Deep Learning for Simultaneous Inference of Hydraulic and Transport Properties

Zhou, Zitong; Zabaras, Nicholas; Tartakovsky, Daniel M.

doi:10.1029/2021WR031438

Title: Deep Learning for Simultaneous Inference of Hydraulic and Transport Properties

Journal Article · Thu Sep 29 00:00:00 EDT 2022 · Water Resources Research

DOI:https://doi.org/10.1029/2021WR031438· OSTI ID:1894737

Zhou, Zitong ^[1];

^[2];

^[1]

Department of Energy Science and Engineering Stanford University Stanford CA USA
Scientific Computing and Artificial Intelligence (SCAI) Laboratory University of Notre Dame South Bend IN USA

Abstract Identification of a heterogeneous conductivity field and reconstruction of a contaminant release history are key aspects of subsurface remediation. These two goals are achieved by combining model predictions with sparse and noisy hydraulic head and concentration measurements. Solution of this inverse problem is notoriously difficult due to, in part, high dimensionality of the parameter space and high computational cost of repeated forward solves. We use a convolutional adversarial autoencoder (CAAE) to parameterize a heterogeneous non‐Gaussian conductivity field via a low‐dimensional latent representation. A three‐dimensional dense convolutional encoder‐decoder (DenseED) network serves as a forward surrogate of the flow and transport model. The CAAE‐DenseED surrogate is fed into the ensemble smoother with multiple data assimilation (ESMDA) algorithm to sample from the Bayesian posterior distribution of the unknown parameters, forming a CAAE‐DenseED‐ESMDA inversion framework. The resulting CAAE‐DenseED‐ESMDA inversion strategy is used to identify a three‐dimensional contaminant source and conductivity field. A comparison of the inversion results from CAAE‐ESMDA with physical flow and transport simulator and from CAAE‐DenseED‐ESMDA shows that the latter yields accurate reconstruction results at the fraction of the computational cost of the former.

View Accepted Manuscript (Publisher)

Cite

Export

Save

Sponsoring Organization:: USDOE

OSTI ID:: 1894737

Journal Information:: Water Resources Research, Journal Name: Water Resources Research Vol. 58 Journal Issue: 10; ISSN 0043-1397

Publisher:: American Geophysical Union (AGU)Copyright Statement

Country of Publication:: United States

Language:: English

References (53)

Iterative filter based estimation of fully 3D heterogeneous fields of permeability and Mualem-van Genuchten parameters Chaudhuri, A.; Franssen, H. -J. Hendricks; Sekhar, M. Advances in Water Resources, Vol. 122 https://doi.org/10.1016/j.advwatres.2018.10.023	journal	December 2018
Groundwater contaminant source identification by a hybrid heuristic approach: GROUNDWATER CONTAMINANT SOURCE IDENTIFICATION Yeh, Hund-Der; Chang, Tung-Hua; Lin, Yu-Chung Water Resources Research, Vol. 43, Issue 9 https://doi.org/10.1029/2005wr004731	journal	September 2007
Geological realism in hydrogeological and geophysical inverse modeling: A review Linde, Niklas; Renard, Philippe; Mukerji, Tapan Advances in Water Resources, Vol. 86 https://doi.org/10.1016/j.advwatres.2015.09.019	journal	December 2015
The Ensemble Kalman Filter: theoretical formulation and practical implementation Evensen, Geir Ocean Dynamics, Vol. 53, Issue 4 https://doi.org/10.1007/s10236-003-0036-9	journal	November 2003
Towards a predictor for CO2 plume migration using deep neural networks Wen, Gege; Tang, Meng; Benson, Sally M. International Journal of Greenhouse Gas Control, Vol. 105 https://doi.org/10.1016/j.ijggc.2020.103223	journal	February 2021
An exploration of coupled surface–subsurface solute transport in a fully integrated catchment model Liggett, Jessica E.; Partington, Daniel; Frei, Sven Journal of Hydrology, Vol. 529 https://doi.org/10.1016/j.jhydrol.2015.09.006	journal	October 2015
Large-Scale Machine Learning with Stochastic Gradient Descent Bottou, Léon Proceedings of COMPSTAT'2010 https://doi.org/10.1007/978-3-7908-2604-3_16	book	January 2010
Inverse methods in hydrogeology: Evolution and recent trends Zhou, Haiyan; Gómez-Hernández, J. Jaime; Li, Liangping Advances in Water Resources, Vol. 63 https://doi.org/10.1016/j.advwatres.2013.10.014	journal	January 2014
Scripting MODFLOW Model Development Using Python and FloPy Bakker, M.; Post, V.; Langevin, C. D. Groundwater, Vol. 54, Issue 5 https://doi.org/10.1111/gwat.12413	journal	March 2016
A Deep-Learning-Based Geological Parameterization for History Matching Complex Models Liu, Yimin; Sun, Wenyue; Durlofsky, Louis J. Mathematical Geosciences, Vol. 51, Issue 6 https://doi.org/10.1007/s11004-019-09794-9	journal	March 2019
Data-space inversion using a recurrent autoencoder for time-series parameterization Jiang, Su; Durlofsky, Louis J. Computational Geosciences, Vol. 25, Issue 1 https://doi.org/10.1007/s10596-020-10014-1	journal	November 2020
1D geological imaging of the subsurface from geophysical data with Bayesian Evidential Learning Michel, Hadrien; Nguyen, Frédéric; Kremer, Thomas Computers & Geosciences, Vol. 138 https://doi.org/10.1016/j.cageo.2020.104456	journal	May 2020
Efficient Bayesian experimental design for contaminant source identification Zhang, Jiangjiang; Zeng, Lingzao; Chen, Cheng Water Resources Research, Vol. 51, Issue 1 https://doi.org/10.1002/2014wr015740	journal	January 2015
A geostatistical approach to contaminant source identification Snodgrass, Mark F.; Kitanidis, Peter K. Water Resources Research, Vol. 33, Issue 4 https://doi.org/10.1029/96wr03753	journal	April 1997
History Matching of a Channelized Reservoir Using a Serial Denoising Autoencoder Integrated with ES-MDA Kim, Sungil; Min, Baehyun; Kwon, Seoyoon Geofluids, Vol. 2019 https://doi.org/10.1155/2019/3280961	journal	April 2019
Gradient-based deterministic inversion of geophysical data with generative adversarial networks: Is it feasible? Laloy, Eric; Linde, Niklas; Ruffino, Cyprien Computers & Geosciences, Vol. 133 https://doi.org/10.1016/j.cageo.2019.104333	journal	December 2019
Bayesian evidential learning: a field validation using push-pull tests Hermans, Thomas; Lesparre, Nolwenn; De Schepper, Guillaume Hydrogeology Journal, Vol. 27, Issue 5 https://doi.org/10.1007/s10040-019-01962-9	journal	March 2019
Kernel Principal Component Analysis for Efficient, Differentiable Parameterization of Multipoint Geostatistics Sarma, Pallav; Durlofsky, Louis J.; Aziz, Khalid Mathematical Geosciences, Vol. 40, Issue 1 https://doi.org/10.1007/s11004-007-9131-7	journal	December 2007
Identification of Contaminant Source Location and Release History in Aquifers Aral, Mustafa M.; Guan, Jiabao; Maslia, Morris L. Journal of Hydrologic Engineering, Vol. 6, Issue 3 https://doi.org/10.1061/(asce)1084-0699(2001)6:3(225)	journal	June 2001
Deep-learning-based surrogate flow modeling and geological parameterization for data assimilation in 3D subsurface flow Tang, Meng; Liu, Yimin; Durlofsky, Louis J. Computer Methods in Applied Mechanics and Engineering, Vol. 376 https://doi.org/10.1016/j.cma.2020.113636	journal	April 2021
Efficient Implementation of an Iterative Ensemble Smoother for Data Assimilation and Reservoir History Matching Evensen, Geir; Raanes, Patrick N.; Stordal, Andreas S. Frontiers in Applied Mathematics and Statistics, Vol. 5 https://doi.org/10.3389/fams.2019.00047	journal	October 2019
Ensemble smoother with multiple data assimilation Emerick, Alexandre A.; Reynolds, Albert C. Computers & Geosciences, Vol. 55 https://doi.org/10.1016/j.cageo.2012.03.011	journal	June 2013
Markov chain Monte Carlo with neural network surrogates: application to contaminant source identification Zhou, Zitong; Tartakovsky, Daniel M. Stochastic Environmental Research and Risk Assessment, Vol. 35, Issue 3 https://doi.org/10.1007/s00477-020-01888-9	journal	October 2020
Joint identification of contaminant source location, initial release time, and initial solute concentration in an aquifer via ensemble Kalman filtering: JOINT IDENTIFICATION OF SOURCE INFORMATION Xu, Teng; Gómez-Hernández, J. Jaime Water Resources Research, Vol. 52, Issue 8 https://doi.org/10.1002/2016wr019111	journal	August 2016
Method of Distributions for Quantification of Geologic Uncertainty in Flow Simulations Yang, Hyung Jun; Boso, Francesca; Tchelepi, Hamdi A. Water Resources Research, Vol. 56, Issue 7 https://doi.org/10.1029/2020WR027643	journal	June 2020
Information geometry of physics-informed statistical manifolds and its use in data assimilation Boso, F.; Tartakovsky, D. M. Journal of Computational Physics, Vol. 467 https://doi.org/10.1016/j.jcp.2022.111438	journal	October 2022
Uncertain Future of Hydrogeology Tartakovsky, Daniel M.; Winter, C. L. Journal of Hydrologic Engineering, Vol. 13, Issue 1 https://doi.org/10.1061/(asce)1084-0699(2008)13:1(37)	journal	January 2008
New Hybrid Optimization Methodology to Identify Pollution Sources Considering the Source Locations and Source Flux as Unknown Leichombam, Sophia; Bhattacharjya, Rajib Kumar Journal of Hazardous, Toxic, and Radioactive Waste, Vol. 23, Issue 1 https://doi.org/10.1061/(ASCE)HZ.2153-5515.0000431	journal	January 2019
Direct forecasting of global and spatial model parameters from dynamic data Park, Jihoon; Caers, Jef Computers & Geosciences, Vol. 143 https://doi.org/10.1016/j.cageo.2020.104567	journal	October 2020
A New Differentiable Parameterization Based on Principal Component Analysis for the Low-Dimensional Representation of Complex Geological Models Vo, Hai X.; Durlofsky, Louis J. Mathematical Geosciences, Vol. 46, Issue 7 https://doi.org/10.1007/s11004-014-9541-2	journal	July 2014
An Improved Implementation of the LBFGS Algorithm for Automatic History Matching Gao, Guohua; Reynolds, A. C. All Days https://doi.org/10.2118/90058-MS	conference	September 2004
An adaptive Gaussian process-based iterative ensemble smoother for data assimilation Ju, Lei; Zhang, Jiangjiang; Meng, Long Advances in Water Resources, Vol. 115 https://doi.org/10.1016/j.advwatres.2018.03.010	journal	May 2018
Integration of Adversarial Autoencoders With Residual Dense Convolutional Networks for Estimation of Non‐Gaussian Hydraulic Conductivities Mo, Shaoxing; Zabaras, Nicholas; Shi, Xiaoqing Water Resources Research, Vol. 56, Issue 2 https://doi.org/10.1029/2019WR026082	journal	February 2020
Multivariate sensitivity analysis of saturated flow through simulated highly heterogeneous groundwater aquifers Winter, C. L.; Guadagnini, A.; Nychka, D. Journal of Computational Physics, Vol. 217, Issue 1 https://doi.org/10.1016/j.jcp.2006.01.047	journal	September 2006
Data-Informed Method of Distributions for Hyperbolic Conservation Laws Boso, Francesca; Tartakovsky, Daniel M. SIAM Journal on Scientific Computing, Vol. 42, Issue 1 https://doi.org/10.1137/19m1260773	journal	January 2020
Simultaneous identification of a contaminant source and hydraulic conductivity via the restart normal-score ensemble Kalman filter Xu, Teng; Gómez-Hernández, J. Jaime Advances in Water Resources, Vol. 112 https://doi.org/10.1016/j.advwatres.2017.12.011	journal	February 2018
A hybrid simulation–optimization approach for solving the areal groundwater pollution source identification problems Ayvaz, M. Tamer Journal of Hydrology, Vol. 538 https://doi.org/10.1016/j.jhydrol.2016.04.008	journal	July 2016
Bayesian multiscale deep generative model for the solution of high-dimensional inverse problems Xia, Yingzhi; Zabaras, Nicholas Journal of Computational Physics, Vol. 455 https://doi.org/10.1016/j.jcp.2022.111008	journal	April 2022
A new framework for experimental design using Bayesian Evidential Learning: The case of wellhead protection area Thibaut, Robin; Laloy, Eric; Hermans, Thomas Journal of Hydrology, Vol. 603 https://doi.org/10.1016/j.jhydrol.2021.126903	journal	December 2021
Recent developments in fast and scalable inverse modeling and data assimilation methods in hydrology Ghorbanidehno, Hojat; Kokkinaki, Amalia; Lee, Jonghyun Journal of Hydrology, Vol. 591 https://doi.org/10.1016/j.jhydrol.2020.125266	journal	December 2020
Improved Characterization of DNAPL Source Zones via Sequential Hydrogeophysical Inversion of Hydraulic‐Head, Self‐Potential and Partitioning Tracer Data Kang, Xueyuan; Kokkinaki, Amalia; Kitanidis, Peter K. Water Resources Research, Vol. 56, Issue 8 https://doi.org/10.1029/2020wr027627	journal	August 2020
History matching geological facies models based on ensemble smoother and deep generative models Canchumuni, Smith W. A.; Emerick, Alexandre A.; Pacheco, Marco Aurélio C. Journal of Petroleum Science and Engineering, Vol. 177 https://doi.org/10.1016/j.petrol.2019.02.037	journal	June 2019
Nonlinear least squares algorithm for identification of hazards White, Robert E.; Wu, Yong Hong Cogent Mathematics, Vol. 2, Issue 1 https://doi.org/10.1080/23311835.2015.1118219	journal	December 2015
Transfer Learning on Multifidelity data Song, Dong H.; Tartakovsky, Daniel M. Journal of Machine Learning for Modeling and Computing, Vol. 3, Issue 1 https://doi.org/10.1615/JMachLearnModelComput.2021038925	journal	January 2022
Modeling complex geological structures with elementary training images and transform-invariant distances Mariethoz, Gregoire; Kelly, Bryce F. J. Water Resources Research, Vol. 47, Issue 7 https://doi.org/10.1029/2011wr010412	journal	July 2011
DRAM: Efficient adaptive MCMC Haario, Heikki; Laine, Marko; Mira, Antonietta Statistics and Computing, Vol. 16, Issue 4 https://doi.org/10.1007/s11222-006-9438-0	journal	December 2006
Inversion using a new low-dimensional representation of complex binary geological media based on a deep neural network Laloy, Eric; Hérault, Romain; Lee, John Advances in Water Resources, Vol. 110 https://doi.org/10.1016/j.advwatres.2017.09.029	journal	December 2017
Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification He, Kaiming; Zhang, Xiangyu; Ren, Shaoqing 2015 IEEE International Conference on Computer Vision (ICCV) https://doi.org/10.1109/ICCV.2015.123	conference	December 2015
Efficient gHMC Reconstruction of Contaminant Release History Barajas-Solano, David A.; Alexander, Francis J.; Anghel, Marian Frontiers in Environmental Science, Vol. 7 https://doi.org/10.3389/fenvs.2019.00149	journal	October 2019
An Adaptive Metropolis Algorithm Haario, Heikki; Saksman, Eero; Tamminen, Johanna Bernoulli, Vol. 7, Issue 2 https://doi.org/10.2307/3318737	journal	April 2001
Learning on dynamic statistical manifolds Boso, F.; Tartakovsky, D. M. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, Vol. 476, Issue 2239 https://doi.org/10.1098/rspa.2020.0213	journal	July 2020
Conditioning Geostatistical Models to Two-Phase Production Data Wu, Zhan; Reynolds, A. C.; Oliver, D. S. SPE Journal, Vol. 4, Issue 02 https://doi.org/10.2118/56855-pa	journal	June 1999
Deep generative models in inversion: The impact of the generator's nonlinearity and development of a new approach based on a variational autoencoder Lopez-Alvis, Jorge; Laloy, Eric; Nguyen, Frédéric Computers & Geosciences, Vol. 152 https://doi.org/10.1016/j.cageo.2021.104762	journal	July 2021

Similar Records

Computationally efficient Bayesian inference for inverse problems.

Technical Report · Mon Oct 01 00:00:00 EDT 2007 · OSTI ID:1894737

Marzouk, Youssef M; Najm, Habib N; Rahn, Larry A

Inverse analysis with variational autoencoders: a comparison of shallow and deep networks

Journal Article · Sat Jan 01 00:00:00 EST 2022 · Journal of Machine Learning for Modeling and Computing · OSTI ID:1894737

Wu, Hao; O'Malley, Daniel; Golden, John Kimbell; +1 more

Learning functional priors and posteriors from data and physics

Journal Article · Thu Feb 17 00:00:00 EST 2022 · Journal of Computational Physics · OSTI ID:1894737

Meng, Xuhui; Yang, Liu; Mao, Zhiping; +2 more

Title: Deep Learning for Simultaneous Inference of Hydraulic and Transport Properties

Citation Formats

References (53)

Similar Records

Related Subjects