A5041957512- Hydraulic Fracturing and Reservoir Analysis
- Seismic Imaging and Inversion Techniques
- Reservoir Engineering and Simulation Methods
- Drilling and Well Engineering
- Image Retrieval and Classification Techniques
- Hydrocarbon exploration and reservoir analysis
- Remote-Sensing Image Classification
- Advanced Data Compression Techniques
- Methane Hydrates and Related Phenomena
- Atmospheric and Environmental Gas Dynamics
- Advanced Chemical Sensor Technologies
- Geochemistry and Geologic Mapping
- Image and Signal Denoising Methods
- Geological Modeling and Analysis
- Geophysical and Geoelectrical Methods
- Oil and Gas Production Techniques
- Spectroscopy and Chemometric Analyses
- Advanced Image Fusion Techniques
- Seismology and Earthquake Studies
- Anomaly Detection Techniques and Applications
- Machine Fault Diagnosis Techniques
- Face recognition and analysis
- Aeolian processes and effects
- Face and Expression Recognition
- Geophysics and Gravity Measurements
Pioneer Natural Resources (United States)
2019-2023
Pioneer (United States)
2019-2023
University of Oklahoma
2013-2015
Oklahoma Biological Survey
2013-2015
United States Geological Survey
2013-2015
The University of Texas at El Paso
2005-2009
During the past decade, size of 3D seismic data volumes and number attributes have increased to extent that it is difficult, if not impossible, for interpreters examine every line time slice. To address this problem, several facies classification algorithms including [Formula: see text]-means, self-organizing maps, generative topographic mapping, support vector machines, Gaussian mixture models, artificial neural networks been successfully used extract features geologic interest from...
Summary Petroleum reservoirs are often associated with multiple target zones or a single zone adjacent to nonproductive intervals. Real-time geosteering therefore becomes important remain in dynamically steer toward target. This requires knowledge of the petrophysical/rock mechanical properties rock surrounding bit. Although logging while drilling can provide this information, cost-effective and almost-real-time solution is lacking. In general, there depth lag, therefore, time delay, between...
Machine learning models have been widely used by geoscientists to accelerate their interpretation and highlight hidden patterns in data. However, as the complexity of model increases, results can become quite challenging. The SHAP technique provides a measure importance each input seismic attributes on model's output. We illustrate value using tree-based machine implementation trained distinguish between Mass Transport Deposits (MTDs) salt facies Gulf Mexico survey. Presentation Date:...
During the past two decades, geoscientists have used machine learning (ML) to produce a more quantitative reservoir characterization and discover hidden patterns in their data. However, as complexity of these models increases, sensitivity results choice input data becomes challenging. Measuring how model uses perform either classification or regression task provides an understanding data-to-geology relationships which indicates confident we are prediction. To provide such insight, ML...
Summary Classification of different lithofacies and petrotypes is one the main objectives modern quantitative seismic interpretation. In this study, we present preliminary results application a proximal support vector machine (PSVM) classification algorithm to data. illustrate PSVM method differentiate limestone from shale in Barnett Shale gas play. The PVSM's low complexity feature compared standard machines could be well exploited data intensive computation such as 3D classification. paper...
Summary In this study, we show an application of estimating total organic carbon (TOC) in a Barnett Shale play from the widely available triple combo logs using support vector machine (SVM). Being nonlinear supervised learning technique, SVM provides superior estimation than traditional multi-linear regression. Using to automatically estimate TOC content limited number pre-existing measurements, proposed method delivers convenient and relatively accurate resource where core measurements...
Summary In this study, we demonstrate the application of an interpretable (or explainable) machine-learning workflow using surface drilling data to identify fracturable, brittle, and productive rock intervals along horizontal laterals in Marcellus Shale. The results are supported by a thorough model-agnostic interpretation input/output relationships make model explainable users. methodology described here can easily be generalized real-time processing for optimal landing laterals, placing...
In this paper we illustrate unsupervised and supervised learning algorithms that accurately classify the lithological variations in 3D seismic data. We demonstrate blind source separation techniques such as principal components (PCA) noise adjusted conjunction with Kohonen Self organizing maps to produce superior classification maps. Further, utilize PCA space training Maximum likelihood (ML) classification. Results ML produces an improved of facies dataset from Anadarko basin central Oklahoma.
In this paper, we propose spatio-spectral processing techniques for the detection of dust storms and automatically finding its transport direction in 5-band NOAA-AVHRR imagery. Previous methods that use simple band math analysis have produced promising results but drawbacks producing consistent when low signal to noise ratio (SNR) images are used. Moreover, seeking automate storm detection, presence clouds vicinity creates a challenge being able distinguish these two types image texture....
Planning and optimizing completion design for hydraulic fracturing require a quantifiable understanding of the spatial distribution brittleness rock other geomechanical properties. Eventually, goal is to maximize stimulated reservoir volume with minimal cost overhead. The compressional shear velocities ([Formula: see text] [Formula: text], respectively) can also be used calculate Young’s modulus, Poisson’s ratio, mechanical In field, sonic logs are not commonly acquired operators often...
Abstract Microseismic datasets typically have relatively low signal‐to‐noise ratio waveforms. To that end, several noise suppression techniques are often applied to improve the of recorded We apply a linear geometric mode decomposition approach microseismic for background suppression. The method optimizes patterns within amplitude–frequency modulated modes and can efficiently distinguish events (signal) from noise. This also split non‐linear dispersive seismic into modes. segmented in...
Abstract The topic of fracture complexity is commonly evoked when discussing hydraulic fracturing unconventional reservoirs. In this context, it typically considered beneficial to successful stimulation, as provides increased surface area, relative single planar fractures. However, in the near-wellbore region (NWR), same complexity, referred tortuosity, can be detrimental placement fluid and proppant. extreme, if not properly identified mitigated, stages may need abandoned which leads...
Abstract This study demonstrates the application of an interpretable (or explainable) machine learning workflow using surface drilling data to identify fracable, brittle and productive rock intervals along horizontal laterals in Marcellus shale. The results are supported by a thorough model-agnostic interpretation input-output relationships make model explainable users. methodology described here can easily be generalized real-time processing for optimal landing laterals, placing fracture...
The E&P community, both in the industry and academia, is painfully aware of challenges complexity performing seismic interpretation reservoir characterization increasingly larger, more intricate, heterogeneous data sets. This increase size coupled with an emphasis on
Summary URTeC 1619856 Conventional reservoirs benefit from a long scientific history that correlates successful plays to seismic measurements through depositional, tectonic, and digenetic models. Unconventional are less well understood, however significantly denser control. Thus, allowing us establish statistical rather than model-based correlations between data, geology, completion strategies. One of the more commonly encountered correlation techniques is based on computer assisted pattern...
This paper studies the effect of different bit rate allocation strategies in JPEG2000 part 2 compression hyperspectral data on results background classification. We compare traditional approach, based high quantizer with distortion optimal (RDO) approach that produces a mean squared error (MSE) sense. The experiments show for relatively low rates both perform excellent and almost similar accuracy (96% at 0.125 bpppb). However, very rates, RDO outperforms (90% 0.0375 bpppb) terms detection....
Summary In this study, we implemented and tested a new processing-based broadband solution for mitigating F-K transform artifacts receiver deghosting in marine environment. The FK has traditionally been used flat cable (constant depth) often times tailored to meet the slanted (variable criteria. Recently, usage of τ − p do- main deterministic deghost operator more prominent with slant deghosting. Irrespective type or used, windowed process is essential due time offset varying character...
Summary This paper explores and discusses the use of generative topographic mapping (GTM) in estimating estimated ultimate recovery (EUR) from geologic, petrophysical, completion parameters, further distinguishing poorly performing wells high productivity simultaneously to quantify ranges explanatory reservoir related parameters that dictate well performance. Using an application on a field dataset prominent shale gas play with over hundred horizontal wells, we demonstrate advantage GTM when...
PreviousNext No AccessInterpretationVolume 7, Issue 3Introduction to special section: Machine learning in seismic data analysisAuthors: Haibin DiTao ZhaoVikram JayaramXinming WuLei HuangGhassan AlRegibJun CaoMauricio Araya-PoloSatinder ChopraSaleh Al-DossaryFangyu LiErwan GloaguenYouzuo LinAnne SolbergHongliu ZengHaibin DiSchlumberger, Exploration and Field Development, Data Analytics Program, Houston, Texas, USA. E-mail: .Search for more papers by this authorEmail the author at [email...