Observer-based output feedback linearizing control strategy for a nitrification–denitrification biofilter

The objective of the present work is to propose an output-feedback control scheme for a wastewater treatment biofilter, in order to regulate the outlet concentrations of nitrate and nitrite below a prescribed level (in accordance with the norms). The nitrification–denitrification biofilter is described by a set of mass balance partial differential equations (PDEs), of hyperbolic or parabolic type, depending on the consideration of diffusion/dispersion phenomena. The design of the output-feedback control strategy follows a late lumping approach, in which the PDE system is considered, and descretized only at a later stage for implementation purposes. This strategy is based on feedback linearization and requires a state estimator, which takes the form of a distributed parameter Luenberger observer. A point of particular attention is the formulation of the model boundary conditions, from the classical Dankwerts conditions to more advanced formulations including dynamic boundary conditions, which has a significant influence on the control performance. The control strategy is tested against model uncertainties and measurement noise in simulation.

► A nitrification–denitrification biofilter modeled by a set of PDEs is considered.
► We put attention in the formulation of the model boundary conditions.
► A late lumping approach is considered to synthesize an observer-based controller.
► Reference tracked when knowledge about the reaction rates appears in the control law.
► Diffusion term improves the controller performance.