This paper proposes to extend the dynamical analysis results on chemical tubular reactors presented in [5] to two linear distributed parameter models. These are in particular the linearized tangent models of two typical bioprocess models: a basic tubular bioreactor model, and a denitrifying biofilter model. The tools used for the analysis are those of the infinite dimensional system theory (e.g., [2]). In the paper we show the existence of solutions for the studied models, and emphasize stability conditions.

Biological denitrification in a fixed bed reactor is a process gaining popularity in the area of water treatment. The advent of more stringent criteria on nitrate and nitrite concentrations of treated water (along with the need to control and optimize the addition of a carbon source needed to denitrify) make the use of dynamical models able to represent the behaviour of the process not only of interest; it is also compulsory for optimizing control. The paper is concerned with the process dynamic modelling, the parameter identification and the control of the fixed bed biofilter with real-life experiments.

Biological denitrification in a fixed-bed reactor is increasingly popular for water treatment. With the advent of more stringent criteria for nitrate and nitrite concentrations in treated water and the need to control and optimize the addition of a carbon source (for denitrification purposes), dynamical models taking into account the process behavior, are of particular interest and become a necessity for control optimization. The control of biotechnological processes is highly complex because the behavior of microorganisms used for various purposes such as microbial growth, metabolite production or consumption of harmful substrates, remains relatively unknown. The aim of this paper is to design a robust control law capable of governing biochemical reactions characterizing the behavior of such reactors, thereby regulating the concentration of residual substrates at the reactor outlet, by acting either on the influent flow rate or on the supplied ethanol concentration.

Biological denitri®cation in a ®xed bed reactor is a process gaining popularity in the area of water treatment. The advent of more stringent criteria on nitrate and nitrite concentrations of treated water (along with the need to control and optimize the addition of a carbon source needed to denitrify) make the use of dynamical models able to represent the behaviour of the process not only of interest; it is also compulsory for optimizing control. The paper is concerned with the process dynamic modelling, the parameter identi®cation and the control of the ®xed bed bio®lter with real-life experiments. # 2000 Elsevier Science Ltd. All rights reserved.

This paper proposes to extend the dynamical analysis results on chemical tubular reactors presented in [3] to two typical bioprocess models : a basic tubular bioreactor model, and a denitrifying biofilter model. The tools used for the analysis are those of the infinite dimensional system theory (e.g. [2]). In the paper we show the existence of solutions for the studied models, and emphasize stability conditions.

Biological denitrification in a fixed bed reactor is a process gaining popularity in the area of water treatment. The advent of more stringent criteria on nitrate and nitrite concentrations of treated water make the use of efficient process control algorithms not only of interest but compulsory for process optimization. The paper is concerned with the design and real-life implementation of an adaptive linearizing controller of a pilot fixed bed biofilter.

This paper is concerned with the identification of the parameters of the distributed parameter model of a fixed bed denitrifying biofilter. It is twofold : first, the analysis of the structural identifiability in order to emphasize which parameters are a priori identifiable; then the identification of the model parameters by using experimental data.

In this paper we discuss the choice of sensor positions for a tubular reactor with a preset number of sensors. Different observability measures, based on the observability matrix, the observability gramian and the Popov-Belevitch-Hautus rank test are considered. The analysis is carried out on the reduced finite-dimensional model of the process. The results are put in perspective with the modal observability properties of the original infinite-dimensional model.

The parametric identification of the model of a fixed bed bioreactor is carried out by use of orthogonal functions. The studied model is nonlinear in the "variables and nonlinear in the parameters, but a hierarchical rewriting of the equations allows to perform a least square identification. The double Walsh orthogonal functions are used for this purpose.

This paper reports on the use of dynamic simulations for submerged packed biofilters. It sheds light on different aspects of modelling such non-conventional systems (biological reaction scheme, deep filtration, hydrodynamics of the filter) as well as on two numerical approaches for solving the resulting partial differential-algebraic equation system (a global approach using the finite difference method connected with a Gear's integration method for differential-algebraic equation systems and an orthogonal collocation approximation approach). Illustrative applications in the case of denitritation and denitrification with pilot plant data comparison sketch the interest of dynamic simulation.