Sitemap

This contribution presents a new method designed to populate unstructured, adaptively refined tetrahedral finite element meshes with highly resolved petrophysical properties from a commercial property modeling tool.

Recommended citation: Mosser, L. J. (2013, September). Mapping Geo-Cellular Models to Unstructured Simulation Grids. In SPE Annual Technical Conference and Exhibition. Society of Petroleum Engineers. https://www.onepetro.org/conference-paper/SPE-167629-STU

Tesselations Stable under Iteration

Published in Discrete Fracture Network Engineering 2014, 2014

This contribution presents a novel stochastic model based on tessellations stable under iteration also referred to as STIT. Such models naturally incorporate important features like fracture truncation and the possibility to extend their definition to arbitrary orientation distributions.

Reconstruction of three-dimensional porous media using generative adversarial neural networks

Published in Physical Review E, 2017

We present a method to reconstruct the solid-void structure of porous media by applying a generative neural network that allows an implicit description of the probability distribution represented by three-dimensional image data sets.

Recommended citation: Mosser, L., Dubrule, O., & Blunt, M. J. (2017). Reconstruction of three-dimensional porous media using generative adversarial neural networks. Physical Review E, 96(4), 043309.' https://link.aps.org/pdf/10.1103/PhysRevE.96.043309

Reconstruction of Three-Dimensional Porous Media: Statistical or Deep Learning Approach?

Published in Book Chapter: Statistical Data Science, 2017

We compare covariance-based approaches to deep generative models for pore-scale image reconstruction.

Recommended citation: Mosser, L., Le Blévec, T., & Dubrule, O. (2018). Reconstruction of Three-Dimensional Porous Media: Statistical or Deep Learning Approach?. Statistical Data Science, 125. https://www.worldscientific.com/doi/abs/10.1142/9781786345400_0008

Well-data-based discrete fracture and matrix modelling and flow-based upscaling of multilayer carbonate reservoir horizons

Published in Geological Society, London, Special Publications, 2017

Discrete fracture and matrix (DFM) homogenization, simultaneously capturing reservoir layers and contained fractures, is an alternative to discrete fracture network (DFN) upscaling. Here, the DFM approach was applied to a fractured carbonate reservoir.

Recommended citation: Milliotte, C., Jonoud, S., Wennberg, O. P., Matthäi, S. K., Jurkiw, A., & Mosser, L. (2018). Well-data-based discrete fracture and matrix modelling and flow-based upscaling of multilayer carbonate reservoir horizons. Geological Society, London, Special Publications, 459(1), 191-210. http://sp.lyellcollection.org/content/459/1/191

Stochastic reconstruction of an oolitic limestone by generative adversarial networks

Published in Transport in Porous Media, 2017

We apply a GAN-based workflow of training and image generation to an oolitic Ketton limestone micro-CT unsegmented gray-level dataset.

Recommended citation: Mosser, L., Dubrule, O., & Blunt, M. J. (2018). Stochastic reconstruction of an oolitic limestone by generative adversarial networks. Transport in Porous Media, 125(1), 81-103. https://link.springer.com/article/10.1007/s11242-018-1039-9

Conditioning of Generative Adversarial Networks for Pore and Reservoir Scale Models

Published in 80th EAGE Conference and Exhibition 2018, 2018

Recently, generative adversarial networks (GAN) have been shown to be a successful method for generating unconditional simulations of pore- and reservoir-scale models. This contribution leverages the differentiable nature of neural networks to extend GANS to the conditional simulation of three-dimensional pore- and reservoir-scale models.

Recommended citation: Mosser, L., Dubrule, O., & Blunt, M. J. (2018). Conditioning of three-dimensional generative adversarial networks for pore and reservoir-scale models. arXiv preprint arXiv:1802.05622. http://www.earthdoc.org/publication/publicationdetails/?publication=92364

Rapid seismic domain transfer: Seismic velocity inversion and modeling using deep generative neural networks

Published in 80th EAGE Conference and Exhibition 2018, 2018

Traditional physics-based approaches to infer sub-surface properties such as fullwaveform inversion or reflectivity inversion are time consuming and computationally expensive. We present a deep-learning technique that eliminates the need for these computationally complex methods by posing the problem as one of domain transfer.

Recommended citation: Mosser, L., Kimman, W., Dramsch, J., Purves, S., De la Fuente Briceño, A., & Ganssle, G. (2018, June). Rapid seismic domain transfer: Seismic velocity inversion and modeling using deep generative neural networks. In 80th EAGE Conference and Exhibition 2018. https://arxiv.org/abs/1805.08826

Stochastic Seismic Waveform Inversion using Generative Adversarial Networks as a Geological Prior

Published in ArXiv, 2018

We combine a generative adversarial network (GAN) representing an a priori model that creates subsurface geological structures and their petrophysical properties, with the numerical solution of the PDE governing the propagation of acoustic waves within the earths interior. We perform Bayesian inversion using an approximate Metropolis-adjusted Langevin algorithm (MALA) to sample from the posterior given seismic observations.

Recommended citation: Mosser, L., Dubrule, O., Blunt, M. J. (2018). Stochastic seismic waveform inversion using generative adversarial networks as a geological prior. arXiv preprint arXiv:1806.03720. https://arxiv.org/abs/1806.03720

talks

Interpore 2017: 3D Reconstruction of Porous Media using Generative Adversarial Networks

Published: April 01, 2017

My first major talk I gave as a PhD student. This highlights our Physical Review E paper discussing 3D deep generative models of porous media using generative adversarial networks to create large three-dimensional images of porous media at pore-scale as well as evaluation using statistical descriptors of random materials. Talk PDF here Poster PDF here

IAMG 2018: Stochastic Simulation with Generative Adversarial Networks

Published: May 01, 2018

A talk I gave on stochastic simulation with GANs for geoscience applications at pore- and reservoir-scale. Includes work on conditioning of generated models to orthogonal cross-sections for 3D porous media and well data at reservoir grid block scale. Talk PDF here

Generative Adversarial Networks as Priors for Inverse Problems

Published: June 01, 2018

A poster I presented at the Gordon Conference on flow and transport in porous media highlighting the use of GANs as general priors for ill-posed inverse problems such as geophysical inversion or hydrocarbon reservoir history matching. Poster PDF here

EAGE 18: Conditioning of Generative Adversarial Networks for Pore and Reservoir-Scale Models

Published: June 05, 2018

This talk focuses on creating spatially conditioned pore- and reservoir-scale models based on generative adversarial networks (GANs). Talk PDF here

PESGB 18: Stochastic Seismic Waveform Inversion Using Generative Adversarial Networks As A Geological Prior

Published: October 01, 2018

We present a probabilistic framework to solve ill-posed inverse problems governed by partial differential equations where a deep generative model is used as a prior on the coefficients governing the evolution of the solution space. A geophysical, seismic inversion problem is presented where the aim is to recover the spatial distribution p-wave velocities of a synthetic subsurface model, given observed acoustic waves at discrete recording stations on the surface. Posterior samples are acquired using a gradient-based approach based on stochastic gradient langeving dynamics. Talk PDF here

Lukas Mosser

Sitemap

Pages

Posts

Reproducing Roeth and Tarantola using Modern Tools

Part 1: Data Generation using Pytorch Distributions

Reproducing Roeth and Tarantola using Modern Tools

Part 2: Forward Modeling Acoustic Waveforms with Devito

publications

talks

teaching