INT. J. BIOAUTOMATION, 2010, 14(1), 37-48 37 Stability Analysis оf Some Nonlinear Anaerobic Digestion Models Ivan Simeonov 1,* , Sette Diop 2 1 Institute of Microbiology, Bulgarian Academy of Sciences 26 Acad. G. Bonchev Str., 1113 Sofia, Bulgaria E-mail: issim@microbio.bas.bg 2 Laboratoire des Signaux et Systèmes Supélec, 91192 Gif sur Yvette cedex, France E-mail: diop@lss.supelec.fr * Corresponding author Received: March 05, 2010 Accepted: April 02, 2010 Published: April 15, 2010 Abstract: The paper deals with local asymptotic stability analysis of some mass balance dynamic models (based on one and on two-stage reaction schemes) of the anaerobic digestion (AD) in CSTR. The equilibrium states for models based on one (with Monod, Contois and Haldane shapes for the specific growth rate) and on two-stage (only with Monod shapes for both the specific growth rate of acidogenic and methanogenic bacterial populations) reaction schemes have been determined solving sets of nonlinear algebraic equations using Maples. Their stability has been analyzed systematically, which provides insight and guidance for AD bioreactors design, operation and control. Keywords: Anaerobic digestion, CSTR, Mathematical modelling, Mass balance non-linear dynamic models, Local stability analysis. Introduction Anaerobic digestion (AD) is an effective biotechnological process for treatment of different agricultural, municipal and industrial wastes [1]. It combines environmental depollution (ecological aspect) with production of renewable energy – biogas, which main component is methane (energetical aspect). AD is a very unstable process in regard to the biogas reactors (digesters) operation. This is due to the complicated interactions between different microbial species as well as of the complex transformations of the organic matter affected by a variety of environmental factors [1]. In this context use of mathematical models is a powerful tool for investigations, optimization and control of the anaerobic biodegradation [2-8]. More than 750 papers in this field are published during the last 30 years (the half of them – during the last 5 years). Depending on the aim of the modelling, a large number of modelling approaches are available. From a general point-of-view, they can be classified within three important classes: mass balance models, consisting of sets of non-linear ordinary differential equations (ODE) and describing more accurately all microbiological and biochemical phenomena during the AD processes; black box models (linear or non-linear); heuristic models, based on qualitative and fuzzy logic