Heuristiques optimisées et robustes de résolution du problème de gestion d'énergie pour les véhicules électriques et hybrides

Abstract : The system studied in this thesis is a hybrid electrical vehicle with two energy sources (fuel cell system and super-capacitor). The first goal is to minimize the fuel consumption whilst satisfying the requested power for each instant, taking into account constraints on the availability and the state of charge of the storage element. The system was modeled as a global optimization problem. The heuristics developped for obtaining the best power split between the two sources and the lower bound consumption computation proposed provide better results than those found in the literature. The second goal of the thesis is the study of the robustness of the solutions in order to minimize the worst-case consumption when perturbation happens or uncertainty is added to the input data. In this study the uncertainty concerns the power required for traction. The objective is to maintain the feasibility of solutions and limit the worst consumption that can happen due to a demand fluctuation. Dedicated heuristics are proposed for solving the identified robust variant of the problem, modeled as a Minimax problem. The solutions provided are less sensitive to the perturbations than the previous ones
Complete list of metadatas

Cited literature [152 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-01094409
Contributor : Jessica Toscano <>
Submitted on : Thursday, December 18, 2014 - 9:49:03 AM
Last modification on : Monday, April 29, 2019 - 3:50:49 PM
Long-term archiving on : Saturday, April 15, 2017 - 8:16:06 AM

Identifiers

  • HAL Id : tel-01094409, version 1

Collections

Citation

Mouloud Guemri. Heuristiques optimisées et robustes de résolution du problème de gestion d'énergie pour les véhicules électriques et hybrides. Traitement du signal et de l'image [eess.SP]. Institut National Polytechnique de Toulouse - INPT, 2013. Français. ⟨tel-01094409⟩

Share

Metrics

Record views

334

Files downloads

279