Integrated internal model control (IIMC) of dissolved oxygen (DO) in bioreactor

Dissolved oxygen (DO) is one of the key elements that influence bioreactor performance and is essential for aerobic bacteria in bioreactors. Hence, any uncontrolled fluctuation in DO concentration may lead to problems in aerobic growth or a decrease in the efficiency of microorganism metabolism, eventually leading to detrimental production and undesirable by-products. The biological process is highly dynamic, making it challenging to design using a conventional PID controller, as any changes cause instability to the controller parameters. The study presented the development of a control strategy for dissolved oxygen (DO) in a bioreactor. In this study, the system identification toolbox in MATLAB was used to simulate a mathematical model of the process based on simulation data running from MINIFORS (MINIFORS, Infors). An artificial neural network (NN) with a Levenberg-Marquardt training algorithm and feed-forward back propagation network was used as a training method.