazil.

Fax: +551937883910

email: maciel@feq.unicamp.br

Compared with growing studies of microorganism populations, few improvements on the development and application of structured models for product formation have appeared. This work presents an alternative dynamic structured model, adapted from a structured growth model and modified with the objective to simulate and to control a high productivity tower bioreactor. The original model of Sacharomyces cerevisae growth was adapted for ethanol production. Reduction techniques were applied in the spheric particle direction (pectin gel) and the reduced model was open loop simulated under disturbances in the inlet variables to define the control set-point and manipulated variable. Because of the slow dynamic, it was necessary to establish through "experimental planning" a relationship among the manipulated variable, feed substrate concentration and the control objective (feedforward model). The controller design involved in a SISO form, the classic (PID), predictive (DMC and GPC) and predictive adaptive (STDMC and STGPC) control algorithms were evaluated isolatedly and coupled to the feedforward model obtained with statistical procedures. The application of reduction techniques was useful to obtain a structured model for applications in simulation and control. The study of control showed that an early action relating to the manipulated variable and the set-point coupled to control feedback algorithms is the best form to quickly stabilize the process.