Pore Structure Characteristics of Tight Carbonate Reservoir in Qingxi Sag, Jiuquan Basin


 The Xiagou Formation is the main tight oil reservoir in Qingxi Sag of Jiuquan Basin. Given the poor physical properties and other factors restricting tight oil exploitation and production in this area, studies should focus on microscopic pore structure characteristics. In this study, a nano-CT scanner, a SEM, and an NMR were used to study the pore structure characteristics of a tight carbonate reservoir in Qingxi Sag, Jiuquan Basin. The Xiagou Formation reservoir mainly consists of gray argillaceous dolomite and dolomitic mudstone. The pore categories are mainly elliptic, irregular, intergranular, and intragranular and mostly filled with clay and carbonate cement. Pore space is small, the intergranular or organic pores are mostly separated, and pore-throat is weakly connected. The throats mostly develop with lamellar and tube bundle-like characteristics and with poor seepage ability. The pore-throats mostly span from nanometer to micrometer sizes, and pore diameters are mainly concentrated in the range of 0.01–0.1 and 1–10 μm. It is a unimodal pattern mainly composed of micropores, or a bimodal regular allocation dominated by micropores supplemented by macropores. The relationship between micropore (<0.1 μm) and macropore (>1 μm) content allocation and mean pore diameter strongly controls the permeability of reservoir rocks. When macropore content reaches more than 85%, or when pore content totals less than 3%, the permeability of a reservoir remarkably increases. At a higher ratio of the average finest throat sectional area and throat-pore of reservoir rock, the throat radius lies closer to the connecting pore radius, pore and throat connectivity improves, and reservoir seepage ability becomes stronger. Based on reservoir capacity and seepage ability, pore structures of the tight carbonate reservoirs in study area are classified into type I (small-pore–thin-throat), type II (thin-pore–thin-throat), and type III (microporous-microthroat) with rock permeability>0.1 mD, 0.05–0.1 mD, and <0.05 mD, respectively. The type I pore structure reservoir should be regarded as an indicator of tight oil “sweet spots” reservoir in the study area.


Introduction
Unconventional resources including shale oil and shale gas have received great attention in the last few decades [1,2]. Tight oil, as another kind of unconventional resources, has become a research hotspot. Numerous scholars concentrate on tight oil reservoirs, especially on the micropore structure of reservoirs [3][4][5][6]. Tight oil refers to the oil resources in tight reservoirs such as tight sandstone, tight carbonate rock, or mixed rock with permeability less than or equal to 0:1 mD under overburden pressure. In the Jiuquan Basin, the portion of Qingxi Sag is small but enriched in oil and contains mainly carbonate rocks and clastic rocks that are deposited in the sag. Well-developed carbonate rocks and fractured reservoirs are the main oil-producing layers in Yumen Oilfield [7][8][9][10][11]. The pore types are primarily composed of elliptical and irregular intergranular pores. The intercrystalline pores and organic pores are mostly isolated and have poor connectivity. On the micron scale, pore-throat sizes vary, showing a shelf shape, a band shape, and a spherical shape. On the nanoscale, we observed more nanoscale pores. These pores are conchlike and spherical on the surface or interior of mineral grains (crystals). Nanoscale spherical micropores have poor connectivity and are isolated in three-dimensional space, and most of them are reservoir spaces. The nanosized short shelf micropores are connected with the microshelf micropores, which are adjacent to the isolated spherical nanopores and act as both pore-throat and pore structures [12]. In recent years, with the in-depth exploration of Qingxi Sag and the improvement of oilfield development technologies, tight oil has become a key exploration and development resource in the region [13][14][15][16]. Liugouzhuang is located in the central part of the sag. Dolomite and mudstone developed widely with a long and consistent hydrocarbon charging history, yielding abundant oil and gas, whereas numerous tight oil reservoirs have occurred around oil and gas reservoirs in the area [17][18][19]. With the continuous exploration of tight reservoirs in Liugouzhuang, Qingxi Sag, Jiuquan Basin, massive geological data and useful practice experiences are accumulated; however, difficulty arises from analyzing the pore structure of the tight oil reservoir in the Xiagou Formation, and the accuracy of pore structure classification and evaluation remains low [20][21][22][23][24][25]. Therefore, this work is aimed at investigating the pore structure characteristics of the Xiagou Formation in Qingxi Sag, Jiuquan Basin, to discuss the comprehensive division and assessment methods of the pore structure in tight oil carbonate reservoirs, improving classification accuracy of "sweet spots" in the reservoir, and to provide guidance for the exploration and development of tight oil carbonate reservoirs. At the same time, the basic types, geological characteristics, resource potential, and exploration status of different types of continental tight oil in China are deeply studied, and some ideas and suggestions are put forward.

Geologic Setting
The Jiuquan Basin is positioned at the eastern boundary of Heli Mountain, the western border of Hongliuxia, the northern boundary of Dabei Mountain, and the southern boundary of Qilian Mountain [26,27]. The entire basin area is approximately 2:2 × 10 4 km 2 . Jiuquan Basin is divided into two depressions by the Jiayuguan Uplift [28,29]; the east to west basin area is divided into two depressions of Jiudong and Jiuxi. The area of the Jiudong depression measures 1:1 × 10 4 km 2 , whereas the Jiuxi depression features a smaller area of 0:68 × 10 4 km 2 and the three mainly sedimentary developed sags of Qingxi, Shida, and Huahai [30,31] ( Figure 1). The Qingxi Sag in the research area is positioned in the western Jiuxi depression and features a rhombic shape distributed along the northwest direction [32]. The northern part of Qingxi Sag is the Altyn Tagh strike-slip fault, the southern part is the North Qilian thrust fault zone, and the eastern part is bounded by the 509 Fault [33][34][35]. The sag comprises four provenance areas, including east, south, west, and north. Three kinds of depositional systems grew in the Lower Cretaceous strata: fan delta, lake, and sublacustrine fan [36,37]. The current area of Qingxi Sag totals 490 km 2 . The Hongnan and Qingnan subsags and the Qingxi low sag are distributed in the sags from west to east. At present, the actually proven reserves measure 1:62 × 10 8 t, accounting for 76% of the total resources in the Jiuxi depression and earning the "small but fat" reputation [38,39]. The Xiagou Formation is located in the center of the sag, where the source and the reservoir of the strata are integrated [26]. This formation not only serves as a hydrocarbon source rock in the basin but also as desirable oil and gas reservoir with a thickness of about 1000 m. Thick and gray glutenite developed at the bottom of the strata. Two kinds of lithologies are interbedded and developed in the lower part of the strata: a grayish green conglomerate and gray dolomitic mudstone [40]. The upper part of the strata features various lithologies with different thicknesses and interbedded deposits with dark mudstone, shale, dolomitic siltstone, and dolomitic mudstone. The top is occupied with gray mudstone and shale with different thicknesses and interbedded deposits. The strata developed with horizontal and trough crossing bedding [12,41] (Figure 2).

Samples and Methods
3.1. Samples. Twenty core samples were collected from wells Long104, Q2-4, Liu4, and Liu102 in Qingxi Sag. All specimens were acquired from the Xiagou Formation reservoir and observed in terms of carbonate lithology. Rock types were dolomitic mudstone, argillaceous dolomite, and dolomitic siltstone. Organic matters initially comprised types I and II, and total organic carbon (TOC) was 0.5%-2.5% with a peak above 4%. Organic matter with TOC > 1:5% occupied over 50%, indicating that the TOC level was medium to high. Rocks from this formation had a Ro of 0.5% to 1.0%, implying that the majority of the source rocks are mature and are at a hydrocarbon production phase.

3.2.
Methods. The MicroMR12-025V nuclear magnetic resonance (NMR) laboratory apparatus (Niumag Electronic Technology Co., Ltd.) was operated at a resonance frequency of 11.826 MHz, a magnet temperature of 32 ± 0:02°C, and an echo interval of 0.30 ms [42,43]. Cylinder specimens were measured to be 50 mm in length and 25 mm in diameter. The primary samples without any treatment were tested at first. Full water samples were evacuated for 4 h in a pressurized vacuum saturator and saturated under pressure for 16 h.
Using highly precise nanocomputed tomography (CT), we observed and quantified nanopore-throats of a tight reservoir [44][45][46][47][48]. Micro-and nanoscale CT scanning also enabled the examination and experiments of pore-throats at the two scales. Noninvasive 3D rock imaging, which facilitates identification of allocation, size, and connectivity of micro- 2 Lithosphere and nanopores, enabled the characterization of microporeand nanopore-throat networks. Herein, nano-CT (highest resolution of 65 nm) approaches were combined. An UltraXRM-L200 nano-CT scanner from the Nanophysics Lab of the China University of Petroleum, Beijing, was employed.
A 3D pore-throat stereogram was plotted by high-resolution nano-CT scanning of 2 samples with a diameter of 65 μm. Scanning electron microscopy (SEM) can uncover pore and interstitial features [47][48][49][50]. Rock specimens were firstly extracted and cleaned. Then, the fresh surface was processed      4 Lithosphere studied reservoir, 1525 porosity and 584 permeability values from 15 wells, such as Liu9, Long104, and Qing2-4 wells, were calculated. As illustrated in Figure 3(a), reservoir porosity in the study area generally reaches less than 10%, mainly between 1% and 7%, and features poor reservoir characteristics. Figure 3(b) shows that reservoir permeability in the study area mainly lies between 0:1 × 10 −3 and 2 × 10 −2 μm 2 ; thus, the permeability of reservoirs is poor. Permeability values indicate that the Xiagou Formation reservoir in Qingxi Sag is strongly compacted and its permeability and porosity are low. Based on the permeability-porosity relationship, Figure 4 shows that with increasing porosity, permeability generally increased, but no mutual control relationship was observed between them. For rock samples with the same porosity, the maximum difference in permeability values was nearly 4 times. For rock samples with the same permeability, maximum difference in porosity values approximated 5%. Thus, including pore-throat size, allocation, and connectivity, the complex pore structure of a tight reservoir more significantly influenced the pore and permeability of the reservoir.

Occurring Characteristics of Movable Fluid in Reservoir.
The distribution of a saturated movable fluid in the study area ranged from 11.42% to 45.52% and incurred an average of 26.73%, indicating the low concentration of the movable fluid in the studied reservoir and the strong heterogeneity of fluid occurrence. A strong positive correlation was found between permeability and movable fluid saturation, with the correlation coefficient reaching 0.7632 ( Figure 5(a)). In comparison, the correlation of porosity with migrating fluid saturation significantly weakened, presenting almost no remarkable interaction ( Figure 5(b)). The role of permeability in controlling movable fluid saturation is much more significant than that of porosity, whereas the development of secondary pores and microcracks and other factors exhibit strong control on reservoir permeability [52][53][54]. Our results indicate that movable fluid saturation is also affected by the growth of secondary pores and microcracks. Better permeability of a reservoir results in a stronger seepage ability of rocks and a smaller bound impact on the movable fluid, which develops in pores larger than the cutoff pore diameter, thus causing more significant saturation of the corresponding

4.2.
Size and Distribution of Pore-Throat. The average pore diameter, which reflects the average pore size of reservoir rocks, was distributed widely and ranged from 0.06 to 11.69 μm (mean = 2:39 μm; Figure 6(a)). The reservoir space is dominated by nanopores and micropores. Median radius ranged from 16.32 to 161.8 μm (mean = 89:08 μm; Figure 6(b)). Median radius is the pore-throat radius representing 50% mercury saturation in the mercury intrusion experiment. A larger median radius indicates better porethroat seepage performance. Therefore, pore-throats in the study area possess poor seepage capacity.  As shown in the overlay of the T 2 spectral curve ( Figure 7), rock pore diameter distribution in the study area is mainly unimodal, bimodal, or single-and double-peaked transition states with pore diameters generally concentrating at 10 μm. The content of unimodal and bimodal pore diameters mainly falls within 0.01-0.1 and 1-10 μm. Pore diameters of single and double peaks in a transitional state were all concentrated at 10 μm but were mostly within 0.01-0.1 μm. The reservoirs in the study area feature a complex pore structure, and the peak and ranges of pore diameters were significantly different among different samples.
Typical unimodal samples, such as Long104-26 samples, can be observed from Figure 8(a). Pore diameters are mainly concentrated within 0.01-0.07 μm, the content peak of pore diameters are approximately 0.03 μm, and the remaining content of pore diameters is notably small. In the research area, rocks with unimodal pore diameter distribution were widely developed. Considering Long104-26 samples as representative, pore diameter distribution was mainly characterized by micropores, followed by mesopores, whereas macropore content was the lowest. The pore-throat type features small reservoir capacity and poor seepage capacity. Typical bimodal samples, such as the Q2-4-04 core, can be observed from Figure 8(b). Pore diameters were mainly concentrated between 0.01-0.22 and 2-10 μm, with a peak pore diameter content of 0.04 μm. Extremely low content was observed between 0.22 and 2 μm. In this study area, rocks with bimodal pore diameter distribution were widely developed. Considering the Q2-4-04 core as representative, pore diameter distribution was mainly comprised of micropores and macropores, whereas mesopore content was extremely low. Pore-throat type exhibited better reservoir capacity and poor seepage ability. Single-and double-peaked transition state samples, such as the Liu4-02 core, can be observed from Figure 8(c). Pore diameters were mainly concentrated in the range of 0.01-0.69 μm, the peak of pore diameter content was 0.02 μm, and a decrease was noted in the percentage of micropore content. Rocks with single-double peaks of transitional pore diameter distribution in the study area were rarely developed, as represented by Liu4-02. Pore diameter distribution was dominated by micropores and mesopores. This pore-throat type presents good reservoir capacity and strong seepage ability.
(1) Intergranular pore implies primary porosity. Intergranular pores refer to the pores between particles, matrix, and cement, pores supported by particles or matrix, and intergranular residual pores with a reservoir property that is not fully filled with matrix and authigenic cement [55,56]. The studied Xiagou Formation is in the late diagenetic stage. As a result of the effects of strong compaction and authigenic cementation, most of the intergranular pores feature smaller pore diameters. Well-preserved intergranular pores occur mostly between organics and brittle minerals and are supported and protected by brittle minerals [52,53]. The intergranular pores in the study area showed good overall connectivity, significantly con-tributing to the permeability of the reservoir. Such pores possess good oil and gas reservoir properties and good seepage ability and are mainly elliptical or irregular, and their diameters span from the nanometer level to the micrometer level (Figures 9(a) and 9(b)).
(2) Dissolution pores occur at the edge of interstitial particles and distributed between particles [57]. Such pore types are mostly irregularly shaped, elongated, or honeycombed due to the remaining undissolved material [52,53]. SEM showed that dissolution not only formed intergranularly dissolved pores, enlarged pores, and throat but also connected the original pores which were not connected to each other. Pore diameter is mostly at the micropore level, effectively improving the reservoir and its seepage capabilities (Figure 9(c)).
(3) The intragranular pores consist of internal primary pores formed from particle deposition, and secondary pores resulting from diagenesis and later transformation [58]. Elliptical and irregular intragranular pores can be observed in the SEM images of the reservoir area. The sizes of pore diameter are mostly at the nanoscale, and most of them are associated with intergranular pores. At the same time, the intragranular pores were also distributed in the easily dissolved minerals, such as dolomites, in the study area. Most intragranular pores were isolated pores, which cannot be well connected with the surrounding pore-throat. Thus, intragranular pores exerted minimal impact on pore and permeability of the reservoir (Figure 9(d)).  8 Lithosphere (4) Intercrystal pores correspond to the pores between carbonate mineral crystals, which are formed by recrystallization and dolomitization [59,60]. In the study area, the intercrystal pores between carbonate rock and clay slime crystal were developed. Results showed the micropore diameters, which were mostly at the nanoscale and were mainly observed in argillaceous dolomite. Intercrystal pores are the basic pores in dolomite and feature micropore diameter. However, these pores are abundant in dolomite and can dissolve and spread into dissolved intercrystal pores, which are also critical in oil and gas reservoirs and percolation (Figures 9(e) and 9(f)).  9 Lithosphere (5) Organic matter pores form during the transition from organic matter to hydrocarbons; such a process will generate residual pores [61]. Organic pore diameter reaches the nanoscale level. In the study area, strip-shaped organic matter was developed in the reservoir, and organic matter pores were partially developed in a small amount of organic matter. The study shows that increased maturity of organic matter will generate more reserved pores after hydrocarbon for-mation. From the laboratory test, the studied Xiagou Formation was mainly at 0.7% to 1.3% maturity, which has no specific benefit to the development of organic matter pores. Most organic matter pores were isolated pores, and their pore diameters were notably smaller than those of intergranular and intragranular pores. Organic matter pores also feature poor pore connectivity, contributing little to the permeability of the reservoir (Figures 9(g) and 9(h)).

Influence of Reservoir Pore-Throat Size on Physical
Properties. Considering 0.1 and 1 μm as boundaries, the studied pores were separated into macropores, mesopores, and micropores. The pores with sizes larger than 1 μm represented macropores, the pores with sizes ranging from 0.1 μm to 1 μm represented mesopores, and the pores with sizes smaller than 0.1 μm represented micropores. By counting the average pore diameter distribution of argillaceous dolomite and dolomitic mudstone (Figure 10), pore content of each interval of dolomitic mudstone showed two ends, whereas micropores and macropores yielded higher contents than mesopores. The pore content of argillaceous dolomite in each distribution range increased with pore diameter and was dominated by macropores. Argillaceous dolomite features a reservoir space dominated by macropores and thus possesses better properties than dolomitic mudstone, which mainly consists of micropores and macropores.
To clarify the influence of pore-throat size on average pore diameter, the correlations between macropores, mesopores, and micropores with average pore diameters were analyzed. The comparative analysis in Figure 11 shows the strong correlation between average pore diameter and macropores (correlation coefficient: 0.5725) and micropores (correlation coefficient: 0.5867). However, a weak correlation was observed between mesopores and average pore diameter, with a correlation coefficient of 0.242. When macropore content increased, average pore diameter increased. When the content of micropores increased, average pore diameter decreased. Variations in mesopore content showed no remarkable regularity on average pore diameter. The contents of micropores and macropores exhibited a strong controlling effect on average pore diameter inside the reservoir. This result is due to the micropores and macropores dominating the reservoir area; mesopore content is controlled by macropore and micropore contents. The reservoir spaces in the study area are dominated by micropores and macropores, and their contents control reservoir capacity.
As shown in Figure 12, with a correlation coefficient of 0.9369, average pore diameter featured a strong correlation with permeability, indicating that average pore diameter directly determines reservoir seepage capacity. At the same time, changes in the content of micropores and macropores significantly affect average pore diameter. Therefore, studies should analyze the correlation between permeability and micropores, mesopores, and macropores.
In Figure 13, we can observe the very strong correlation between the content of macropores and micropores with permeability, with correlation coefficients of 0.5851 and 0.5886, respectively. However, a weak correlation was noted between mesopore content and permeability, with a correlation coefficient of 0.2712. With increasing macropore content, permeability notably increased until the content reached more than 85%. The explanation for this observation is as follows. When macropore content accounted for a sufficient proportion of pore diameter, the connectivity of pore-throats will naturally improve and permeability will increase (Figure 13(a)). When the content of micropores measured less than 3%, the permeability of the reservoir notably increased because the micropores represent the throat size of the reservoir. When the content of micropores is low, the content of a small throat naturally declines, and its binding effect on pores also weakens, whereas reservoir permeability increases (Figure 13(c)). When mesopore content totaled less than 7%, reservoir permeability was divided into two types: the first type indicates that permeability increased notably, whereas the other indicates no change. This result was attributed to the fixed mesopore content, whereas the contents of micropores and macropores determine permeability (Figure 13(b)). Therefore, the configuration relationship between macropores and micropores in the reservoir decides the physical properties in the study area. When macropore content reaches higher than 85%, or when micropore content measures lower than 3%, the permeability of the reservoir improves.

Impact of Pore-Throat Distribution Characteristics on
Reservoir Physical Properties. The pore-throat sorting coefficient of the samples from the Xiagou Formation in the study area ranged from 1.39 to 3.12, indicating a medium-to-poor sorting. The correlation between movable fluid saturation and sorting coefficient was strong, with a coefficient of 0.753 (Figure 14(a)). The sorting coefficient and movable fluid saturation are exponential. With increasing sorting coefficient, mobile fluid saturation will increase. However, when sorting coefficient continually increases to some  12 Lithosphere extent, movable fluid saturation will decrease. This result is attributed to the pore sizes, which are dominated by micropores and nanopores, in the tight reservoir of the study area. At a very low pore-throat sorting coefficient, the small throats will account for most of the throats and exert no or little effect on seepage. Thus, this explains why the movable fluid saturation is very small at this time. With increasing sorting coefficient, the extent of pore-throat distribution was wider, and the percent of macroporous throat and saturation of movable fluid increased. When the sorting coefficient is extremely high, pore-throat heterogeneity will be enhanced, pore structure will be more complex, and saturation of the movable fluid will be naturally decreased [62]. Therefore, when the sorting coefficient is very low or high, the value of movable fluid saturation will show no considerable enlargement when sorting coefficient reaches between 1.8 and 2.5, that is, at a wide distribution of pore-throats and a high ratio of mesopore to macropore content of reservoir rock, saturation of movable fluid will reach the highest, and a high content of movable fluid in the reservoir will be observed. As depicted in Figure 14(b), the correlation coefficient between sorting coefficient and permeability has reached 0.5582, indicating a strong correlation. The sorting coefficient and porosity are exponential. The change regulation between permeability and sorting coefficient is similar to that between movable fluid saturation and the sorting coefficient. With an increasing sorting coefficient, the permeability of the reservoir will increase and then decrease. The sorting coefficient reflects the distribution regulation of pore-throats in the reservoir. At a low sorting coefficient, pore-throat sorting will be moderate, and the pore-throat will be small. Thus, reservoir permeability will naturally be low. When the sorting coefficient increases, pore-throat distribution will widen, and mesopore and macropore contents, permeability of the reservoir, and permeability will increase. When sorting difference is poor, micropores, mesopores, and macropores are all distributed, but mesopores and macropores are bound by tiny throats where the fluid cannot flow freely, and reservoir permeability decreases. Porosity characterizes the reservoir capac-ity of reservoir rocks, which are naturally not significantly related to the distribution characteristics of pore-throats.
Skewness is the symmetry parameter of pore-throat frequency distribution, which reflects the relative position of modes, and is called coarse skewness when the mode is in the coarse-pore end and fine skewness when the mode is in the fine-pore end; improved coarse skewness was observed for the studied reservoir [63][64][65]. The movable fluid saturation and skewness are strongly correlated, with a coefficient of 0.6212 (Figure 15(a)). This result is explained as follows: when skewness increases, the proportion of macropores in the pores becomes larger than that of the micropores. Thus, the saturation of the movable fluid also increases. Correlation was observed between skewness and permeability, with a correlation coefficient of 0.4938 (Figure 15(b)). With an increasing skewness value, macropore content increased, favoring seepage of the reservoir. No notable correlation was detected between skewness and porosity because the type of the studied reservoir space is dominated by macropores or mesopores and micropores, which cannot affect the size of the total reservoir space. Compared with the sorting coefficient, skewness exerts a weak influence on effective porosity of the movable fluid. The controlling effect of pore-throat allocation on effective porosity is reflected in the uniformity of pore-throats, but it does not mean that the bigger the pore size, the better. Thus, the configuration relationship of pore-throats plays an important role.

Influence of Reservoir Pore-Throat Connectivity
Characteristics on Reservoir Physical Properties. Movable fluid saturation and permeability are strongly correlated with mercury removal efficiency at coefficients of 0.7068 and 0.7869, respectively ( Figure 16). When mercury removal efficiency increased, saturation of the movable fluid and permeability increased. During mercury removal, the finer throats first emptied the mercury, destroying the connectivity of the entire pore space; as a result, a portion of the pores and throats became separated, and no mercury was emitted any more, reflecting the favorable regulating effect of a small throat on the pore fluid [52,63]. When rock mercury 13 Lithosphere removal efficiency was low, the development of fine porethroats in the rock was relatively high, and the saturation of the movable fluid and the permeability in the reservoir were naturally low. When mercury removal efficiency increases, the proportion of macropore-throats and mesopore-throats, movable fluid saturation, and permeability will increase. Mercury removal efficiency exhibited good correlation with movable fluid saturation and permeability, which can   14 Lithosphere indirectly uncover the size of mobile fluid saturation and permeability in the reservoir. Notably, when micropores and macropores were mainly developed in the reservoir, low mercury removal efficiency was observed from the reservoir, but saturation of the movable fluid in the reservoir was high because the micropores impeded mercury removal, and at the same time, macropores enlarged the saturation of mobile fluid. Therefore, the level of connectivity is crucial in the physical properties of the reservoir.

Characteristics of Reservoir Pore-Throat Network System.
Porosities of the Liu4-06 and Long102-06 samples reached 2.47% and 2.21%, respectively, with permeabilities of 0.143 and 0.059 mD. The two samples presented similar porosity, but the difference in permeability was nearly 3 times. To confirm the differences in pore-throat connectivity between the two samples, a 3D nano-CT imaging method was performed. Throat-pore ratio is defined as the ratio of the rock throat radius to its connected pore radius. A larger throatpore ratio indicates a larger throat radius, a smaller pore radius, a closer connection of pore and throat radius, and a stronger rock seepage ability. Figure 17(a) shows the histogram of the throat-pore ratio for the two samples, Liu4-06 and Long102-06, in which the Liu4-06 sample featured three throat ratio intervals between 0 and 0.2, 0.2 and 0.4, and 0.4 and 0.6 and a percentage content smaller than that of the Long102-06 sample. The percentage of sample No. 4-06-06 was higher than that of the Long102-06 samples in the throat-pore ratio range of 0.6 to 0.8 and 0.8 to 1.0, whereas the percentage content of the Liu4-06 sample was significantly higher than that of the Long102-06 sample. The throat radius of the Liu4-06 sample is much closer to the size of the connected pore radius than that of the Long102-06 sample; the pore-throat also showed good connectivity. This paper next compared and analyzed the distribution of the finest sectional area of the throat of the Liu4-06 and Long102-06 samples. Figure 17(b) shows that the content of the finest sectional area of the throat of the Liu4-06 sample was notably higher than that of the Long102-06 sample in two intervals over 80 μm 2 , and the content in four intervals   15 Lithosphere of less than 80 μm 2 was generally lower than that of the Long102-06 sample. This result indicates that compared with the Long102-06 sample, the proportion of the large throat content in the Liu4-06 sample is higher, thus favoring fluid flow in the rock pore and indicating good pore-throat connectivity.
As observed from the 3D pore-throat connectivity diagram of the two samples, Liu4-06 and Long102-06 (Figure 18), the same color in the local area represents the interconnected pore-throats connected with each other. In the local pore-throat interconnection area of the Liu4-06 sample, volume is visibly higher than that of the Long102-06 sample and is widely distributed in the reservoir. Thus, Liu4-06 features higher permeability than that of the Long102-06 sample.
In summary, a large throat-pore ratio with a large sectional area at the finest part of the throat and with good local pore-throat connectivity and pore-throat connection area is widely distributed in the reservoir. The pore-throats will exhibit good connectivity and high permeability which is caused by the higher volume of the local pore-throat interconnection area and wider distribution in the reservoir. This observation primarily explains the similar porosities of samples Liu4-06 and Long102-06 but the remarkably higher permeability of Liu4-06.

Classification of Reservoir Pore
Structure. Based on permeability, movable fluid saturation, sorting coefficient, and mercury removal efficiency, pore structures were classified into three types ( Table 1). The influencing factors of the physical properties of various pore structures were analyzed from a diagenetic perspective.
5.5.1. Type I. This pore structure is good and is a small-porethin-throat type. Movable fluid saturation exceeds 20%, mercury removing efficiency is 20%-30%, permeability is above 0.1 mD, and sorting coefficient is higher than 1.5 (Table 1). Maximum mercury saturation of this pore structure neared 70%, displacement pressure was small, and mercury injection curves follow a more remarkable platform, indicating the  16 Lithosphere more concentrated effective throat distribution of rocks and good configuration relationship of the pore-throat. The T 2 curve shows single-and double-peak transition states, indicating that this type of pore structure contains a certain number of macropores ( Figure 19). A few clay minerals exist in this type of pore structure. Most pore types include intergranular and dissolved pores. The pore types are mainly controlled by lithology and diagenesis, while organic matter types and abundance contribute little to the pore types. The contact type is point or linear. Given the large amount of mudstone developed in the bottom section of the Xiagou Formation reservoir, a certain amount of organic acid will be produced simultaneously in mature hydrocarbon generation of organic matter. These acidic components will preferentially enter the reservoir adjacent to the mudstone, resulting in the dissolution of unstable components; this phenomenon is conducive to reservoir rock dissolution. Therefore, the type of pore structure is strongly affected by dissolution. Under the background of the strong compaction of the Xiagou Formation reservoir, dissolution is pivotal in strengthening the connectivity and permeability of pore space. As observed in microscopic observation, feldspar and other minerals dissolved to form intragranular pores, and dissolution occurred along the cleavage and was followed by a honeycomb-like pattern. The intergranularly dissolved pores, which were formed by the dissolution of unstable components between particles, can be observed from the field of view with a pore diameter of 10-200 μm and with almost no particle filling the pores. The microfracture is well developed and features a mainly diagenetic fracture as the mineral components of sediments in the study area will undergo a series of changes given that self-burying and compaction and recrystallization will cause rock particles, which are conducive to microfracture formation, to expand and contract. The microfractures that developed along the grain boundaries can interconnect previously unconnected pores and significantly enhance the seepage capacity of the reservoir, whereas the large number of developed microfractures also act as an effective reservoir space (Figures 20(a) and 20(b)). Therefore, the type I pore structure features a large reservoir capacity and a strong seepage ability.

Type II.
This thin-pore-thin-throat type features a medium-to-poor structure. Its features include a movable fluid saturation of 10%-20%, a mercury removal efficiency of 10%-20%, a permeability basically of 0.05-0.1 mD, and a sorting coefficient of 1.0-1.5 ( Table 1). The displacement pressure of this type of pore structure is notably higher than that of type I. The slope of the mercury injection curve increased and notably shifted to the right and upward, and the range of the effective radius of the throat was reduced. Maximum mercury saturation decreased to about 60%. The T 2 spectral curve is bimodal. The macropore content of this pore structure decreased, the micropore content increased gradually, and the pore-throat configuration is more complicated than that of type I (Figure 19).
Type II mainly develops intergranular and intercrystal pores. Mineral filling of pores is enhanced, and mineral coatings develop on the surface of particles, enhancing the viscous force of the pore-throat surface to the fluid and narrowing the throat. The pores also exhibit poor connectivity. Owing to the presence of acidic medium, such as hydrocarbons, the local environment was acidic, and cementation became more developed, thus remarkably reducing pore space. Microfractures developed in small amounts with weak dissolution, intercrystal pores were mostly isolated, and the fluid inside the pores cannot participate in the flow and contributed little to reservoir seepage (Figures 20(c) and 20(d)). Therefore, the type II pore structure possesses better reservoir capacity and poor seepage.

Type III.
This very poor pore structure is a microporemicrothroat type. Movable fluid saturation is below 10%, mercury removal efficiency is 5%-10%, permeability is below 0.05 mD, and sorting coefficient is lower than 1.0 ( Table 1). The displacing pressure of this type is significantly large, indicating that the maximum pore-throat radius is significantly less than the first two types. The slope of the mercury injection curve is also very large and visibly offsets to the upper right. The pore-throat effective radius distribution range was dispersed, and maximum mercury saturation reached less than 60%. NMR T 2 spectrum distribution showed a unimodal curve, indicating the very low macropore content and very high micropore content ( Figure 19).   Figure 19: Relationship between mercury injection and nuclear magnetic resonance characteristics of three types of pore structures in the Xiagou Formation, Qingxi Sag. (a) Relationship between mercury saturation and pressure in different pore structures. (b) Relationship between pore-throat radius and porosity components in different pore structures. 19 Lithosphere This type contains undeveloped microfractures and is dominated by clay pores and micropores. Given the development of fine and silty sandstone with a high content of matrix in the Xiagou Formation reservoir in the study area, the fine grain size of sediments and buried depth in the geologic historical period resulted in compaction and densification of the reservoir, whereas the tight contact of particles worsened sorting. Clay minerals filled the pores and throats and not only divided intergranular pores into abundant binding pores, resulting in a narrower reservoir space, but also filled the throat into a bundle-like type, binding the flow of fluid in the pores and reducing seepage performance and also causing difficult dissolution (Figures 20(e) and 20(f)). Therefore, the type III pore structure features poor reservoir capacity and poor seepage ability.

Conclusions
(1) The dolomitic mudstones and argillaceous dolomites of the Xiagou Formation reservoir in Qingxi Sag are the main reservoirs of tight oil. Porosity is generally less than 10% and mainly ranges from 1% to 7%. Permeability mainly lies between 0:1 × 10 −3 and 2 × 10 −3 μm 2 . Compaction transformation action is strong, and porosity, permeability, and storage capacity are low. Strong heterogeneity of tight reservoirs and a complicated pore-throat configuration relationship are the main reasons leading to low storage capacity and permeability (2) Intergranular, dissolution, intragranular, intercrystal, and organic matter pores were detected in the Xiagou Formation reservoir in Qingxi Sag. Among these pores, the intergranular and dissolution pores provide the main reservoir space. However, the intragranular and organic matter pores are mostly independent, which hardly contributed to the pore rate or permeation of the reservoir. Throats consist mainly of tube-shaped throats and sheet throats, and slightly of necking throats, dotted throats, and bend sheet throats, which are inconducive to fluid flow in the pores (3) In the studied formation, the average pore diameter of the reservoir totals 2.39 μm. Storage space is dominated by nanopores and micropores. Argillaceous dolomite is dominated by macropores, whereas dolomitic mudstone is dominated by micropores and macropores. Rock pore-throats are distributed either unimodally, which is mainly composed by micropores, or bimodally, which is dominated by micropores and supplemented by macropores. The configuration relationship between the contents of macropores and micropores bears importance and exerts a strong control on reservoir permeability. Overall, large throat-pore ratio, large finest sectional area of the throat, good connectivity of the local pore-throat, and pore-throat connectivity in the region are distributed among a wide range of rocks. Stronger connectivity of the pore-throat indicates higher permeability (4) Considering storage capacity and permeability as standard, the pore structure of the Xiagou Formation tight oil reservoir was categorized into three types. The type I pore structure features a movable fluid saturation of more than 20% and a permeability of more than 0.1 mD. Dissolving and microfracture development significantly enhance the size and connection of pore-throats. Movable fluid saturation of type II pore structures is between 10% and 20%, and permeability is generally between 0.05 and 0.1 mD. Dissolution is weak, few microfractures are developed, and common clay mineral filling occurs, reducing the pore space. Movable fluid saturation of the type III pore structure measures less than 10%, and permeability is basically below 0.05 mD. Strong development of clay mineral filling inhibits fluid flow, notably reducing mobile fluid saturation of the tight oil reservoir. Therefore, the type I pore structure with high permeability, large ratio of macropores, and medium pore-throat allocation presents the best pore-throat configuration, a large reservoir performance, and strong permeating ability. Therefore, the type I pore structure reservoir should be regarded as a key area for surveying of the tight oil reservoir "sweet spots" in the study area

Data Availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflicts of Interest
There are no conflicts of interest with respect to the results of this paper.