Nos tutelles

CNRS UPS Enseeiht

Rechercher




Accueil > Actualités

Integrated Design methods for Multi-energy Systems Under Uncertainty

par Laurence Laffont - publié le , mis à jour le

Hugo RADET’s thesis defense, intitled "Integrated Design methods for Multi-energy Systems Under Uncertainty", supervized by Xavier ROBOAM and Bruno SARENI, will be taking place on Thursday, March 3rd 2022 at 02:00pm in room C002 of the ENSEEITH.

Link for visio : https://inp-toulouse-fr.zoom.us/j/97026986286?pwd=L2Z6WlBISEo4cStsclNYczVQZkJTZz09

Reunion ID : 970 2698 6286
Passcode : 460480

Jury :

Mrs Florence OSSART, Rapporteur (GEEPS, Univ. Sorbonne)

Mr Robin ROCHE, Rapporteur (FEMTO-ST, UTBM)

Mr Bruno FRANÇOIS, Reviewer (L2EP, Centrale Lille)

Mr Pierre HAESSIG, Reviewer (IETR, CentraleSupélec)

Mr Maurizio REPETTO, Reviewer (PEIC, Politecnico di Torino)

Mr Vincent DEBUSSCHERE, Reviewer (G2Elab, Grenoble INP)

M. Xavier ROBOAM, Thesis supervizer (LAPLACE, Toulouse)

M. Bruno SARENI, Thesis supervizer (LAPLACE, Toulouse)

Abstract :

With the growing integration of variable renewable energy (VRE) sources into the conventional power grid, the concept of distributed energy systems (DES) has emerged : the energy is produced close to the end-user, and flexibility means are included in the system to supply energy demands at any time. Among the considered options, storage systems along with multi-energy strategies tend to be promising directions to mitigate the production variability by coupling the energy carriers with each other.
Planning the design of such systems is a challenging task because the problem displays multiple facets that are difficult for policy- and decision-makers to address in a systemic manner. Also, decisions are made while many parameters remain uncertain (e.g., future investment costs, energy prices, demands and production) as their values progressively unfold over time. Therefore, mathematical tools are often needed to provide decision support regarding several techno-economic requirements : the problem is usually expressed in the form of an optimization problem where decision variables are the sizes of the equipment.
This work addresses this issue by developing a generic framework to assess and compare different design and operation strategies for multi-energy systems. Then, three critical questions are tackled using this framework. In the first part of the thesis, the deterministic design model is built. Solving such a model is fast and allows running parametric analysis to assess the value of multi-energy systems and seasonal storage to supply residential customers with a high share of solar production.
Then, the second part of this work addresses the design of DES under uncertainty. To this end, two design methods based on stochastic programming are developed : one relies on mathematical programming and the other uses a metaheuristic algorithm. To solve the problem in a reasonable time, these methods are usually based on simplified versions of the problem. In particular, sizing values are computed assuming perfect foresight of the operation strategy for a given scenario. The main objective of this part is to challenge this hypothesis by jointly evaluating the design solutions with realistic operation policies which only have access to past and current information. In addition, this work aims at further investigating the close relationship between operation and design. Should the operation strategy used to design the system and the one used in real-time be strictly identical ? This part attempts to clarify this point.
Finally, the last part of this work deals with the dynamic design of DES. In this case, the model takes technology replacement due to aging into account, so multiple design decisions must be made over the horizon. Unlike the majority of studies, the optimization model includes the impact of the operation over system lifetimes : the latter are not fixed a priori, but they depend on the way technologies are operated over time. The aware aging method (which comes from the literature) is then compared with two heuristic design strategies based on single representative years.
All the previous methodological developments are applied to a DES which may include a set of hydrogen units (i.e., fuel cell, electrolyzer and storage tank) where the cogenerated heat can be recovered to supply thermal energy demands.