skip to main content


Title: Solving multiphysics-based inverse problems with learned surrogates and constraints
Abstract

Solving multiphysics-based inverse problems for geological carbon storage monitoring can be challenging when multimodal time-lapse data are expensive to collect and costly to simulate numerically. We overcome these challenges by combining computationally cheap learned surrogates with learned constraints. Not only does this combination lead to vastly improved inversions for the important fluid-flow property, permeability, it also provides a natural platform for inverting multimodal data including well measurements and active-source time-lapse seismic data. By adding a learned constraint, we arrive at a computationally feasible inversion approach that remains accurate. This is accomplished by including a trained deep neural network, known as a normalizing flow, which forces the model iterates to remain in-distribution, thereby safeguarding the accuracy of trained Fourier neural operators that act as surrogates for the computationally expensive multiphase flow simulations involving partial differential equation solves. By means of carefully selected experiments, centered around the problem of geological carbon storage, we demonstrate the efficacy of the proposed constrained optimization method on two different data modalities, namely time-lapse well and time-lapse seismic data. While permeability inversions from both these two modalities have their pluses and minuses, their joint inversion benefits from either, yielding valuable superior permeability inversions and CO2plume predictions near, and far away, from the monitoring wells.

 
more » « less
NSF-PAR ID:
10468650
Author(s) / Creator(s):
; ; ;
Publisher / Repository:
Springer Science + Business Media
Date Published:
Journal Name:
Advanced Modeling and Simulation in Engineering Sciences
Volume:
10
Issue:
1
ISSN:
2213-7467
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. We present the Seismic Laboratory for Imaging and Modeling/Monitoring open-source software framework for computational geophysics and, more generally, inverse problems involving the wave equation (e.g., seismic and medical ultrasound), regularization with learned priors, and learned neural surrogates for multiphase flow simulations. By integrating multiple layers of abstraction, the software is designed to be both readable and scalable, allowing researchers to easily formulate problems in an abstract fashion while exploiting the latest developments in high-performance computing. The design principles and their benefits are illustrated and demonstrated by means of building a scalable prototype for permeability inversion from time-lapse crosswell seismic data, which, aside from coupling of wave physics and multiphase flow, involves machine learning. 
    more » « less
  2. Abstract The growing amount of seismic data necessitates efficient and effective methods to monitor earthquakes. Current methods are computationally expensive, ineffective under noisy environments, or labor intensive. We leverage advances in machine learning to propose an improved solution, ArrayConvNet—a convolutional neural network that uses continuous array data from a seismic network to seamlessly detect and localize events, without the intermediate steps of phase detection, association, travel-time calculation, and inversion. When testing this methodology with events at Hawai‘i, we achieve 99.4% accuracy and predict hypocenter locations within a few kilometers of the U.S. Geological Survey catalog. We demonstrate that training with relocated earthquakes reduces localization errors significantly. We outline several ways to improve the model, including enhanced data augmentation and use of relocated offshore earthquakes recorded by ocean-bottom seismometers. Application to continuous records shows that our algorithm detects 690% as many earthquakes as the published catalog, and 125% as many events than the Hawaiian Volcano Observatory internal catalog. Because of the enhanced detection sensitivity, localization granularity, and minimal computation costs, our solution is valuable, particularly for real-time earthquake monitoring. 
    more » « less
  3. Geologic carbon storage represents one of the few truly scalable technologies capable of reducing the CO 2 concentration in the atmosphere. While this technology has the potential to scale, its success hinges on our ability to mitigate its risks. An important aspect of risk mitigation concerns assurances that the injected CO 2 remains within the storage complex. Among the different monitoring modalities, seismic imaging stands out due to its ability to attain high-resolution and high-fidelity images. However, these superior features come at prohibitive costs and time-intensive efforts that potentially render extensive seismic monitoring undesirable. To overcome this shortcoming, we present a methodology in which time-lapse images are created by inverting nonreplicated time-lapse monitoring data jointly. By no longer insisting on replication of the surveys to obtain high-fidelity time-lapse images and differences, extreme costs and time-consuming labor are averted. To demonstrate our approach, hundreds of realistic synthetic noisy time-lapse seismic data sets are simulated that contain imprints of regular CO 2 plumes and irregular plumes that leak. These time-lapse data sets are subsequently inverted to produce time-lapse difference images that are used to train a deep neural classifier. The testing results show that the classifier is capable of detecting CO 2 leakage automatically on unseen data with reasonable accuracy. We consider the use of this classifier as a first step in the development of an automatic workflow designed to handle the large number of continuously monitored CO 2 injection sites needed to help combat climate change. 
    more » « less
  4. In many mechanistic medical, biological, physical, and engineered spatiotemporal dynamic models the numerical solution of partial differential equations (PDEs), especially for diffusion, fluid flow and mechanical relaxation, can make simulations impractically slow. Biological models of tissues and organs often require the simultaneous calculation of the spatial variation of concentration of dozens of diffusing chemical species. One clinical example where rapid calculation of a diffusing field is of use is the estimation of oxygen gradients in the retina, based on imaging of the retinal vasculature, to guide surgical interventions in diabetic retinopathy. Furthermore, the ability to predict blood perfusion and oxygenation may one day guide clinical interventions in diverse settings, i.e., from stent placement in treating heart disease to BOLD fMRI interpretation in evaluating cognitive function (Xie et al., 2019 ; Lee et al., 2020 ). Since the quasi-steady-state solutions required for fast-diffusing chemical species like oxygen are particularly computationally costly, we consider the use of a neural network to provide an approximate solution to the steady-state diffusion equation. Machine learning surrogates, neural networks trained to provide approximate solutions to such complicated numerical problems, can often provide speed-ups of several orders of magnitude compared to direct calculation. Surrogates of PDEs could enable use of larger and more detailed models than are possible with direct calculation and can make including such simulations in real-time or near-real time workflows practical. Creating a surrogate requires running the direct calculation tens of thousands of times to generate training data and then training the neural network, both of which are computationally expensive. Often the practical applications of such models require thousands to millions of replica simulations, for example for parameter identification and uncertainty quantification, each of which gains speed from surrogate use and rapidly recovers the up-front costs of surrogate generation. We use a Convolutional Neural Network to approximate the stationary solution to the diffusion equation in the case of two equal-diameter, circular, constant-value sources located at random positions in a two-dimensional square domain with absorbing boundary conditions. Such a configuration caricatures the chemical concentration field of a fast-diffusing species like oxygen in a tissue with two parallel blood vessels in a cross section perpendicular to the two blood vessels. To improve convergence during training, we apply a training approach that uses roll-back to reject stochastic changes to the network that increase the loss function. The trained neural network approximation is about 1000 times faster than the direct calculation for individual replicas. Because different applications will have different criteria for acceptable approximation accuracy, we discuss a variety of loss functions and accuracy estimators that can help select the best network for a particular application. We briefly discuss some of the issues we encountered with overfitting, mismapping of the field values and the geometrical conditions that lead to large absolute and relative errors in the approximate solution. 
    more » « less
  5. The energy transition to meet net-zero emissions by 2050 has created demand for underground caverns needed to safely store CO2, hydrocarbon, hydrogen, and wastewater. Salt domes are ideal for underground storage needs because of their low permeability and affordable costs, which makes them the preferred choice for large-scale storage projects like the US Strategic Petroleum Reserves. However, the uneven upward movement of salt spines can create drilling problems and breach cavern integrity, releasing harmful gases into overlying aquifers and endangering nearby communities. Here, we present a novel application of data-driven geophysical methods combined with machine learning that improves salt dome characterization during feasibility studies for site selection and potentially advances the effectiveness of current early-warning systems. We utilize long-term, non-invasive seismic monitoring to investigate deformation processes at the Sorrento salt dome in Louisiana. We developed a hybrid autoencoder model and applied it to an 8-month dataset from a nodal array deployed in 2020, to produce a high-fidelity microearthquake catalog. Our hybrid model outperformed traditional event detection techniques and other neural network detectors. Seismic signals from storms, rock bursts, trains, aircraft, and other anthropogenic sources were identified. Clusters of microearthquakes were observed along two N-S trends referred to as Boundary Shear Zones (BSZ), along which we infer that salt spines are moving differentially. Time-lapse sonar surveys were used to confirm variations in propagation rates within salt spines and assess deformation within individual caverns. Seismicity along one BSZ is linked with a well failure incident that created a 30-ft wide crater at the surface in 2021. This study introduces a novel method for mapping spatial and temporal variations in salt shear zones and provides insights into the subsurface processes that can compromise the safety and lifetime of underground storage sites. 
    more » « less