Facies classification using machine learning

Hall, Brendon

doi:10.1190/tle35100906.1

Cited by 177 publications

(78 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The incorporation of multi-well logs and lithological report made unsupervised machine learning facies classification possible. Lithological logs such as gamma ray, porosity, density, sonic are used to generate initial principal component analysis model (Hall, 2016). The facies classification is then filtered to be sandstone and shale sequence as the research focuses on sand reservoir interval.…”

Section: B Facies Classification and Well Correlationmentioning

confidence: 99%

Integrated reservoir characterization of low resistivity thin beds using three-dimensional modeling for natural gas exploration

Lim¹,

Lo²

2018

BGSM

View full text Add to dashboard Cite

A natural gas reservoir was discovered at approximately 3 km (TVDSS) through first vertical wildcat. Four subsequent wildcats were drilled in deviated trajectory to assess hydrocarbon distribution with no success. Resistivity log response from hydrocarbon interval appeared as low value low contrast. Seismic acquired onshore with high degree of static variation resulted in low frequency, unable to delineate thin sand interval efficiently. Several hypotheses were formed to explain the failed discovery. First, geological structure is complex due to local tectonic deformities created faults that compartmentalized the reservoir. Second, hydrocarbon charge and migration pathway might be underestimated. Third, presence of high conductive mineral might affect the resistivity log acquisition. An integration of three-dimensional enhanced seismic horizon and fault interpretation, unsupervised machine learning in facies classification, petrophysical data conditioning, rock physics cross validation, and three-dimensional static modeling is used to provide clearer insight on the natural gas play.

show abstract

Section: B Facies Classification and Well Correlationmentioning

confidence: 99%

Integrated reservoir characterization of low resistivity thin beds using three-dimensional modeling for natural gas exploration

Lim¹,

Lo²

2018

BGSM

View full text Add to dashboard Cite

show abstract

“…Through logging, it is possible to obtain a detailed description of rock formations at different depth levels by measuring a wide variety of rock properties. From this premises, the goal of facies classification is to detect the facies present at a given depth in the ground, from the analysis of a set of well log measurements obtained at the considered depth (Hall, 2016).…”

Section: Problem Formulationmentioning

confidence: 99%

“…In this context Hall (2016) proposed a Geophysical Tutorial where he showed a simple application of machine learning techniques for facies classification. In particular he used a small dataset of seven wireline logs and associated interpreted facies extracted from ten wells of the Hugoton gas field in southwest Kansas (Dubois et al, 2007), in order to predict geologic facies in two additional wells based on wireline measurements.…”

Section: Introductionmentioning

confidence: 99%

A machine learning approach to facies classification using well logs

Bestagini

Lipari

Tubaro

2017

SEG Technical Program Expanded Abstracts 2017

View full text Add to dashboard Cite

In this work we describe a machine learning pipeline for facies classification based on wireline logging measurements. The algorithm has been designed to work even with a relatively small training set and amount of features. The method is based on a gradient boosting classifier which demonstrated to be effective in such a circumstance. A key aspect of the algorithm is feature augmentation, which resulted in a significant boost in accuracy. The algorithm has been tested also through participation to the SEG machine learning contest.

show abstract

“…After training the network with synthetic data, the velocities can be topographically estimated by the trained network with the new cross-well data. In recent years, most ML-based methods have focused mainly on pattern recognition in seismic attributes (Zeng, 2004;Zhao et al, 2015) and facies classifications in well logs (Hall and Brendon, 2016). In the work of Guillen et al (2015), the authors proposed a novel workflow to detect salt bodies based on seismic attributes in a supervised learning method.…”

Section: Introductionmentioning

confidence: 99%

Deep-learning inversion: A next-generation seismic velocity model building method

2019

View full text Add to dashboard Cite

Seismic velocity is one of the most important parameters used in seismic exploration.Accurate velocity models are key prerequisites for reverse-time migration and other high-resolution seismic imaging techniques. Such velocity information has traditionally been derived by tomography or full-waveform inversion (FWI), which are time consuming and computationally expensive, and they rely heavily on human interaction and quality control. We investigate a novel method based on the supervised deep fully convolutional neural network (FCN) for velocity-model building (VMB) directly from raw seismograms. Unlike the conventional inversion method based on physical models, the supervised deep-learning methods are based on big-data training rather than prior-knowledge assumptions. During the training stage, the network establishes a nonlinear projection from the multi-shot seismic data to the corresponding velocity models.During the prediction stage, the trained network can be used to estimate the velocity models from the new input seismic data. One key characteristic of the deep-learning method is that it can automatically extract multi-layer useful features without the need for human-curated activities and initial velocity setup. The data-driven method usually requires more time during the training stage, and actual predictions take less time, with only seconds needed. Therefore, the computational time of geophysical inversions, including real-time inversions, can be dramatically reduced once a good generalized network is built. By using numerical experiments on synthetic models, the promising performances of our proposed method are shown in comparison with conventional FWI even when the input data are in more realistic scenarios. Discussions on the deep-learning methods, training dataset, lack of low frequencies, and advantages and disadvantages of the new method are also provided.

show abstract

Facies classification using machine learning

Cited by 177 publications

References 2 publications

Integrated reservoir characterization of low resistivity thin beds using three-dimensional modeling for natural gas exploration

Integrated reservoir characterization of low resistivity thin beds using three-dimensional modeling for natural gas exploration

A machine learning approach to facies classification using well logs

Deep-learning inversion: A next-generation seismic velocity model building method

Contact Info

Product

Resources

About