Ordos Basin with rich resources is one of the main natural gas-producing areas in China (Zou et al., 2012b). Benxi Formation in the basin is mainly filled with Lagoon-tidal flat, with lagoon iron aluminum mudstone at the bottom and tidal flat sandstone, mudstone, coal seam, thin limestone lens, barrier sand bar, and shallow water delta sediments upward. Its strata are widely distributed and the hydrocarbon source rocks with stable thickness generating here, which is an important gas source stratum in the Ordos basin. Meanwhile, two sets of large-scale sand bodies in the Pangou member and Jinci member also developed here (Figure 1C). The study area is located in the northeast of the Yishaan slope in Ordos Basin, as shown in Figure 1A, B

In the present study, a total of 11 tight sandstone reservoir column samples were drilled from different exploration wells for various tests and comprehensive analysis. The location distribution of sampling wells and their basic parameters are shown in Figure 1B and Table 1 respectively.

11 column samples with a required size of 25 mm×50 mm were drilled from 11 sandstone samples, which were prepared for NMR and overburden permeability experiments. The remaining samples were broken into blocks of different sizes and prepared for CTS, SEM and, HPMI experiments.

The cast thin section (CTS) samples were prepared in the Langfang Branch of China Petrol. Explor. Dev+ Research Institute and then identified and analyzed via polarizing microscope technology in China University of Mining and Technology (CUMT), Xuzhou to investigate the petrological characteristics including reservoir deposit composition, characters of detrital particles, and pore-throat characters. The SEM observation was completed in the Advanced Analysis and Calculation Center of CUMT. In this study, all 11 samples about 0.5 cm long for each were gold-coated and dried and then the subsequent analysis was conducted with FEI quantatm 250 equipment to investigate the characteristics of mineral, pores, and reservoir morphology.

Sandstone samples with a size of about 4 mm were selected for high-pressure mercury intrusion experiments by using Micro metrics Auto pore Ⅳ9520 (0.1–60000 psi) according to the Chinese Standard SY/T5346-2005 to investigate the pore-throat structure characters. Firstly, the sample is dried in an oven at 70°C for 12 h, and then the test is carried out. A total of 108 pressure points were collected under the control of the computer during the whole experiment process. The upper-pressure limit was 200 MPa, as a result, the pore throat radium of 3 nm–1000 μm can be measured.

11 plunger samples were selected to conduct T ₂ spectrum signal measurement under saturated centrifugal state by MacroMR12-150H-1 tester produced by Newmark testing company, Suzhou, China. The specific operation flow and parameter setting during the experiment are shown in Hu et al. (2020) (Lai et al., 2018).

Then, the AP-608 overlying pressure porosimeter produced by American core Temco Products Co., Ltd. was used to conduct pressure-sensitive experiments on column samples (Figure 2), aiming to measure the permeability of samples. The experimental method is called the unsteady-state pressure drop method, of which the experimental gas is high-purity nitrogen. The displacement pressure was provided by plunger pumps and the confining pressure was provided by separate confining pressure pumps. Under the condition of keeping the displacement pressure unchanged, the change of effective stress was simulated by changing the confining pressure. In this way, the effective stress was increased to 3, 10, 15, 20, 25, 30, 40, and 50 MPa. At the same time, to avoid the influence of pressurization time and the change of total volume, it took 30 min to measure each pressure point, and the time interval of 30 s between the two pressure points was maintained.

Through CTS and SEM analysis, it is found that the reservoir rocks are mainly quartz arenite and sublitharenite. quartz is mainly single crystal quartz (75.8%–89%, with an average of 81.79%). Almost free of feldspar. The content of rock fragments is low (3%–16.1%, with an average of 9.14%), and it mainly consists of rigid rock fragments such as magmatic rock and quartzite (Figure 3A) (with an average of 5.65%) and a small amount of ductile rock fragments such as slate, phyllite, and mica (Figure 3B,C) (with an average of 3.48%). The content of interstitial material ranges from 6.5% to 11.2%, with an average of 9.07%. It mainly consists of clay minerals (4.65%), quartz cement following (2.91%), and a small amount of carbonate and iron (average 1.13% and 0.47% respectively) (Figures 3D–F). The particles with good sorting are in subround-subangular shape and their size distribution of main particles range are different. The cement types are mainly poreed and enlarged pore type.

According to the analysis of CTS and SEM, various pores including residual primary intergranular pores, intergranular dissolved pores, intragranular dissolved pores, intergranular pores, and microfractures exist in the reservoir of the study area (Figure 3). Previous studies demonstrated that mercury injection parameter R₅₀ of sandstone pore structure shows a significant correlation with reservoir porosity and permeability (Zhong et al., 2020). In the present study, based on the mercury injection experimental parameter R₅₀ and the morphological characteristics of the capillary pressure curve of each sample, the pore structures are divided into three categories, type I, type II, and type Ⅲ. The type I pore-throat structure reservoirs are composed of quartz arenite in coarse grain and giant grain size. The pore types appear in the form of multiple pore combinations, including residual intergranular pores, intergranular dissolved pores, intragranular dissolved pores, and intergranular pores, however, it is dominated by the intergranular pores and the throat is mainly necked and curved lamellar throat. The type II pore-throat structure reservoirs mainly consist of quartz arenite in medium-grained, with the pore types similar to the type I reservoir, however, the pores are mainly dissolution pores, with the curved lamellar throats. The type Ⅲ pore-throat structure reservoirs are composed of quartz arenite and sublitharenite, fine-silt grained. The pores are mainly intragranular dissolved pores and intergranular pores, tubular throat. The pore structure characteristics of different types of reservoirs are shown in Table 2.

In terms of characterizing reservoir pore size and its distribution, The NMR technology shows higher accuracy than high-pressure mercury injection technology (Loucks et al., 2012; Li et al., 2019; Hu et al., 2020b). A large amount of research work has been conducted on pore classification based on pore size (Zou et al., 2012a; Zou et al., 2012b; Hu et al., 2020a). This research adopts the previous schemes of pore size division to divide pores into micropores, micro mesopores, mesopores, and macropores (Lai et al., 2016), and the corresponding T ₂ values are ≤ 1 m, 1 ≤ T ₂ ≤ 10 m, 10 ≤ T ₂ ≤ 100 m and 100 ≤ T ₂ ≤ 1000 m respectively (Figure 4A). By comparing the pore distribution characteristics of type I, II, and Ⅲ reservoirs at different stages, it is founded that with the deterioration of reservoir physical properties, the proportion of reservoir pore volume at different stages changes significantly. The proportion of macro pores decreases sharply from the average of 32.84% of type I reservoirs to the average of 1.63% of type Ⅲ reservoirs; The proportions of micro mesopores and micropores in type I and type II reservoirs are low, but they increase rapidly to an average of 25.16% and 36.77% respectively in type Ⅲ reservoirs. The proportion of mesopores increases first and then decreases, and shows the highest proportion in type II reservoirs, accounting for 78.98%.

The DP-P test result of 11 sandstone samples is shown in Figure 5A. It appears that as the effective stress increases from 0 MPa to 50MPa, the permeability shows a good exponential function decreasing trend. To investigate the influence of dynamic effective stress on permeability quantitatively, the dimensionless parameter D _K (permeability loss rat) was introduced to evaluate the permeability sensitivity caused by stress. The D _K can be calculated through the formula as follows: D k = k 0 − k i k 0 × 100 % (10) where k ₀ is the permeability of sandstone samples under initial confining pressure, mD; K _i is the permeability of sandstone samples after applying confining pressure for i times, mD.

The variation of permeability damage rate of different samples with different effective stress is shown in Figure 5B. When effective stress reaches 50MPa, the permeability loss rate of samples ranges from 81.51% to 99.74%, with an average of 93.01%, which indicates that effective stress can compress reservoir pores, resulting in great changes in reservoir pore structure, and the permeability response of samples with different pore structures differs greatly to the same effective stress.

The pore compression coefficients of 11 sandstone samples under different effective stress are calculated, as shown in Figure 6A. When effective stress is 3Mpa, the pore compression coefficients of sandstone samples are between 0.054 and 0.125, with an average of 0.084. When effective stress increases to 50MPa, the pore compression coefficients are between 0.013 and 0.040, with an average of 0.021. The pore compression coefficients decrease with the increase of the confining pressure. The comparison analysis of the pore compressibility characteristics among the three types of reservoirs (Figure 6B) indicates that under the same effective stress, the type Ⅲ reservoir shows the largest average pore compressibility coefficient, and the type I reservoir has the smallest average pore compressibility coefficient, which indicates that the response sensitivity of pore compressibility to the change of effective stress increases from type I to type Ⅲ reservoirs.

The effect of effective stress on pore compressibility among the tested samples shows obvious stages. During the stage of effective stress less than10MPa, with the increase of effective stress, the pore compressibility of samples decreases sharply. However, when effective stress reaches 10MPa, the change of pore compressibility coefficient to the increase of stress is not obvious. The unit stress-strain rate of reservoir pore volume decreases gradually and the pore compressibility coefficient tends to be stable, which reflects that pores are more difficult to be compressed in the later stage of compression. It is probably because pores and fractures of sandstone samples used to show high compressibility under the low-pressure stage, during that period, the large, medium-sized pores and micro-fractures tend to close firstly with the increasing effective pressure, thereby resulting in the obvious decrease of sandstone permeability (Figure 5A). However, with the continuous increase of the effective stress, the decreasing rate of pore-fracture compressibility slows down to a very obvious extent. It is mainly because in this stage, the increase of stress can only cause the closure of some small pores, which leads to the slight change in the pore compressibility coefficient of sandstone during the later stage of stress increase (Figure 6A).

The deformation of sandstone reservoir medium resulted from the deformation and destruction of skeleton particles and interstitial materials under effective stress can lead to the change of reservoir pore-throat morphology and loss of reservoir space, which is macroscopically reflected in the compressibility of reservoirs (Shan and Zhou, 2020). The compression deformation of sandstone reservoirs is not a single complete elastic deformation or ductile deformation, but an extremely complex elastic-ductile deformation process, which is jointly controlled by the sandstone deposit composition, contact relationship of detrital particles, arrangement of the detrital particles, and the mode of cementation and type of cement.

When tight sandstone is compressed, deformation happens in the throat firstly (Ruan and Wang, 2002), and large and medium-sized pores and micro-fractures also close at first with the increase of pressure. Slate, phyllite soft rock fragments and kaolinite, illite, and other clay minerals filled in reservoir throat are deformed and damaged by compression, resulting in the occupancy of the pore space and the great reduction of the pore volume. All of those make the reservoirs more compact (Fjaer et al., 2008), and make the permeability of sandstone decrease significantly. At this stage (stage I), the sandstone reservoir is dominated by viscoplastic failure and deformation of ductile rock fragment and clay minerals. During this stage, the sandstone reservoir shows high permeability damage and pore compressibility. (Figure 13). After the compression of the first stage, reservoir rigid quartz particles are in close contact. With the further increase of effective stress, the closure of the reservoir throat will not continue, the damage increment of reservoir permeability decreases and small-scale non-ideal elastic deformation of rigid quartz particles occurs in sandstone reservoirs, which cause a small amount of loss of intergranular pore space (stage II) (Figure 13), thereby presenting a small pore compression coefficient. The compression deformation of sandstone reservoirs without ductile material mainly happens in stage II.