Knowledge-based System for Supervision and Control of a Fed-Batch Fermentation of BtT

L. Valdez-Castro, A. Loza-Vásquez, J.M. Rodríguez-Morales, and J. Barrera-Cortés (Mexico)

Keywords

knowledge based systems, supervision and control, recurrent neural networks, fedbatch fermentation, Bacillus thuringiensis.

Abstract

This paper presents a knowledge based system (KBS) applied to control a fed-batch fermentation of Bacillus thuringiensis (Bt). Bt is a microorganism that synthesizes toxic proteins commonly used as pesticides to fight insects in immature stages. The yield and toxicity of proteins depend on the appropriate control of the operating conditions and other factors such as specific growth rate. The KBS is in charge of supervising the fulfilment of operating conditions and of controlling the specific growth rate of Bt. The latter is controlled by the manipulation of the nutrient feeding rate, using a neural network model able to predict the nutrient concentration in the fermentor. The KBS was carried out in the Labwindows/CVI environment and was validated reproducing a fed-batch fermentation of Bt. Results were satisfactory because it was possible to increase Bt production by 30% and protein toxicity by 10%.

Important Links:



Go Back