Qualitative Analysis of a Three-Stage Anaerobic Digestion Model with Microbial Decay
Radhouane Fekih-Salem · 2026
Read Qualitative Analysis of a Three-Stage Anaerobic Digestion Model with Microbial Decay by Radhouane Fekih-Salem (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-18818.txt.
Introduction
Fekih-Salem extends a three stage anaerobic digestion model by adding microbial mortality to hydrolysis. The model covers hydrolysis, acidogenesis, and methanogenesis, each with its own decay rates, and captures substrate inhibition. Two hydrolysis forms are studied, a first order form and a biomass based form, with the latter splitting washout from mortality. The author proves the solutions are well posed, above zero, and bounded, finds all equilibria, and checks their stability.
Catalog record
| CREATOR | Radhouane Fekih-Salem |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37936 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.18818
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Qualitative Analysis of a Three-Stage Anaerobic Digestion Model with Microbial Decay Radhouane Fekih-Salem Abstract We extend a three-stage anaerobic digestion model by incorporating microbial mortality into the hydrolysis of particulate organic matter. The model describes hydrolysis, acidogenesis, and methanogenesis, each with distinct dilution and decay rates, and accounts for non-monotonic growth in order to capture substrate inhibition. Two hydrolysis mechanisms are considered: a first-order formulation and a biomass-dependent one. The latter distinguishes hydraulic washout from intrinsic mortality and leads to different persistence conditions. For both models, we establish well-posedness and prove the positivity and boundedness of solutions. We then characterize all equilibria and analyze their local stability with respect to key operating parameters. The inclusion of microbial mortality provides a more general and biologically relevant framework, thereby enriching the qualitative dynamics compared to the classical AM2 model. 1 Introduction Anaerobic digestion (AD) is a biological process used for the treatment of organic wastes, with the additional benefit of producing renewable energy in the form of biogas. During this process, complex organic matter is successively transformed through hydrolysis, acidogenesis, and methanogenesis into methane and carbon dioxide. Despite its wide applications, AD remains a complex process due to the coupling between biological and physicochemical phenomena and the difficulty of model calibration. To address this complexity, several mathematical models have been proposed, ranging from detailed to reduced formulations [ 2 , 10 ] . Reduced models, in particular, aim to capture the essential dynamics while remaining mathematically tractable. In this context, simplified models involving two or three microbial populations have been widely studied [ 8 , 10 ] . When each microbial growth rate depends only on its own substrate, the system exhibits a cascade structure and describes a commensalistic relationship, in which one species benefits from another without affecting it [ 1 , 2 , 3 , 7 , 9 ] . In contrast, when the growth of one species depends on an intermediate substrate produced by another, the model captures a syntrophic relationship, in which both populations are interdependent for their survival [ 5 , 6 , 8 , 10 ] . These models exhibit rich dynamical behaviors, including coexistence equilibria, multistability, and oscillations arising from nonlinear interactions. In particular, the emergence of limit cycles through Hopf bifurcations has been reported in [ 10 ] , and further investigated in [ 5 , 6 ] by incorporating additional effects such as inhibition and microbial decay. These results highlight the importance of biological mechanisms and operating conditions in shaping the qualitative behavior of AD systems. In this work, we revisit this three-stage framework by incorporating microbial mortality, which leads to a more realistic description of population dynamics and significantly affects persistence and stability properties. The originality of this study lies in the combined consideration of mortality effects and non-monotonic growth kinetics, as well as in the comparison of two modeling approaches for the hydrolysis step. More precisely, we consider a chemostat model describing the sequential transformation of a particulate substrate X_{0} into a soluble substrate S_{1} through hydrolysis, followed by its conversion into intermediate products S_{2} by acidogenic bacteria X_{1} , and finally into biogas by methanogenic bacteria X_{2} . The system operates in a continuous reactor of constant volume with dilution rate D . Two hydrolysis mechanisms are investigated. The first assumes a constant enzymatic activity, leading to first-order kinetics r_{0}=k_{\rm hyd}X_{0}, where hydrolysis depends only on the substrate concentration. The second explicitly involves the acidogenic biomass, resulting in biomass-dependent kinetics r_{0}=\mu_{0}(X_{0})\,X_{1}, which allows a clear distinction between hydraulic washout and intrinsic mortality, and may induce different persistence thresholds for the microbial populations. The main objective of this paper is to analyze how the inclusion of microbial mortality and the choice of hydrolysis modeling affect the qualitative behavior of the system. In particular, we study the existence and stability of equilibria and highlight the differences between models with and without an explicit hydrolytic microbial compartment. The paper is organized as follows. Section 2 introduces the general mathematical model. Sections 3 and 4 are devoted to the analysis of the models without and with a hydrolytic microbial compartment, respectively. Section 5 concludes the paper. Technical proofs related to the well-posedness of the model, as well as the positivity and boundedness of solutions, are given in Appendix 0.A.1 . Further analytical results concerning the existence and multiplicity of equilibria for the model with a microbial hydrolytic compartment, together with their local stability analysis, are provided in Appendices 0.A.2 and 0.A.3 , respectively. 2 Mathematical model To analyze the effects of microbial mortality and the hydrolysis mechanism on anaerobic digestion, we consider the following chemostat model describing the sequential transformation of substrates by microbial populations: \left\{\begin{aligned} \dot{X}_{0}&=DX_{0}^{in}-\alpha_{0}DX_{0}-r_{0},\\ \dot{S}_{1}&=D(S_{1}^{in}-S_{1})+k_{0}r_{0}-k_{1}\mu_{1}(S_{1})X_{1},\\ \dot{X}_{1}&=(\mu_{1}(S_{1})-D_{1})X_{1},\\ \dot{S}_{2}&=D(S_{2}^{in}-S_{2})+k_{2}\mu_{1}(S_{1})X_{1}-k_{3}\mu_{2}(S_{2})X_{2},\\ \dot{X}_{2}&=(\mu_{2}(S_{2})-D_{2})X_{2}.\end{aligned}\right. (1) Here, X_{0} denotes the slowly biodegradable substrate, S_{i} the soluble substrates, and X_{i} ( i=1,2 ) the microbial populations. The functions \mu_{i} represent the specific growth rates, S_{i}^{in} are influent concentrations, and D is the dilution rate. The effective removal rates of microbial populations account for both hydraulic washout and intrinsic mortality: D_{i}=\alpha_{i}D+a_{i},\qquad i=1,2, (2) where \alpha_{i}\in[0,1] represent fractions subject to washout, and a_{i}\geq 0 are the intrinsic mortality rates. This formulation decouples the hydraulic retention time {\rm HRT}=1/D from the solid retention time {\rm SRT}=1/(\alpha_{i}D) [ 1 , 9 ] . The coefficients k_{i} , i=0,\ldots,3 , denote the yield factors, constrained by 1\geq k_{0},\quad k_{1}\geq 1+k_{2},\quad k_{3}\geq 1, ensuring mass conservation. Two hydrolysis mechanisms are considered: r_{0}=k_{\rm hyd}X_{0}\quad\text{(first-order)},\qquad r_{0}=\mu_{0}(X_{0})X_{1}\quad\text{(biomass-dependent)}, allowing a comparison of models with and without an explicit hydrolytic microbial compartment, and capturing differences between washout and intrinsic mortality effects. The growth functions satisfy the following general assumptions (continuously differentiable, \mathcal{C}^{1} ): H1: \mu_{1}(S_{1}) is monotone increasing with \mu_{1}(0)=0 . H2: \mu_{2}(S_{2}) is non-monotone, with \mu_{2}(0)=\mu_{2}(+\infty)=0 ; there exists S_{2}^{m}>0 such that \mu_{2}^{\prime}(S_{2})>0\text{ for }0<S_{2}<S_{2}^{m},\quad\mu_{2}^{\prime}(S_{2})<0\text{ for }S_{2}>S_{2}^{m}. H3: \mu_{0}(0)=0 , \mu_{0} is increasing and concave on \mathbb{R}_{+} . Assumption H3 is satisfied, for example, when \mu_{0} is linear or of Monod type. As a preliminary step, we establish that the model is biologically well-posed, in the sense that all concentrations remain nonnegative and bounded over time. Proposition 1 All solutions of system ( 1 ) with nonnegative initial conditions remain nonnegative and bounded for all t\geq 0 . Let Z=k_{0}X_{0}+S_{1}+S_{2}+(k_{1}-k_{2})X_{1}+k_{3}X_{2} denote the total mass density in the chemostat, and define the corresponding input quantity by Z^{in}=k_{0}X_{0}^{in}+S_{1}^{in}+S_{2}^{in} . Then, the set \Omega=\left\{(X_{0},S_{1},X_{1},S_{2},X_{2})\in\mathbb{R}_{+}^{5}:Z\leq\tfrac{D}{D_{\min}}Z^{in}\right\} is positively invariant and is a global attractor for system ( 1 ), where D_{\min}=\min(\alpha_{0}D,D_{1},D_{2}) . 3 Model without a hydrolytic microbial compartment We consider the case where hydrolysis follows first-order kinetics r_{0}=k_{\rm hyd}X_{0}, corresponding to a model without an explicit hydrolytic microbial compartment. In this case, the dynamics of X_{0} are decoupled from the rest of the system and satisfy a linear equation. Hence, X_{0} globally converges to X_{0}^{*}=\tfrac{D}{k_{\rm hyd}+\alpha_{0}D}X_{0}^{in}. (3) As a consequence, the asymptotic behavior of system ( 1 ) is governed by the reduced four-dimensional subsystem in (S_{1},X_{1},S_{2},X_{2}) . At equilibrium X_{0}=X_{0}^{*} , system ( 1 ) reduces to the following AM2-type model [ 1 , 9 ] : \left\{\begin{aligned} \dot{S}_{1}&=D(S_{1}^{in*}-S_{1})-k_{1}\mu_{1}(S_{1})X_{1},\\ \dot{X}_{1}&=(\mu_{1}(S_{1})-D_{1})X_{1},\\ \dot{S}_{2}&=D(S_{2}^{in}-S_{2})+k_{2}\mu_{1}(S_{1})X_{1}-k_{3}\mu_{2}(S_{2})X_{2},\\ \dot{X}_{2}&=(\mu_{2}(S_{2})-D_{2})X_{2},\end{aligned}\right. (4) where the effective input concentration of S_{1} is S_{1}^{in*}=S_{1}^{in}+\tfrac{k_{0}k_{\rm hyd}}{k_{\rm hyd}+\alpha_{0}D}X_{0}^{in}. (5) Each equilibrium F=(S_{1}^{*},X_{1}^{*},S_{2}^{*},X_{2}^{*}) of system ( 4