Fast self-supervised learning for reconstruction of 3D seismic data

Yinshuo Li, Wenkai Lu, We Cao

International Meeting for Applied Geoscience and Energy (IMAGE)

... extractor, and a skip connection. The global waveform extractor consists of real fast Fourier transform, complex convolutional layer, complex PReLU...

Joint inversion of multi-height gravity and vertical gradient via physics-informed neural network

Yinshuo Li, Wenkai Lu, Cao Song

International Meeting for Applied Geoscience and Energy (IMAGE)

...]. The inversion model is based on convolutional layers. Since the 3D convolution neural network is computationally heavy, this abstract proposed to reduce...

Time-lapse full-waveform inversion by model order reduction using radial basis function

Haipeng Li, Robert G. Clapp

International Meeting for Applied Geoscience and Energy (IMAGE)

...Time-lapse full-waveform inversion by model order reduction using radial basis function Haipeng Li, Robert G. Clapp Time-lapse full-waveform...

Seismic facies segmentation via mask-assisted transformer

Jinlong Huo, Naihao Liu, Zhiguo Wang, Yang Yang, Yijie Zhang, Jinghuai Gao

International Meeting for Applied Geoscience and Energy (IMAGE)

... on the characteristics of seismic reflectors. The application of using convolutional neural networks (CNNs) in seismic facies segmentation is growing rapidly. However, CNN...

Automatic detection of DAS-recorded microseismic fracture reflections

Youfang Liu, Ivan Lim, Chen Ning, Kurt Nihei

International Meeting for Applied Geoscience and Energy (IMAGE)

... The workflow can be further automated through deep learning methods. For instance, we can utilize the deep convolutional neural networks to perform the point...

Unsupervised machine learning for seismic facies classification using a 3D grid approach

David Manzano, Edgar Galvan, Dan Ferdinand Fernandez

International Meeting for Applied Geoscience and Energy (IMAGE)

... convolutional neural networks (CNNs) to directly apply to 3D seismic volumes, providing better spatial awareness and automatic feature extraction based...

Optimized transparent boundary conditions for wave propagation

G. Roncoroni, B. Arntsen, E. Forte, M. Pipan

International Meeting for Applied Geoscience and Energy (IMAGE)

... topographies have been derived to accurately model wave propagation (Ruud & Hestholm, 2001). These formulations are crucial for representing...

Bi-directional LSTM-based non-causal deconvolution

G. Roncoroni, I. Deiana, E. Forte, M. Pipan

International Meeting for Applied Geoscience and Energy (IMAGE)

...). To generate the training dataset, we used a modified convolutional model defined as: where tracei and traceo represent the input and the reference output...

Generating high-quality labels for deep learning CO2 monitoring using local orthogonalization

Shuang Gao, Sergey Fomel, Yangkang Chen

International Meeting for Applied Geoscience and Energy (IMAGE)

... label generation. We implement a modified 3D U-Net deep learning model to interpret the seismic attributes associated with CO2 injections for complex...

Innovative Deep Autoencoder and Machine Learning Algorithms Applied in Production Metering for Sucker-Rod Pumping Wells

Peng Yi, Xiong Chunming, Zhang Jianjun, Zhang Yashun, Gan Qinming, Xu Guojian, Zhang Xishun, Zhao Ruidong, Shi Junfeng, Liu Meng, Wang Cai, Chen Guanhong

Unconventional Resources Technology Conference (URTEC)

.... The machine-learning model contains two neural networks: first, a deep autoencoder to extract the feature representations from all the dynamometer...

Using Machine Learning to Automate FDI Analysis

Reid Thompson, Lance Legel, Thomas Hanlon

Unconventional Resources Technology Conference (URTEC)

... is an automated stage detection model. The core of the stage detection model is a onedimensional deep convolutional U-net neural network with residual layers...

Understanding the Seismic Wavelet; Steven G. Henry; Search and Discovery Article #40028 (2001)

Search and Discovery.com

Introduction to Deep Learning: Part I

Hongbo Zhou, Lasse Amundsen, Martin Landrø

GEO ExPro Magazine

... of some objective or loss function on a training set of examples. Loss functions express the misfit between the predictions of the model being...

Horizon detection with CNN-based multiscale volumetric flattening

Jesse Lomask

International Meeting for Applied Geoscience and Energy (IMAGE)

... combines the power of Convolutional Neural Network (CNNs) and traditional geophysical inversion methods, flattening a seismic volume into a pseudo...

Novel application of machine learning assisted fault interpretation to delineate earthquake risk from saltwater disposal in the Midland Basin

Niven Shumaker, Mohammed Afia

International Meeting for Applied Geoscience and Energy (IMAGE)

... seismic survey using a 3D convolutional neural for edge pixels in a 2D data array. Lines that pass through network (Abubakar et al. 2022). Fault points...

An unsupervised intelligent stacking velocity analysis based on clustering

Lide Wang, Xingrong Xu, Jie Wu, Huahui Zeng, Yundong Yong, Yanxiang Wang

International Meeting for Applied Geoscience and Energy (IMAGE)

...-consuming and has poor noise immunity, requiring an initial model given by manual. With the development of machine learning technology, the deep...

AVA attribute estimation from misaligned seismic gathers using U-Net

Ammar Ghanim, Ricard Durall, Norman Ettrich

International Meeting for Applied Geoscience and Energy (IMAGE)

... are mainly caused by inaccuracies in the velocity model used during migration. Aligning these reflections in seismic normal moveout (NMO) corrected...

Large Mudstone-Nucleus Sandstone Spheroids in Submarine Channel Deposits: NOTES

Daniel J. Stanley

Journal of Sedimentary Research (SEPM)

...-balls (Dzulynski, et al., 1957) and convolutional balls (Dott and Howard, 1962) may be cited as examples. The above-mentioned ball structures, however...

Transfer Learning with Multiple Aggregated Source Models in Unconventional Reservoirs

J. Cornelio, S. Mohd Razak, Y. Cho, H-H. Liu, R. Vaidya, B. Jafarpour

Unconventional Resources Technology Conference (URTEC)

... Oilfield in the Middle East. Society of Petroleum Engineers. Mohd Razak S, Jafarpour B. (2020a) Convolutional neural networks (CNN) for feature-based model...

Automatic facies classification using convolutional neural network for three-dimensional outcrop data: Application to the outcrop of the mass-transport deposit

Ryusei Sato, Kazuki Kikuchi, and Hajime Naruse

AAPG Bulletin

... point clouds used as training data for the convolutional neural network (CNN) model. (A, C) Original point cloud used as training data for the CNN model...

Abstract: Variable-factor S-transform for Time-frequency Decomposition, Deconvolution, and Noise Attenuation; #90172 (2014)

Todor I. Todorov, Gary F. Margrave

Search and Discovery.com

... to the physical phenomena of the seismic wave propagation in the earth over the traditional stationary convolutional model. Margrave and Lamoureux...

Abstract: Post-stack Inversion of the Hussar Low Frequency Seismic Data; #90187 (2014)

Patricia E. Gavotti, Don C. Lawton, Gary F. Margrave, and J. Helen Isaac

Search and Discovery.com

... on the convolutional model of the seismic trace according to the equation 1: , (1) where S is the seismic trace, W is the wavelet, R is the reflectivity and N...

Seismic Meta-Attributes and the Illumination of the Internal Reservoir Architecture of a Deepwater Synthetic Channel Model, #41267 (2014)

Staffan Van Dyke, Renjun Wen

Search and Discovery.com

...Seismic Meta-Attributes and the Illumination of the Internal Reservoir Architecture of a Deepwater Synthetic Channel Model, #41267 (2014) Staffan...

Source location using physics-informed neural networks with hard constraints

Xinquan Huang, Tariq Alkhalifah

International Meeting for Applied Geoscience and Energy (IMAGE)

... allows for direct image extraction of the subsurface using inverse Fourier transform. Numerical tests on the Overthrust model demonstrate...

Reliable uncertainty estimation for seismic interpretation with prediction switches

Ryan Benkert, Mohit Prabhushankar, Ghassan AlRegib

International Meeting for Applied Geoscience and Energy (IMAGE)

... the model “forgets” seismic traces during training. In our empirical analysis, we find that our method is significantly more fine-grained than existing state...