Speaker: Ayoub Belhachmi
Date: Thursday 13th of November 2025, 1:15pm
Abstract:
Generating a valid mesh from geological interfaces is a difficult task due to the complexity of the geometrical configurations encountered in geosciences. Furthermore, uncertainties regarding the locations or even the existence of some geological discontinuities call for robust local mesh updating techniques. These methods allow for the local update of the model, instead of remeshing the entire model.
This seminar will explore techniques for local updating of tetrahedral meshes, with a particular focus on the insertion of finite surfaces such as faults in existing multi-material meshes. I will present recent progress in fault insertion using the open-source remeshing library MMG. In this approach, the finite surface to be inserted is defined as the intersection of two level set functions: one describing the surface itself, and another describing its boundary. The two level set functions can be obtained via interpolation, subject to a smoothness criteria. In the second part of the seminar, I will introduce an anisotropic smoothing energy for geological data interpolation. This high-order smoothing energy is discretized using linear finite elements via a mixed finite element formulation.
- Details
- Category: Seminar
Speaker: Nour Alawieh
Date: Thursday 29th of January 2026, 1:15pm
Abstract:
Geothermal reservoirs are often hosted in highly fractured porous rocks, where fractures control fluid flow while the surrounding matrix acts as the main heat storage. Modeling coupled fluid flow and heat transfer in such systems is challenging due to the strong contrast between fracture and matrix properties and the geometric complexity of fracture networks. Existing modeling approaches involve a trade-off between accuracy and computational efficiency: implicit models rely on upscaled properties and are computationally efficient but may overlook local fracture–matrix interactions, whereas explicit models such as Discrete Fracture Matrix (DFM) provide high accuracy at the expense of high computational cost.
In this seminar, I present an efficient hybrid modeling framework for flow and heat transfer in fractured porous media, with applications to geothermal energy. A fully explicit DFM model is first introduced as a reference. Then, a Discrete Fracture Network–Dual Porosity (DFNDP) model is proposed in this work, in which fluid flow is restricted to the fracture network while heat exchange with the surrounding matrix is represented through a semi-empirical exchange coefficient. The DFNDP model is validated against the DFM reference under various flow conditions and fracture densities. Results show that the hybrid approach accurately reproduces DFM heat transfer simulations, particularly in advection-dominated and highly fractured systems, while reducing computational cost significantly. These results indicate that the DFNDP model provides a reliable and efficient alternative for simulating heat transfer in fractured geothermal reservoirs.
- Details
- Category: Seminar
Speaker: Erwan Gloagen
Date: Friday 7th of November 2025, 1:15pm
Abstract:
Le site des lagunes de Mercier, Québec, Canada, est un site contaminé géré par pompage et traitement. Le gouvernement du Québec souhaite faire une rénovation de l’usine et en profiter pour optimiser les débits de pompage. Pour ce faire, nous avons effectué une analyse des données existantes et proposé ensuite une série de mesures ciblées : sismique réflexion, forages, diagraphies et slug tests multi-niveaux. L’ensemble de ces données a permis de réalise une modélisation stochastique des propriétés hydrauliques de l’aquifères et de le caler sur les données en régimes permanent et transitoire. Ces jumeaux numériques permettent de tester différents scénarios et de les classer selon des critères environnementaux, économiques et sociétaux.
- Details
- Category: Seminar
Speaker: Charles Dapogny
Date: Thursday 22nd of January 2026, 1:15pm
Abstract:
L’optimisation de formes est une discipline au confluent des mathématiques, de la physique et du calcul scientifique, qui suscite un engouement croissant au sein des milieux académique et industriel. En quelques mots, il s’agit d’optimiser une fonction objectif, dépendant de la “forme" (qui suivant les applications peut représenter une structure mécanique, un domaine fluide, etc.), sous certaines contraintes. Dans les applications, ces fonctions dépendent de la physique en jeu par l’intermédiaire de la solution d’équations aux dérivées partielles posées sur la forme. Ces problèmes soulèvent de nombreuses difficultés spécifiques, liées par exemple au calcul des “dérivées" des critères d’optimisation par rapport au domaine, à la représentation numérique de la forme, etc.
L’objectif de cette présentation est de brosser un panorama succinct (et biaisé) de ce domaine en pleine ébullition. Une première partie traitera de quelques généralités et présentera les principaux paradigmes d’optimisation de formes, à savoir l’optimisation paramétrique (ou contrôle optimal), l’optimisation géométrique, et l’optimisation topologique. On introduira ensuite quelques ingrédients théoriques pour l’étude des problèmes d’optimisation de formes, et notamment la méthode de l’état adjoint permettant de calculer les dérivées des critères d’optimisation par rapport au design et de leur donner une structure “exploitable" en pratique. Dans une troisième partie, on discutera de quelques points essentiels touchant à l’implémentation numérique de ces méthodes. Notamment, on expliquera comment décrire la forme optimisée et son évolution au cours du processus d’optimisation. On présentera finalement quelques applications récentes de ces techniques en mécanique des structures ou bien en mécanique des fluides.
- Details
- Category: Seminar
Speaker: Roberto Cilli
Date: Thursday 25th of September 2025, 1:15pm
Abstract:
The seminar will first discuss a downstream application of a pretrained SentenceBERT + Random Forest (RF) for the lithological classification of texts from borehole logs. A custom lightweight architecture relying on a single Transformer module is proposed to handle contextual and positional relationships of each lithological text description. The proposed method achieves an accuracy gain of approximately 10% compared to a RF fed with SentenceBERT embeddings. Finally, a comparison between benchmark uncertainty quantification (UQ) algorithms, including Bayesian NN (Blundell et al. 2015), MAPIE Conformal Learning (Taquet et al. 2023) and Deep Ensemble (Lakshminarayanan et al. 2016) is shown. Preliminary results indicate that the Deep Ensemble UQ method seems the most reliable while still feasible in low-resource computing environments.
- Details
- Category: Seminar
Speaker: Zhixiang Guo
Date: Thursday 8th of January 2026, 1:15pm
Abstract:
In recent years, deep learning has achieved transformative progress across many domains, with foundation models demonstrating remarkable generalization under few-shot and even zero-shot settings. However, training such models from scratch is often impractical in geophysics due to limited labeled data and constrained computational resources. To address this, we investigate an efficient adaptation strategy that fine-tunes foundation models pretrained in other domains for geophysical tasks, aiming for a cost-effective pathway to leverage large models. Within our adaptation framework, the adapted models consistently exhibit stronger generalization and improved performance over conventional deep learning baselines on multiple small, labeled datasets. These findings provide new insights for deploying foundation models in geophysics. This seminar also surveys recent advances in deep learning for seismic interpretation and inversion, and reports my ongoing progress on diffusion-based implicit structural modeling jointly constrained by faults and horizons.
- Details
- Category: Seminar