Skip to Main content Skip to Navigation
Theses

Modélisation de la croissance des plantes en interaction avec la ressource en eau et contrôle optimal de l'irrigation

Abstract : This thesis is part of the project Digiplante in the MAS laboratory of Ecole Centrale Paris, in collaboration with the research unit G-water in the IRSTEA. In a context of scarcity of water resources, it is important to better promote the use of water in agricultural production. When designing decision support tools for farmers or farm managers, it is necessary to develop a quantitative approach to describe the interaction of water resources with crop production. In this context, the objective of this thesis is to propose a multidisciplinary methodology based on modeling the soilplant- atmosphere interaction with coupled models and on the strategy optimization for managing irrigation. The first contribution of this thesis is the development and study of plant-soil coupled models at different scales. In comparison with the Pilote model that simulates the crop yield based on the harvest index and the empirical description of the leaf area index LAI, we proposed two other levels of modeling for the plant part and their coupling with the soil module of Pilote. On the one hand, we have introduced a compartment model which allows both a finer management of the allocation and better adaptation to the experimental data. On the other hand, the coupling was also carried out with the GreenLab model which describes the architectural growth of the plant. This has necessitated the development of a continuous version of GreenLab which facilitates in particular the integration of the effect of water stress. A set of experimental data has allowed the parametrization of these models under different irrigation regimes and thus different levels of water stress. The second major contribution was the development of a methodological framework for the optimal control of water supplies in both deterministic and stochastic case. The proposed resolution is based on dynamic programming that can handle different optimization criterion either without or under constraints (for example regulatory or logistical constraints, such as irrigation quotas). An iterative version of the dynamic programming algorithm has also been proposed to improve the performance of the numerical solution, but also in order to deal with dynamical systems with delay (which can occur in some formulations of senescence in growth models). Our generic methodology could be applied to concrete problems, and allows the development of decision support tools in agriculture for a more efficient water resources management.
Document type :
Theses
Complete list of metadatas

Cited literature [234 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00909261
Contributor : Abes Star :  Contact
Submitted on : Tuesday, November 26, 2013 - 10:22:09 AM
Last modification on : Wednesday, October 14, 2020 - 3:55:03 AM
Long-term archiving on: : Thursday, February 27, 2014 - 6:55:14 AM

File

these_zhongping-li.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-00909261, version 1

Collections

Citation

Zhongping Li. Modélisation de la croissance des plantes en interaction avec la ressource en eau et contrôle optimal de l'irrigation. Autre. Ecole Centrale Paris, 2013. Français. ⟨NNT : 2013ECAP0003⟩. ⟨tel-00909261⟩

Share

Metrics

Record views

879

Files downloads

22557