Introduction

Offshore oil and gas resources constitute approximately one-third of the global supply, thus making them an essential alternative energy source (Pinder 2001). The exploration of offshore oil and gas necessitates substantial investment, with accurate and efficient evaluation of physical parameters and reservoir classification playing a crucial role in determining subsequent exploration and development decisions. Various techniques, such as cores, thin sections, scanning electron microscopy (SEM), and well-logging, are utilized to characterize rock reservoirs. These techniques enable the analysis of crucial rock properties including composition, structure, and porosity (AL-Dujaili 2023; Zhu et al. 2023). However, it is important to acknowledge that these methods have their limitations. Traditional classification approaches heavily rely on well logging and laboratory analysis of core samples, which offer limited descriptive capability at the micro-scale level. Additionally, traditional methods rely on geology-based parameters derived from well logging, which may be insufficient for comprehensive classification, particularly during the early stages of exploration or when only cuttings are available (Lai et al. 2019). As a result, these limitations render traditional reservoir classification methods somewhat inadequate in effectively describing and evaluating sandstone reservoirs.

With the advancement in micro-CT technology and numerical methods, the utilization of 3D imaging for digital rock analysis has emerged as a crucial approach for calculating rock physical properties (Jia et al. 2007; Karpyn et al. 2010; Singh et al. 2011; Landry et al. 2014; AlRatrout et al. 2017). Digital rock analysis involves the application of micro-CT scanning on rock images and employing high-performance fluid parallel computing methods. This method offers a more convenient, rapid, accurate, and cost-effective analytical approach, which holds considerable potential and applications in the field of petroleum exploration and development (Bultreys et al. 2016; Blunt et al. 2013; Wildenschild and Sheppard 2013; Qajar and Arns 2022). However, traditional single-scale digital rock methods employ a 1-micron resolution for CT scanning, which only covers a physical size of a few millimeters. Consequently, these methods fail to capture nanoscale pore sizes and sub-millimeter intergranular pore structures present in wellbore core samples. This limitation hampers their accuracy in assessing pore structures within heterogeneous reservoirs (Chhatre et al. 2018), thereby directly impacting reserve estimation and reservoir simulation. To overcome this limitation, Okabe and Blunt (2007) proposed a reconstruction technique that integrates micro-CT images, which capture larger pores, with high-resolution images obtained from 2D thin sections through statistical simulation. Additionally, Mohebi et al. (2009) introduced a method for merging low and high-resolution models, aiming to enhance the overall resolution of reconstructed images. Nonetheless, these methods are computationally demanding, and the generated outcomes may not accurately represent pore structures at the wellbore core scale (Wang et al. 2019; Walczak et al. 2015). Furthermore, there has been limited research conducted on reservoir classification using the digital rock method.

This study utilizes a multi-scale digital rock scanning technique, employing micro-CT scanning (at resolutions of 35 μm and 2 μm) and SEM scanning (at resolutions of 65 nm and 6.5 nm), on the sandstone reservoir cores of a specific offshore oil field. A novel classification methodology has been devised for the sandstone reservoir, encompassing qualitative classification of rock particle contact relationships and pore types based on analysis of micro-CT images, along with quantitative classification of pore throat characteristics at different scales from micro-CT and SEM images. The introduction of this innovative approach displays promising potential in supporting forthcoming exploration and development endeavors within the field.

Method

The reservoir classification method, based on the quantitative analysis of digital rock at multiple scales, follows the following steps for evaluation:

  1. 1.

    Micro-CT scanning at 35 μm Resolution: Core plug samples are scanned using a micro-CT scanner with a resolution of 35 μm. This allows the capture of sub-millimeter scale porosity between particles.

  2. 2.

    Micro-CT scanning at 2 μm Resolution: Micro-samples are scanned with a micro-CT device at a resolution of 2 μm to detect smaller, micron-scale pores.

  3. 3.

    Analysis of micro-CT scanned images: The obtained microscopic pore images from micro-CT scanning are carefully observed. By analyzing the contact relationships and the cementation methods, the types of pores are qualitatively assessed and classified.

  4. 4.

    SEM scanning at two different magnifications: Micro-samples are subjected to SEM scanning at two different magnifications (500 × and 5000×), equivalent to resolutions of 65 nm and 6.5 nm, respectively. This step aims to identify nanometer-scale pores.

  5. 5.

    Calculation of porosity and pore characteristics: Using the digital rock models generated at each scale previously, calculations are performed to determine porosity and various pore characteristic parameters at its corresponding scale, such as maximum pore radius, median pore radius, and the proportion of connected pores. The porosities measured at different scales are combined and upscaled to the total porosity at the core-scale.

  6. 6.

    Reservoir classification combining qualitative and quantitative analysis: The reservoir classification method is formulated by combining the qualitative evaluation of CT scan images with the quantitative analysis of multi-scale pore characteristic parameters. This method is then utilized to classify and assess the reservoir. Figure 1 depicts the flowchart outlining the reservoir classification process.

Fig. 1
figure 1

Flowchart of the reservoir classification

Full size image

Micro-CT and SEM equipment

In response to the limitations of traditional core plug single-scale digital rock methods in capturing nanoscale pores and large intergranular pores with physical sizes ranging from 1 to 3 mm, this paper employs a multi-scale scanning approach to acquire digital rock models from different scales (Xiong et al. 2023a, b; Xu et al. 2022). The multi-scale core scanning equipment utilized in this study consists of micro-CT and SEM.

Micro-CT, an imaging technique utilizing synchrotron radiation to generate high-energy X-rays, is employed in this study. The specific micro-CT device utilized is the nanoVoxel-3502E, manufactured by QingNeng SanYing, a renowned CT manufacturer in China. This equipment is equipped with high-resolution detectors and high-precision X-ray tubes, enabling fast and high-spatial-resolution imaging through an advanced data acquisition system. It is capable of generating high-resolution three-dimensional images of samples, making it suitable for analyzing their microstructures. This device provides a field of view with dimensions of 1920 × 1920 × 1920 pixels and employs a micro-focus X-ray source with a maximum output of 300 kV. The micro-CT system consists of a radiation source, sample stage, and signal receiver. Through emission, attenuation, and reception of X-rays, it captures sample projection images and reconstructs them to create three-dimensional structures with resolutions as fine as 0.5 μm. This cutting-edge technology facilitates the transition from microscopic rock analysis to impactful applications such as formation evaluation, digital cuttings, enhance the recovery, etc. (Xiong et al. 2023a, b; Jin et al. 2023).

TESCAN MIRA3 XMH, the state-of-the-art scanning electron microscope (SEM) of the newest generation, has been specifically engineered to achieve high-performance imaging and analysis of diverse sample types. In this particular investigation, it is utilized to acquire intricate details regarding nanoscale porosity and pore distribution. The exceptional capabilities and versatility of this instrument perfectly meet the demands for high-resolution imaging of rock samples, thereby enabling comprehensive characterization and evaluation of their internal pore structures.

The methodology applied in this study involves initially performing a CT scan of the entire core plug sample with a resolution of 35 μm to capture intergranular porosity at the sub-millimeter level. Following this, three micro-samples measuring 2-3 mm in size are drilled from the core plug sample and scanned with a CT at a resolution of 2 μm to capture smaller micron-scale pores. Additionally, 5 mm micro-samples are extracted from three distinct locations within the core plug sample and scanned using SEM at magnifications of 500 × and 5000×, respectively, to capture nanoscale pores. The SEM results complement and validate the micro-CT findings, providing more precise data for multi-scale analysis. Note both micro-CT and SEM are capable of identifying various constituents within a skeletal structure, including grains (such as quartz-feldspar grain), impurities, cementing materials (such as clay), and so forth.

Qualitative classification based on CT images

The interparticle contact associations within sedimentary particles have a direct impact on the porosity and permeability of a reservoir, thereby dictating its quality. Employing a CT scan image with a resolution of 2 μm enables the observation of intricate contact relationships among rock particles and the identification of various pore types. These contact associations and pore types undergo changes due to compaction and diagenesis, thereby manifesting distinct pore types and enabling the qualitative classification of reservoir quality (Pittman 1979; Lai et al. 2016; Salah et al. 2023).

Particle contact types

In the study, four primary types of particle contact relationships are utilized, corresponding to four distinct quality reservoirs.

Point contact: Describes sedimentary particles that only touch at small points. Point contact typically occurs in more mature sandstones, indicating lesser compaction during deposition (Torrente 2016; Cheng and Yang 2018; Horng et al. 2023). This structure retains higher porosity and favorable permeability due to increased interparticle spacing (Niu et al. 2023). Reservoirs predominantly characterized by this contact relationship are classified as Type I reservoirs.

Line contact: Involves particles making contact along lines or elongated surfaces. This contact relationship suggests a certain level of compaction in the rock while still maintaining relatively good porosity. Line contact is more frequently observed in select sandstones and coarser sediments. Reservoirs primarily exhibiting this contact relationship are classified as Type II reservoirs.

Concavo–convex contact: Refers to particle contact surfaces that are uneven. Concavo–convex contact increases the apparent contact area between particles, enhancing mechanical connectivity and friction. As a result, it influences the mobility of reservoir fluids and reduces the extractability of the reservoir. Reservoirs predominantly displaying this contact relationship are classified as Type III reservoirs.

Sutured line contact: Describes particles that are cemented together, tightly fitting their surfaces. The space between particles is filled, resulting in reduced flow channels, increased flow resistance, and subsequently decreased porosity and permeability of the reservoir. Reservoirs mainly characterized by this contact relationship are classified as Type IV reservoirs.

Pore types

Four main types of pore structures are utilized, each corresponding to a distinct quality of reservoir. The first is intact intergranular pores, which are pores that formed during the initial deposition of the rock and have remained preserved to this day. These reservoirs are generally shallow and have been impacted less by compaction, with particle contact predominately being point contact. Infilling materials are also less common, with most of the original intergranular pores being fully preserved. Reservoirs with such porosity are classified as Type One reservoirs.

Residual intergranular pores are the second type and refer to the original intergranular pores that have not been filled with cementing materials or infill. These primary porosity features typically indicate favorable reservoir characteristics, as they provide relatively large spaces for fluid storage and permeation pathways. Reservoirs dominated by this type of porosity are classified as Type Two reservoirs.

The third type is intergranular dissolution pores, which are pores that form due to the enlargement of spaces between sedimentary particles as a result of chemical dissolution. As a secondary porosity feature, intergranular dissolution pores can substantially improve the quality of reservoirs, particularly in instances where primary pores may not be well-developed. These pores are critical for enhancing reservoir quality and are classified as Type Three reservoirs.

Finally, intragranular dissolution pores refer to pores formed within particles due to chemical dissolution. Although these pores can increase porosity, they have a relatively small impact on permeability because these pores are often isolated and do not affect fluid flow as directly as intergranular pores. Reservoirs dominated by this type of porosity are classified as Type Four reservoirs.

Quantitative classification based on CT and SEM images

Multi-scale porosity calculation

X-ray CT scanning technology employs synchrotron radiation to generate high-energy X-rays. Different materials exhibit varying levels of X-ray absorption. The CT number represents the gray value of each pixel in the CT image, which is directly related to the linear attenuation coefficient of the corresponding minerals (Knoll 2000; Attix 2004). Various image processing methods are then applied, including three-dimensional reconstruction, image denoising, and intelligent image segmentation. In this study, we utilize the WEKA software for two-phase image segmentation (Blunt et al. 2013; Sun et al. 2017).

Due to limitations in resolution and physical size, conventional single-scale DR analysis struggles to capture pore information in core plugs. Therefore, an upscaling method has been devised to address this issue, integrating porosity data from four scales with resolutions of 6.5 nm and 65 nm using scanning electron microscopy (SEM), as well as 2 μm and 35 μm using micro-computed tomography (micro-CT), respectively. The entire core, with a 1-inch diameter and 2-inch length, is initially scanned using micro-CT at a resolution of 35 μm to acquire comprehensive information. Subsequently, representative sub-samples measuring 2–3 mm are extracted from the core for further analysis using micro-CT at a 2 μm resolution. Following this, micro-samples (0.13–0.15 mm) and nano-samples (13-15 μm) are derived from the sub-samples for SEM analysis at resolutions of 65 nm and 6.5 nm, respectively.

Porosity at each scale is determined by summing the pore pixels through segmentation. It is assumed that the rock pixels encompass unresolved porosity, which is elucidated through finer-scale analysis. The upscaling calculation for porosity at the core-scale can be expressed using Eq. (1) to Eq. (3) (Xiong et al. 2023a, b):

$${\varphi }_{{\text{nano}}}={f}_{2\mathrm{\mu m}}{\varphi }_{65{\text{nm}}}+{\left(1-{\varphi }_{65{\text{nm}}}\right)*\varphi }_{6.5{\text{nm}}}$$
(1)
$${\varphi }_{{\text{micro}}}={{f}_{35\mathrm{\mu m}}\varphi }_{2\mathrm{\mu m}}+{\left(1-{\varphi }_{2\mathrm{\mu m}}\right)*\varphi }_{{\text{nano}}}$$
(2)
$${\varphi }_{{\text{total}}}={\varphi }_{35\mathrm{\mu m}}+{\left(1-{\varphi }_{35\mathrm{\mu m}}\right)*\varphi }_{{\text{micro}}}$$
(3)

where ϕnano represents the porosity of pores smaller than 2 μm, ϕmicro represents the microporosity for pore sizes smaller than 35 μm, ϕtotal is the total digital porosity of the core sample, ϕ6.5 nm represents the digital porosity using SEM with a resolution of 6.5 nm, ϕ65nm represents the digital porosity using SEM with a resolution of 65 nm, ϕ2μm represents the digital porosity using micro-CT with a resolution of 2 μm, ϕ35μm represents the digital porosity using micro-CT with a resolution of 35 μm, f2μm is the ratio of the pore volume greater than 2 μm to the total pore volume at the resolution of 65 nm, and f35μm is the ratio of the pore volume greater than 35 μm to the total pore volume at the resolution of 2 μm.

Pore characteristic calculation

After acquiring 2D and 3D images of rocks, the pore morphology model is utilized to extract the topology of the pore network and the distribution of pore throats from CT and SEM images (Walczak et al. 2015). The use of the pore morphology model offers several advantages in extracting the topology and distribution. It accurately represents intricate pore structures, handles various types of pores, and enables quantitative analysis of pore throats. Additionally, the model is computationally efficient, allowing for the analysis of large datasets, and aligns with the non-destructive nature of CT and SEM imaging techniques. Further detailed explanations of the fundamental concepts and techniques can be found in previous works (Liu et al. 2012; Al-Kharusi and Blunt 2007; Hilpert and Miller 2001).

The pore morphology model operates through a systematic two-step process aimed at characterizing the intricate structure of pore spaces within a material:

  1. 1.

    Constructing a List of Maximal Inscribed Spheres:

    • Initialization: The model starts by considering each void point within the material as a potential nucleus for an expanding sphere.

    • Sphere Expansion: An expanding sphere is centered at each void point, growing until it touches either the material’s rock matrix or its boundary, effectively becoming an inscribed sphere within the pore space.

    • Sorting Inscribed Spheres: The resulting inscribed spheres are sorted in descending order based on their radii, allowing for efficient subsequent analysis.

    • Identifying Neighbors: For each inscribed sphere, a search is conducted to identify smaller or equally sized inscribed spheres that overlap with it, indicating neighboring spheres. Spheres completely contained within larger spheres are removed from consideration, leaving only the maximal inscribed spheres—those that cannot be further expanded without intersecting with another sphere.

  2. 2.

    Assembling Spheres into Chains:

    • Establishing Auxiliary Index Pairs: An auxiliary index pair (node, rank) is assigned to each maximal inscribed sphere, facilitating subsequent processing.

    • Traversing Maximal Inscribed Sphere List:** The sorted list of maximal inscribed spheres is traversed again to identify chains between neighboring maximal inscribed spheres.

    • Chain Assembly: This traversal process links the maximal inscribed spheres together, forming chains where end node spheres represent pores, and intermediate spheres represent throats connecting these pores.

    • Simplification of Pore and Throat Morphology: Based on the morphological attributes of the original pore space, pores are simplified as spherical entities; while, throats are represented as cylindrical structures.

  3. 3.

    Parameter Determination:

    • Segmentation of Pore Throats: The pores and throats, now represented as simplified geometric shapes, are segmented.

    • Analysis of Morphological Parameters: Various parameters such as pore size distribution, average pore throat ratio, and average pore radius are determined by analyzing the lengths of links between spheres and the radii of inscribed spheres.

In essence, this model provides a robust framework for quantifying the complex geometry of pore spaces within materials, enabling comprehensive analysis and characterization crucial for various applications in fields like materials science, geology, and porous media engineering.

Results and discussion

This research focused on a selection of 87 core samples taken from a sandstone reservoir in an offshore oil field. The reservoir’s lithology comprises various sedimentary rock types, including siltstone, fine sandstone, medium sandstone, coarse sandstone, and conglomeratic sandstone.

Sample selection

The Weixinan Sag is situated in the Beibu Gulf Basin, China, covering an approximate area of 3800 km2. Within this region lies the WZX oilfield. The sedimentary deposits found in the basin can be categorized from bottom to top as Changliu (Ec), Liushagang (El), Weizhou (Ew), Xiayang (Nx), Jiaowei (Nj), Dengloujiao (Nd), Wanglougang (Nw), and Quaternary (Q), as shown in Fig. 2.

Fig. 2
figure 2

Locations of the WZX oilfield and wells on the Southeast Slope Belt of the Weixinan Sag, Beibu Gulf Basin

Full size image

The Beibu Gulf Basin has undergone three episodes of rift formation during the early Paleocene, late Paleocene, and late Eocene periods. This study specifically focuses on the third member of the Liushagang Formation (El3) as the primary research target. The clastic reservoirs within the El3 consist mainly of lithic sandstone and feldspar lithic sandstone. These reservoirs primarily contain coarse and fragmented sediment derived from nearby sources, with multiple layers of sandbodies deposited on top of one another. These sandbodies display prominent planar heterogeneity, and common lithologies found within them include conglomerate, coarse sandstone with gravel, and fine sandstone.

Qualitative classification based on micro-CT images

Analysis of particle contact types

Figure 3 depicts various contact relationships observed through 2 μm resolution CT scanning in the study samples. The choice of contact type, including point, point-line, concavo-convex-suture line, and suture line contacts, depends on several criteria that are unique to the sample under observation described in section "Particle contact types". Point contacts (Fig. 3a) and point-line contacts (Fig. 3b) are more common in loose sediments or sedimentary rocks, particularly in reservoirs having abundant intergranular porosity. These types of contact can confer good storage space (porosity) and connectivity (permeability) because of the intergranular pore throats present in these reservoirs.

Fig. 3
figure 3

Particle contact types for samples from WZX oilfield in Beibu Gulf Basin by micro-CT (2 μm)

Full size image

Concavo–convex contacts (Fig. 3c) and sutured contacts (Fig. 3d), on the other hand, are more frequently observed in deeply buried rocks (at a depth of 4000 m below the seafloor or even deeper in China South Sea) that have experienced strong compaction. In these rocks, primary porosity has mostly disappeared, giving way to secondary dissolution pores. Both storage space and connectivity tend to be relatively poor in these cases. Such contact relationships tend to appear in finer clastic particles and are typically found in deeper oil reservoirs.

Analysis of pore types

Figure 4 illustrates representative images of the primary pore types found in the examined samples using CT scanning with a resolution of 2 μm. The first type of pore, referred to as intact intergranular pores (Fig. 4a), originates from the sedimentary process of the rock and has not experienced substantial diagenetic alterations such as dissolution or cementation. These intact intergranular pores reflect the initial porosity of the rock as it was initially deposited. In the process of burial and multiple diagenetic stages, the preservation of the initial porosity in reservoirs can be attributed to several factors. Firstly, some rock types exhibit inherent resistance to dissolution or cementation, which helps maintain their porosity. Secondly, the diagenetic processes that occur during burial, such as compaction or cementation, may not completely obliterate the pores. Instead, they may selectively affect certain parts of the rock, allowing for the preservation of porosity in other areas. Additionally, the presence of certain minerals or organic material can act as pore fillers or stabilizers, preventing complete pore occlusion. Finally, regional variations in diagenesis or localized fluid flow can result in differential diagenetic effects, leading to heterogeneity in porosity preservation. Together, these factors contribute to the preservation of the original porosity in reservoirs despite the burial and diagenesis processes.

Fig. 4
figure 4

Pores types for samples from WZX oilfield in Beibu Gulf Basin by micro-CT (2 μm)

Full size image

The second type, Residual Intergranular Pores (Fig. 4b), arises from the transformation of the rock caused by mechanical compaction and cementation subsequent to its deposition. The original pores are partially filled or modified, resulting in residual intergranular porosity.

The third and fourth types of pores, Intergranular Dissolution Pores (Fig. 4c) and Intragranular Dissolution Pores (Fig. 4d), emerge as a consequence of diagenetic evolution. This evolution predominantly involves various dissolution and fracturing processes affecting feldspar particles and cementing materials. These secondary pores form after the deposition of the rock and have a significant impact on its porosity and permeability.

By observing and analyzing the four primary types of particle contact relationships and pore structures in rock samples, it is possible to make a qualitative assessment of reservoir quality.

Quantitative classification based on micro-CT and SEM images

Through the use of CT and SEM scans conducted at various scales on core samples and subsequent threshold segmentation of the pore-rock skeleton in the scanned images, it becomes feasible to obtain porosity measurements at sub-millimeter, micron, and nanometer scales for the core samples (as demonstrated in Fig. 5). These measurements allow for the determination of the total porosity of the core samples by employing the multi-scale porosity calculation method outlined in section "Pore characteristic calculation".

Fig. 5
figure 5

Sample’s Multi-Scale Porosity for samples from WZX oilfield in Beibu Gulf Basin (Blue represents the pores, and gray represents the rock skeleton)

Full size image

Utilizing the three-dimensional model derived from CT scan images at a resolution of 2 μm, it becomes feasible to calculate key characteristics of the pore-rock skeleton, including the ratio of connected micron-scale porosity, the maximum pore radius, and the median pore radius. Figure 6 visually presents images of both connected and unconnected pores found in three distinct lithological samples.

Fig. 6
figure 6

Micron-scale connected and unconnected pores for samples from WZX oilfield in Beibu Gulf Basin (red represents connected pores, blue represents unconnected pores)

Full size image

By merging the three-dimensional representations of connected and unconnected pores with their corresponding grayscale CT scan images, it becomes evident that the connected pores predominantly consist of original intergranular pores that take on strip and flake-like spatial formations. On the other hand, the unconnected pores mainly correspond to intragranular dissolution pores and intergranular dissolution pores.

This observation indirectly serves as validation for the reliability of the method employed in qualitatively classifying reservoirs based on the relationships between particle contacts. The findings suggest that the method effectively identifies the types of pores present in the reservoirs, based on their respective origins and spatial configurations.

Reservoir classification results

Through the analysis of contact relationships and pore types observed in CT images, it is possible to make an initial qualitative assessment of reservoir quality. Subsequently, by integrating the outcomes of multi-dimensional quantitative analysis of core samples, more precise pore characteristics and reservoir properties can be derived. A clustering method is used for the 87 samples in order to divide the reservoir into various types. Big data clustering is the process of grouping together similar data points in a large dataset. Clustering allows us to identify patterns and structures within the data that may not be easily discernible through mere observation or basic statistical analysis. By utilizing the clustering algorithms, we can process and analyze vast amounts of data in a scalable and efficient manner. This can help us gain insights into complex systems, identify anomalies and outliers, and make informed decisions based on the patterns and trends that emerge from the data. Big data clustering is increasingly important in oil and gas fields for reservoir classification, as it enables us to make sense of the massive quantities of data generated by modern-day technologies (Ali et al. 2023a, b; Al-Anazi and Gates 2010; Konoshonkin et al. 2020). Figure 7 depicts the results of clustering analysis conducted on a dataset comprising 87 samples. The analysis effectively showcases the presence of four distinct types of clusters, demonstrating strong agreement with the qualitative evaluation performed. This alignment between the quantitative clustering outcomes and the qualitative assessment highlights the reliability and accuracy of the clustering methodology employed in this study.

Fig. 7
figure 7

Reservoir classification of pore characteristics based on the clustering for samples from WZX oilfield in Beibu Gulf Basin (colors represent various reservoir classifications)

Full size image

This study entails a thorough analysis of the pore structure exhibited by the 87 samples subjected to investigation. The objective of the study is to discern specific parameters that efficiently capture and depict the intricate attributes of the pores. The purpose of this endeavor is to facilitate the classification of the reservoir based on quality levels, including high quality, good quality, poor quality, and very poor quality. To this end, we have summarized and presented the pore characteristics corresponding to each quality level in the form of Tables 1, 2, 3, and 4, respectively.

Table 1 Pore characteristics for 7 samples from reservoir classification of high quality
Full size table
Table 2 Pore characteristics for 18 samples from reservoir classification of good quality
Full size table
Table 3 Pore characteristics for 20 samples from reservoir classification of poor quality
Full size table
Table 4 Pore characteristics for 42 samples from reservoir classification of very poor quality
Full size table

Based on the analysis of diagenesis and microscopic pore structure parameters in various samples, it has been observed that reservoirs characterized by weak compaction and less than 5% lithic fragments content exhibit highly developed primary intergranular pores, thus representing high-quality reservoirs. Reservoirs with moderate compaction and less than 10% lithic fragments content retain a certain amount of primary intergranular porosity, demonstrating well-developed reservoir characteristics. On the other hand, reservoirs subjected to strong dissolution experience a further reduction in primary porosity, resulting in the development of a mixture of primary and secondary dissolution pores, making them poor reservoirs. Reservoirs characterized by strong compaction and tight cementation primarily possess secondary dissolution pores with limited pore development, rendering them extremely poor reservoirs.

In order to properly classify the samples quantitatively, a series of well-defined criteria and corresponding parameter ranges have been established from the clustering, tailored to the unique context of assessing reservoirs in a specific offshore oil field. These criteria and parameter ranges are comprehensively summarized in Table 5, serving as a valuable reference for researchers and industry professionals alike in conducting comparable analyses in similar contexts. The establishment of such specific and standardized evaluation methods is essential to ensuring accurate and reliable assessments of reservoirs, ultimately contributing significant value to oil exploration and production efforts.

Table 5 Reservoir classification method for sandstone reservoirs
Full size table

Local operational team has indicated that the high-quality reservoirs display well-developed intergranular pores, characterized by larger pore radii and strong connectivity, which promotes high and consistent productivity. In contrast, good reservoirs exhibit relatively fewer preserved intergranular pores, characterized by smaller pore radii and near-network connectivity, resulting in moderate and consistent productivity levels. Poor reservoirs showcase limited residual intergranular pores, smaller and unevenly distributed pore radii, and moderate pore connectivity, resulting in low and unpredictable productivity. Extremely poor reservoirs are characterized by the absence of intergranular pores, with very small and unevenly distributed pore radii, poor pore connectivity, leading to no productivity. These extreme cases may even exhibit no or minimal productivity even after acid fracturing attempts.

By utilizing the combination of the CT image micro-geological understanding and digital quantitative analysis of core samples, a reliable classification and evaluation method can be developed for such reservoir. In previous studies, conventional techniques including core analysis, thin-section analysis, scanning electron microscopy (SEM), and well-logging have been widely employed to characterize reservoirs. However, these methods have certain limitations, such as providing limited descriptive capability at the micro-scale level and insufficient data for comprehensive reservoir classification. Previous investigations utilizing the digital rock method have often focused on single-scale analysis, which may not fully encapsulate the multifaceted complexity and heterogeneity present within reservoirs. Moreover, the application of digital rock in the context of reservoir classification remains relatively uncommon and underexplored. By addressing these existing gaps, the proposed method facilitates the quantitative description and evaluation of reservoirs. This enables the prioritization of high-quality reservoirs for development purposes, allowing for a more efficient allocation of resources and maximizing overall productivity. Meanwhile, lower-quality reservoirs can be subjected to approaches that can optimize development strategies or the application of other enhanced recovery techniques. As a result, this approach proves to be highly effective in improving exploration success rates and development efficiency.

Conclusions

This study represents an integration of detailed digital rock data alongside microscopic geological understanding, leading to improved reservoir classification accuracy and an enhanced understanding of these formations. The inclusion of various resolution technologies promotes a comprehensive approach to reservoir characterization, thereby enhancing exploration and development strategies in sandstone reservoir exploration, as well as other relevant fields.

  1. 1.

    The study has developed a novel approach for the classification of sandstone reservoirs using a multi-scale digital rock analysis method. This method integrates qualitative and quantitative evaluations of microscopic characteristics and geological properties, utilizing data from micro-CT (35 μm and 2 μm resolutions) and SEM (65 nm and 6.5 nm resolutions).

  2. 2.

    The application of the multi-scale digital rock analysis technique in this study enables a comprehensive understanding of reservoirs by identifying and interpreting microscopic geological features such as particle contact relationships, cementation, and various types of pores. The analysis of porosity, pore structure, and pore connectivity at different scales has facilitated a detailed characterization of reservoir features and variations, enhancing the understanding of formation mechanisms and evolutionary history.

Acknowledging and addressing certain limitations is crucial and it is necessary to identify areas for future research. The current classification method used for clustering into four types lacks a quantitative ranking system. It would be beneficial to develop a score system based on expert scoring to rank reservoirs. This ranking system can then be utilized to optimize production strategies, determine well placement, and plan field development. Additionally, future research should expand the application of this method to other reservoirs to cater to broader interests and enhance its applicability.