Output Feedback Linearization of Turbidostats after Time Scaling

Abstract

Turbidostats are a class of bioreactors gaining interest due to the recent availability of microscale and small-scale devices for characterization and scalingup of the biotechnological systems relevant in the biotech and pharma industries. The goal is to keep cell density constant in continuous operation. Thus, the control law, i.e., the substrate feeding strategy, must guarantee global or semiglobal convergence to an equilibrium point. However, their control is difficult due to the uncertain, time varying, and nonlinear nature of the processes involved. In this brief, we propose an adaptive control law that globally stabilizes the desired biomass setpoint. Furthermore, in a certain region of the state space, the controller linearizes the dynamic behavior after some time scaling. By this way, the orbits of the closedloop system are imposed by the designer. The intrinsic integral action of the gain adaptation rejects the parameter uncertainties. Moreover, the controller implementation only assumes the biomass concentration to be measured. Both the simulated and experimental results show the performance of the controller.