Deep Learning for Refined Lithology Identification of Sandstone Microscopic Images
2
Citation
40
Reference
10
Related Paper
Citation Trend
Abstract:
Refined lithology identification is an essential task, often constrained by the subjectivity and low efficiency of classical methods. Computer-aided automatic identification, while useful, has seldom been specifically geared toward refined lithology identification. In this study, we introduce Rock-ViT, an innovative machine learning approach. Its architecture, enhanced with supervised contrastive loss and rooted in visual Transformer principles, markedly improves accuracy in identifying complex lithological patterns. To this end, we have collected public datasets and implemented data augmentation, aiming to validate our method using sandstone as a focal point. The results demonstrate that Rock-ViT achieves superior accuracy and effectiveness in the refined lithology identification of sandstone. Rock-ViT presents a new perspective and a feasible approach for detailed lithological analysis, offering fresh insights and innovative solutions in geological analysis.Keywords:
Lithology
Identification
Through the research on types and characteristics of lithologic reservoirs of Fuyu oil layer in Honggangbei area, Songliao Basin, such types of lithologic reservoirs is identified from up-dip pinch-out sandstone reservoirs, abandoned channel screened reservoirs, local-oscillation channel sandbody reservoirs, fault-lithologic reservoirs, fault-structural nose reservoirs. Its characteristics are complicated distribution of oil and water, rapid lateral change of reservoirs and monosandbody pool-controlling. Through fine dissection of developed reservoirs, the results show that genetic monosandbody traps are minim oil-controlling units, oil-accumulating units and main controlling factors for lithologic reservoirs of Honggangbei, which it is controlled by vertical layer and planar sedimentary microfacies. Spatial complicated superposition of vertical multi-layer and planar multi-branch monosandbody traps and mis-traps is essential reason of complicated distribution of oil and water.
Lithology
Water cut
Cite
Citations (0)
Protolith
Cite
Citations (0)
There are 4 kinds of oil-and-gas reservoir in Xujiaweizi zone:fault-lithology reservoir,fault block reservoir,tectonic-lithology reservoir and lithology reservoir.The main target stratum is Putaohua oil layer.According to my research,the main styles of oil-and-gas reservior are fault-lithology researvior and fault block researvior,not the pure lithology researvior.Based on reservoir styles and characteristics,through analyzing the distributed law and pool forming period,I summarize 4 kinds of pool forming pattern of hydrocarbon migration and accumulation and confirm main control factors of pool forming and distribution.
Lithology
Stratum
Fault block
Petroleum reservoir
Cite
Citations (0)
The discovery of volcanic reservoir in Qingshen gas field of Xujiaweizi area in north Songliao basin supplies a new kind of reservoir for this area.In order to further explore and develop the volcanic gas pool,it is absolutely necessary to study geologic features of volcanic reservoir.The characteristics of lithology and lithofacies about the local volcanic reservoir are studied through multiple methods.It will play an important role for exploration and development of volcanic gas pool.
Lithology
Natural gas field
Cite
Citations (2)
Shehejie Formation of Dongying Sag develops many lithologic reservoirs.The formation and distribution of these lithologic reservoirs are controlled by many graphic factors,including the types of structural belts,the types of dispositional system tracts,the types of reservoir facies,the conditions of source rock and the condition of the temperature-pressure,etc.Through the comprehensive analysis of the characteristics of the lithologic reservoirs in Shehejie Formation of Dongying Sag,it is held that whether the lithologic traps can form oil reservoirs is usually controlled by four major factors:hydrocarbon migration pathways,power condition of fluid flow,physical properties of reservoirs,and the sealing condition of sandbodies.But the formation of the lithologic reservoirs in Niuzhuang Subsag is controlled by three factors.
Lithology
Petroleum reservoir
Cite
Citations (0)
Among lithologic identification methods, ANN has been used widely because its identification result is very credible. In this paper, a test on JIN66 well by method of BP ANN to identify well lithology has been made, and a method to improving convergence speed for BP neural network been explored.
Lithology
Identification
Cite
Citations (11)
Here is a method of interpreting well logs for formations having lithologies that are mixtures of silica, limestone, dolomite, anhydrite, and clay, or that are mixtures of any two specified minerals and clay. Analysis is made for clay content, matrix density, porosity, hydrocarbon saturation, and secondary-porosity index. Introduction During the past several years interpretation methods have been developed to obtain porosity and lithology from combinations of neutron, density, and sonic logs. One of these is the Dual Mineral Method which uses a crossplot of the neutron- and density-log data to arrive at values of porosity and apparent matrix density of the formation. The sonic log is used to indicate zones of secondary porosity and to help define the lithology. The Dual Mineral Method has been in use for a number of years. It gives good results in clean, liquid-saturated formations or clean, gas-bearing formations of known lithology. However, its use is restricted to these cases. This paper describes a more general method, based on the Dual Mineral technique, which has been developed for the interpretation of formations having lithologies that are mixtures of silica, limestone, dolomite, anhydrite, and clay, or that are mixtures of any two specified minerals plus clay. This new method takes account of both formation shaliness and hydrocarbon effects. In a preinterpretation procedure, log corrections are made, and certain parameters are evaluated. By study of crossplots of the log readings over intervals within lithologic units, the main mineral constituents of the rock matrix are identified and the corresponding properties are evaluated. properties are evaluated. Level-by-level computations are then made. Using the values of clay properties determined during preinterpretation, clay content is evaluated at each preinterpretation, clay content is evaluated at each level, and corrections are made for clay content and hydrocarbon effects. For the hydrocarbon correction, a value of hydrocarbon density is used, based on field information or crossplot study. Values of porosity, apparent matrix density, and water saturation are computed. Safeguards are included to minimize the possible effect of adverse borehole conditions on possible effect of adverse borehole conditions on some log readings. A moved-hydrocarbon index can also be calculated. The logs used for this method include a density log, a gamma ray-neutron log, a sonic log, and an induction log or a Laterolog, preferably with an SP curve. The addition of an Rxo log (Microlaterolog or Proximity log) and the value of the hydrocarbon Proximity log) and the value of the hydrocarbon density provides a reliable evaluation of hydrocarbon effect. The sonic log is used to determine lithology in the M-N plot, to evaluate secondary porosity, and to provide a limiting value of porosity in clay-free formations that are caved. The method can still be used without an Rxo log or sonic log, but the reliability of some of the results is diminished. The method is used in a computer program called CORIBAND. A preinterpretation pass is made in the computer, and depth plots and frequency crossplots are produced to determine parameters needed for the final computations. JPT P. 995
Lithology
Anhydrite
Matrix (chemical analysis)
Formation evaluation
Cite
Citations (51)
Lithology
Stratum
Petroleum reservoir
Cite
Citations (0)
Dawangzhuang area in Raoyang sag is abundant in oil and gas resources and it has favorable structural setting.Sandbody developed in the meandering river channel of Dongying Formation,and reservoir features in medium porosity and medium permeability.The first member of Dongying Formation is a regional cap rock,fault is a major channel for vertical oil and gas migration,oil and gas is migrated terracedly and in the shape of T or reticulately.The area has favorable conditions for the formation of lower-generating-upper reservoiring type sandstone lithologic reservoir.It has favorable conditions for hydrocarbon preservation,after oil and gas are accumulated at Neogene Minghuazhen period,because structural movement gets weakened and fault is not developed.Two typical reservoir-forming patterns exist in the area:one is the pattern of channel sandstone lithologic and updip pinchout reservoir in the negative area of downthrow block of contemporaneous fault;the other is the channel sandstone lithologic and lateral pinchout reservoir in the flank of Dawangzhuang structure.The formation and distribution of channel sandstone reservoir of Dongying Formation is mainly controlled by oil and gas resources,structures and lithology.Oil reservoir is mainly lithologic and structural-lithologic.
Lithology
Petroleum reservoir
Neogene
Fault block
Cite
Citations (0)
Based on initial potential data of nearly 4000 production wells and static geologic material including the structure, logging interpretation, sedimentary microfacies and types of abandoned wells, oil—water distribution regularity in the Putaohua oil layer in the Sanzhao Sag, Songliao Basin, was systematically analyzed. Single local (structure, lithologic) trap controls final hydrocarbon accumulations and distribution of oil—water in the tertiary structure. Distribution regularity of oil—water is unifying in the single local (structure, lithologic) trap. Distribution regularity of oil—water is different in adjacent and different single local (structure, lithologic) traps. Layers of oil—water distribution are controlled by distribution characteristics of reservoir, and boundaries of oil—water distribution are controlled by the array modes of fault and sandbody. In similar conditions, oil-control capacities are weakened from horst fault-block reservoirs→antithetic normal fault structural nose reservoirs→herringbone antithetic normal fault reservoirs→fault—microstructure reservoirs→fault—lithologic reservoirs, lithologic reservoirs→ other reservoirs.
Lithology
Horst
Fault block
Cite
Citations (0)