Increasing the dilution rate can globally stabilize two-step biological systems
Introduction
In most of continuous cultures, it is well known that increasing the dilution rate (or equivalently reducing the residence time inside the reactors) can destabilize the dynamics, in the sense that it enlarges the attraction basin of the wash-out equilibrium. This can be easily shown on the classical mathematical model of the chemostat, whatever the kinetics includes inhibition or not (see [1]). For growth inhibited by the substrate, bi-stability systematically occurs for large values of the input concentration of substrate. This feature has practical impacts on positive equilibrium (when it exists) because it cannot be globally stable, and the dynamics can conduct the system to the wash-out of the biomass, when the state belongs to the attraction basin of the washout equilibrium. Ways to guarantee a global stability is either to fix a lower dilution rate, which is penalizing for the performance of the process, or to control the dilution rate with a feedback loop, which temporarily diminishes the dilution rate when the state is far from the positive steady state [2], [3], [4], [5]. In any case, the removal has to be reduced at a certain stage.
Here we consider a more complex reaction scheme that are two-step systems, as met for instance in many models representing the anaerobic digestion process [6] or the nitrification process [7]. For these systems we show that there exist situations presenting a bi-stability for which increasing (and not decreasing) the dilution rate also leads the system to a globally asymptotically stable steady state, in opposition to classical stabilizing practices. In such a way, we can treat more matter per unit time during the transient than when decreasing the dilution rate.
In the paper, we denote by the set of non-negative numbers and by the set of positive numbers.
Let us consider the general mathematical model of a two-step mass-balance biological process, given by the following equations: where the parameter denotes the dilution rate.
This model is presented under the original form proposed in [6]. The first reaction involves a microbial species of concentration which grows on a substrate of concentration with a monotonic specific rate . The incoming flow fed the culture with substrate of concentration . The second reaction involves a second microbial species of concentration which grows on another substrate of concentration , with a specific growth rate denotes . This reaction is also fed by the first one which produces the second substrate. In addition, the incoming flow rate may contain (or not) substrate of concentration . The parameter reflects the fact that the effective dilution rate of the biomass is impacted by a retention inside the tank, differently to the abiotic resource. Here, the yield coefficients of the transformations of substrate into biomass (), and of the production of substrate for the second reaction by the first one, have been all kept equal to (this is always possible without any loss of generality, by a right choice of the concentration units).
In many biological systems, such as the anaerobic digestion or the nitrification processes, it is often reported in the literature that the second reaction is inhibited by large values of , which amounts to consider the following hypotheses.
Assumption 1 The functions , belong to and fulfill the following properties. is increasing on with . There exists such that is increasing on and decreasing on , with and .
The model (1) has a cascade structure: the first reaction is independent of the second one and the sub-system follows the classical (mono-specific) chemostat model. However, the sub-system is more complex to study as it receives substrate from the first reaction and is non-monotonic. This model and some of its variants has been already well studied in the literature [6], [8], [9], depending on the operating parameters . In particular, it has been shown that the dynamics may exhibit a multiple-stability, and the complete operating diagram has been established in [7], [10]. The purpose of the present work is to complement those studies, investigating how to adapt the value of the dilution rate to ensure a global stability of the dynamics. For sake of completeness, we first recall in the next section the set of possible asymptotic behaviors of the model.
Section snippets
Stability analysis
Let us first denote, for convenience,
We define the break-even concentration associated to the first reaction as the function
Then, we define the following quantity that is playing an important role in the analysis of the equilibria, as an “effective” input concentration for the second reaction.
As plotted in Fig. 1, we define also the break-even concentrations ,
Wash-out avoidance
In this section, we consider situations for which the attraction basin of equilibria with wash-out of biomass 1 or 2 or both is non empty. According to Proposition 1, this happens in cases (1) to (5). We study now how to play only with the value of the dilution rate to move to case (6).
Consider the domains which are the sets of operating parameters that
Simulations
Typical instances of functions that fulfill Assumption 1 are given by the Monod expression for the first reaction and the Haldane one for the second for which one has
Then, the break even concentrations defined in Section 2 have the expressions for the Monod function (9), and for the Haldane function (10)
Fig. 2 gives the values of the parameters chosen for the numerical
Conclusion
This study reveals the role played by the sum of the break-even concentrations, as the function , in the counter-intuitive phenomenon of increasing the dilution rate to stabilize a two-steps bio-process model. More precisely, we show that when this function is non-monotonic on its domain, this phenomenon occurs provided that the input concentration of substrate of the second reaction is null of not too large. This result provides a new way to stabilize such processes in
CRediT authorship contribution statement
J. Harmand: Realized the simulations and figures. A. Rapaport: Realized the simulations and figures.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
The work has been achieved in the framework of the PHC program ‘TOURNESOL’ 2018–19 between France and Belgium-Wallonia. The authors are very grateful to Professor Tewfik Sari for fruitful discussions.
References (17)
- et al.
Saturated PI control of continuous bioreactors with Haldane kinetics
Chem. Eng. Sci.
(2012) - et al.
Steady-state multiplicity of two-step biological conversion systems with general kinetics
Math. Biosci.
(2010) - et al.
Determination of appropriate operating strategies for anaerobic digestion systems
Biochem. Eng. J.
(2010) - et al.
Advanced control of anaerobic digestion processes through disturbances monitoring
Water Res.
(1999) - et al.
The Chemostat: Mathematical Theory of Microorganism Cultures
(2017) - et al.
On-Line Estimation and Adaptive Control of Bioreactors
(1990) - et al.
Control of unstable bioreactor using fuzzy-tuned PI controller
Bioprocess Eng.
(1999) - et al.
Robust regulation of a class of partially observed nonlinear continuous bioreactors
J. Process Control
(2002)
Cited by (3)
Global stability of a continuous bioreactor model under persistent variation of the dilution rate
2023, Mathematical Biosciences and EngineeringThe operating diagram for a two-step anaerobic digestion model
2021, Nonlinear Dynamics
- 1
Jérôme Harmand, Alain Rapaport and Denis Dochain have equally contributed to the results presented in the paper.