The buried rocks are the direct materials to uncover the geological formation, such as grain size, sedimentary structure, facies and depositional sequence. Normally, the only available real rocks are cores, but cores cannot offer uncored section information. Due to the low-resolution and non-image, the conventional logging programs are not able to resolve the grain size or sufficiently identify the centimeter-scale lithostrome. The borehole electrical imaging tools sample the formation every 0.2 inch, measure variations in micro-resistivity along the borehole wall and leave no doubt that the conductivity contrast is sufficient to reflect the grain size of rocks. The borehole electrical images provide a wealth of geologic data obtainable only from cores previously. Even though spatial resolution of features visible on the images (down to 0.2 inch) is lower than that on a core, borehole images can offer major geologic information where core is absent [1,2,3,4,5]. However, the generated image essentially provides numerical data in the form of log curves which may be useful to quantify the images and classify grain size of rocks . The only related literature is that of Sovich (1993), who picked conductivity thresholds of Formation MicroScanner (FMS) image to quantitatively define the facies of cobbles, sand matrix, silty sand and non-reservoir rock . His purpose was to estimate the proportion of grain sizes roughly by the conductivity threshold values, but not to classify grain size. Then, to know the grain size from borehole electrical images, the usual approach is only qualitative by the geological and logging workers, not automatic.
In Lower Triassic Baikouquan Formation, Mahu Depression, Junggar Basin, fan-delta deposits were developed, and the lithology is mainly conglomerate. As one of the most important petroleum formations, the borehole electrical images are acquired to realize the vertical change rules of grain sizes. The qualitative way is obviously too slow and with subjective factors in a number of identification tasks. A new automatic method is warranted to provide a solution to the location and grain size of rocks and to better distinguish rock type in complex lithologies. Taking Baikouquan Formation as an example, the goal of this study is to propose an automatic and fast way to classify various grain sizes of rocks. First, borehole electrical images of various grain sizes are qualitatively compared with cores. Second, four categories of grain sizes are discriminated by averages and variances of the gray values of electrical images. Third, three kinds of gravels or sands are classified by frequency distribution of gray values. Furthermore, auto-classified result is compared with that of core descriptions and experimental grain size distributions in a cored interval of one well.
2 Geological setting
Located in northwestern China, Junggar Basin is one of the most significant oil producing regions in this country (Figure 1a). Several large-scale oil-fields have been built in six hydrocarbon generation depressions. As one of the principal hydrocarbon accumulation sags, “hundred-mails oil-fields” are being established around Mahu Depression [7,8]. Tectonically, the Mahu Depression is located at the front of a foreland in the northwestern margin of the Junggar Basin (Figure 1b), which developed due to the collision and subduction between the Junggar–Turpan and Kazakhstan Plates during the Late Paleozoic [9,10]. Sediments were denuded in the northwestern margin and transported by gravity and flow into the Mahu Depression, which takes the Carboniferous Formation as the basement. These numerous coarse sediments accumulated on the slope to form alluvial fans, fan-deltas or fluvial delta deposits from Permian to Quaternary. In Lower Triassic Baikouquan Formation, the coarse grains stacked as fan-deltas, named Zhongguai, Karamay, Huangyangquan, Xiazijie, Xiayan and Dabasong Fan-delta around the depression (Figure 1c) [11,12,13]. The data used in this article are from Huangyangquan and Xiazijie Fan-delta. In addition, later tectonic activities were weak in this formation, which resulted in poor natural fractures in conglomerates.
Since 2010, the Baikouquan Formation gradually became a key petroleum exploration target zone, and multiple high-yielding intervals were found in conglomerates. Abbreviated as T1b, Baikouquan Formation is divided into three members, T1b1, T1b2 and T1b3 upwards in stratigraphy. With a thickness of 30–50 m, T1b1 is mostly thick gray, grayish-green and brown conglomerates; with a thickness of 60–100 m, T1b2 is mainly thick grayish-green conglomerates embedded thin brown mudstones; with a thickness of 40–90 m, T1b3 is dominantly thick gray sandstones and gray-brown mudstones embedded thin grayish-green conglomerates (Figure 2) [14,15,16]. The coarse grains are major lithology of reservoir.
3 Databases and methodologies
Borehole electrical imaging logs have been used to resolve lithologies, sedimentary structures, depositional microfacies and environments in petroleum geology successfully [17,18,19,20,21]. To understand the depositional system of Baikouquan Formation, borehole electrical imaging tools were run in 21 wells. The following two kinds of tools were used: (1) Fullbore Formation MicroImager (FMI) served by Schlumberger Corporation, with 196 buttons distributed on eight pads in contact with the borehole wall; and (2) Extended-Range MicroImager (XRMI) developed by Halliburton Corporation, with 150 buttons distributed on six pads. A number of resistivity data are converted into electrical images by the following two filters: (1) static normalization, rescaling resistivity to the highest and lowest values of the logging run; and (2) dynamic normalization, rescaling the data in a narrow interval (1 m in this article). For the higher contrast, the dynamic image artificially enhances the geological textures . The images are scaled in gray values ranging from 0 to 255, and bright indicates high resistivity and dark-brown color indicates high conductivity (Figure 3). In this article, the gray values of two-dimensional dynamic electrical images are extracted to classify grain sizes of rocks.
Cores were obtained from 15 wells where the borehole electrical imaging tools were run. Cores were descripted and the diameters of gravels were measured by straightedge to classify the grain sizes. Cores were photographed or scanned into computers to digitally confirm the characteristics of lithology. The core works were done independently of the image analyses and compared later with the auto-classification of grain sizes. Based on the diameter range, grain size of rocks in the Baikouquan Formation has been classified in many literature studies, which is divided into four categories (gravel, sand, silt and clay) and nine classifications (boulders, cobbles, pebbles, granules, coarse sandstones, medium sandstones, fine sandstones, siltstones and mudstones) (Table 1) [22,23]. Individual boulders are rarely observed in core samples as well as in electrical images. Hence, boulders will not be considered in this study.
Diameters of boulders are even larger than those of cores and they are hard to be observed in electrical images.
Ultrahigh-resolution digital information derived from borehole electrical images provides detailed grain size characteristics and plays a key role in understanding the depositional environments of the overall formation. Taking the fan-delta deposits of Lower Triassic Baikouquan Formation in Mahu Depression for example, the method using borehole electrical images to auto-classify grain size of rocks requires knowledge of primary lithological features and comprises three stages in this article.
The first stage, comparison of borehole electrical images with descriptions and photographs of cores, involves (1) eliminating the depth error between images and cores and (2) summarizing the borehole electrical image feature of the four categories and eight classifications qualitatively. This is conducive to the general sight for different grain size rocks.
The second stage, extraction of the gray values of dynamic borehole electrical images, consists of (1) calculating the average and variance of gray values at each depth point and (2) discriminating four categories of grain sizes using averages and variances. At individual depth point, average of grays is the sum of all grays divided by the total number of values, and variance of grays is defined by the quadratic sum of difference between gray and average divided by the total number of values:
The third stage, based on the above identified result, embraces (1) calculating frequency distribution of 25 gray class intervals and (2) dividing three classifications of grain sizes in gravels or sands according to the gray frequency distribution. The gray frequencies of each depth are defined by the number of values falling into each interval which was then divided by the total number of values:
4.1 Comparison of borehole electrical images with cores
Individual gravelly grain is very obvious on the core cylindrical surfaces and their photos. According to the laboratory quantitative identification of clastic components, volcanic tuff, felsite, andesite, granite and rhyolite are the dominated rock types of gravelly grains . These grains within different volcanic materials present different conductivity, which corresponds to different gray values in borehole electrical images. Therefore, the gravels are elaborate “mottled” in two-dimensional electrical images. Benefitting from the ultrahigh-resolution, the grain sizes are clear to be classified in the visualized images. The spots within different gray values indicate the approximate gravelly grains: the bigger the spots, the coarser the gravels (Figures 4a, b, 5a, b and 6a, b) [24,25,26,27].
The sands transported by braided rivers or distributary channels of the fan-delta system are usually well sorted. The sizes of sands range from 0.06 to 2 mm, which is below the resolution of borehole electrical imaging tools. As a result, the individual sandy grain could not be observed in electrical images. Being different from gravels, the images of sands are homogeneous bright colored. However, the coarse sandstones are bright (Figure 7a and b), medium sandstones are bright-yellow (Figure 8a and b) and fine sandstones are yellow (Figure 9a and b). It means that the coarser the sandy grains, the brighter the images [24,25,26,27].
The silts and clays accumulated in quiet water have well conductivity, and silty and muddy intervals have medium-low and the lowest resistivity, respectively. Correspondingly, electrical images of both are homogeneous brown (Figure 10a and b) and dark-brown colored, respectively (Figure 11a and b) [24,25,26,27].
4.2 Auto-discrimination of four categories
In general, the gravelly formations have low conductivity, whereas silts and clays have low resistivity in open-hole conventional logs [28,29]. The borehole electrical images are composed of a number of gray points, which can be extracted gray averages and variances of images at each depth point. For the whole Baikouquan Formation in one well, gray average and variance curves are processed to auto-discriminate four grain size categories.
Electrical images of gravels are inhomogeneous mottled, mixed with various gray values. The averages of gravelly grays are medium and variances are high (Figures 4c, 5c and 6c). Electrical images of sands are homogeneous bright or yellow colored, within similar gray values. Averages and variances of sandy grays are both middle-low (Figures 7c, 8c and 9c). Electrical images of silts are homogeneous brown colored, and gray averages are high and variances are low (Figure 10c). Images of clays are homogeneous dark-brown colored, and gray averages are the highest and variances are low (Figure 11c). For the aforementioned features, gray averages and variances of gravels, sands, silts and clays are drawn in cross-plot, which is effective to auto-discriminate four grain size categories of rocks (Figure 12). The gray variances of gravels are about larger than 3 × 103, whereas those of other grains are lower than this threshold. The gray averages of sands are less than 170, those of silts are between 160 and 200 and those of clays are larger than 200.
4.3 Auto-classification of gravels or sands
After grain sizes of rocks are distinguished into four categories in the previous step, the third step is auto-classifying grain size in gravels or sands internally. The corresponding two-dimensional gray frequency profiles of eight grain size classifications are plotted (Figures 4d, 5d, 6d, 7d, 8d, 9d, 10d and 11d). In addition, at each depth point, the one-dimensional frequency distribution curves are also derived from frequency profiles (Figure 13). These two-dimensional profiles and one-dimensional curves could be used to auto-classify three kinds of gravels or sands.
In gravels, the gray frequency curves of cobbles are unimodal. The gray values of peaks are less than 50, and the corresponding frequency percentages are larger than 15% (Figures 4d and 13a). The gray frequency curves of pebbles are unimodal or bimodal, one peak at low gray, less than 50, and another peak at high gray values, larger than 200. The frequency percentages of both peaks are between 10% and 20% (Figures 5d and 13b). The gray frequency curves of granules are without obvious peaks; gray values are widespread; and the almost gray frequency percentages are less than 10% (Figures 6d and 13c). All these features are shown in the cross-plot of gray values and frequency percentages of peaks (Figure 14a). Centralized regions of cobbles, pebbles and granules are distinguished from each other.
In sands, the gray frequency curves are unimodal. However, the gray values of peaks are different from each other: the primary gray values of coarse sandstones are less than 50 (Figures 7d and 13d); the main gray values of medium sandstones are between 50 and 160 (Figures 8d and 13e); the dominated gray values of fine sandstones range from 160 to 240 (Figures 8d and 13f). In the cross-plot of gray values and frequency percentages of peaks, these three types of sandstones are differentiated by the distribution of dominated gray values easily (Figure 14b).
Borehole electrical images have been primarily used for the qualitative recognition of grain sizes and depositional structures. The excellent resolution of the borehole electrical imaging log makes it possible to discriminate features that are not seen in conventional logs. The ability to discriminate grain size leads to the capability to quantify the image through gray values of digital images. The auto-classified criterion of different grain size rocks using borehole electrical images is summarized in Table 2. Four categories of grain sizes are auto-discriminated by the averages and variances of gray values. Three kinds of gravels or sands could be auto-classified by the frequency distribution of the gray values. This auto-classified method has been successfully applied to identify the grain sizes of fan-delta deposits in Baikouquan Formation, Mahu Depression. Taking 2775.45–2777.59 m interval of Well FN10 for example, the photographs of core, grain size description of core, dynamic borehole electrical image, gray average and variance curves, two-dimensional gray frequency profile and auto-classified result are shown in Figure 15, and six experimental grain size distributions are given in Figure 16.
Auto-classified criterion of different grain size rocks using borehole electrical images
|Categories||Average||Variance||Classification||Peak’s gray||Peak’s frequency (%)|
|Gravel||50–175||>3 × 103||Cobble||<50||>15|
|Pebble||<50 or >200||10–20|
|Sand||0–170||1 × 103–3 × 103||Coarse||<50|
|Silt||160–200||1 × 103–3 × 103|
|Clay||200–255||<1 × 103|
Average and variance of grays are used to auto-discriminate four categories. Frequency distributions of gray values are used to auto-classify three kinds in gravels or sands.
The first channel shows the core photographs, and core recovery was near 100%, despite that very few cores are broken and lost by sampling (Figure 15). The second channel shows the core description result which is made in the core library (Figure 15). Because of the understanding level of grains, the observer considered that the gravels are all pebbles. Besides the depth error with logs, the depth of artificial results is not entire. The third channel is the depth of experimental grain size distributions. The fourth channel is the logging depth. The fifth channel shows the borehole electrical images collected by XRMI (Figure 15). The average and variance curves of image’s grays are shown in the sixth channel (Figure 15). According to the auto-discriminated criterion of the categories, silts, sands and gravels are identified, which is shown in the seventh channel (Figure 15). The result of auto-discriminated categories is close to that of core description: in the gravelly interval of cores, the rocks are gravels as well (2776.12–2777.18 m and 2777.31–2777.58 m). The eighth channel gives the two-dimensional gray frequency profiles of images (Figure 15). Cobbles, pebbles, granules, coarse sandstones and medium sandstones are recognized, respectively, at different intervals in the ninth channel, according to the auto-classified criterion of three kinds of gravels or sands (Table 2). In addition, the grain sizes of six samples were analyzed (Figure 16). Grain size results are gained from the grain size distributions, and they are in accordance with the auto-classified results (Figure 15). Therefore, same as experimental grain size distribution results and more elaborate than core description results, the auto-classified results demonstrate the availability and feasibility of this method.
Compared to the result of core description, the auto-classified result has the following two advantages: (1) no qualitative influence of core observer’s knowledge; and (2) integral and continuous depth in the imaging logging running intervals. In addition, the method of this study would be applied to the near provenance conglomeratic deposits, such as alluvial fans and fan-deltas. For the short transportation distance, the grain sizes of gravels within different mineral components are usually larger than the resolution of borehole electrical images. Their conductivity must be different from that of interstitial matters filled in the spaces of grains. These borehole electrical images of conglomeratic formations are visualized to identify the grain sizes of rocks, and digital attributes derived from high-quality images can be certainly used to auto-classify grain sizes by this proposed method.
Taking Baikouquan Formation in Mahu Depression for example, an automatic method is proposed to classify the grain size of various rocks using borehole electrical images, which includes comparing electrical images with cores, auto-discriminating four categories roughly and auto-classifying three kinds of gravels or sands finely. The gravels are “mottled” in two-dimensional borehole electrical images, and the bigger the spots, the coarser the gravels. The images of sands are homogeneous bright colored, and the coarser the sandy grains, the brighter the images. The electrical images of silts and clays are homogeneous brown and dark-brown colored.
The gray averages and variances of electrical images are calculated to auto-discriminate gravels, sands, silts and clays. The variances of gray values of gravels are high, whereas those of sands are medium usually. The variances of gray values of silts and clays are both low, but the gray averages of silts are between 160 and 200, and gray averages of clays are larger than 200.
The two-dimensional gray frequency profiles and one-dimensional gray frequency distribution curves of borehole electrical images are counted to auto-classify cobbles, pebbles and granules, or coarse sandstones, medium sandstones and fine sandstones. The peaks’ gray values of cobbles are less than 50 and frequencies are larger than 15%. The peaks’ gray values of pebbles are less than 50 or larger than 200 and frequencies are between 10% and 20%. The almost gray frequencies of granules are less than 10%. The dominated gray values of coarse sandstones, medium sandstones and fine sandstones are less than 50, between 50 and 160 and ranged from 160 to 240, respectively. In a cored interval, the availability and feasibility of the proposed method are verified by core descriptions and experimental grain-size distributions, and it can be used to auto-classify grain size of rocks in near provenance conglomeratic environments.
The authors are grateful to anonymous reviewers for their constructive reviews on the manuscript and the editors for carefully revising the manuscript. This research was financially supported by the Hubei Provincial Natural Science Foundation of China (No. 2019CFB343), the Scientific Research Project of Hubei Provincial Department of Education (No. Q20181310), the Open Fund of Key Laboratory of Exploration Technologies for Oil and Gas Resources (Yangtze University), Ministry of Education (No. K2018-21), the National Science and Technology Major Project (No. 2016ZX05027-002-007) and the National Naturel Science Foundation of China (No. 41772094). These supports are gratefully acknowledged.
Author contributions: Rui Yuan produced the borehole electrical images and frequency profiles and prepared the manuscript with contributions from all co-authors. Rui Zhu proposed the method, guided the thought of this study and provided the geological knowledge. Xuguang Guo, Lei Zhang, Zhiyuan An and Jun Wu are oilfield researchers and offered the initial borehole electrical imaging logging data and core scanned photographs. Kang Zhao processed part data. The authors applied the sequence-determines-credit approach for the sequence of authors.
Trouilier JC, Delhomme JP, Carlin S, Anxionnaz H. Thin-bed reservoir analysis from borehole electrical images. San Antonio: SPE 19578. 1989;61–72.
Bourke LT. Sedimentological borehole image analysis in clastic rocks: a systematic approach to interpretation. Geol Soc Spec Publ. 1992;65:31–42.
Xu CM, Cronin TP, McGinness TE, Steer B. Middle Atokan sediment gravity flows in the Red Oak field, Arkoma Basin, Oklahoma: a sedimentary analysis using electrical borehole images and wireline logs. AAPG Bull. 2009;93(1):1–29.
Thompson L. Atlas of borehole imagery, 2nd edn. AAPG/Datapages Discovery Series No. 13, Tulsa, Oklahoma, USA; 2009.
Nian T, Jiang ZX, Wang GW, Xiao CW, He WJ, Fei LY, et al. Characterization of braided river-delta facies in the tarim basin lower cretaceous: application of borehole image logs with comparative outcrops and cores. Mar Pet Geol. 2018;97(6):1–23.
- Export Citation
Nian T, Jiang ZX, Wang GW, Xiao CW, He WJ, Fei LY, et al. Characterization of braided river-delta facies in the tarim basin lower cretaceous: application of borehole image logs with comparative outcrops and cores. Mar Pet Geol. 2018;97(6):1–23.)| false 10.1016/j.marpetgeo.2018.06.024
Sovich JP, Newberry B. Quantitative applications of borehole imaging. SPWLA 34th Annual Logging Symposium, 1993, FFF.
Kuang LC, Tang Y, Lei DW, Wu T, Qu JH. Exploration of fan-controlled large-area lithologic oil reservoirs of Triassic Baikouquan Formation in slope zone of Mahu Depression in Junggar Basin. China Pet Exploration. 2014;19(6):14–23, (in Chinese with English abstract).
Wu F, Xi YP, Fan QC, Yao C, Cong LL, Zhang FS, et al. Influence of spatial distribution of pores on NMR transverse relaxation time in pebbly sandstone. J Magnet. 2019;24(4):1–13.
He DF, Yin C, Du SK, Shi X, Ma HS. Characteristics of structural segmentation of foreland thrust belts – a case study of the fault belts in the northwestern margin of Junggar Basin. Earth Sci Front. 2004;11(9):91–101, (in Chinese with English abstract).
Lei ZY, Lu B, Wei YJ, Zhang LP, Shi X. Tectonic evolution and development and distribution of fans on northwestern edge of Junggar Basin. Oil Gas Geol. 2005;26(1):86–91, (in Chinese with English abstract).
Zhang SC, Zou NN, Shi JA, Chang QS, Lu XC, Chen B. Depositional model of the Triassic Baikouquan Formation in Mabei area of Junggar Basin. Oil Gas Geol. 2015;36(4):640–50, (in Chinese with English abstract).
Zou NN, Shi JA, Zhang DQ, Ma CY, Zhang SC, Lu XC. Fan delta depositional model of Triassic Baikouquan Formation in Mabei Area, NW Junggar Basin. Acta Sedimentol Sin. 2015;33(3):607–15, (in Chinese with English abstract).
Yuan R, Zhu R, Qu JH, Wu J, You XC, Sun YQ, et al. Utilizing Integrated Prediction Error Filter Analysis (INPEFA) to divide base-level cycle of fan-deltas: a case study of the Triassic Baikouquan Formation in Mabei slope area, Mahu Depression, Junggar Basin, China. Open Geosci. 2018;10(1):79–86.
- Export Citation
Yuan R, Zhu R, Qu JH, Wu J, You XC, Sun YQ, et al. Utilizing Integrated Prediction Error Filter Analysis (INPEFA) to divide base-level cycle of fan-deltas: a case study of the Triassic Baikouquan Formation in Mabei slope area, Mahu Depression, Junggar Basin, China. Open Geosci. 2018;10(1):79–86.)| false 10.1515/geo-2018-0007
Xiao M, Yuan XJ, Wu ST, Cao ZL, Tang Y, Xie ZR, et al. Conglomerate reservoir characteristics of and main controlling factors for the Baikouquan Formation, Mahu sag, Junggar Basin. Earth Sci Front. 2019;26(1):212–24, (in Chinese with English abstract).
Yuan R, Zhu R, Qu JH, You XC, Wu J, Huang YF. Abnormal open-hole natural gamma ray (GR) log in Baikouquan Formation of Xiazijie Fan-delta, Mahu Depression, Junggar Basin, China. Open Geosci. 2018;10(1):844–54.
Tang Y, Xu Y, Li YZ, Wang LB. Sedimentation model and exploration significance of large-scaled shallow retrogradation fan delta in Mahu Sag. Xinjiang Pet Geol. 2018;39(1):16–22, (in Chinese with English abstract).
Donselaar ME, Schmidt JM. Integration of outcrop and borehole image logs for high-resolution facies interpretation: example from a fluvial fan in the Ebro basin, Spain. Sedimentology. 2005;52:1021–42.
Slatt RM, Davis RJ. Calibrating borehole image and dipmeter logs with outcrops and behind-outcrop cores: Case studies and applications to deep-water deposits. In: Pöppelreiter M, García-Carballido C, Kraaijveld M, editors. Dipmeter and borehole image log technology. vol. 92. Tulsa, Oklahoma, USA: AAPG Memoir; 2010. 167–94.
Folkestad A, Veselovsky Z, Roberts P. Utilising borehole image logs to interpret delta to estuarine system: a case study of the subsurface Lower Jurassic Cook Formation in the Norwegian northern North Sea. Mar Pet Geol. 2012;29:255–75.
Xu CM, Gehenn JM, Zhao DH, Xie GY, Teng MK. The fluvial and lacustrine sedimentary systems and stratigraphic correlation in the Upper Triassic Xujiahe Formation in Sichuan Basin, China. AAPG Bull. 2015;99(11):2023–41.
Brekke H, MacEachern JA, Roenitz T, Dashtgard SE. The use of microresistivity image logs for facies interpretations: an example in point-bar deposits of the McMurray Formation, Alberta, Canada. AAPG Bull. 2017;101(5):655–82.
Zhang CM, Wang XL, Zhu R, Qu JH, Pan J, An ZY. Lithofacies classification of Baikouquan formation in Mahu sag, Junggar basin. Xinjiang Pet Geol. 2016;37(5):606–14, (in Chinese with English abstract).
Qu JH, Zhang L, Wu J, You XC. Characteristics of sandy conglomerate reservoirs and controlling factors on physical properties of Baikouquan Formation in the western slope of Mahu Sag, Junggar Basin. Xinjiang Pet Geol. 2017;38(1):1–6, (in Chinese with English abstract).
Zhou LX. Application of Fullbore Formation Microimager (FMI) to study of glutenite sedimentary structures in Jiyang depression. Xinjiang Pet Geol. 2008;29(5):654–6, (in Chinese with English abstract).
Yan JP, Cai JG, Zhao HM, Li ZZ, Xu GH. Application of electrical image logging in the study of sedimentary characteristics in sandy conglomerates. Pet Explorat Dev. 2011;38(4):15–16, (in Chinese with English abstract).
Yuan R, Zhang CM, Tang Y, Qu JH, Guo XD, Sun YQ, et al. Utilizing borehole electrical images to interpret lithofacies of fan-delta: a case study of Lower Triassic Baikouquan Formation in Mahu Depression, Junggar Basin, China. Open Geosci. 2017;9(1):539–53.
Luo XP, Pang X, Su DX, Lu H, Zhang N, Wang G. Recognition of complicated sandy conglomerate reservoir based on micro-resistivity imaging logging: a case study of Baikouquan Formation in western slope of Mahu Sag, Junggar Basin. Xinjiang Pet Geol. 2018;39(3):345–51, (in Chinese with English abstract).
Schlumberger. Log interpretation principles/applications. Sugar Land, Texas, USA: Schlumberger; 1991.
Darling T. Well logging and formation evaluation. Burlington, Massachusetts, USA: Elsevier; 2005.