Modeling and system identification of an unconventional bioreactor used for single cell protein production

In this study a model for an unconventional bioreactor, a U-loop fermentor is presented and compared to experimental data from a series of pH tracer pulse experiments. A good agreement between model and data, with very little error and computation time could be achieved when numerical diffusion combined with a carefully determined variable finite volume discretization method is employed to solve the system of partial differential model equations. The approach also overcame the obstacle posed by a constraint related to a complex pressure profile. An analysis of stability, controllability and observability if the reactor model is extended to include manipulated inputs, measured outputs and a complex validated bio-kinetic model revealed unstable process dynamics but good observability and controllability, even if very few sensors were functioning. A thorough investigation of the linearization effect suggested to perceive this result with caution but encouragement.