Modeling and control of microbial ecosystems and bioprocesses

Modeling "for understanding"

Modeling is widely used to better understand bioprocesses used in biorefinery, characterize waste (qualification and quantification of resources, fractionation of organic matter...) or to formalize the dynamics of systems studied at the LBE. A large range of models, without methodological a priori, is implemented (linear/non-linear, static/dynamic, stochastic/deterministic systems, in small or large populations, digital twins, machine learning...). An example of using a combination of modeling and mathematical analysis for investigating model properties and testing hypotheses is shown in Figure 1.
Modeling is also integrated to have virtual devices via simulation, especially to test and evaluate strategies and control laws (Figure 2).

Figure 1 : Qualitative mathematical analysis may help to test hypotheses about model structures. As long as specific processes such as for instance mortality, are not included in the model of an assemblage of N species, even in the presence of species interactions, the classical "recource-competition model" in batch shows that the "best" pure culture is always "more efficient" than any mixed culture [1].

Figure 2 : Model prediction and validation with the ALBA (ALgae-BActeria) model for a pilot-scale (3.8 m2) high rate algae-bacteria pond treating agricultural digestate [2].

Modeling "for the decision"

Besides models dedicated to a better understanding or to simulations of our systems, simpler models are used to solve optimization problems and for the synthesis of control laws. Coupled with available sensors, these control laws can be used to stabilize these dynamic systems as well as to optimize their operation.

Instrumentation

Our modeling activities and developments of optimization strategies for biotechnological systems are complemented by an important activity in the development and evaluation of new sensors, devices and data processing for matter characterization. In Figure 3, an example of the use of chemometrics is shown.
On these aspects, collaborations with actors of the economic sector allow a fast dissemination of research results and make it possible to provide answers to major challenges in biotechnologies.

Figure 3 : An example of chemometrics: estimation of the water content in waste using NIRS data, multivariate statistics and signal processing [3].