Machine Learning Assisted Interpretation: Integrated Fault Prediction Extraction Case Study from the Groningen Gas Field, Netherlands

Seismic fault detection using machine learning techniques, in particular the convolution neural network (CNN), is becoming a widely accepted practice in the field of seismic interpretation. Machine learning algorithms are trained to mimic the capabilities of an experienced interpreter by recognizing patterns within seismic data and classifying them. Regardless of the method of seismic fault detection, interpretation or extraction of 3D fault representations from edge evidence or fault probability volumes is routine. Extracted fault representations are important to the understanding of the subsurface geology and are a critical input to upstream workflows including structural framework definition, static reservoir and petroleum system modeling, and well planning and de-risking activities. Efforts to automate the detection and extraction of geological features from seismic data have evolved in line with advances in computer algorithms, hardware, and machine learning techniques. We have developed an assisted fault interpretation workflow for seismic fault detection and extraction, demonstrated through a case study from the Groningen gas field of the Upper Permian, Dutch Rotliegend; a heavily faulted, subsalt gas field located onshore, NE Netherlands. Supervised using interpreter-led labeling, we apply a 2D multi-CNN to detect faults within a 3D pre-stack depth migrated seismic dataset. After prediction, we apply a geometric evaluation of predicted faults, using a principal component analysis (PCA) to produce geometric attribute representations (strike azimuth and planarity) of the fault prediction. Strike azimuth and planarity attributes are used to validate and automatically extract consistent 3D fault geometries, providing geological context to the interpreter and input to dependent workflows more efficiently.

Machine Learning as a Silent Observer of Advanced Geoscience Interpretation

Machine learning is proving its successes in all fields of life including medical, automotive, planning, engineering, etc. In the world of geoscience, ML showed impressive results in seismic fault interpretation, advance seismic attributes analysis, facies classification, and geobodies extraction such as channels, carbonates, and salt, etc. One of the challenges faced in geoscience is the availability of label data which is one of the most time-consuming requirements in supervised deep learning. In this paper, an advanced learning approach is proposed for geoscience where the machine observes the seismic interpretation activities and learns simultaneously as the interpretation progresses. Initial testing showed that through the proposed method along with transfer learning, machine learning performance is highly effective, and the machine accurately predicts features requiring minor post prediction filtering to be accepted as the optimal interpretation.

Ultra-Deep LWD Electromagnetic Directional Resistivity for Waterflood Mapping – A Game Changer

Abu Dhabi's thick Lower Cretaceous carbonate reservoirs experience injection water overriding oil. The water is held above the oil by negative capillary pressure until a horizontal borehole placed at the reservoir base creates a small pressure drawdown. This causes the water above to slump unpredictably towards the horizontal producer, increasing water cut and eventually killing the well under natural lift after a moderate amount of oil production. Water slumping is difficult to forecast using the reservoir model. This paper showcases the successful deployment of an ultra-deep electromagnetic directional resistivity (UDDE) instrument to map injection water movement. The UDDE instrument selected for the 6-in. horizontal hole was a 4¾-in. OD multifrequency tool with configurable transmitter-to-receiver spacings. Pre-well modeling using hybrid deterministic 1D resistivity inversions was conducted for the candidate well to investigate the resistivity tool's ability to identify water slumping at distances 60-100 ft TVD above the planned well trajectory. The inversions aided the selection of optimum operating frequencies, transmitter-to-receiver spacings and BHA configuration. During operations, multiple 1D and 3D inversions were run in the cloud real time during drilling to provide simultaneous deep and shallow resistivity inversions for early identification of the water fronts and structural changes, and near wellbore changes to geosteer and maximize reservoir contact in the complex layered reservoir. Real-time 1D and 3D deep inversion results indicated the resistivity tool had a depth of reliable waterflood detection of more than 80 ft. While drilling, an interpreted subseismic fault was encountered which appeared to influence how water moved in the reservoir. Water slumped closest through the sub-seismic fault towards the well path. Past the fault, the waterfront receded upwards away from the well bore. The data proved useful for updating the static model, providing a snapshot of water flood areas, reservoir tops and faults with throw, helping to optimize the completion design to defer water production and enhance oil production. Furthermore, it captured resistivities of target, underlying and overlying reservoirs to integrate with other geology and geophysics data for better reservoir and fluid characterization near the drilled area. The positive results of this case study encouraged field-wide implementation of this technology for waterflood mapping. The information provided allowed petroleum engineers to adjust the completion design to delay water breakthrough. This proactive approach to waterflood field management improves cumulative oil production and recovery factors according to mechanistic models which have been built and tested.

Interacting inclined strike-slip faults in a layered medium

Two inclined, interacting, strike-slip faults, both buried, situated in a viscoelastic layer, resting on and in welded contact with a viscoelastic half space, representing the lithosphere-asthenosphere system, is considered. Solutions are obtained for the displacements, stresses and strains, using a technique involving the use of Green’s functions and integral transforms, for three possible cases - the case when no fault is slipping, the case when one fault is slipping and the other is locked and the case when both the faults are slipping. The effect of sudden movement across one fault on the shear stress near the fault itself and near the other faults has been investigated. Some situations are identified where a sudden movement across one fault results in the release of shear stress near the other fault, reducing the possibility of seismic movements across it. Other situations are also identified where a sudden movement across one fault increases the possibility of seismic fault movements. A detail study may lead to an estimation of the time span between two consecutive seismic events near the mid points of the faults. It is expected that such studies may be useful in understanding the mechanism of earthquake processes and may be identified as an earthquake precursor.

Deformation Mechanisms in Orogenic Gold Systems During Aseismic Periods: Microstructural Evidence from the Central Victorian Gold Deposits, Southeast Australia

In many orogenic gold deposits, gold is located in quartz veins. Understanding vein development at the microstructural scale may therefore provide insights into processes influencing the distribution of gold, its morphology, and its relationship to faulting. We present evidence that deformation processes during aseismic periods produce characteristic quartz microstructures and crystallographic preferred orientations, which are observed across multiple deposits and orogenic events. Quartz veins comprise a matrix of coarse, subidiomorphic, and columnar grains overprinted by finer-grained quartz seams subparallel to the fault trace, which suggests an initial stage of cataclastic deformation. The fine-grained quartz domains are characterized by well-oriented quartz c-axis clusters and girdles oriented parallel to the maximum extension direction, which reveals that fluid-enhanced pressure solution occurred subsequent to grain refinement. Coarser anhedral gold is associated with primary quartz, whereas fine-grained, “dusty” gold trails are found within the fine-grained quartz seams, revealing a link between aseismic deformation and gold morphology. These distinct quartz and gold morphologies, observed at both micro- and macroscale, suggest that both seismic fault-valving and aseismic deformation processes are both important controls on gold distribution.

A local binary patterns/variance operator based on guided filtering for seismic fault detection

Aiming at suppressing noise interference, improving the fault detection ability of seismic data, fully excavating the effective information in seismic data, and further improving the accuracy of fault detection, this study proposes a seismic fault detection method that combines the local binary pattern/variance (LBP/VAR) operator with guided filtering. The proposed method combines the advantages of LBP/VAR and guided filtering to remove noise from seismic data, and can simultaneously smooth the data and preserve linear features. When compared with several existing methods (coherent operator, LBP/VAR operator, LBP/VAR operator based on median filtering, and Canny operator based on guided filtering), the proposed method exhibits a better SNR, a better ability to identify small faults, and robustness to noise. This novel algorithm can control the balance between noise attenuation and effective signal preservation as well as effectively detect faults in seismic data. Therefore, the proposed method effectively improves the fault identification accuracy, facilitates the gas-bearing analysis of the structure, provides guidance for the actual well location deployment of the project, and has important practical significance for oil and gas exploration and development.