Introduction

DYMECOS (DYnamical  Microbial and Environmental eCOSystems) is an Inria Associated Team with Chile (2014-2015)

French team
    Principal investigator: Alain Rapaport

Chilean teams
    Principal investigator: Hector Ramirez

Other partners
  •     Escuela de Ingeniería Bioquímica, Pontificia Universidad Católica de Valparaíso (PUCV).
  •     Bioprocesses and Natural Resources Management research line of the Communication and Innovation Research and Innovation Center (CIRIC), INRIA Chile.
  •     Lemon team (INRIA/Univ. Montpellier II)


Left to right: Alain Rapaport, Hector Ramirez, Pedro Gajardo, Gonzalo Ruiz



Context of the cooperation

Natural or reconstituted microbial ecosystems are often very complex (high diversity, interactions within and between species, coupling with spatial processes: niches, aggregation, biofilm...). We bet that “simple models” (in the sense that they are manageable analytically and in a computer) of these ecosystems can explain their main functions, mainly concerning degradation and conversion, in a decision-making perspective (i.e. bioprocess control) without an exhaustive description of all microbial actors.

 

Our objectives is then develop, from expert knowledge and experimental observations made by micro-biologists, models that are simple enough to carry out the determination of "control laws", but realistic enough to be validated on real processes. One of the difficulties is to identify the limits of the validity of these models (especially in terms of population size, and of prediction of the coexistence of species). This requires necessarily a proper mathematical analysis, as well as the development of adapted simulation tools.

 

A recent trend focuses on the enhancement of secondary outputs of bio-processes. This is typically the case of the biogas production in the treatment of anaerobic solid waste-and wastewater. Practitioners look for strategies that drive the systems to a desired target in terms of both the degradation of the organic matter and the biogas production. This naturally leads to optimization problems in terms of design (e.g does a network of interconnected tanks run better than a single one?), and in terms of feedback control of a given process.


Scientific program: Modelling, Analysis and Simulation of Microbial Ecosytems and Natural Resources

Challenge 1: Parsimonious representations of spatial heterogeneity in liquid media for natural and industrial bioprocesses.

In the consideration of large volumes to be treated by bioprocesses (such as lake purification, landfills or micro-algae cultures...), the effects of spatial heterogeneity cannot be neglected  when the media are not perfectly mixed. We focus on the modelling with interconnected compartments along with the following objectives:

  • numerical comparison with faithful fluid dynamics models (based on numerical Navier-Stokes or St-Venant equations) and characterisation of reasonable approximations by a network of interconnected tanks,
  • extensions of the studies of the chemostat model to micro-algae dynamics, taking into consideration the light resource.
Challenge 2: Optimal control for the piloting of bioprocesses

The increasing concerns of valorisation of the production of biogas while conducting anaerobic digestion processes to a given target lead to revisit the problems of piloting bioreactors. Instead of the stabilization around a nominal point, we seek for optimal control strategies for maximizing the biogas production during transient or to reach as fast the best nominal set point. More precisely, we investigate:

  • optimal feedback strategies for bioprocesses with several interconnected tanks, or when the substrate is present in different forms (unsolubilized versus solubilized)
  • feasible sets with the help of Viability theory. Ideally, we also aim at characterizing “risky” zones, with “worst-case” approach (a.e. minimizing the “crisis” time).