Log interpretation for lithology and fluid identification using deep neural network combined with MAHAKIL in a tight sandstone reservoir

He, Mei; Gu, Hanming; Wan, Huan

doi:10.1016/j.petrol.2020.107498

Cited by 51 publications

(11 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Machine learning is a suitable approach for automated zonation and characterization of multi-dimensional data, with several examples of its successful application to geophysical logs, such as seismic velocity, resistivity, gamma ray, and average neutron density porosity (e.g., Raeesi et al 2012;Grana et al 2017;Caté et al 2017;He Communicated by: H. Babaie Fukishima Renewable Energy Institute, National Institute of Advanced Industrial Science and Technology, Koriyama, Fukushima, Japan et al 2020;Feng 2020). Machine-learning classification of geophysical data can help with objectively classifying the subsurface's physical properties and has also been used to identify natural-resource reservoirs (e.g., oil/gas).…”

Section: Introductionmentioning

confidence: 99%

Characterization of hydrothermal alteration along geothermal wells using unsupervised machine-learning analysis of X-ray powder diffraction data

et al. 2021

View full text Add to dashboard Cite

Zonal distribution of hydrothermal alteration in and around geothermal fields is important for understanding the hydrothermal environment. In this study, we assessed the performance of three unsupervised classification algorithms—K-mean clustering, the Gaussian mixture model, and agglomerative clustering—in automated categorization of alteration minerals along wells. As quantitative data for classification, we focused on the quartz indices of alteration minerals obtained from rock cuttings, which were calculated from X-ray powder diffraction measurements. The classification algorithms were first examined by applying synthetic data and then applied to data on rock cuttings obtained from two wells in the Hachimantai geothermal field in Japan. Of the three algorithms, our results showed that the Gaussian mixture model provides classes that are reliable and relatively easy to interpret. Furthermore, an integrated interpretation of different classification results provided more detailed features buried within the quartz indices. Application to the Hachimantai geothermal field data showed that lithological boundaries underpin the data and revealed the lateral connection between wells. The method’s performance is underscored by its ability to interpret multi-component data related to quartz indices.

show abstract

Section: Introductionmentioning

confidence: 99%

Characterization of hydrothermal alteration along geothermal wells using unsupervised machine-learning analysis of X-ray powder diffraction data

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Thus, a full description of the lacustrine shale reservoir is necessary to be associated with their lithofacies. Many classification systems originate from (1) logging, core explanation, and diagenesis [53]; (2) deep neural network [54]; and (3) lithological characteristics and rock sections. According to the marine, transitional, and lacustrine shale lithofacies classification, in this study, combined with the previous lithofacies division scheme, the 4 Geofluids result of XRD is associated with the value of OM to form the lithofacies classification in the study area [55,56].…”

Section: Resultsmentioning

confidence: 99%

Investigation into the Pore Structure and Multifractal Characteristics of Shale Reservoirs through N2 Adsorption: An Application in the Triassic Yanchang Formation, Ordos Basin, China

Liang

Jiang

et al. 2021

Geofluids

View full text Add to dashboard Cite

To understand the pore structure and heterogeneity of pore size distribution (PSD) is essential for revealing fluid mechanics and evaluating the utilization of unconventional resources. In this study, there are multiple shale examples collected from the Chang 7 section in the Ordos Basin for the investigation was conducted on the basis of various experiments on total organic carbon (TOC), X-ray diffraction (XRD), and nitrogen gas adsorption, through scanning electron microscopy (SEM) and multifractal method. The multifractal characteristic parameters, including the width of singularity spectra ( Δ α ), Hurst exponent ( H ), D 1 / D 0 , and nitrogen gas adsorption, were used to find out about the characteristics of pore development and to quantify the complexity and heterogeneity of pore structure. Depending on the exact mineral composition, the Yanchang Formation of Chang 7 shales is classified into either silty mudstone (SM) or muddy siltstone lithofacies (MS). According to the investigative results, the Chang 7 lacustrine shale features a complex pore system with the pores ranging from 1.5 to 10 nm in diameter. Besides, mesopores contribute significantly to the total pore volume (TPV) and total surface area (TSA). As for TPV and TSA of the SM lithofacies in the samples under investigation, they are nearly 1.09–1.78 and 0.80–1.72 times greater as compared to the MS lithofacies samples. The dominant types of reservoir spaces include organic matter (OM) pore and interparticle pore which are related to inorganic minerals. The value of Δ α is higher for MS lithofacies than for SM lithofacies, indicating a greater heterogeneity of PSD in the MS lithofacies. The pore structure of MS lithofacies is determined mainly by TOC and siliceous mineral content, whereas the influencing factors for SM lithofacies are TOC and clay mineral content. There is a significant relationship between multifractal parameters and pore structure parameters for both SM and MS lithofacies. The TOC of SM and MS lithofacies exhibits a close correlation with Δ α , suggesting that the pores in organic matter are dominated by those nanopores with a complex and heterogeneous pore structure. The rock composition of the lithofacies can affect Δ α to a varying extent, which means that the minerals have an evident impact on the heterogeneity of MS and SM lithofacies.

show abstract

“…Parameter P 1/2 was adopted for drawing a map on the probability paper to identify oil and water strata and estimate Sw, a method called the "normal probability distribution method". Let P=Rt × Φm, where Rt is replaced by deep detection resistivity, Φ is obtained from porosity logging, and cementation index m can be estimated by using a statistical method (Pan et al, 2019;He et al, 2020;Ajaz et al, 2021). p values were calculated for each reservoir, and P 1/2 was plotted against the cumulative frequency on a probability paper.…”

Section: Methodsmentioning

confidence: 99%

The relation of the “four properties” and fluid identification of the carboniferous weathering crust volcanic reservoir in the Shixi Oilfield, Junggar Basin, China

Qin²,

Xie

et al. 2022

Front. Earth Sci.

View full text Add to dashboard Cite

This study addresses the poorly understood physical properties of the Shixi Oilfield reservoir, which consists of a weathered Carboniferous volcanic rocks with strong heterogeneity and in which logging identification and evaluation are difficult. Using the lithology, lithofacies, and reservoir space characteristics of volcanic materials, this comprehensive study uses core, well logging, mud logging, and production testing data to analyze the relationship among the lithology, physical properties, electrical properties, and oil-bearing properties (referred to as the “four properties”) of weathered Carboniferous volcanic crust in addition to fluid identification. 1) The lithology of Carboniferous volcanic crust is dominated by breccia lava, agglomerate, banded lava, and compact tuff, and the lithofacies are mainly effusive facies. Secondary pores and tectonic fissures are important reservoir spaces, and the corrosion-fracture pores are significant for reservoir properties. 2) The “four properties” of volcanic reservoirs in the study area have clear relationships. On this basis, data on the electrical properties of the material, such as interval transit time, density, and neutron, were used to establish a logging interpretation model of the properties and oil saturation of the volcanic rock. 3) Using the resistivity-porosity cross-plot method, normal probability distribution method, and Rt/Rxo-Rt cross-plot method, volcanic reservoir fluids were identified with coincidence rates of 80%, 63.63%, and 63.63%, respectively. The cross-plot method determines lower limits of the reservoir’s physical properties and oil saturation, yielding porosity>9%, permeability>0.2 mD, and oil saturation>45%.

show abstract

Log interpretation for lithology and fluid identification using deep neural network combined with MAHAKIL in a tight sandstone reservoir

Cited by 51 publications

References 23 publications

Characterization of hydrothermal alteration along geothermal wells using unsupervised machine-learning analysis of X-ray powder diffraction data

Characterization of hydrothermal alteration along geothermal wells using unsupervised machine-learning analysis of X-ray powder diffraction data

Investigation into the Pore Structure and Multifractal Characteristics of Shale Reservoirs through N2 Adsorption: An Application in the Triassic Yanchang Formation, Ordos Basin, China

The relation of the “four properties” and fluid identification of the carboniferous weathering crust volcanic reservoir in the Shixi Oilfield, Junggar Basin, China

Contact Info

Product

Resources

About