Numerical Methods Based Strategy and Particle Filter State Estimation For Bio Process Control

OLGA QUINTERO MONTOYA; GUSTAVO SCAGLIA; FERNANDO DI SCIASCIO; VICENTE MUT

China

Congreso; IEEE INTERNATIONAL CONFERENCE ON INDUSTRIAL TECHNOLOGY,IEEE ICIT2008; 2008

IEEE INTERNATIONAL CONFERENCE ON INDUSTRIAL TECHNOLOGY,IEEE ICIT2008

Biotechnological processes represents a challenge into the control field, due to its highly non linearity. Particularly, continuous alcoholic fermentation by Zymomonas Mobilis presents high ethanol performance, but it shows highly oscillatory behavior in process variables, because of the influence of inhibition dynamics (rate of ethanol concentration Z and I) over Biomass, Substrate and Product concentrations; it is a significantly challenge from control perspective. In this work, a new and significantly solution for control of biotechnological variables into the fermentation process, based on numerical methods and linear algebra is proposed, also a state estimator of Biomass into the control loop is presented, using to reach this aim, the Non Linear Filtering based on the recursive application of Bayes rule and Monte Carlo techniques. Specifically, the recursive Bayesian Filters SIR (Sampling Importance Resampling) are used, including different kinds of resampling. This methodology, allows to develop the controller design trough the use of dynamic analysis and without the use of complex calculations and the use of a tested state estimators for Biomass in Z.M