Data‐Worth Analysis for Heterogeneous Subsurface Structure Identification With a Stochastic Deep Learning Framework
TLDR
In this paper , the authors extended a recently developed stage-wise stochastic deep learning inversion framework by coupling it with nonisothermal flow and transport simulations to estimate subsurface structures.Abstract:
Reliable characterization of subsurface structures is essential for earth sciences and related applications. Data assimilation‐based identification frameworks can reasonably estimate subsurface structures using available lithological (e.g., borehole core, well log) and dynamic (e.g., hydraulic head, solute concentration) observations. However, a reasonable selection of the observation type and frequency is essential for accurate structure identification. To achieve this, we extended a recently developed stage‐wise stochastic deep learning inversion framework by coupling it with non‐isothermal flow and transport simulations. With the extended framework, the worth of three common observations (hydraulic head, concentration, and temperature) are compared under different observation noise and frequency. The framework combines the emerging deep‐learning (DL)‐based framework with the traditional stochastic approaches. This combination makes it possible to simultaneously compare the ability of these two methods to assimilate observation data. Our results show that including at least one type of dynamic observation strongly improves subsurface structure identifiability and reduces the uncertainty. However, the DL‐based framework is able to identify subsurface structures more accurately than stochastic identification methods under the same scenarios. Assimilation of certain types of dynamic observations could reduce the prediction error for related dynamic responses, but not necessarily for other uncorrelated dynamic responses. Observation data worth is affected by the observation noise and frequency. High observation noise increases the uncertainty of the prediction and reduces the estimation accuracy. However, the higher observation frequency can significantly improve the temporal dynamic information of observations. This information can compensate for negative impacts of high observation noise.read more
Citations
More filters
Journal ArticleDOI
Subsurface sedimentary structure identification using deep learning: A review
Chuanjun Zhan,Zhenxue Dai,Zhijie Yang,Xiaoying Zhang,Ziqi Ma,Hung Vo Thanh,Mohamad Reza Soltanian +6 more
TL;DR: In this article , a review of deep learning-based methods for the detection of subsurface sedimentary structures is presented, and the limitations and challenges of existing methods are summarized.
Journal ArticleDOI
Supervised deep learning-based paradigm to screen the enhanced oil recovery scenarios
TL;DR: In this article , a deep learning-based classifier consisting of a one-dimensional (1D) convolutional neural network, long short-term memory (LSTM), and densely connected neural network layers is proposed to select the best EOR method based on the reservoir's rock and fluid properties (depth, porosity, permeability, gravity, viscosity), and temperature).
Journal ArticleDOI
Combining Artificial Neural Network and Seeker Optimization Algorithm for Predicting Compression Capacity of Concrete-Filled Steel Tube Column
TL;DR: In this paper , a conventional artificial neural network (ANN) is hybridized with a metaheuristic algorithm called the seeker optimization algorithm (SOA) for estimating axial compression capacity of circular concrete-filled steel tube columns.
Journal ArticleDOI
Sensitivity study of multi-field information maps of typical landslides in mining areas based on transfer learning
TL;DR: Li et al. as mentioned in this paper analyzed the main influencing factors of the landslide in the coal mine area and established the landslide sensitivity evaluation model based on transfer learning, which can achieve a transfer effect higher than 0.95.
Journal ArticleDOI
New Fuzzy-Heuristic Methodology for Analyzing Compression Load Capacity of Composite Columns
Bizhan Karimi Sharafshadeh,Mohammad Javad Ketabdari,Farhood Azarsina,Mohammad Mohammadi Amiri,Moncef L. Nehdi +4 more
TL;DR: In this paper , the adaptive neuro-fuzzy inference system (ANFIS) was trained using four metaheuristic techniques, namely earthworm algorithm (EWA), particle swarm optimization (PSO), salp swarm algorithm (SSA), and teaching learning-based optimization (TLBO).
References
More filters
ReportDOI
TOUGH2 User's Guide Version 2
TL;DR: This report is a self-contained guide to application of Tough2 to subsurface flow problems, and gives a technical description of the TOUGH2 code, including a discussion of the physical processes modeled, and the mathematical and numerical methods used.
Journal ArticleDOI
A natural gradient experiment on solute transport in a sand aquifer: Spatial variability of hydraulic conductivity and its role in the dispersion process
TL;DR: The Borden aquifer was examined in great detail by conducting permeability measurements on a series of cores taken along two cross sections, one along and the other transverse to the mean flow direction as discussed by the authors.
Journal ArticleDOI
TOUGHREACT-A simulation program for non-isothermal multiphase reactive geochemical transport in variably saturated geologic media: Applications to geothermal injectivity and CO2 geological sequestration
TL;DR: This work examines ways in which the chemical composition of reinjected waters can be modified to improve reservoir performance by maintaining or even enhancing injectivity, and uses recent European studies as a starting point to explore chemically induced effects of fluid circulation in the geothermal systems.
Journal ArticleDOI
Geostatistical Software Library and User's Guide
TL;DR: This volume is a library of programs aimed at three major areas of geostatistics: quantifying spatial variability (variograms), generalized linear regression techniques (kriging), and stochastic simulation.
Journal ArticleDOI
An integrated finite difference method for analyzing fluid flow in porous media
TL;DR: The integrated finite difference method (IFDM) as mentioned in this paper is a powerful numerical technique for solving problems of groundwater flow in porous media, which combines the advantages of an integral formulation with the simplicity of finite difference gradients and is convenient for handling multidimensional heterogeneous systems composed of isotropic materials.