Adaptive linearizing control of bioreactors

We present developments concerning the synthesis of single-input, single-output (SISO) and multi-input, multi-output (MIMO) adaptive linearizing algorithms for the operation of bioreactors. Results are illustrated for the baker´s yeast fermentation process. The synthesis of these nonlinear control laws is performed by employing differential geometry techniques with system linearization by state feedback. The controller design includes a step of order reduction of the process state model. The adaptive feature comes from the online estimation of the required process time varying parameters. The adaptive algorithm proposed enforces a desired and pre-set second order convergence dynamics. Formulating the estimator on this basis leaves the user with the choice of two simple and intuitive tuning parameters with physical meaning-a damping coefficient and a natural period of oscillation.