Retained hydrocarbons in lacustrine shale are one important type of shale oil. Shale oil is the oil that occurs in large sets of organic-rich shale strata. It can be divided into two types in narrow and broad senses. In a narrow sense, shale oil refers to the oil that is retained in shale layers and has not been expelled and is stored in situ (in fact, there is also micro-scale percolation from organic layers to the adjacent reservoir layers). On a certain scale, it can be considered that shale layers themselves are both the source rocks and the reservoirs, that is, shale oil in integrated sources and reservoirs. It is primarily characterized by the in-situ retention of hydrocarbons. Retained hydrocarbons in shale are the focus of this paper. The so-called integration of sources and reservoirs does not mean that the source rocks are the reservoirs in an absolute sense. High-resolution images and component analysis reveal that sources and reservoirs frequently alternate in thin interbeds in shale formations, and the source and reservoir laminae also interbed and separate from each other. Only from the perspective of the scale of fracturing operation (larger in scale from a dozen meters to tens of meters), the sources and the reservoirs are integrated, also known as coexistence of sources and reservoirs. In a broad sense, shale oil generally refers to not only the shale oil in the narrow sense, but also the oil in tight sandstone and tight carbonate rock layers within a shale formation (namely interbedded shale oil and interlayer shale oil)^[1,2,3], there is a certain degree of hydrocarbon migration.

China is rich in shale oil resources and shale oil is a new proven oil and gas domain with huge resources potential after shale gas. Since 2010, many new understandings have been obtained on the geological characteristics and sweet spot enrichment laws of lacustrine shale oil: (1) Complex in mineral compositions, the shale has a high content of brittle minerals such as quartz, feldspar, calcite, and dolomite, and an average clay content of 20%-40%, overthrow the traditional understanding that the clay content of shale is generally greater than 50%^[4,5,6]. (2) The application of high-resolution scanning electron microscopy, nano-micro CT scanning, high-pressure mercury intrusion and other technologies further allow us to get deeper understandings on the micro- and nano-pore types and three-dimensional structural characteristics, and it is recognized that organic pores, inorganic pores, micro-fractures, and bedding fractures are important storage spaces in shale^[7,8]. (3) The shale is characterized by organic matter of moderately high abundance, diverse types, and moderate thermal evolution level, lower than the maturity of marine shale oil layers in North America^[9]. (4) The shale formations are oil-bearing universally but strong in heterogeneity, and have no clear oil-water boundary^{[1, 7]}. (5) It is proposed that organic matter abundance, free hydrocarbon content, porosity, bedding and fractures, formation pressure, shale brittleness, and horizontal stress difference are important factors affecting the enrichment and high yield of shale oil^[10,11,12]. (6) The shale oil accumulation theory of “favorable lithofacies + preservation conditions” has been put forward, and two types of shale oil source-reservoir assemblages, thin interbedded type and source-reservoir in one type, have been identified^[13], and the standard for lithofacies classification by “organic matter abundance-rock sedimentary structure-mineral composition” has been established. Based on a comparative analysis of geological characteristics and single well production, shale oil enrichment and high-yield models such as “high brittleness + high porosity, high abundance of organic matter + horizontal fractures, medium organic matter abundance + laminar structure + felsic shale facies” have been put forward^{[14,15,16,17]}. The above understandings have played an important role in guiding the exploration of lacustrine shale oil in China^{[3-8, 18-21]}. In the second member of Kongdian Formation and Shahejie Formation of the Paleogene in the Bohai Bay Basin, the Paleogene Hetaoyuan Formation in the Biyang Sag of the Nanxiang Basin, the Cretaceous Qingshankou Formation in the Songliao Basin, the Permian Lucaogou Formation in the Junggar Basin, multiple horizontal wells have obtained industrial oil flow and reached annual cumulative production of 10 000 tons, individually. However, compared with marine shale oil in North America, the lacustrine shale oil reservoirs in China have more complex structural conditions, stronger stratigraphic heterogeneity, and lower thermal evolution degree, so they are more difficult to be developed economically. There are many understanding and technical challenges in the way to economic development of shale oil, among which sweet spot enrichment law and high-yield geological conditions of shale oil is one of the basic problems waiting to be answered urgently.

The Cangdong Sag belongs to the Cenozoic fault lacustrine basin in the Bohai Bay Basin, eastern China. The second member of the Paleogene Kongdian Formation (shortened as “Kong 2” hereafter) is a typical representative of retained petroleum in lacustrine shale, with a wide distribution, large thickness, and active oil and gas shows. Multiple vertical and horizontal wells deployed in the early stage have obtained industrial oil flow. Through systematic geological analysis in the early stage, “seven properties” of the Kong 2 Member shale oil have been sorted out^{[3-4, 6, 22-30]}, but the current understandings can no longer meet the needs of economic exploration and development of shale oil. Especially, it is more and more important to deepen the understanding on shale oil enrichment law and high-yield geological conditions. For this reason, based on previous research results and analysis data of three wells G108-8, GD12, and GD14, together with TOC, pyrolysis of drilling cuttings, and fracture and actual production dynamic data of more than 30 new wells drilled there, the enrichment law, model, and high-yield geological conditions of retained hydrocarbons in lacustrine shale have been further sorted out in this study, in the hope to provide basis for the optimization of horizontal well deployment and design of fracturing stages and clusters in this area and adjacent areas.

The Cangdong Sag is one of the oil-rich sags in the Bohai Bay Basin. It is located in southwestern Huanghua depression, sandwiched between the Cangxian Uplift in the west, the Xuhei Uplift in the east, the Kongdian Uplift in the north and the Dongguang Uplift in the south. It is a continental fault lacustrine basin formed by regional stretching in the Cenozoic^[31,32,33] (Fig. 1). During the sedimentary period of the Kong 2 Member, the Cangdong Sag was an inland fresh-brackish water close lacustrine basin in an elliptical shape^[34]. With clastic materials collecting from the four major provenances, Kongdian Uplift, Cangxian Uplift, Dongguang Uplift, and Xuhei Uplift, the sedimentary facies belts appeared as the outer, middle and inner rings from the edge to the middle of the lacustrine basin^[35]. The outer ring is the zone where conventional sandstone dominated by braided river delta front facies develops; the middle ring is the zone where tight rocks like silty mudstone, silt sandstone and limestone of distal front-prodelta of braided river delta facies come up; and the inner ring is the zone where fine-grained sediments such as thick-layer dark mud shale with thin siltstone and dolomite interlayers develop. Previous studies have shown that the Kong 2 Member is a complete three-order sequence (SQEk₂₎. According to the different vertical stratigraphic stacking styles and differences in evolution laws, it can be further divided into four four-order sequences from bottom up (SQEk₂⁴-SQEk₂¹), of which SQEk₂³- SQEk₂¹ are the lacustrine transgression systems tract-highstand systems tract^[24]. During this period, a set of shale mainly composed of fine-grained sedimentary rocks less than 0.0625 mm in grain size developed in the middle of the lacustrine basin, which laid the material foundation for the formation and enrichment of large-scale shale oil.

The fine-grained section of the Kong 2 Member in the Cangdong Sag has the characteristics of high-frequency laminae, various types and high abundance of organic matter, complex lithologies, high content of brittle minerals, and low clay content. The 338.5 m high-abundance shale section of Well G108-8 has an average lamina density of 33 layers/dm, and invisible lamina density from microscopic section observation and field emission scanning electron microscope AmicScan of 1 100 layers/dm. It can be observed that micron- nano-level feldspar, quartz, dolomite, clay and other mineral laminae stack over each other longitudinally. The pyrolysis and TOC tests of 941 samples from the Kong 2 Member reveal that the shale of this member has mainly Type I and Type II₁ organic matters. Among these samples, samples with Type I kerogen account for more than 70%, and samples with mainly Type II₁ kerogen account for more than 20%. These samples have a maximum TOC value of 12.92%, and an average TOC of 4.87%, indicating that the source rock is of good to very good quality^[36]. The fine-grained section of the Kong 2 Member is complex in mineral composition, and mainly composed of terrigenous debris, carbonate, clay and other minerals (analcite, pyrite, siderite, etc.). Among the minerals, terrigenous debris takes up 34% on average, and is mostly clay-grade quartz and feldspar; carbonates account for 34%, and are composed of mainly clay-grade calcite and dolomite; clay minerals account for 16% on average; and pyrite, siderite, analcime and other minerals take up 16%. Clearly, the fine-grained sediment has no dominant minerals and is characterized by high brittle mineral contents and low clay contents. According to the mineral composition and structural characteristics, the fine-grained section of the Kong 2 Member can be classified into felsic shale, mixed shale and limy dolomitic shale^{[3, 22]}.

2.1.1. TOC and S₁, OSI

2.1.1.1. Relationship between TOC and movable hydrocarbons

The high yield of retained hydrocarbons in lacustrine shale mainly depends on the content of movable hydrocarbon S_f in the stratum:

where, the K_a value takes 100 mg/g according to the experimental data. In the TOC-S₁^*correlation diagram, the distance L of any pyrolysis test point P to the straight line S₁^*=TOC (the straight line indicates that the lower limit of movable hydrocarbon is 100 mg/g) parallel to the Y axis is S_f (Fig. 2a, 2b).

Through analysis of pyrolysis data of 12468 cuttings from 33 wells such as GY1-1-1H and GY2-1-1H in the Cangdong Sag (Fig. 2a, 2b), it is found that the data points (representing better geological conditions) on the outer envelope line in the S₁ and TOC correlation diagram mainly show the characteristics of rapid rise and stable high values ^{​​[3, 37-41]}. The TOC values ​​at the inflection points are both 2% to 4% (Fig. 2b), which is the peak value of movable hydrocarbon content in a single well, and the peak value of movable hydrocarbon in this interval increases with the increase of TOC value, that is, the inflection point of the outer envelope curve shifts to the right with the increase of TOC value (Fig. 2b). It can be seen that when the TOC near the inflection point of the outer envelope line (TOC value of 2% to 4%) is the most enriched range of movable hydrocarbons.

2.1.1.2. Relationship between TOC and OSI

Studies at home and abroad have confirmed that S₁^*/TOC (movable oil index, shortened as OSI) can reflect the movable efficiency of retained hydrocarbons in shale strata, and has a good application effect in oil-bearing property evaluation^[42,43]. Through analyzing the relationship between TOC and OSI of 33 wells such as GY1-1-1H and GY2-1-1H in Cangdong Sag (Fig. 2c, 2d), it is found that the outer envelopes of all wells have the characteristics of right skew distribution. Taking the horizontal well GY1-1-1H as an example (Fig. 2c), when TOC is less than 2.7%, the OSI value gradually increases as the TOC value increases; when the TOC is 2.7%, the OSI value reaches the maximum of 600 mg/g; when TOC > 2.7%, the OSI value decreases with the increase of TOC value. This is because the amount of retained hydrocarbons in the shale strata has reached the upper limit of storage capacity, the S₁^* no longer increases (the stable high value section in Fig. 2b). With the increase of TOC value further, the OSI value shows a decreasing trend. From the distribution characteristics of the outer envelope lines of the relationships between TOC and OSI of 10 example wells, it can be seen that the OSI peak values mainly appear when the TOC value is about 3%.

It can be seen from the relationships between TOC and S_f and OSI that a moderate TOC value is conducive to the enrichment of movable retained hydrocarbons in shale stratum and the high production of shale oil. Especially when the TOC value is 2%-4% and at the intermediate value of 3%, both the S_f value and OSI value are high, indicating this TOC range is the most favorable organic matter abundance range for the enrichment of shale oil.

2.1.2. TOC and brittleness of shale layer

Organic matter comes in interbeds with inorganic minerals, and occupies a certain proportion of the rock volume. It is one of the important factors affecting the brittleness of shale. The brittleness of over 1000 core samples were analyzed based on their XRD composition data by taking the brittleness of quartz as a benchmark, giving other minerals different coefficients to indicate their brittleness^[44], and considering the influence of organic matter on the brittleness of the rock. In this way, the brittleness index BI₁ estimation formula (2) was established. The brittleness index BI₂ formula (3) was obtained from the mechanical parameter method:

The brittleness indexes BI₁ and BI₂ of Well G108-8, GD12, and GD14 calculated by the two methods are well correlated, with correlations of over 0.80, indicating that the rock brittleness reflected by contents of minerals and mechanical parameters have good consistency (Fig. 3a). Therefore, the brittle minerals from XRD analysis can also accurately characterize the brittleness of shale, and XRD analysis of brittle minerals is easy to operate. The correlations between brittleness index BI₁ and TOC and clay content of 189 samples from Well W16 were examined through statistics (Fig. 3b, 3c), and the results show: most samples with TOC value of less than 4% or clay content of less than 20% have BI₁ of > 50% and are highly brittle shale. In contrast, most samples with TOC value of greater than 6% or clay content of greater than 40% have BI₁ of <40%, and are lowly brittle shale. It can be seen that the TOC value is negatively correlated with the brittleness index on the whole, but some samples with higher clay content (with the ratio of TOC to clay minerals in volume of less than 0.3, that is, the ratio of the volume of organic matter in shale to the volume of clay is less than 0.3) are not highly brittle shale (Fig. 3b).

The above mentioned research shows that as the TOC value increases, the S_f value and OSI value first increase and then decrease, but the brittleness gradually decreases. When the TOC value is 2%-4%, the oiliness and brittleness reach the best match, and the shale is not only rich in oil but also favorable for fracturing, and thus turning out to be the optimal sweet spot. When the TOC value increases further (more than 6%), the shale decreases in brittleness gradually and turns poorer in engineering quality.

By comparing the sizes of lacustrine basin (the area of ​​the fine-grained sedimentary area represents the distribution range of the ancient lacustrine basin) and sediment supply intensities (expressed by the spread range of terrigenous debris input into the lacustrine basin, this paper takes the influence of provenance supply intensity in the short-axis direction as an example for statistics) in 21 different depositional periods (21 small layer units), it is found that the matching between lacustrine basin size and sediment supply intensity has a controlling effect on the brittle mineral content, clay content and TOC value.

(1) Moderate size of lacustrine basin and moderate sediment coverage. For example, during the deposition of No. 2 sub-layer, the fine-grained sedimentary area (short axis) in the east-west direction was 18.5 km wide (indicated by A), and the distance of terrigenous clastics into the lacustrine in the east-west direction was 8.6 km, accounting for 46% of the lacustrine basin width. The two were in moderate match. The lacustrine shale of this layer has an average clay content of 13.9%, average TOC value of 3.68%, average felsic content of 28.8%, and average brittleness index of 44.6%, meeting the standard of TOC value of 2%-4% and clay content of less than 20% for high-quality shale oil sweet spot (Table 1). The moderate match of lacustrine basin size and sediment supply intensity resulting in the best match of shale TOC and brittle mineral content, is the internal controlling factor for the development of shale oil sweet spot (Table 1).

(2) Small lacustrine basin but high sediment supply intensity. For example, during the deposition of No. 8 sub-layer, the fine-grained sedimentary area in the east-west direction was 16.2 km wide, while the distance of terrigenous clastics into the lacustrine in the east-west direction was 10.8 km, accounting for 67% of the lacustrine basin width. The lacustrine shale of this layer has an average clay content of 10.8%, average TOC value of only 1.22%, and average felsic content of 36.9%. Although conducive to fracturing, this shale layer with low abundance of organic matter has meager hydrocarbon generated, not enough to enrich the shale itself.

(3) Large lacustrine basin and low sediment supply intensity. For example, during the deposition of No. 17 sub-layer, the fine-grained sedimentary area in the east-west direction was 19.2 km wide, and the distance of terrigenous clastics into the lacustrine in the east-west direction was 3.5 km, accounting for 18% of the lacustrine basin width. This layer of lacustrine shale has an average clay content of 27.8%, a maximum TOC of 12.4%, an average TOC value of 6.62%, and a maximum of felsic content of 12.4%. Although this shale layer has high total hydrocarbon generation quantity (expelled hydro-carbons + retained hydrocarbons), it has high retained adsorbed hydrocarbons and relatively small amount of free movable hydrocarbons, and is not conducive to fracturing (with the brittleness index of only 36.2% on average) (Table 1).

In short, only when the size of lacustrine basin and the input distance of sediment (indicated by the B/A value) are in moderate match, can a reasonable configuration of clay content, TOC value and brittle minerals be formed (Table 1), thereby achieving the best match of hydrocarbon generation amount and fracability. The production practice of the Kong 2 Member in the Cangdong Sag shows that when the ratio of the size of lacustrine basin to the input distance of sediment is about 40%-60%, the hydrocarbon generation amount and fracability are well matched, but different lacustrine basins may have some differences due to differences in other geological conditions.

Through thermal simulation and kerogen swelling experiments on samples from Well G77 (2106.8 m, TOC value of 5.24%), the proportions of kerogen adsorbed hydrocarbons, retained movable hydrocarbons, and expelled hydrocarbons in different thermal evolution stages were quantitatively characterized (Table 2 and Fig. 4). The thermal simulation experiment was done with a semi-open direct pressure hydrocarbon generation and expulsion simulator close to geological conditions. During the experiment, the temperature was increased from 300 °C to 630 °C at an interval of 25-30 °C. After the sample was heated to a temperature, the R_o value of kerogen in the residual rock sample was measured to obtain hydrocarbon production rate data under different maturity conditions. The residual sample after the completion of thermal simulation experiment was prepared into kerogen to carry out swelling experiment. Five kinds of solvent, n-tetradecane, o-xylene, acetic acid, isopropanol, and ethanol covering the solubility range of kerogen and common oil and gas components were used in the experiment, a bell-shaped swelling curve was obtained and was converted into the amount of retained hydrocarbons. Combined with hydrocarbon yield data from thermal simulation experiment, the ratios of kerogen adsorbed hydrocarbons, retained movable hydrocarbons, and expelled hydrocarbons in different thermal evolution stages were calculated (Table 2, Fig. 4). When the R_o value is 0.6%-1.2%, the hydrocarbon generated by kerogen not only satisfies its own adsorption and dissolution, but also has 7.4%-60.0% of movable hydrocarbon. When the R_o value is 0.7%-1.0%, the retained movable hydrocarbon content is the highest, exceeding 40% of the total hydrocarbon generated.

Statistics on the relationship between daily oil production and burial depth (vertical depth) of shale oil layers in 15 vertical wells and 16 horizontal wells in Cangdong Sag show that (Fig. 5): in the depth range of 2500-3900 m, regardless of vertical well or horizontal well, there is a positive correlation between daily oil production and burial depth. That is, the greater the burial depth, the better the overall oil test effect is. The layers with daily oil production of more than 10 t are all more than 3300 m deep. Moreover, if a layer is more than 3700 m deep, it produces a certain amount of gas, which can increase the gas-oil ratio of retained hydrocarbons, reduce the viscosity and density of retained hydrocarbons, and facilitate the seepage of hydrocarbons. Beyond the burial depth of 3 900 m, the daily oil production decreases as the burial depth increases, and the amount of expelled hydrocarbons increases and the amount of retained hydrocarbons decreases, which are consistent with the results of thermal simulation hydrocarbon generation and expulsion experiment. The horizontal wells of GD1701H and GD1702H deployed have a vertical depth of about 3750 m. After fracturing, they have obtained high-yield oil flows, with an average daily oil production of about 5-6 times that of vertical wells. To date, they have a cumulative oil production of 8404 m³ and 11336 m³ respectively.

It can be seen that the moderate thermal maturity (at R_o value of 0.7%-1.0%, corresponding to the burial depth of 3200-4300 m) is the best matching interval between hydrocarbon generation and oil adsorption by organic matter during the thermal evolution of shale. In this interval of thermal maturity, the amount of movable retained hydrocarbon is large, and the oil production of formation testing is usually high. Therefore, this interval is the sweet spot of thermal evolution for shale oil enrichment.

By examining the Kong 2 Member in multiple wells such as G108-8 and GD12 of Cangdong Sag, we found that the porosity, pore volume and specific surface area of the ​​shale strata have three typical evolution stages with burial depth (Fig. 6). (1) The first stage is mainly the early diagenesis stage B, when compaction took dominance, and with the increase of depth, the porosity, pore volume and specific surface area gradually decreased. (2) The second stage is mainly the middle diagenetic A stage, when the shale formation reached R_o value of 0.5%-0.9% and the burial depth of 2900-3850 m, and the porosity, pore volume and specific surface area of shale strata turned from the gradual decrease in the early stage to gradual increase. This is mainly because a large amount of organic matter was converted to hydrocarbons at this stage, resulting in pressure increase from hydrocarbon generation which prohibited the mechanical compaction of the strata on one hand, and the hydrocarbon generation would create a large number of organic pores on the other hand. If 35% of kerogen in source rock with a TOC value of 7% is consumed, organic pores will increase by about 5%^[45]. The second is that a large amount of organic acids and CO₂-rich acidic fluids can be formed in the hydrocarbon generation process of the organic-rich layer, accounting for 2%-12% of organic matter content totally. The acidic fluids have a strong dissolution effect on carbonates, feldspar and other minerals, so a large number of dissolution pores would be created. Especially the easily soluble minerals adjacent to kerogen, are likely to be dissolved to form lath, harbor, and irregular shape dissolved pores at the edges, thereby increasing the porosity by 4.5%- 10.0% at the maximum^[46]. In addition, the acidic fluids formed during the hydrocarbon generation process would promote the recrystallization of carbonate minerals and the conversion of smectite, I/S mixed layer into illite, thus the storage space of shale strata would be further increased, with the porosity reaching up to 10.25%. (3) In the third stage, after the burial depth exceeded 3850 m, acidic fluids decreased, clay mineral conversion was completed, carbonate recrystallization became weaker, and stratigraphic compaction once again took dominance, and the reservoir space gradually reduced with the increase of depth. It can be seen that a moderate diagenetic evolution stage is conducive to the formation of more storage space in organic-rich shale strata^[47], especially in the middle diagenetic A stage (at the burial depth of 3200- 4300 m), a large number of organic pores, dissolved pores, intercrystalline pores, etc. could be produced.

According to the genesis, there are two types of fractures in shale strata. One type is fractures formed by tectonic stress. Strong tectonic movement may give rise to large fracture zones, which penetrate the roof and floor caprocks and thus damage the caprock condition for the enrichment and accumulation of shale oil. At the same time, tectonic stress would derive a large number of secondary fissures and micro-fractures, which extend only inside the shale strata and are distributed regionally. The other type is the fracture formed by various physical or chemical actions, mainly including interlayer fracture, abnormally high pressure fracture, dissolved fracture, mineral phase transformation fracture, dehydration fracture, etc. These types of microfractures are in limited distribution, strong heterogeneity, and small scale.

It can be seen from the test production of shale oil (all data from vertical wells) and the distance to fault in the Cangdong Sag (Fig. 7) that as the distance between the test interval and the fault increases, the overall oil test production tends to decrease in general. For example, Well G1608 is only 150 m from the fault and the stratum has a slight anticline structural background (with fracture zone likely to occur), the well had an oil production of up to 47.1 t/d in formation testing, and cumulative oil production of 1540.68 t during 105 days of production test. The (micro) fractures derived from this type of fault extend only inside the shale strata in the longitudinal direction, without destructing the overlying mudstone caprock. Therefore, the smaller the distance between the well and the fault, the greater the development degree of micro-fractures derived from the fault structure, and the higher the oil production in formation testing is. However, wells 200-300 m from the fault differ widely in oil production during formation test. For example, Well GD13 is about 275 m from the fault, and had an oil production of only 4.52 t/d; while Well KN9 also about 275 m from the fault had an oil production of 13.27 t/d, which has further confirmed that the development of natural fractures is only one of the important factors that controls the enrichment of shale oil in local intervals, but other factors (such as lithology, reservoir properties, and structure) also affect the degree of shale oil enrichment.

Due to the low porosity and low permeability of shale strata, the natural fractures can not only serve as storage space, but more importantly, communicate matrix pores to facilitate hydrocarbon seepage. However, the fractures shouldn’t be too large; otherwise they may damage the caprock conditions of shale oil enrichment and cause the loss of hydrocarbons. The shale interval with natural fractures only within it is more conducive to the seepage and preservation of retained hydrocarbons.

Based on comprehensive analysis of the retained hydrocarbon enrichment conditions in the Kong2 Member shale of Cangdong Sag, a model of retained hydrocarbon enrichment in the lacustrine shale has been established (Fig. 8). In the case that moderate size of ancient lacustrine basin matches with moderate sediment supply intensity (Model M2), shale layers characterized by medium organic matter abundance (TOC value of 2%-4%), higher content of fine-grained felsic and carbonate and other brittle minerals (more than 40%), lower clay content (less than 20%), and high-frequency laminar structure would more likely be formed. The shale is dominated by frequent interbeds of felsic shale, mixed shale and limy dolomitic shale. When the shale stratum reaches the burial depth of 3200-4300 m (middle diagenesis stage A) and R_o value of 0.6%-1.2%, it would generate a large amount of hydrocarbons, resulting in evident crossover effect (with high OSI value). At the same time, the shale has larger storage space and higher brittleness, so this stage is the most favorable for the enrichment of retained hydrocarbons, and shale in this stage is the first choice for shale oil development at present.

When the ancient lacustrine basin was relatively small and the sediment supply was strong (Model M1), the shale with high brittleness and low organic matter abundance composed of mainly felsic shale, mixed shale and psammite would be formed. This kind of shale, with lower organic matter abundance and thus insufficient hydrocarbon generated, is lower in oil content generally. But the psammitic layers very next to the high-abundance shale sections often have higher oil content (with very obvious crossover effect) due to direct charging of hydrocarbons near source rock. Moreover, the psammitic layers have favorable geological conditions for high oil yield such as good physical properties and high brittleness, so they are the main targets in the development of interbedded shale oil or interlayer shale oil (Fig. 8).

When the ancient lacustrine basin was large and the sediment supply intensity was low (model M3), the shale layers with high organic matter abundance and low content of brittle minerals, consisting of laminar clayey shale intercalated with felsic shale, mixed shale, and limy-dolomitic shale would be deposited. This type of shale has a relatively large amount of hydrocarbon generated and has a medium-high crossover effect. After satisfying its own adsorption and storage, a large amount of hydrocarbons generated by the shale would migrate out of the shale to form conventional oil reservoirs. But the type of shale has a higher content of retained adsorbed hydrocarbons and a slightly poor fracturing effect. They can be taken as a reserve resource type for shale oil exploration and development. The interval immediately next to high-quality and high-abundance limy-dolomitic shale can also form a good shale oil enrichment layer due to near source rock hydrocarbon charging and good storage conditions (Fig. 8).

4.1.1. Comprehensive evaluation of sweet spots and deployment of horizontal wells

“Moderate TOC is conducive to the formation of retained hydrocarbons in shale and high brittleness of shale layers”, so

moderate TOC is taken as a principle for layer selection. “Moderate sediment coverage would make shale oil sweet spots”, so moderate sediment coverage is taken as a principle for zone selection. “Moderate thermal evolution is favorable for enrichment of retained hydrocarbons in shale”, “moderate degree of diagenetic evolution is conducive to the development of shale reservoir space”, and “moderate development of natural fractures is good for the seepage and preservation of shale oil”, so moderate thermal evolution, moderate degree of diagenetic evolution and moderate development of natural fractures are taken as principles for evaluating enrichment conditions and productivity. The above five principles and the retained hydrocarbon enrichment model in shale have effectively guided the selection of sweet spot sections in longitudinal direction and sweet spot zones on the plane, providing important scientific support for the deployment of horizontal wells. Taking the Ek₂-C1 development layers in the Kong 2 Member of Well GD14 as an example (Fig. 9a), this interval can be divided into 4 sub-layers. The sub-layer 1 is dominated by M3 model; the sub-layers 2 and 3 and the upper part of the sub-layer 4 are dominated by M2 model, and the lower part of sub-layer 4 is dominated by M3 model. Furthermore, the middle and lower part of sub-layer 1 and the middle and upper parts of sub-layer 3 were picked out as the best sweet spot sections, followed by the sub-layers 2 and 4.

On the basis of longitudinal sweet spot section division, the OSI, brittleness index, porosity from nuclear magnetic resonance logging, R_o, and formation thickness etc were used to comprehensively evaluate the Ek₂-C1 sweet spot zone on the plane (Fig. 9b). The Class I sweet spot zones covering an area of ​​136 km², mainly distributed in the Guandong area and the lower part of the Kongxi Slope, were selected. Several horizontal wells such as GD1701H were first deployed in the Guandong area.

At present, the deployment of shale oil horizontal wells in Guandong area follows the principle of overall deployment, integration of evaluation and production construction, integration of underground and ground facilities, maximization of resource utilization, and replacement of resource development. The fault block is taken as the basic unit for overall planning and design. The best among high-yield blocks are selected for development in priority. In consideration of the distribution of surface water supply pipelines and large ditches, a three-dimensional cross well pattern of horizontal wells has been built. The deployment follows the idea of replacement development by well cluster sites, layers, and blocks to achieve orderly sustained production. Taking the horizontal well deployment plan of the Ek₂-C1 development layers in Guandong area as an example (Fig. 9c), the horizontal wells usually have an angle of more than 45° between the horizontal section and the primary stress direction, and a distance between the entry point and the fault of 150-200 m, and an average well spacing of 200 m. The wells with long horizontal section and short horizontal section are arranged alternately. A total of 61 horizontal wells have been deployed, and they are expected to produce geological reserves of 3 000×10⁴ t.

4.1.2. Practical effects

4.1.2.1. Two scientific experimental wells GD1701H and GD1702H achieved high and stable production

In the favorable exploration area of ​​Guandong area, two horizontal wells GD1701H and GD1702H targeting Ek₂-C1 were drilled. Well GD1701H had a highest daily oil production of 75.9 m³ and daily gas production of 5200 m³. By March 1, 2020, it has been producing by natural flow for 200 days and by pumping for 370 days, with a cumulative oil production of 8576 m³ (Table 3). Well GD1702H had a maximum daily oil production of 61 m³ and daily gas production of 5947 m³. By March 1, 2020, it has been producing in natural flow for 633 days, with a cumulative oil production of 11,603 m³ (Table 3).

4.1.2.2. Industrial development of the first shale oil well group

The well group composed of 3 shale oil horizontal wells, GY2-1-1H, GY2-1-2H, and GY2-1-3H. All the three wells targeted the C1-2 layer of Ek₂¹SQ (9) (dominated by Model 2), and had flowback rate of less than 1% when oil began to be produced. Well GY2-1-1H had an initial daily oil production of 24.56 m³. By March 1, 2020, it had produced 2449 m³ of oil cumulatively (Table 3). Well GY2-1-2H had an initial daily oil production of 17.83 m³, and a cumulative oil production of 2832 m³ by March 1th 2020 (Table 3 and Fig. 10). Well GY2-1-3H had an initial daily oil production of 23.37 m³, and a cumulative oil production of 2108 m³ by March 1, 2020 (Table 3). At present, this well group has a stable daily oil production of 40-50 m³, and cumulative oil production of 7390 m³.

4.2.1. Comprehensive evaluation of sweet spots

The above five principles and enrichment models of lacustrine shale retained hydrocarbons were applied to the comprehensive evaluation of shale oil and gas sweet spots in the Sha 3 Member of the Qikou Sag. Taking the Sha 3 Member in Well F38X1 as an example, it was divided into 18 sub-layers in 8 development layers of C1-C8, and models M2 and M3 take dominance. Of them, sub-layers 1, 2, 3, 4, 8, 9, 10, 11, 12 are the most favorable sweet spots, with a cumulative thickness of 400 m (Fig. 11).

Based on the selection of longitudinal sweet spot sections, OSI, brittleness index, porosity from NMR logging, R_o, formation thickness and seismic attributes etc were used to comprehensively evaluate planar sweet spot zones of C1 development layer (Fig. 12), Class I sweep spot zone of 48.6 km², Class II sweep spot zone of 53.9 km², and Class III sweep spot zone of 175.2 km² were sorted out.

4.2.2. Practical effects

The shale oil and gas exploration of the Sha 3 Member in the Qikou Sag has made important discoveries. The shale oil and gas research of this area started in 2017, and currently the Sha 3¹ submember in Well F38X1 obtained high oil production after seawater-based fracturing, opening a new prospect for shale oil exploration (Fig. 12). Well F38X1 had a maximum daily oil production of 50.1 t and daily gas production of 17 240 m³, and a cumulative oil production of 762.7 m³. At the same time, a horizontal well QY10-1-1H was deployed in the Class I sweet spot zone, targeting C1 development layers. After being fractured, this well had a highest daily oil production of 102 t (Fig. 12). The results of horizontal well drilling for shale oil in the deep lacustrine facies zone of ​​the Qikou Sag have confirmed that the fine- grained section of the Paleogene lacustrine flooding period in the Qikou Sag also has huge shale oil and gas exploration potential.

The enrichment of retained movable hydrocarbons in lacustrine shale is jointly affected by the organic matter abundance (TOC value), thermal maturity (R_o value), diagenetic evolution, development of natural fractures, and the ratio of lacustrine basin size to input distance of sediment (B/A value). The case that all the factors are moderate is the best for shale oil enrichment, that is, TOC values fall between 2% and 4%, R_o values between 0.7% and 1.0% (at the burial depth of 3200-4300 m), diagenetic evolution in the middle diagenetic stage A, natural fractures are developed but not destroying the roof and floor sealing conditions, and B/A values between 40% and 60%. Too high or too low of the indexes are not conducive to shale oil enrichment.

The principle of five moderates is mainly applicable to the retained hydrocarbons in lacustrine shale, but for the tight oil in narrow sense with evident migration (belonging to shale oil in broad sense), the higher the TOC value and the higher the thermal maturity, the more favorable it is for its enrichment, it is also beneficial to conventional oil and gas exploration. Compared with tight oil in narrow sense, lacustrine shale retained hydrocarbons have larger distribution area, larger scale of resources, and are more difficult to develop. The principle of five moderates for enrichment of shale oil can guide the development of this type of shale oil.

By coupling and matching the five major factors, sweet spots were sorted out and horizontal well target layer and sweet spot zones were picked in the Kong 2 Member of the Cangdong Sag and the Sha 3 Member in the Qikou Sag to guide the deployment of wells, realizing the industrial development of retained hydrocarbons in the Cangdong Sag and an important breakthrough of shale oil in the Qikou Sag. The two scientific experimental wells in the Kong 2 Member have produced over 2×10⁴ m³ of oil cumulatively, and the first shale oil well group has a stable production of 40-50 m³/d. The results of exploration practice show that the principle of five moderates can effectively guide the exploration and development of retained hydrocarbons in lacustrine shale.

BI₁—brittleness index based on XRD mineral composition, %;

BI₂—brittleness index based on mechanical parameters, %;

GR—natural gamma, API;

K_a—adsorption by unit organic matter, mg/g;

OSI—movable oil index, mg/g;

PR_sd—normalized Poisson’s ratio of rock, %;

R_M2R6—609.6 mm (2 ft) longitudinal resolution, 1524 mm (120 in) radial detection depth array induction logging curve, Ω·m;

R_M2RX—609.6 mm (2 ft) longitudinal resolution, 3048 mm (120 in) radial detection depth of array induction logging curve, Ω·m;

S_f—retained movable hydrocarbon amount, mg/g;

SP—spontaneous potential, mV;

S₁—pyrolysis free hydrocarbon content, mg/g;

S₁^*—pyrolysis retained hydrocarbon content, mg/g;

TOC—residual total organic carbon content, %;

∆t—compensated acoustic time difference, μs/ m;

V^*_o—volume occupied by organic matter, %;

V^*_qa, V^*_fe, V^*_ca, V^*_do, V^*_an, V^*_cl—contents of quartz, feldspar, calcite, dolomite, analcime and clay in volume % from XRD analysis after organic matter correction;

YM_sd—the elastic modulus of rock after normalization, %;

ρ—compensated density, g/cm³.