Fed-batch Optimization of PHB Synthesis through Mechanistic, Cybernetic and Neural Approaches

Abstract

Despite its superiority over chemically synthesized petroleum-based polymers, poly-?-hydroxybutyrate (PHB) has been less successful commercially. A prime reason is the low productivity of microbial processes for PHB. High fermentation efficiency requires good modelling and optimization. Neither classical mechanistic models nor the recent cybernetic models have resulted in sufficiently high yields of PHB. So a neural network description has been proposed here. Relative to the other two approaches, neural optimization doubled the maximum PHB concentration in fed-batch fermentation with Ralstonia eutropha, the most commonly employed organism for PHB production, and it consumed less of the substrates. This advantage and their model-free nature make neural networks an attractive technique to enhance PHB productivity.

Keywords

• PHB
• Fed-batch fermentation
• Neural kinetics
• Optimum dispersion