- Home
- Conferences
- Conference Proceedings
- Conferences
EAGE 2020 Annual Conference & Exhibition Online
- Conference date: December 8-11, 2020
- Location: Online
- Published: 08 December 2020
361 - 368 of 368 results
-
-
A Robust Phase-Only Reflection Full Waveform Inversion with Multi-Channel Local Correlation
More LessSummaryA phase-only reflection FWI algorithm is proposed using multi-channel local cross correlation objective function. This method focuses on the phase difference between the observed and synthetic datasets and can provide high fidelity and high resolution model parameters. With proper constraints in both data (i.e., energy weighting) and model (i.e., dynamic minimum total-variation) domains, this new algorithm produce superior models with high resolution which resolves many geological features like sand injectites and volcanic sills; it also produce a big uplift to the velocity model when it is applied to the Santos project which significantly improves the image of salt bodies and pre-salt structures.
-
-
-
Development of Guideline for Monitoring, Measurement and Verification (MMV) of CO2 Storage in Malaysia
By S. Mohd AminSummaryMonetization of the high CO2 fields require economical removal and safe storage of the CO2, since gas flaring is no longer an option. Geological storage sites such as depleted gas reservoirs provide readily available sites for the storage of CO2. In order to develop these fields, there is a need to manage the health, safety and environment (HSE) elements. The main issue for high CO2 fields are their contribution to overall Climate Change as high CO2 from these fields will contribute to the overall Greenhouse Gas (GHG) inventory to the atmosphere.
Leakage is the main concern in any Carbon Capture and Storage (CCS) project. To monitor any potential leakage in CCS fields, there is a need of Measurement, Monitoring and Verification (MMV) program. This guideline will provide the minimum requirement as per practiced by industry for planning, implementation and audit process for MMV process specifically for any CO2 storage projects in depleted oil and gas reservoirs. This program is designed to establish the framework and procedures that will ultimately define the MMV plan. The main objectives are to screen and evaluate available monitoring technologies and develop technologies based on program requirements and cost effectiveness.
-
-
-
Fully Reversible Neural Networks for Large-Scale 3D Seismic Horizon Tracking
More LessSummaryNeural networks are a successful tool for horizon tracking for the interpretation of seismic images. So far, most research works with 2D inputs. In 3D, most work is restricted to relatively small 3D inputs because of memory limitations. Training a neural network typically requires the storage of the network states for every layer. This becomes a problem for deep networks and large data inputs. We avoid this problem by employing recently introduced fully reversible convolutional networks that require storage of the network states for a few layers only. Therefore, we can use much larger 3D input data than in the case of non-reversible networks. A field data example illustrates that fully reversible neural networks are suitable for horizon tracking and allow the input size to increase by a least an order of magnitude, such that we can also learn from structures with a larger length scale.
-
-
-
Turbidite Fan Interpretation in 3D Seismic Data by Point Cloud Segmentation Using Machine Learning
Authors Q. Corlay, V. Demyanov, D. McCarthy and D. ArnoldSummaryThis paper presents a novel method for interpreting turbidite fan in 3D seismic through a new Point Cloud Segmentation-based workflow. Fan detection is a key issue for oil and gas purposes. Indeed, turbidite systems forms excellent hydrocarbon reservoirs and traps. Turbidite deposits are hard to pick in the seismic due to their complex 3D geomorphology. The proposed method relies on a two-steps workflow. First, a point selection to create a seismic point cloud by filtering the relevant seismic data using seismic attributes such as amplitude and coherency. Then, a point segmentation to individualize the geological features present in the point cloud using an unsupervised density-based algorithm, DBSCAN. The outcome of this method is a fully segmented seismic point cloud, meaning that every selected points of the 3D seismic is either associated to a cluster of points or classified as noise. The individualization of geological features into clusters provides a better visualization of their 3D morphology. Thus, the method provides a clear delimitation and interpretation of the fan and at the same time provides a better understanding of its surrounding features. In addition, the method offers the advantages of being fast to compute and robust to noise.
-
-
-
Effect of Asphaltene Deposition on the Wettability Alteration of Sandstone Reservoir Rock
Authors J. Taheri-Shakib, M. Rajabi-Kochi and S.A. HosseiniSummaryIn this study, the effect of two asphaltene samples on the wettability properties of sandstone reservoir rock of SOF field was investigated. For this purpose, the relative permeability and contact angle of sandstone before/after asphaltene deposition are evaluated. Each asphaltene sample shows a different behavior on the oil-wetting of reservoir rock with the changes made on contact angle and relative permeability graphs. Also, the elemental analysis of asphaltene and zeta potential of sandstone interface with adsorbed asphaltene is determined. The asphaltene sample containing higher sulfur (S) and oxygen (O) causes asphaltene acidity, and increasing interaction with the rock due to the presence of sulfur-containing functional groups such as thiophene and sulfide and hydroxyl oxygen-containing function groups along with carbonyl group forms carboxylic acid. Moreover, nitrogen reduces the tendency of asphaltene surface adsorption. Based on the zeta potential, the existence of acidic compounds in adsorbed asphaltene make the zeta more negative and shifts towards the oil-wet, whereas the nitrogen makes the surface charge positive.
-
-
-
Mining-Assisted Heavy Oil Production (MAHOP)
Authors S. Canbolat, H. Ozturk and S. AkinSummaryThis research aims to investigate and compare the ultimate recovery from the largest oil reserve in Turkey (1.85 billion barrels) using a new method called mining-assisted heavy oil production (MAHOP) with conventional SAGD. Tunnels will be excavated from the surface to the reservoir. Fan-shaped up holes will then be drilled in the reservoir from the tunnels.
Heavy oil production through these tunnels will be explored using SAGD method. Several numerical models have been designed using CMG’s STARS simulator. Since the fan wells are opened vertically and at certain intervals along the tunnel, both a tight vertical fracturing of these wells and a separate fracture network formed by micro fractures in the vicinity of the fan holes are formed.
The validation of these hypotheses has been conducted in CMG which showed that MAHOP gave better results compared to conventional SAGD where two horizontal wells are used. MAHOP gave better recovery values with less steam oil ratios. With the results of the simulation study a laboratory model was designed. Experimental operational parameters using three different wettability cases were simulated to observe recovery by considering several possible physical effects such as steam distillation and in-situ upgrading. Saturation and pressure distributions were also obtained.
-
-
-
Forward Stratigraphic Modeling Conditioning and Integration with Geostatistics
Authors A. Miller and S. CourtadeSummaryGeostatistical reservoir modeling techniques are often insufficient for describing reservoir geometry and properties distribution because they do not place geologic features according to the natural order and interactions that created them. Furthermore, they tend to disrupt naturally occurring permeability structures.
Geologic process modeling (GPM), is a forward stratigraphic modeling (FSM) technology that uses quantitative deterministic physical principles to model erosion, transport, and deposition of clastic sediments, as well as carbonate growth and redistribution. These models reproduce reservoir architectures and property distributions in a more realistic way than traditional reservoir modeling techniques.
Conditioning these forward models to honor the well and seismic data for a specific dataset is possible, but the computational effort makes it impractical to honor all the details at the reservoir characterization scale. An alternative approach involves integrating the forward model and the interpreted data using geostatistical techniques. In this work, we discuss the details of the FSM conditioning and the integration of the FSM with geostatistic workflows, illustrating them with examples from field data.
-
-
-
Five Years of Experience Applying AI in the Subsurface Data Management Domain
Authors H. Blondelle and A. JunejaSummaryAfter 5 years of implementation of Machine Learning to support the document indexing performed by the data manager, this presentation makes a stepping point about their capabilities.
Different types of indexing are reviewed, from documents classification to enhanced reading of the graphical content thanks to computer vision and lessons learnt are derived.
-