Optimal Control of Fedbatch Bioreactors by Iterative Dynamic Programming

Obtaining the optimal control policy of a fedbatch bioreactor is a difficult task due to the low sensitivity of the feed rate on the objective function. In this work, iterative dynamic programming was selected for identification of optimal control policies of two fedbatch bioreactors for citric acid production. In the first application the raw material is molasses, and in the second is whey. The kinetic models for these processes were obtained based on experimental data available in literature. Iterative dynamic programming (IDP) has proven to be one of the best techniques in establishing the optimal control policies for typical chemical engineering problems. For the proposed applications, the influence of several IDP parameters (e.g. the number of time stages, the number of state and control grids points) on the optimal solution was studied. The best values of these parameters are particular to each application. Key-Words: Optimal control, Iterative dynamic programming, Fedbatch bioreactors, Kinetic models, Citric acid