Training-image based geostatistical inversion using a spatial generative adversarial neural network

TL;DR

This paper introduces a spatial GAN-based method for efficient probabilistic geostatistical inversion, capable of generating high-dimensional realizations quickly after training, and effectively incorporating conditioning data.

Contribution

The paper presents a novel spatial GAN approach that enables fast, low-dimensional parameterization for geostatistical inversion, improving computational efficiency over traditional methods.

Findings

SGAN can generate millions of pixels/voxels in seconds after hours of training.

The method effectively explores the posterior distribution in 2D inversion cases.

SGAN-based inversion accurately recovers models in 3D hydraulic tomography simulations.

Abstract

Probabilistic inversion within a multiple-point statistics framework is often computationally prohibitive for high-dimensional problems. To partly address this, we introduce and evaluate a new training-image based inversion approach for complex geologic media. Our approach relies on a deep neural network of the generative adversarial network (GAN) type. After training using a training image (TI), our proposed spatial GAN (SGAN) can quickly generate 2D and 3D unconditional realizations. A key characteristic of our SGAN is that it defines a (very) low-dimensional parameterization, thereby allowing for efficient probabilistic inversion using state-of-the-art Markov chain Monte Carlo (MCMC) methods. In addition, available direct conditioning data can be incorporated within the inversion. Several 2D and 3D categorical TIs are first used to analyze the performance of our SGAN for unconditional geostatistical simulation. Training our deep network can take several hours. After training, realizations containing a few millions of pixels/voxels can be produced in a matter of seconds. This makes it especially useful for simulating many thousands of realizations (e.g., for MCMC inversion) as the relative cost of the training per realization diminishes with the considered number of realizations. Synthetic inversion case studies involving 2D steady-state flow and 3D transient hydraulic tomography with and without direct conditioning data are used to illustrate the effectiveness of our proposed SGAN-based inversion. For the 2D case, the inversion rapidly explores the posterior model distribution. For the 3D case, the inversion recovers model realizations that fit the data close to the target level and visually resemble the true model well.