Estimation de l’état interne d’une batterie lithium-ion à l’aide d’un modèle électrochimique

Abstract : Developed in the nineties, lithium batteries have colonized our environment in less than thirty years and they keep spreading faster and faster. Powerful, efficient, light and compact, this technology remains hazardous. In order to limit the danger and slow the aging of lithium cells, most of such batteries embed a management system. The latter needs to access some internal states, which are not directly measurable. This thesis intends to estimate these variables using a nonlinear observer, which is based on an electrochemical model. The behavior of the battery is driven by the transportation phenomenon of its main electrochemical species. We therefore built a finite dimensional electrochemical model of these adapted to estimation. It relies on the spatial discretization of the partial differential equations, which describe these transportation phenomena. It also formulates some assumptions, such as the fact that an electrode globally behaves like a single particle of its active material. The obtained state space model has affine dynamics and a nonlinear output. Among the existing observers for such systems that we are aware of, none can be applied directly to the developed model. Hence, we developed new ones whose stability is guaranteed provided a linear matrix inequality holds, which is used to construct the observation gain. We then confront these observers to experimental data acquired on commercialized batteries. The obtained results are encouraging and the observer seems to be a fair compromise between physical meaning and numerical complexity
Complete list of metadatas

Cited literature [58 references]  Display  Hide  Download

https://hal.univ-lorraine.fr/tel-02096451
Contributor : Thèses Ul <>
Submitted on : Thursday, April 11, 2019 - 12:51:23 PM
Last modification on : Thursday, July 18, 2019 - 10:40:21 AM
Long-term archiving on : Friday, July 12, 2019 - 1:36:09 PM

File

DDOC_T_2019_0008_BLONDEL.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-02096451, version 1

Collections

Citation

Pierre Blondel. Estimation de l’état interne d’une batterie lithium-ion à l’aide d’un modèle électrochimique. Automatique / Robotique. Université de Lorraine, 2019. Français. ⟨NNT : 2019LORR0008⟩. ⟨tel-02096451⟩

Share

Metrics

Record views

103

Files downloads

270