En poursuivant votre navigation sur ce site, vous acceptez le dépôt de cookies dans votre navigateur. (En savoir plus)

PhD Student M/F : Distributed Optimal Control for Complex Systems

This offer is available in the following languages:
Français - Anglais

Date Limite Candidature : jeudi 18 août 2022

Assurez-vous que votre profil candidat soit correctement renseigné avant de postuler. Les informations de votre profil complètent celles associées à chaque candidature. Afin d’augmenter votre visibilité sur notre Portail Emploi et ainsi permettre aux recruteurs de consulter votre profil candidat, vous avez la possibilité de déposer votre CV dans notre CVThèque en un clic !

General information

Reference : UMR5269-NADMIC-016
Workplace : GRENOBLE
Date of publication : Thursday, July 28, 2022
Scientific Responsible name : DELINCHANT Benoit
Type of Contract : PhD Student contract / Thesis offer
Contract Period : 36 months
Start date of the thesis : 1 October 2022
Proportion of work : Full time
Remuneration : 2 135,00 € gross monthly

Description of the thesis topic

Context and Objectives:
The wide deployment of connected and controllable assets at the individual scale incurs new challenges in the operation and management of legacy systems in which they are integrated. Thus, many infrastructures can be considered as system of systems or complex systems in short. In such frameworks, decisions taken at the individual level may have a significant impact on the overall system's behaviour. It is especially true for all the networked environments such as energy infrastructures or transportation. As an example, the push for small scale renewable based generation endanger the operations of the macroscopic electrical grid with additional uncertainty and variability. Flexibility shall then be leveraged at the individual scale with an appropriate coordination that ultimately ensures the global balance between the supply and demand of energy.
The objective of this project is the implementation of a methodology for distributed control in such complex systems. Mathematically, an equilibrium shall be reached at the global system scale while taking into account individual objectives and potential constraints. A specific attention will be paid to the information and coordination signals exchanged between the different stakeholders. The developed methods will highly rely on distributed optimization, game theory and iterative processes.
At the individual system scale, we will adapt a Hybrid-AI based optimization methodology for local decision making.
From traditional controllers to Hybrid AI controller:
Conventional management strategies for power and energy systems rely on Model-Based Optimal Control to compute the best operation profiles of heterogeneous equipments (e.g. generators, energy storage, adjustable loads, etc) to minimize sets of technical, economical or environmental objectives. Those approaches suffer from sensitivity to uncertainties on the model approximations and parameters along with the prediction errors of environment variables such as load or renewable generation profiles. The Hybrid AI approach lies on the use of a coarse controller to generate the main management actions - e.g. : rule based strategy to schedule the charge/discharge profile of an energy storage system in a building/microgrid. Such predictive controllers are also computing the expected system state that could be room temperatures, voltage levels or battery state of charge. Once the setpoints are sent to the controlled systems, it is then possible to measure the deviations between the actual state values and the prediction. Those deviations can then be used after an appropriate training phase to adjust the main management actions and reach better system performances.

Work Context

The G2Elab Laboratory - UMR5269 - is a joint research unit attached to the CNRS, Grenoble INP and Grenoble Alpes University, located on the scientific campus of Grenoble, in the GreEn-ER building, where it occupies 5,000 m². It also has three other sites on which three experimental platforms hosting research teams are deployed: Minatec, the Laboratory of Magnetic Metrology in Weak Fields in Herbeys and several experimental rooms in the CNRS polygon. The Grenoble Electrical Engineering Laboratory (G2Elab) is composed of 60 permanent staff, 40 ITAs, 120 PhD students and 80 interns, post-docs and visiting researchers.
The successful candidate will be registered for a 3-year PhD, the candidate will receive a double PhD degree from the University of Grenoble Alpes jointly with Nanyang Technological University (NTU), Singapore.

Constraints and risks

Applicants should hold a MsC in Applied Mathematics or Computer Science
Knowledge and strong interest in optimal control and/or optimisation and/or Machine Learning will be appreciated.
Simulation/modelling and development tools – Matlab, Python, Java, C, etc; Mathematical Language Programming and experience in optimization solvers would be a plus – GAMS, Julia, YALMIP, CPLEX, GUROBI, GLPK, among others.
Strong analytical and communication skills, ability to present clearly and concisely. Good writing and oral skills for efficient communication in international conferences and publications in scientific journals.
Ability to work in an international environment, learn from experienced researchers and transfer knowledge.

We talk about it on Twitter!