State observation and adaptive linearizing control for distributed parameter (bio)chemical reactors (Q2755391)

The observer and control algorithm for a non-isothermal biochemical process based on the so-called late lumping approach is proposed. From general mass and energy balance considerations, the dynamic nonlinear distributed parameter model of a tubular reactor with axial dispersion is deduced. Based on this model the asymptotic observer is developed; it reconstructs the evolution of the distribution of concentrations without involving any knowledge about process kinetics. The reaction rates are estimated on-line. A set of adaptive (in order to deal with the badly known kinetics) linearizing control algorithms are designed. The control scheme is implemented on a fixed-bed biofilter used for the treatment of drinkable water. Special experiments are carried out to illustrate the response of the controller and the observer to various practical stimuli.