Geological conditions, through their influence on reservoir performance, structural characteristics, geochemical properties, and fluid dynamics, collectively determine the hydrocarbon potential of tight sandstone gas reservoirs. Detailed geological analysis and comprehensive evaluation are crucial in predicting and developing these gas reservoirs. The tight sandstone samples analyzed in this study were collected from the Hangjinqi block, China (Ashraf et al., 2022; Tan et al., 2024). As illustrated in Figure 1, the Hangjinqi block in the northern Ordos Basin is characterized by Paleozoic-Mesozoic transitional sedimentary systems, with significant reservoirs concentrated in the Lower Permian Shanxi Formation tight sandstones (80–120 m thick) and the Triassic Yanchang Formation lacustrine shales. The Shanxi Formation consists of fluvial-deltaic sandstones interbedded with coal seams, which are characterized by strong diagenetic alteration and low permeability (0.1–2.0 mD). The reservoir porosity ranges from 4% to 12%, mainly consisting of secondary dissolution pores, with gas-bearing intervals (6–15 m thick) located at depths of 2,800–3,500 m. Stratigraphically, the block is characterized by a gentle SW-dipping monocline (<1°), intersected by NW-SE trending faults (20–50 m displacement), forming subtle lithologic-structural traps.

Hydrocarbon accumulation is derived from the Carboniferous Benxi Formation coal measures (TOC 2.5%–5.8%, Ro 1.2%–2.0%), and is sealed by thick Shihezi Formation mudstones. The reservoirs exhibit significant heterogeneity, with logging responses showing “low resistivity (tens to hundreds Ω·m), low gamma-ray (20–45 API), and high acoustic impedance” in gas-saturated zones. Challenges stem from complex pore structures (microfractures and dissolution vugs) and reservoir anisotropy, necessitating advanced NMR logging for fluid identification. NMR T₂ spectra indicate gas-bearing intervals through delayed peak positions (>50 ms); however, fluid differentiation remains ambiguous due to overlapping pore size and fluid property effects. Production depends on hydraulic fracturing to improve connectivity in these low-porosity, ultra-tight sandstones.

Table 1 summarizes the physical properties of the tight sandstone samples. These samples included key tight sandstone core samples from the Xinzhao block. To ensure the quality of the experiment, the samples were processed into cylinders with dimensions of 25 mm in diameter and 50 mm in height, the roughness of two end surface of rock were confined into ±0.05mm, and the deviation of diameter did not exceed ±0.05 mm. Following established experimental protocols, porosity and permeability were measured using the overpressure pore permeability system, the formation temperature and pressure were calculated from the sampling depth and logging data. Furthermore, samples exhibiting higher porosity and greater permeability were selected, as these properties suggest potential productive layers of significant research value. Consequently, 13 representative core samples were ultimately chosen for the experiment based on these criteria. Images of samples are provided in Figure 2.

This study utilizes the MesoMR23-060HeI high-temperature, high-pressure nuclear magnetic resonance (NMR) online displacement system. The system incorporates a 0.3 T magnetic field strength, achieving operational parameters of 150°C maximum temperature and 70 MPa confining pressure, with minimum aperture resolution of 10 nm and liquid flow measurement accuracy of 0.01 mL/min. This combination of technical specifications enables precise control over experimental parameters during multiphase displacement processes. Figure 3 presents a schematic diagram of the dynamic NMR experimental setup, comprising a core holder, three precision pumps; and a computerized control system.

To achieve the research objective, thirteen samples with good lithology consistency and porosity variation ranging from 2.9% to 15.3% were selected from the core of key exploration Wells in the target area for experiment, the experimental design of this study comprises three main components: analysis of the effects of varying temperature and pressure conditions on the NMR response characteristics of tight sandstone, investigation of the NMR response characteristics of methane gas in tight sandstone under different saturation pressures, and examination of the NMR response characteristics of gas and water in tight sandstone under varying water flooding pressures.

The specific process of the experiment is as follows:

(1) The samples were placed in a constant-temperature oven at 80°C for 12 h to remove residual moisture. The dried samples were then positioned in the sample holder of the high-temperature, high-pressure, multi-dimensional NMR analyzer for testing. The temperature was gradually increased to the formation temperature of the core, and nitrogen was pressurized to the formation pressure. Subsequently, the T₂ spectra and T₁–T₂ two-dimensional spectra were recorded to obtain the baseline signal.

The experimental parameters for NMR test are listed in Table 2. The T₂ spectra of the sample was obtained by mathematical inversion calculation using the CPMG sequence, while the T₁-T₂ maps were obtained using the SR-CPMG sequence.

Figure 4 shows the data processing steps after conducting the NMR experiment. Details will be discussed in later section.

Due to the distinct relaxation mechanisms between gas and water, methane typically has a longer relaxation time than water, influenced by surface relaxation. Under varying methane gas saturation pressures, compressibility can lead to changes in the gas content within pores. At low pressures, gas may occupy larger pores, while at high pressures, it may enter smaller ones. Moreover, pressure fluctuations can significantly alter the morphology and opening of small pores, impacting gas migration and distribution. As shown in Figure 5, the T₂ spectra and two-dimensional NMR results of tight sandstone samples clearly exhibit pressure-dependent characteristics, indicating significant changes in gas occurrence and pore structure. Experimental findings reveal that with increased methane gas saturation pressure, the distribution of the T₂ spectra and the characteristics of two-dimensional NMR maps undergo significant evolution.

At a low saturation pressure (10 MPa), the main peak of the T₂ spectra lies between 0.2 and 1 ms, with several secondary peaks in the long relaxation region. In comparison with higher saturation pressures, free gas occupies larger pore spaces, while gas in smaller pores is relatively less. The overall NMR signal is much lower than at higher saturation pressures. As the methane saturation pressure increases to 20 MPa, methane gas gradually adsorbs onto the mineral surface, increasing the proportion of adsorbed gas. According to the Langmuir adsorption model, the amount of adsorbed gas increases with pressure. At this stage, the gas content of the sample rises significantly, and the T₂ spectra tends to a single peak, indicating that most of the methane gas has transitioned into an adsorbed or restricted diffusion state. Additionally, as the methane saturation pressure increases, the main peak shifts toward longer T₂ times. When the sample saturation pressure increases to 30 MPa, the T₂ spectra shape remains similar to that at 20 MPa, but the peak in the short relaxation section increases.

As illustrated in Figure 6, two-dimensional NMR maps offer more direct insights into fluid states. The signals are predominantly distributed in three regions: T₂ ranging from 0.09 to 0.3 ms, T₁/T₂ from 100 to 1,000; T₂ from 1 to 4 ms, T₁/T₂ from 25 to 100; and T₂ from 30 to 200 ms, T₁/T₂ from 1 to 10. During pressure increases, the signal noticeably rises, marking the methane gas presence area. As methane gas compresses, the free gas region within the T₂ spectra diminishes, whereas the adsorbed gas region enlarges. At high saturation pressure (30 MPa), the NMR maps exhibit typical expansion in the low T₂ value region in the T₁-T₂ plot, indicating a substantial rise in the adsorbed phase proportion of methane gas, with restricted diffusion of gas within the pores.

The T₂ spectra of water-saturated tight sandstone demonstrates significant characteristic variations under different temperature and pressure conditions. These changes stem from temperature/pressure-induced modifications in fluid viscosity and pore structure, which subsequently influence the T₂ spectra’s distribution patterns. For tight sandstones with inherent low porosity-permeability and compromised signal-to-noise ratios, understanding these pressure-temperature interactions becomes particularly critical. In order to better reflect the influence of the parameter variation band, samples 1 and 2 with the largest porosity were selected for one-dimensional NMR experiments under different temperature and pressure conditions.

Figure 7 presents the T₂ spectra of sample 2 under various pressures. Due to the incompressibility of water, the T₂ spectra shows less significant changes compared to methane gas in Figure 5, as the pressure increased, the peak height rose only slightly, but the width narrowed slightly. The primary alteration occurs in the short relaxation section, which shifts rightward as pressure increases. This shift is likely caused by higher pressure expanding the pore spaces of smaller components, allowing for longer fluid relaxation times in larger pores. Pressure changes may induce deformations in pore shapes, influencing the flow paths and relaxation behaviors of the fluid. Additionally, due to the non-uniqueness of the Laplace transform, directly inverting experimental data to obtain the T₂ distribution often leads to multiple solutions and noise amplification, which may also have an impact on the experimental results.

In relative terms, temperature increases lead to more pronounced changes in the T₂ spectra. As depicted in Figure 8, with rising temperature, the main peak of the T₂ spectra becomes noticeably smaller and shifts to the right. This is mainly because higher temperatures reduce the viscosity of the fluid within pores, making fluid molecules more active and thus affecting the fluid’s relaxation properties. Additionally, high temperatures may cause microscopic changes in the rock’s pore structure, leading to pore expansion and consequently a rightward shift in the T₂ spectra. By comparing T₂ spectra under different temperature and pressure conditions, the changes in the pore structure of water-saturated tight sandstone and their effects on fluid dynamics can be revealed. This understanding is crucial for comprehending the reservoir properties of rocks and predicting their performance under various geological conditions.

Figure 9 demonstrates the evolution of the T₂ spectra during the water flooding process. As discussed in Section 3.1, at low pressures, methane in the sample predominantly exists as free gas, with the T₂ spectra exhibiting a characteristic peak at short relaxation times, corresponding to the free gas phase in tight sandstone micropores. With increased regional pressure and water injection, the main peak of the T₂ spectra expands and shifts toward the long relaxation time range, indicating a reduction in the proportion of free gas, while adsorbed and dissolved gases gradually dominate. This change reflects the substantial impact of water invasion on the gas phase within the pores. As the displacement pressure rises, the peak area increases significantly, with the main peak expanding markedly and shifting rightward, indicating that, at this stage, micropores are filled with water under displacement pressure, and the pore space is expanding.

It is apparent that, owing to the issue of signal overlap, the analysis of changes in residual methane gas during the displacement process through the T₂ spectra is challenging, which necessitates further analysis using the T₁-T₂ spectra. As shown in Figure10, the relaxation times of methane gas in the T₁-T₂ spectra demonstrate significant variations with changes in water displacement pressure. At low displacement pressures, methane gas exhibits shorter T₁ and T₂ relaxation times. As the water displacement pressure increases, the intensity within the short T₁ and T₂ regions of the spectra gradually increases, indicating that more methane gas is converted into adsorbed or dissolved states. At elevated water displacement pressures, the filling effect of water on pore spaces becomes more pronounced, predominantly filling capillary pores. Three distinct regions can be identified through comparison: T₂ between 0.1 and 0.5 ms, T₁/T₂ between 1 and 100 as the clay signal zone; T₂ between 0.5 and 10 ms, T₁/T₂ between 1 and 20 as the capillary water zone; and T₂ between 0.1 and 0.5 ms, T₁/T₂ between 100 and 6,000 as the capillary residual gas/adsorbed gas zone.

Based on the aforementioned results and the T₁-T₂ spectra of the dried and centrifuged samples, we performed a detailed analysis of the component signal positions. Table 3 illustrates the T₁-T₂ spectra of six samples under saturated, centrifuged, and dried conditions. A quantitative analysis of the component positions was conducted through comparative studies.

For example, in Table 3 No.1a, after saturation, the NMR T₂-T₁ spectra of the sample reveals two distinct component signal responses. The primary peak is located at T₂ = 2.9936 ms and T₁/T₂ = 2.9935, contributing to 98.5% of the total signal. A secondary peak appears at T₂ = 0.51 ms and T₁/T₂ = 0.024, contributing to 1.5%. Following centrifugation, as depicted in Table 3 No.1b, the signal intensity at T₂ = 2.9936 ms and T₁/T₂ = 2.9935 decreases significantly. The T₂ spectra transitions from a triplet to a doublet, with the signal peak at T₂ = 2.99 ms diminishing, and the proportion of the main peak area decreasing to 93.6%. The most significant decrease occurs in the T₁/T₂ range from 240.39 to 2,682.65, where the signal disappears. In Table 3 No.1c, after drying, the signal amplitude in the T₂ > 0.2154 ms region is substantially reduced. With the exception of a small peak at T₂ = 0.8031 ms with a signal intensity of 364, the signal amplitude approaches zero. Table 4 is an overall summary for the variations in T₂ and T₁/T₂ ratios for all samples after water saturation, centrifugation and drying.

Based on experimental data comparative analysis and literature research, the workflow begins with conducting T₂-T₂ 2D NMR experiments on multiple samples using an NMR analyzer to acquire raw data. Following this, the raw data files are batch-imported into MATLAB, where preprocessing steps such as baseline correction (denoising and drift removal) and signal normalization (amplitude standardization) are applied. The processed data are then visualized as 2D plots and contour maps, after which projection data, such as spectral peaks, are extracted and exported individually for each sample. Subsequently, Python is utilized to synthesize and fuse the exported projection datasets basing on the average brightness of each signal zone, integrating multi-sample information into a unified format. Finally, the synthesized data are imported into OriginPro for detailed graphical refinement, including axis labeling and color mapping adjustments, see Figure 4. As a result, the two-dimensional nuclear magnetic gas-water identification map of tight sandstone as shown in Figure 11 can be obtained.

By comparing NMR spectra under different experimental conditions, a gas-water identification chart is created. The 2D NMR chart offers both an intuitive visualization of the gas-water interface and enables effective differentiation of NMR responses across distinct substance types, thereby supporting petrophysical characterization in geological analyses. Compared to conventional sandstone or carbonate rocks with higher porosity and permeability, due to the more complex pore space and tighter constraints on pore fluids, the gas-water boundary in two-dimensional plots is relatively indistinct, and signal separation characteristics are less pronounced. Additionally, the complexity of the pore structure means that nuclear magnetic resonance responses may exhibit lower signal intensity and shorter T1 and T2 relaxation times. Greater attention is required for the accuracy of the experiment, the inversion method, and the experimental details. To accurately identify the plot, meticulous handling of these signals is necessary to distinctly and accurately differentiate the characteristics of tight sandstone gas and water.