Fully connected deep network: An improved method to predict TOC of shale reservoirs from well logs

2021 ◽  
pp. 105205
Author(s):  
Dongyu Zheng ◽  
Sixuan Wu ◽  
Mingcai Hou
Keyword(s):  
Well Logs ◽  
Improved Method ◽  
Deep Network ◽  
Shale Reservoirs ◽  
Fully Connected

scholarly journals Low-Rank Deep Convolutional Neural Network for Multitask Learning

2019 ◽  
Vol 2019 ◽  
pp. 1-10
Author(s):  
Fang Su ◽  
Hai-Yang Shang ◽  
Jing-Yan Wang
Keyword(s):  
Language Processing ◽  
Back Propagation ◽  
Multitask Learning ◽  
Low Rank ◽  
Learning Problem ◽  
Convolutional Network ◽  
Deep Network ◽  
Benchmark Datasets ◽  
Deep Cnn ◽  
Fully Connected

In this paper, we propose a novel multitask learning method based on the deep convolutional network. The proposed deep network has four convolutional layers, three max-pooling layers, and two parallel fully connected layers. To adjust the deep network to multitask learning problem, we propose to learn a low-rank deep network so that the relation among different tasks can be explored. We proposed to minimize the number of independent parameter rows of one fully connected layer to explore the relations among different tasks, which is measured by the nuclear norm of the parameter of one fully connected layer, and seek a low-rank parameter matrix. Meanwhile, we also propose to regularize another fully connected layer by sparsity penalty so that the useful features learned by the lower layers can be selected. The learning problem is solved by an iterative algorithm based on gradient descent and back-propagation algorithms. The proposed algorithm is evaluated over benchmark datasets of multiple face attribute prediction, multitask natural language processing, and joint economics index predictions. The evaluation results show the advantage of the low-rank deep CNN model over multitask problems.

Part-based multi-task deep network for autonomous indoor drone navigation

2020 ◽  
Vol 42 (16) ◽  
pp. 3243-3253 ◽  
Author(s):  
Xiangzhu Zhang ◽  
Lijia Zhang ◽  
Hailong Pei ◽  
Frank L. Lewis
Keyword(s):  
Autonomous Navigation ◽  
Indoor Navigation ◽  
Model Parameters ◽  
Feature Map ◽  
Deep Network ◽  
Navigation Algorithm ◽  
Region Division ◽  
Localization And Mapping ◽  
Previous State ◽  
Fully Connected

Two common methods exist for solving indoor autonomous navigation and obstacle-avoidance problems using monocular vision: the traditional simultaneous localization and mapping (SLAM) method, which requires complex hardware, heavy calculations, and is prone to errors in low texture or dynamic environments; and deep-learning algorithms, which use the fully connected layer for classification or regression, resulting in more model parameters and easy over-fitting. Among the latter ones, the most advanced indoor navigation algorithm divides a single image frame into multiple parts for prediction, resulting in doubled reasoning time. To solve these problems, we propose a multi-task deep network based on feature map region division for monocular indoor autonomous navigation. We divide the feature map instead of the original image to avoid repeated information processing. To reduce model parameters, we use convolution instead of the fully connected layer to predict the navigable probability of the left, middle, and right parts. We propose that the linear velocity is determined by combining three prediction probabilities to reduce collision risk. Experimental evaluation shows that the proposed method is nine times smaller than the previous state-of-the-art methods; further, its processing speed and navigation capability increase more than five and 1.6 times, respectively.

A Novel 3D U-Net Deep Network with Paralleling Structure for Stroke Lesion Image Segmentation

2020 ◽  
Vol 10 (3) ◽  
pp. 724-730
Author(s):  
Chunjiang Fan ◽  
Zijian Wang ◽  
Gang Li ◽  
Jian Luo ◽  
Yang Cao ◽  
...  
Keyword(s):  
Image Segmentation ◽  
Ct Scan ◽  
Improved Method ◽  
Feature Maps ◽  
Stroke Patients ◽  
Lesion Location ◽  
Deep Network ◽  
Valid Data ◽  
Spatial Dimensions ◽  
Stroke Lesion

Image segmentation technologies play a crucial role in medical diagnosis. This paper proposed a novel paralleling structure based on conventional 3D U-net deep network for improving the performance of CT image segmentation. In our model architecture, a new connection channel from analysis path to synthesis path was constructed for exploiting feature maps from deep spatial dimensions. 60 CT scan images of stroke patients were collected for lesion location. Finally, there were 36 valid data were selected for further analysis. The improved method led to better achievement for this task, which segment stroke CT scan images into healthy parts and injury parts. The performance on the test set obtained by our method was compared with other state-of-art U-net models, to demonstrate the effectiveness of our architecture. Furthermore, the result verified that paralleling structure was useful for the convergence of loss curve.

The Unconventional Shale Reservoirs of Jafurah Basin: An Integrated Petrophysical Evaluation Using Cores and Advanced Well Logs

2017 ◽  
Author(s):  
Ghadeer Al-Sulami ◽  
Mohammed Boudjatit ◽  
Mohammed Al-Duhailan ◽  
Salvatore Di Simone
Keyword(s):  
Well Logs ◽  
Shale Reservoirs ◽  
Petrophysical Evaluation

IMPROVED LOG EVALUATION IN A LITHOLOGICALLY COMPLEX, GLAUCONITIC RESERVOIR—A CASE STUDY

1997 ◽  
Vol 37 (1) ◽  
pp. 786
Author(s):  
Y.J. Zhang ◽  
P.A. Lollback ◽  
H.A. Salisch
Keyword(s):  
Well Logs ◽  
Improved Method ◽  
Response Characteristics ◽  
Porosity And Permeability ◽  
Subsequent Elimination ◽  
Variable Content ◽  
Environmental Corrections ◽  
Permeability Evaluation

This paper describes a case study of the application of an improved method of formation evaluation from well logs in a pilot area of the Mardie Greensand reservoirs in the Carnarvon Basin of Western Australia. They are lithologically complex reservoirs with a high and highly variable content of glauconite and extensive micro- porosity. These facts, in addition to the presence of other lithological components, make traditional log analysis, in particular the estimation of log-derived values of permeability, difficult if not impossible. The aim of this project was mainly to determine electrofacies and evaluate porosity and permeability from conventional well logs in this area. The sequential steps in the log evaluation of these glauconite-rich reservoirs were as follows: log quality control (borehole environmental corrections and depth matching); analysis of the log response characteristics; determination of litho- parameters used to identify the electrofacies; identification of the so-called hard streaks and their subsequent elimination for the purpose of reading log responses largely unaffected by these horizons; electrofacies identification and classification; porosity and permeability evaluation. The paper presents examples from several wells in the pilot area of the Mardie Greensand to illustrate this study.

scholarly journals On the Effect of the Activation Function on the Distribution of Hidden Nodes in a Deep Network

2019 ◽  
Vol 31 (12) ◽  
pp. 2562-2580 ◽  
Author(s):  
Philip M. Long ◽  
Hanie Sedghi
Keyword(s):  
Hidden Variables ◽  
Joint Probability ◽  
Activation Function ◽  
Joint Probability Distribution ◽  
Minimal Set ◽  
Deep Network ◽  
Random Weights ◽  
Weight Initialization ◽  
Fully Connected ◽  
Hidden Nodes

We analyze the joint probability distribution on the lengths of the vectors of hidden variables in different layers of a fully connected deep network, when the weights and biases are chosen randomly according to gaussian distributions. We show that if the activation function [Formula: see text] satisfies a minimal set of assumptions, satisfied by all activation functions that we know that are used in practice, then, as the width of the network gets large, the “length process” converges in probability to a length map that is determined as a simple function of the variances of the random weights and biases and the activation function [Formula: see text]. We also show that this convergence may fail for [Formula: see text] that violate our assumptions. We show how to use this analysis to choose the variance of weight initialization, depending on the activation function, so that hidden variables maintain a consistent scale throughout the network.

scholarly journals Ensemble Learning for Predicting TOC from Well-Logs of the Unconventional Goldwyer Shale

Energies ◽  
2021 ◽  
Vol 15 (1) ◽  
pp. 216
Author(s):  
Partha Pratim Mandal ◽  
Reza Rezaee ◽  
Irina Emelyanova
Keyword(s):  
Ensemble Learning ◽  
Supervised Machine Learning ◽  
Well Logs ◽  
Estimation Accuracy ◽  
Canning Basin ◽  
The Core ◽  
Data Points ◽  
Shale Reservoirs ◽  
Petrophysical Logs ◽  
Rock Eval

Precise estimation of total organic carbon (TOC) is extremely important for the successful characterization of an unconventional shale reservoir. Indirect traditional continuous TOC prediction methods from well-logs fail to provide accurate TOC in complex and heterogeneous shale reservoirs. A workflow is proposed to predict a continuous TOC profile from well-logs through various ensemble learning regression models in the Goldwyer shale formation of the Canning Basin, WA. A total of 283 TOC data points from ten wells is available from the Rock-Eval analysis of the core specimen where each sample point contains three to five petrophysical logs. The core TOC varies largely, ranging from 0.16 wt % to 4.47 wt % with an average of 1.20 wt %. In addition to the conventional MLR method, four supervised machine learning methods, i.e., ANN, RF, SVM, and GB are trained, validated, and tested for continuous TOC prediction using the ensemble learning approach. To ensure robust TOC prediction, an aggregated model predictor is designed by combining the four ensemble-based models. The model achieved estimation accuracy with R2 value of 87%. Careful data preparation and feature selection, reconstruction of corrupted or missing logs, and the ensemble learning implementation and optimization have improved TOC prediction accuracy significantly compared to a single model approach.

Rock classification in the Haynesville Shale based on petrophysical and elastic properties estimated from well logs

2015 ◽  
Vol 3 (1) ◽  
pp. SA65-SA75 ◽  
Author(s):  
Mehrnoosh Saneifar ◽  
Alvaro Aranibar ◽  
Zoya Heidari
Keyword(s):  
Elastic Properties ◽  
Physical Models ◽  
Well Logs ◽  
Rock Properties ◽  
Fracture Treatment ◽  
Well Log ◽  
Rock Classification ◽  
Classification Techniques ◽  
Shale Reservoirs ◽  
Organic Shale

Rock classification can enhance fracture treatment design for successful field developments in organic-shale reservoirs. The petrophysical and elastic properties of formations are important to consider when selecting the best candidate zones for fracture treatment. Rock classification techniques based on well logs can be advantageous compared to conventional ones based on cores, and they enable depth-by-depth formation characterization. We developed and evaluated three rock classification techniques in organic-shale formations that incorporate well logs and well-log-based estimates of elastic properties, petrophysical properties, mineralogy, and organic richness. The three rock classification techniques include (1) a 3D crossplot analysis of organic richness, volumetric concentrations of minerals, and rock brittleness index, (2) an unsupervised artificial neural network (ANN), built from an input of well logs, and (3) an unsupervised ANN, constructed using an input of well-log-based estimates of petrophysical, compositional, and elastic properties. A so-called self-consistent approximation rock-physics model is used to estimate elastic rock properties. This model enables assessment of the elastic properties based on the well-log-derived estimates of mineralogy and shapes of rock components, in the absence of acoustic-wave velocity logs. Finally, we apply the three proposed techniques to the Haynesville Shale for rock classification. We verify the identified rock types using thin-section images and previously identified lithofacies. We determined that well logs can be directly used for rock classification instead of petrophysical, compositional, and elastic properties obtained from well-log interpretation. Direct use of well logs, instead of well-log-derived properties, can reduce uncertainty associated with the physical models used to estimate elastic moduli and petrophysical/compositional properties. The three proposed well-log-based rock classification techniques can potentially enhance fracture treatment for production from complex organic-shale reservoirs through (1) detecting the best candidate zones for fracture treatment and (2) optimizing the number of required fracture stages.

An Integrated Experimental Workflow for Formation Water Characterization in Shale Reservoirs: A Case Study of the Bazhenov Formation

SPE Journal ◽  
2021 ◽  
pp. 1-16
Author(s):  
Ekaterina S. Kazak ◽  
Andrey V. Kazak ◽  
Felix Bilek
Keyword(s):  
Isotopic Composition ◽  
Bound Water ◽  
Clay Content ◽  
Isotopic Analysis ◽  
Well Logs ◽  
Formation Water ◽  
Laboratory Methods ◽  
Rock Samples ◽  
Water Origin ◽  
Shale Reservoirs

Summary In this study, we aim to develop a new integrated solution for determining the formation water content and salinity for petrophysical characterization. The workflow includes three core components: the evaporation method (EM) with isotopic analysis, analysis of aqueous extracts, and cation exchange capacity (CEC) study. The EM serves to quickly and accurately measure the contents of both free and loosely clay-bound water. The isotopic composition confirms the origin and genesis of the formation water. Chemical analysis of aqueous extracts gives the lower limit of sodium chloride (NaCl) salinity. The CEC describes rock-fluid interactions. The workflow is applicable for tight reservoir rock samples, including shales and source rocks. A representative collection of rock samples is formed based on the petrophysical interpretation of well logs from a complex source rock of the Bazhenov Formation (BF; Western Siberia, Russia). The EM employs the retort principle but delivers much more accurate and reliable results. The suite of auxiliary laboratory methods includes derivatography, Rock-Eval pyrolysis, and X-ray diffraction (XRD) analysis. Water extracts from the rock samples at natural humidity deliver a lower bound for mineralization (salinity) of formation water. Isotopic analysis of the evaporated water samples covered δ18O and δ2H. A modified alcoholic ammonium chloride [(NH4Cl)Alc] method provides the CEC and exchangeable cation concentration of the rock samples with low carbonate content. The studied rock samples had residual formation water up to 4.3 wt%, including free up to 3.9 wt% and loosely clay-bound water up to 0.96 wt%. The latter correlates well to the clay content. The estimated formation water salinity reached tens of grams per liter. At the same time, the isotopic composition confirmed the formation genesis at high depth and generally matched with that of the region's deep stratal waters. The content of chemically bound water reached 6.40 wt% and exceeded both free and loosely bound water contents. The analysis of isotopic composition proved the formation water origin. The CEC fell in the range of 1.5 to 4.73 cmol/kg and depended on the clay content. In this study, we take a qualitative step toward quantifying formation water in shale reservoirs. The research effort delivered an integrated workflow for reliable determination of formation water content, salinity lower bound, and water origin. The results fill the knowledge gaps in the petrophysical interpretation of well logs and general reservoir characterization and reserve estimation. The research novelty uses a unique suite of laboratory methods adapted for tight shale rocks holding less than 1 wt% of water.

Sign in / Sign up

Export Citation Format

Share Document