Tight and Shale Oil Exploration: A Review of the Global Experience and a Case of West Siberia

Abstract

Shale and tight oil reservoirs, with horizontal wells and hydraulic fractures, typically have a recovery ratio of around 10%. The exploration of tight oil and shale in North America has proven economically viable, thanks to advancements, such as horizontal wells, hydraulic fracturing, and other enhanced oil recovery techniques. Taking inspiration from the global experience (the North American shale experience), the exploration and development of the West Siberian tight and shale reserves was more focused on the reported best practices of the exploration of North American shale. In this study, the advance in the specific areas of shale and tight oil exploration was considered, with more emphasis placed on the progress in the exploration of West Siberian shales. According to the review literature, thermal enhanced recovery methods capable of converting organic matter into hydrocarbons were studied more than other methods of enhanced oil recovery. Aligned with global trends, there has been a growing focus on research aiming to integrate data-driven approaches and pore-scale simulations to enhance recovery from tight and shale formations. Multiple pilot studies have showcased promising prospects for implementing multistage hydraulic fracturing. Nevertheless, there are limited pilot studies dedicated to enhanced oil recovery methods for West Siberian shale.

1. Introduction

The energy transition is a crucial topic in global energy policy. To shift towards cleaner sources of energy, a gradual phasing out of highly carbon-intensive energy sources and their replacement with less carbon-intensive alternatives will be necessary [1,2]. Current economic growth projections for the developing world are expected to increase [1]. Consequently, a reliable, cheap, and readily available source of energy will be required. To meet the anticipated rise in fossil-based energy demand in both the developing and advanced world, there is a need to explore the existing reserves of fossil-based energy. At present, most conventional reservoirs are in their later stages of production [3]. Hence, the exploitation of fossil-based energy has focused on unconventional reserves, such as heavy oils, tight oil deposits, shales, and gas hydrates.

Tight oil and shale reserves are characterized by ultra-low permeability of less than 0.1 mD [4,5]. This almost impermeable rock matrix restricts the flow of fluid and makes it difficult for the accumulated hydrocarbons to be retrieved. The ultra-low permeability characteristic of tight oil reservoirs limits the oil and gas flow area to only around a meter from the well, and the rest of the valuable fluids are trapped in the farther non-stimulated regions, which are hard to reach via the depletion regime. Thus, a small initial stimulated reservoir volume (SRV) is the main problem for shales [6]. A shale resource system can be defined as a continuous organic-rich source rock or rocks that serve dual purposes, functioning as both a source and a reservoir for petroleum (oil and gas) production or, alternatively, as a means to charge and seal petroleum in adjacent, continuous organic-lean intervals. This system involves primary migration processes, confined to movement solely within the source interval, as well as secondary migration, which extends into non-source horizons that are juxtaposed to the source rock(s) [7]. The study of North American shale plays is well documented and thoroughly reviewed in the literature as compared to other shale plays around the world [8]. Although the deposits of tight and shale oil are enormous in West Siberia, they are still in the earliest stages of development. Most of the few ongoing projects include horizontal wells drilling and hydraulic fracturing in the fields of the Bazhenov formation (BF), which is a deep shale reservoir (2500–3000 m) characterized by high organic content (up to 40% of kerogen) and low permeability at the micro or nano Darcy scale. Kalmykov et al. [9] reported that West Siberian shales and tight oil reservoirs possess unique characteristics compared to Bakken formations. These characteristics include varying amounts of carbonates, feldspar, and pyrite. Additionally, the oil shale of the West Siberian formation can be described as sediments saturated with hydrocarbons and heteroatomic compounds, where organic matter plays a significant role in the composition of the rocks.

From the early 2000s, the exploration of tight oil and shale reservoirs increased dramatically due to the application of two advanced, improved oil recovery methods: horizontal well drilling and hydraulic fracturing (HF). While the drilling of horizontal wells increases the drainage area of the well, hydraulic fracturing is required to improve the flow of the accumulated fluids from the matrix towards the wellbore for recovery. Since then, the development of the world’s shales and tight oil reservoirs has fundamentally included increasing the simulated reservoir volume (SRV) by creating a network of drainable fractures through hydraulic fracturing or multistage hydraulic fracturing (MSHF). The principle of hydraulic fracturing involves the injection of large amounts of prepared fluids into the well, providing increased reservoir pressure that exceeds the rock strength at certain positions, literally tearing the rock apart and creating fractures [10]. The creation of large SRV via multistage hydraulic fracturing is usually enough for the stable production of shale gas. However, even massive MSHF leads to the recovery of only up to 10% of shale oil, with a rapid decrease in the flow rate during the first few months of well exploitation. The main reason for that is the low relative permeability of the rock to oil and the depletion of the fracture net, which very slowly refills with the upcoming fluid [11]. Therefore, enhanced oil recovery (EOR) methods are implemented to achieve high shale oil recovery rates following one of these main scenarios: thermal treatment, gas injection, water injection, or chemical-based fluids.

Numerous studies have investigated the different methods of improving oil and gas recovery from shale and tight reservoirs. EOR techniques for shale formations can be directed towards two objectives: firstly, extracting the oil that is already present by enhancing its mobility and/or by creating additional matrix permeability, which can be achieved through hydraulic fracturing, gas injection, or chemical EOR; secondly, transforming kerogen into mobile hydrocarbons through thermal EOR. The effectiveness of these methods varies at different stages of the thermal maturity of kerogen. For immature organic matter and intensive conversion of kerogen, thermal EOR is preferable.

It is significant to note that the real effect of any EOR treatment in the field is usually less or much less than the one achieved in the lab. This is due to the large difference in a rock volume treated by an EOR agent in a laboratory experiment and in the field and the heterogeneity of a real reservoir, e.g., variations of permeability between layers or the presence of natural fractures that might cause breakthroughs. An example presented by Burrows et al. [8] showed that the laboratory test reported oil recovery of up to 100%, while no significant recovery was observed in the field pilot study. This underscores the significance of bridging the gap between laboratory results and field-scale applications through the utilization of numerical simulations and sensitivity analysis. It emphasizes the importance of carefully selecting parameters for enhanced oil recovery and conducting field pilot studies to validate the accuracy of the chosen parameters. Nevertheless, numerical modeling is susceptible to the risk of inaccurate predictions stemming from model simplification and incorrect data inputs. Therefore, it is imperative to foster synergy by incorporating experimental studies, numerical modeling, and field pilots. This approach enables the development of optimal strategies for shale and tight oil explorations, as it allows for a comprehensive understanding of reservoir behavior and facilitates the identification of the most effective techniques and approaches (Figure 1).

This review aims to explore the recent progress in shale and tight oil development in West Siberia by considering global experiences and advancements as reference points. The focus of the review revolves around three key areas: experimental analysis, numerical modeling, and pilot studies. While there is a wide range of experimental work dedicated to various aspects of shale exploration, such as petrophysical, geochemical, and geomechanical analyses, this review specifically focuses on experimental methods for enhanced oil recovery: thermal and gas recovery techniques. The section on numerical modeling in this review incorporates all advancements related to modeling enhanced oil recovery methods and fluid flows, ranging from the pore scale to the field scale. The analysis encompasses both conventional modeling approaches, as well as intelligent modeling techniques. Finally, the review presents recent and significant field pilots conducted in the West Siberian region and globally.

2. Experimental EOR Methods

Due to the fact that the application of hydraulic fracturing and horizontal wells increases the recovery factor of shale and tight oil reserves, enhanced oil recovery methods are very crucial in increasing the recovery factor for these types of reservoirs. Hence, many studies over the years have been dedicated to optimizing EOR methods applicable to conventional reservoirs for tight oil and shale formations. These studies include thermal enhanced oil recovery methods, gas enhanced oil recovery methods, chemical enhanced oil recovery methods, and hybrid methods. Depending on the geological and petrophysical characteristics of the formation, one method might be applicable, while other methods might not be applicable. There are numerous laboratory and numerical investigations conducted globally on shale and tight reservoirs. Some of these studies are pilot studies and commercial application of this technology. Hydraulic fracturing, thermal and gas-based EOR methods are the most popular methods experimented on at the field scale for shales and tight reservoirs.

Gas injection is a well-known technology employed by oil and gas companies globally across various types of formations, resulting in a significant increase in cumulative oil recovery [12]. Its applicability to shales and tight fields is a straightforward choice for EOR. Nevertheless, initial pilot tests of gas injection in shales unveiled several challenges, including early breakthroughs, inadequate sweep efficiency, and insignificant increase in the recovery factor. These field studies highlight the necessity for a customized approach to achieve successful gas EOR performance in shales. One of the most significant choices to make when designing a gas EOR method for field application is the choice of gas to inject. CO₂ and natural (hydrocarbon) gas are the most preferable and frequently used EOR gases in EOR for shales. Both carbon dioxide (CO₂) and natural gas are utilized as miscible agents, leading to enhanced oil recovery through favorable solubility, oil swelling, and diffusion. The injection of these gases has demonstrated efficiency in shales, with certain pilot cases reporting up to a 70% increase in oil recovery [13,14]. Even if, in some cases, other gases (e.g., ethane) might appear to be a more efficient option, it is usually a consequence of the gas source availability near the field. The economic profitability of the additional oil recovery from gas should not be compromised by its production and transportation to the field. In scenarios where the preferred gases are unavailable for various reasons, less miscible agents, such as lean gas or nitrogen, can be employed as EOR agents for shales, albeit resulting in lower oil recovery [8,15,16]. Only when gas injection is not a viable option do enterprises shift their focus towards less desirable water-based fluids.

Cyclic injection (or huff-n-puff mode) is considered to be a more efficient well operation regime for shales and low-permeable reservoirs than continuous flooding due to the poor fracture net connection between wells [8,17]. Huff-n-puff mode allows the injected solvent to propagate deep into the matrix and interact with the oil for a longer period of soaking. An example of successful implementation of the huff-n-puff mode is gas EOR projects on Eagle ford shales [18]. Gas and water-based liquid injections are extensively studied EOR technologies for shales, both in laboratory settings and, to some extent, in the field. Several comprehensive reviews have been conducted to summarize the global experience with gas EOR for shales [8,17,19]. In general, water-based fluids are not very useful in shales and tight formations, unlike conventional fluids. This is because of the ultra-low permeability of the reservoir rocks, clay swelling, and other effects that have already been described, which cause low injectivity, decreased permeability, and permanent damage to the near-wellbore zone [20,21]. For some reservoirs, the calculated economics could be positive if a water-based fluid such as surfactant is used as an EOR agent. Low costs of transportation, preparation, and injection and a slight production increase confirmed by a few field tests promise an optimistic scenario.

Another aspect of successful EOR performance in the field is the surface area for the treatment and the time of contact between oil and an EOR agent within the reservoir. The rock matrix in shales is almost impermeable, and the only way to effectively increase the contact area and provide pressure propagation is to create a large SRV zone by applying hydraulic fracturing, as stated above. A higher degree of fracture branching throughout the reservoir is anticipated to enhance the effectiveness of EOR, regardless of the chosen injected fluid. However, operational parameters, such as injection pressure, re-pressurization control, injection and soaking durations, production periods for huff-n-puff treatments, and the number of cycles, play a crucial role in establishing the necessary oil–fluid contact.

Thermal treatment distinguishes itself from other EOR methods for shales, as it relies on increasing the temperature within the reservoir to convert solid organic matter into synthetic oil [22,23,24]. Thermal EOR techniques for shales can be classified into two distinct technologies: air injection, known as in situ combustion (ISC), or hot-fluid injection. Hot-fluid injection, such as hot-water injection (including overheated vapor and supercritical water) or other agent injections, involves preheating a specific fluid, typically water, to deliver heat into the formation. These methods have established efficacy in enhancing oil recovery in light, heavy, and viscous oil formations.

In this subsection, an overview of recent advances in experimental research on thermal and gas EOR methods for shales and tight oil reservoirs globally is presented. In addition, the recent experimental studies for West Siberian shale and tight reservoirs are discussed.

2.1. Advances in the Thermal Treatment of Shales and Tight Reservoirs

Thermal treatment of reservoirs is aimed at improving the mobility properties of the in situ oil [25]. It mainly involves the application of heat to the in situ oil, thereby decreasing its density and viscosity and consequently increasing its mobility. The application of thermal EOR methods to shales requires the consideration of all physical and chemical characteristics of the reservoir and the in situ fluids. Thermal treatment of shales and tight oil reservoirs can be categorized into four main groups, based on the source of heating—in situ combustion, hot-fluid injection, electrical heating, and electromagnetic heating methods. Considering the mode of injection and the source of heat, the thermal methods that have received the most extensive research attention encompass continuous steam injection, cyclic steam injection, steam-assisted gravity drainage (SAGD), vapor extractions (VAPEX), and in situ combustion [26]. The order of listing does not represent the importance or popularity of these thermal methods. This subsection presents the experimental advances in the application of ISC and hot-fluid injection EOR methods to shale and tight oil reserves. First, an overview of the global experience is presented; then, the advances in the studies related to West Siberian shales and tight reservoirs are presented.

2.1.1. In Situ Combustion and Heating EOR

• The Global Experience:

In situ combustion technologies have been reported to have been developed and actively implemented in the USA since 1916 and have since been tested for heavy oil, shales, and tight oil reservoirs [27]. Since then, numerous experiments and pilots have been reported. At the inception of the shale revolution, the application of in situ combustion to shale oil formation began to gain recognition. In situ combustion is a technology that involves the ignition of combustion in a reservoir and controlling the combustion front to generate heat and increase the mobility of trapped oil or shale oil [28]. The process involves the use of hydrocarbons as a fuel to heat up the reservoir and enhance or speed up chemo-physical processes in hydrocarbon-bearing zones. In situ combustion has become a popular in situ conversion process due to the possibility of generating the required heat needed for increasing the mobility of heavy oil, as well as converting kerogen into oil by burning coke [29,30]. While the application of in situ combustion to heavy oil and bitumen sands is well reported in literature, the complexity of the process required comprehensive research for its application to a specific shale and tight oil-bearing formation as compared to other enhanced oil recovery methods.

In situ combustion can be performed in two ways: forward or reverse, depending on the movement of the combustion front within the reservoir. Forward combustion occurs when air flows in the same direction as the combustion front, while reverse combustion happens when the airflow opposes the movement of the combustion front. The commonly preferred approach is forward combustion due to its stability. The reverse combustion can cause spontaneous ignition in the reservoir, which is unstable and difficult to control, rendering it more complex to be implemented for enhanced oil recovery [31]. Both forward and reverse combustion can be carried out in two modes: dry and wet. In the dry mode, only dry air is injected, whereas in wet combustion, a mixture of air and water are injected. In the process of in situ combustion, many reactions take place simultaneously—some at low temperatures and others at higher temperatures. In terms of implementation, there have been many reported techniques of in situ combustion. These include Top–Down ISC, Bottom–Up ISC, Toe-to-Heel Air Injection (THAI), Combustion Override Split-Production (COSP), and Combustion-Assisted Gravity Drainage (CAGD). A summary of the descriptions of the methods is presented in

Table 1. Techniques of implementing ISC.

Method	Description
Top-Down ISC	In this method, air or oxygen is injected from the top of the reservoir, and combustion progresses vertically downward [32,33].
Toe-to-Heel Air Injection (THAI)	THAI is a specific implementation of ISC, in which air is injected at the “toe” of a horizontal well, while producing oil from the “heel” of the well. The combustion front moves towards the heel, and the produced oil flows towards the toe of the well. This method is limited to a specific range of reservoir thicknesses and early injection gas (air) breakthrough [34,35].
Combustion-Assisted Gravity Drainage (CAGD)	CAGD combines ISC with gravity drainage. Air or oxygen is injected into the reservoir, and the combustion front heats the oil, reducing its viscosity. Gravity then assists in the drainage and recovery of the mobilized oil [36,37].
Combustion Override, Split production, Horizontal-well (COSH)	COSH is a technique that combines steam injection and ISC. Steam is first injected to heat the reservoir, followed by air or oxygen injection to initiate combustion. The combustion front then overrides the steam chamber, displacing and producing the oil [38].

.

Most of the in situ combustion studies presented in the literature were performed for heavy oil and bitumen samples. The implementation of in situ combustion for shales is purposefully for the conversion of shale oil into light synthetic oil for production. Oil shale has low thermal conductivity and permeability, which makes it inefficient in transmitting the injected air and the propagation of heat. In in situ methods, the porosity and permeability of the shale within the geologic formation play a crucial role in enabling heat transfer. In shales and tight reservoirs, the heat propagation is enhanced by hydraulic fracturing and horizontal drilling to improve permeability and increase the reaction area. Sun et al. [39] presented an improved method of in situ combustion know as self-pyrolysis. In the study, a topochemical reaction occurred inside the oil shale under preheated air, which provides sufficient heat for subsequent pyrolysis with any additional external heat. The method is reported to be a low-energy-cost strategy for oil shale pyrolysis. Chen et al. [40] presented a comparison of different gases for the extraction of the organic matter. The gases studies are air, N₂, Ar, and CO₂. The results showed that after heating the tube furnace up to a temperature of 500 °C, the porosity and permeability of all the studied samples increased. For the different gases, the order of performance from best to worse is air, N₂, CO₂, and Ar.

• Advances in Experimental Studies of ISC on West Siberian Shale and Tight Reservoirs:

The first in situ combustion technology in Russia is reported to have been pioneered by Dubrovay and Sheinman [41]. The study by Dubrovay and Sheinman [39] was mainly dedicated to coal gasification. Since then, in situ combustion technology has been applied to heavy oil and shale formations.

Bondarenko et al. [42] investigated the process of thermal decomposition of kerogen by combustion. In order to comprehensively characterize the thermal decomposition of kerogen, three distinct sets of experiments were carried out. Firstly, open-system pyrolysis and kinetics experiments were conducted to facilitate geochemical analysis and examine the pyrolysis kinetics. Secondly, simultaneous thermal analysis experiments were performed to investigate the thermal effects and determine the temperature intervals associated with low temperatures. Lastly, high-temperature oxidation and thermomicroscopy experiments were employed to visually explore the process of kerogen conversion during both pyrolysis and oxidation stages. From the open-system pyrolysis, it was concluded that fixing the frequency factor to $2\times {10}^{14}{\mathrm{s}}^{-1}$ during the derivation of the kinetic parameters and using a spacing of 1 $\mathrm{k}\mathrm{c}\mathrm{a}\mathrm{l}/\mathrm{m}\mathrm{o}\mathrm{l}\mathrm{e}$ in the discrete activation energy distribution yields a more stable solution compared to traditional approach results. Furthermore, fractures were observed on the samples during oxidation at 450 °C.

Thermal treatment studies by Gorshkov and Khomyakov [43] demonstrated a potential increase in the porosity and permeability of West Siberian cores samples when subjected to thermal treatment. In the study, two siliceous–argillaceous core samples with high kerogen and thermal maturity of $\mathrm{T}\mathrm{m}\mathrm{a}\mathrm{x}<429$ °C were compared. The results show a significant increase in the open porosity of the core samples. One of the core samples recorded a sharp increase in porosity when the sample was heated above 150 °C, while the other core sample only recorded an increase in the porosity when the heating temperature was above 250 °C. The average percentage increase in porosity for the two core samples is approximately 17%, while the permeability increases by an average of five times. Similar results of increase in the porosity and permeability of samples of West Siberian origin were reported by Mukhametdinova et al. [44]. Mukhametdinova et al. [44] stated that the porosity of the Bazhenov rock samples increased by 9%, while the permeability of the samples increased, on average, for 1 mD. In the study, rock samples of different West Siberian origin with TOC between 2 to 18 wt% were used. A direct correlation between the increase in porosity and permeability and the combustion temperature was reported. Ponomarev et al. [45] conducted thermal treatment of source rock with and without the effect of a magnetic field. The results show that when the source rock was heated in the presence of a magnetic field, the total content of naphthene is 87%, different from when the source rocks were heated in the absence of a magnetic field at the same temperature. Hence, the study concluded that magnetic catalysts could be added in the thermal treatment of source rocks. Similar studies were conducted by Kovaleva et al. [46] on source rock samples from the Bazhenov and Domanik formations. The study highlighted the difference in heating rates based on the electromagnetic field type. The Bazhenov Formation samples rapidly reached 300 °C within 100 s when exposed to microwave treatment, while radio frequency exposure took 12 min to reach 200 °C. In contrast, the carbonate samples from the Domanik Formation heated up gradually in the microwave field and reached lower temperatures (150 °C) in the radio frequency field compared to the Bazhenov Formation samples. Fracture formations were observed during the heating process.

2.1.2. Hot-Fluid Injection (HFI)

• The Global Experience:

The injection of fluid has been applied to many conventional reservoirs, with reported success in its implementation. Hot-fluid injections are mainly applied to heavy oil reservoirs, with the aim of increasing the fluid mobility in such reservoirs. The injection of hot fluid consists of mainly injection steam into the reservoir to heat up the formation and the in situ fluid to increase the fluid mobility. Some studies have reported the application of hot fluids for shale and tight oil reservoirs [47,48,49]. With respect to shales, the thermal EOR methods applied must be capable of providing the adequate heat required to convert the kerogen mobile hydrocarbons. Hence, not all the thermal methods that are applicable to a conventional reservoir could be applied to shale and tight formations [42,43,50,51]. For generation of hydrocarbons, a temperature above 300 °C is required. The temperature for pyrolysis lies between 300 °C and 500 °C. Initially, the kerogen is converted into bitumen and, later, converted into lighter hydrocarbons [52].

A high yield of liquid hydrocarbons has been reported from other studies where subcritical water was injected into oil shales [53,54,55,56]. Sun et al. [56] emphasized that it takes a long duration for the kerogen to be completely transformed into bitumen. The optimal extraction time was recorded to be approximately 250 h. Other researchers have used thermal treatment coupled with other techniques to enhance the oil generation from kerogen. Washburn et al. [57] presented hydropyrolysis in the presence of overburden formation by using a novel uniaxial confinement clamp. The experiment was conducted on Woodford shale samples. During the artificial maturation process, experiments conducted under confined stress produced more fractures than the unconfined pores. By mimicking the overburden pressure in the reservoirs with the confining pressure, one should expect a higher increase in porosity and permeability during hydropyrolysis in the reservoir. Meanwhile, Kar and Hascakir [48] presented a comparative analysis between in situ combustion and hot-water injection to shales. The study showed that the combustion provides the optimal energy required for the decomposition of the kerogen and is more economical. Furthermore, the heat generated is higher and leads to substantially higher improvement of the permeability and porosity through the formation of new micro-fractures. The different hot-liquid injection studies for shale and tight reservoirs are presented in

Table 2. Summary of different types of hot fluids injected for pyrolysis.

Formation	Injected Fluid	Max Temperature (°C)
Kentucky shale [47]	Supercritical toluene	460
Natih B Formation–Oman [49]	Supercritical water	400
Domanik formation [58]	Supercritical Water	374 at 24.6 MPa
Huadian oil shale [56]	Subcritical water	300
Huadian oil shale [59]	Subcritical FeCl2 solution	350
Huadian oil shale [60]	Subcritical CaCl2 solution	350
Huadian oil shale [61]	Subcritical FeCl3 solution	350
Longkou Liangjia [62]	Supercritical water	425

.

• Advances in Experimental Studies of HFI on West Siberian Shale and Tight Reservoirs:

Popov et al. [63] employed a hydrous pyrolysis technique followed by open-system pyrolysis to investigate the recovery of kerogen from Bazhenov core samples. The study reported that the temperature of the hydrocarbon generation window is in the range of 400 °C to 480 °C for samples with high kerogen maturity and 325 °C to 350 °C for samples with low kerogen maturity. Subsequent open pyrolysis revealed a recovery up to 82%. However, unlike previous thermal treatment methods, no significant increase in porosity was observed (0.88%). Similarly, the increase in permeability was only observed above 400 °C, which signifies the point of initiation of core destruction. The change in porosity ($\phi$) and permeability ($K$) as a function of exposure time ($t$) during the hydropyrolysis was presented as $\phi =1.2956·t^{0.1628}$ and $K=1.2709·e^{0.046·3t},$ respectively. Hydrothermal treatment of both the Domanik and Bazhenov formations by Stennikov et al. [64] showed that synthetic oil could be generated from these formations at 300 °C. Karamov et al. [65] studied the evolution of organic porosity for West Siberian shales by an open system in a temperature range of 350–450 °C. The porosity evolution was measured with broad ion beam polishing and scanning electron microscopy (SEM). This work showed that the thermal treatment led to the formation of two kinds of organic pores: shrinkage and spongy. The shrinkage pores are the pores in which hydrocarbon accumulates, while the spongy pores are the smaller pores that are formed due to the transformation of kerogen to hydrocarbons. The shrinkage pores are visible at around 350 °C to 390 °C, while the spongy pores are formed at temperatures above 400 °C. The formation of the spongy pores increased the porosity from 2.3% to 2.9%. Laboratory investigations of subcritical water in Bazhenov tight oil formation were presented by Turakhanov et al. [66] as a method of in situ shale retorting. The experiment showed that the initiation of the conversion of kerogen to mobile hydrocarbon was recorded when the temperature was increased to 350 °C. The results showed a transformation of bitumen and kerogen to mobile hydrocarbons and additional immobile char. This study concluded that the cyclic injection of subcritical water is effective for the in situ conversion process for the presented West Siberian reservoir.

2.1.3. State-of-the-Art of the Application of Thermal EOR Methods to West Siberian Shale and Tight Reservoirs

With respect to in situ combustion, on the one hand, very low formation permeability of shales and tight reservoirs may reduce the well injectivity of air (injected gas) [42,67]. On the other hand, a huge surface area of the grains due to low permeability will enhance the reactivity of oil with the injected gas. Furthermore, low reservoir porosity of shales and tight reservoirs could result in considerable thermal losses from heating the rock matrix, hence, interfering with the development of a high-temperature combustion front required for the conversion of kerogen to oil. On the brighter side, the high clay concentrations of the West Siberian formation serve as a catalyst for the combustion by decreasing activation energy. The in situ combustion process provides the optimal heat ideal for thermal breakdown of kerogen, which will aid in the production of substantial unconventional hydrocarbon reserves [42,67].

From the analysis, it can be seen that no studies were dedicated to the influence of the hydraulic fractures on the propagation front during in situ combustion. The network of fractures are reported to be very important by controlling both the propagation of the injected gas and the heat; hence, the complexity of heat propagation through fracture networks requires additional investigation for West Siberian shales and tight reservoirs. The thermal methods, such as HPAI, combustion, and injection of supercritical water, are the highly popular experimental studied methods for West Siberian shales and tight reservoirs. However, the field implementation of these technologies will require an overhaul of additional technologies and studies, such as new well designs and pilot studies [67,68]. A summary of the experimental studies of the application of thermal methods for West Siberian shales and tight reservoirs is presented in

Table 3. Thermal treatment experiments of West Siberian shales and tight reservoirs.

Method	Reservoir	TOC (wt%)	Kerogen Type	Max Temperature (°C)	Permeability Change	Porosity Change	Reference
Air injection	Bazhenov	8.42–17.42	II	446	-	-	Bondarenko et al. [42]
Heating	Bazhenov	-	-	350	x8	-	Gorshkov and Khomyakov [43]
Hydrothermal treatment	Bazhenov	3.91–18.48	II–III	300	-	-	Kovaleva et al. [46]
Hydrothermal treatment	Bazhenov	11	II	400	x77	x2.4	Popov et al. [63]
Hydrothermal treatment	Domanik and Bazhenov	6	-	300	-	-	Stennikov et al. [64]
Cyclic Hydrothermal treatment	Bazhenov	1.82–14.11	II–III	350	-	x3	Turakhanov et al. [66]

.

2.2. Gas Treatment of Shale and Tight Reservoirs

Gas-enhanced oil recovery methods involve the process of injecting gas into oil-bearing formations with the aim of improving the physical properties of the oil and increasing its mobility in the pore space. A wide range of mechanisms have been reported to control the recovery process of gas injection in tight oil formations. These mechanisms include displacement (gas drive) due to depressurization, change in the physical characteristics of the oil due to gas–oil interactions (diffusion, oil swelling, viscosity change, IFT reduction), and changes in the physical properties of the reservoirs (wettability alternation, permeability changes, and porosity change) [69,70,71,72,73,74].

• The Global Experience:

N₂, CO₂, lean gas, and rich gas are among the gases investigated to improve oil recovery from tight and shale formations. Both miscible and immiscible displacements have been recorded for shale formations. At injection pressures lower than the minimum miscibility pressure (MMP), the oil recovery is controlled by the gas drive (displacement) mechanism, while at pressures above the MMP, the process is controlled by both gas drive and diffusion [75,76]. One other important factor from the experimental investigation of shale is the noticeable variation in results with the oil and gas contact surface area. Li and Sheng [77] reported that the higher the surface area of the cores is, the higher the recovery factor. This trend is attributed to the higher contact area between the in situ oil and the injected gas [78,79,80].

The performance of CO₂ was also studied by different researchers [81,82]. Recent studies by Liu et al. [83] showed that oil recovery by CO₂ huff-n-puff was mainly controlled by gas drive. Due to this, oil in macro pores was efficiently displaced, while oil in micro- and submicron pores was hardly displaced. Ding et al. [84] indicated that the performance of CO₂ huff-n-puff is significantly reduced for low-permeability samples. To improve the performance of oil recovery in smaller pores of tight reservoirs, the effect of diffusion has to be enhanced. This means more optimal time should be provided to facilitate the miscibility of the injected solvent and the reservoir oil during the shut period [81].

Reactive gases could also be injected, with the purpose of reacting with the rock minerals to increase the porosity and permeability of the rock matrix [74]. Similar effects can be observed during long-term EOR operations with those gases or CO₂ sequestration projects. At the same time, adsorption on the pore surface is a primary physical trapping mechanism for shale and tight reservoirs. Thus, the CO₂ storage capacity of shales largely depends on their geochemical and mineral composition; the amount of clay minerals and kerogen, which are the main adsorption sites for CO₂; and pore size distribution [85].

Laboratory investigation by Luo et al. [74] showed that the injection of supercritical CO₂ could transform the pore shape of shales. The experiment was conducted on cores from Qaidam and Ordos Basins at a pressure of 10 MPa and a temperature of 50 °C. The results show that the dissolution of clay and carbonate minerals due to the reaction between the supercritical CO₂ (ScCO₂) and shale led to the increase in the micropore structure of the shales. Prior investigation of the effect of supercritical CO₂ in shale reservoirs by Yin et al. [69] confirmed the changes in the physical and structural properties of shales. The experiment was conducted by exposing shale samples to supercritical CO₂ for 30 days at a pressure of 16 MPa and a temperature of 40 °C. The physical and structural properties of the samples before and after the experiment were examined by SEM, N₂ adsorption, X-ray diffraction, and Fourier-transform infrared spectroscopy FTIR analysis. Results show no significant change to the pore shapes with the micropores impacted more than the larger pores. In addition, the surface roughness of the shales decreased from complex to a more regular surface roughness. These changes to the pore structure are mainly attributed to the dissolution effect of ScCO₂. This modification could positively impact the fluid storage capacity of shales and the flow pattern of fluids. The understanding of the effect of CO₂ injection into shale was further highlighted by Pan et al. [86]. Two samples each from marine and terrestrial sources were used in the studies. The effect of supercritical CO₂ on the pore structure was higher than that of the subcritical CO_2.

In addition, the number of micro- and mesopores with a diameter of 0.3–20 nm of the marine samples were reduced significantly after exposure to supercritical CO₂. In contrast, the number of the mesopores increased after exposure to supercritical CO₂. This increased the overall permeability and porosity of the shales. However, it was also recorded that the opposite variation trend was observed for the terrestrial shale samples. A similar investigation by Zhou et al. [87] reported an increase in macro pores from 27.15% to 45.80% after the exposure of shales to supercritical CO₂. These observations show that exposure of shale to supercritical CO₂ converts the existing micron and submicron pores into micropores, hence, increasing the porosity, permeability, and pore volume of the shale formation. The process of conversion of the micron and submicron pores into meso- and macropores is both pressure-dependent and phase-dependent. The higher the pressure is, the higher the rate of pore alteration. Similarly, ScCO₂ has a higher conversion rate as compared to SubCO₂ [73,87,88,89,90,91].

Mineral matrix and pore size alteration together lead to great changes in geomechanical properties, e.g., significant reduction of Young’s modulus [92,93] or Brazilian splitting strength [94], which makes geomechanics one of the most important parameters while planning any CO₂ projects in shales or any other reservoirs. Higher temperature and pressure conditions of the reservoir also influence the level of changes in the mineral matrix and pore size, hence, geomechanical properties [85].

Cudjoe et al. [95] employed NMR measurements of core samples from different depths. The results show a lower recovery factor in samples with higher clay content. Furthermore, higher recovery was observed in the inorganic pores (56%) as compared to the organic pores (48%). The low recovery factor in the organic pores is due to the trapping mechanism/absorption of the injected solvent (methane) by the organic matter in place. Comparative experiments of oil recovery with different gases show that CO₂ out performed other gases [96,97,98].

Yu and Sheng [99] conducted a comparative analysis between the huff-n-puff technique and flooding on shale core samples. N₂ was used as the injected gas in this study. For the same operational period, the huff-n-puff technique produced a higher recovery factor as compared to the continuous flooding technique. However, it was emphasized that the recovery factor between the continuous flooding and huff-n-puff were similar until the period of gas breakthrough was encountered in the continuous flooding experiment.

It is a widely agreed phenomenon in both experiments and field pilots that gas injected above the minimum miscibility pressure produced a higher recovery factor as compared to gas injection below the minimum miscibility pressure [74]. Different studies have presented various methodologies of obtaining the minimum miscibility pressure for a given injected gas and respective reservoir oil. The traditional MMP measurement involves the use of cores from the reservoirs through a flooding system at different pressures until a recovery factor close to 100% is recorded. In recent times, microfluidics technology has been adopted to simplify the filtration process by creating a microfluidic analogue to be used. Teklu et al. [73] presented a mathematical model of determining the MMP not only as a function of the fluid interactions, but also as a function of the physical proprieties of the pores. Flow in submicron pores is different from the flow in meso- and macropores: there are changes in the dew-point pressure, interfacial tension (IFT), and minimum miscibility pressure (MMP). Results showed that the MMP in pores below 10 nm is lower than the MMP in larger pores. This decrease in the MMP is due to confinement effects. In addition, the lower MMP was observed for CH₄, N₂, and CO as compared to CO₂. Teklu et al. [73] showed that for all the ranges of simulated pores, gas mixture (inert and hydrocarbon gases) recorded higher values of MMP, while the lowest recorded MMP was related to hydrocarbon gases. Furthermore, most of the simulated gases (or gas mixtures) remained constant for pore radii above 10 nm. The MMP of hydrocarbon gases (lean gas and rich gas) is lower than that of CO₂ and inert gases N₂.

Pilot studies have also shown that for tight and shale reservoirs with very high heterogeneity, the process of continuous flooding could lead to short circuiting, which is the principle when the injected fluid flows through a preferred relatively higher permeability, hence leading to fingering. While many studies have shown that the huff-n-puff method of injection reduces the probability of the occurrence of short circuits, this technique still requires a long waiting time for the injected solvent to interact with the oil due to the low permeability of the tight oil and shale reservoirs. In addition, the recovery in the huff-n-puff method decreases with the number of cycles [18,20,100]. To overcome this problem, single-well alternating production (SWAP) and single-well alternating simultaneous injection and production (SWASIP) were investigated and presented by many studies [73,89,101]. The methods are based on a single-tubing well completion technique with multiple valves that can be activated remotely. This allows the injection and production from the spatially alternating fracture sections in a single-well alternating production (SWAP) [102]. Numerical investigations have shown a better performance of SWAP when compared to huff-n-puff and SWASIP approaches [74,89]. The schematic of the injection and production mode of the SWAP approach is presented in Figure 2.

• Advances in Experimental studies of gas injection on West Siberian Shales:

Shilov et al. [76] investigated the efficiency of huff-n-puff rich-gas injection to improve the recovery factor of organic matter bearing tight oil samples from West Siberian origins. The experimental investigation employed the huff-n-puff technique under reservoir conditions. Five cycles of huff-n-puff were conducted on 18 different core samples as different total organic content (TOC), porosity, permeability, and dimensions. The results show an increase in the recovery factor with increasing pressure, which is attributed to the increase in miscibility. In contrast, the initial TOC content has a negative effect on the recovery factor due to absorption of the injected gas by the kerogen in the rock matrix. Furthermore, a higher recovery factor was recorded for the experimental results conducted at a pressure above the minimum miscibility pressure. A correlation among the porosity, the initial TOC, and the recovery factor for the core samples was presented (Figure 3). By arranging the core samples in an order of increasing porosity, a correlation among the recovery factor, initial TOC, and the porosity was established: an increase in recovery factor with increasing porosity and decreasing TOC (Figure 3).

State-of-the-Art of the Application of Gas EOR Methods to West Siberian Shale and Tight Reservoirs

There is not extensive literature supporting the application of gas injection for West Siberian shale and tight oil formations. Numerous studies documented in the global literature highlight the effectiveness of various gases when utilized in tight oil formations [69,75,78]. Moreover, the injection of CO₂ into tight oil reservoirs has emerged as a promising technique for enhancing oil recovery, as confirmed by several investigations. Remarkably, this method not only boosts oil production, but also serves as a means of geo-sequestration for CO₂, addressing environmental concerns associated with greenhouse gas emissions [96]. A comparison of the discussed EOR methods is presented at the end of this section in

Table 4. Comparison of the review of EOR methods for Shales and Tight Reservoirs.

Method	Advantage	Disadvantage
In situ combustion	Air is readily available. Versatility: can be applied to light oil (HPAI), heavy oil, tight oil, and shale reservoirs [42,104]. No heat lost to the wellbore. Provides optimal heat required for in situ conversion and upgrading of organic matter and bitumen [42]. Geothermal energy generation [105].	Relatively difficult to implement on a commercial scale [106]. Gravity override (Top–Bottom ISC) [107]. Gas channeling (Top–Bottom ISC) [107]. Inability to reproduce high recovery factors at the field scale as reported experimentally [108]. Corrosion [109].
Hot-fluid Injection	Well-established technology. Suitable for in situ upgrading [63,66].	Heat loss to wellbore [110,111]. Clay swelling (hot-water injection) [112]. Temperature is limited to up to 400 °C [63,66].
Gas injection	Well-known technology. Geo-sequestration when CO2 is used as the solvent [113]. Supercritical injection of CO2 can react with clays in shales and improve the permeability and porosity of such formations [63].	Not all gases are readily available. Gas breakthrough. Corrosion [110,111].

.

3. Numerical Modeling of Shale and Tight Reservoirs

3.1. Core Reservoir Scale: Hydrodynamic Modeling

Numerical modeling of core-scale experiment serves as the bridge between laboratory experiments and up-scaling a model to the reservoir scale. The core-scale simulations allow the validation of the model parameters on laboratory experimental data through history matching. These parameters include the effective permeability and the relative permeability of oil, water, and gas phases, and the diffusion coefficient of the injected solvents. Matched parameters could be readily incorporated into up-scaled reservoir models for field simulations. Likewise, the reservoir-scale model is vital for the designing of production methods to be implemented in field pilots or field-scale productions. Mukhina et al. [24] conducted a comprehensive numerical simulation of a West Siberian reservoir to determine the best EOR mechanism to improve oil recovery. The EOR methods presented in this study are hot-water injection, gas injection, and surfactant injection.

The study concluded that the injection of heated agents was the most effective method due to the ability to successfully reduce the oil viscosity, desorption of heavier and lighter hydrocarbons, and increase in the matrix permeability and thermal cracking of kerogen. Shilov et al. [76] presented a core-scale simulation of West Siberian shale samples. The numerical model was developed using a radial grid, a similar method to that of Alharthy et al. [81] and Li et al. [114]. The model showed that the re-injection of associated gas is an effective method of increasing the recovery factor of tight reservoirs. Numerical modeling of gas injection into other tight and shale geological formations has demonstrated that CO₂ is a very efficient agent for increasing the recovery factors of such fields [90]. One of the underlying parameters to increase the oil recovery during gas injection is the minimum miscibility pressure [115].

Jia and Sheng [109] presented a numerical model of huff-n-puff air injection in shales. The simple Cartesian grid model was used in the simulation with real input data from the Eagle Ford formation. The results show that the huff-n-puff air injection outperforms the continuous air injection technique to ensure the accurate temperature distribution and flow of fluid into the fractures local grid refinement around the fractures. The most common method is to refine the grids based on logarithmic spacing [116,117,118]. Zhu et al. [119] presented an analysis of injection temperature, TOC, and surface area by hydrodynamic modeling in the studies of hot-nitrogen injection in a shale of Songlia Basin. It was reported that the increase in temperature increased the rate of pyrolysis and enhanced the mobility of oil. Similarly, Shuai et al. [120] presented a numerical modeling of pyrolysis by the injection of supercritical CO₂. Diffusive and advection effects on the pyrolysis were investigated. The simulations showed that Sc CO₂ could be an effective agent for the conversion of organic matter to liquid hydrocarbons in shale formations. The sensitive operating parameters are the temperature and the retention time. Junira et al. [121] presented the method of selecting the appropriate gas for gas injection in tight and shale reservoirs. The numerical results indicate that gases such as CO₂, C₂H₆, and C₃H₈ are better suited for reservoir fluid with a higher gas-to-oil ratio (GOR), while gases such as N₂ and CH₄ are suited for reservoirs with a lower GOR. Furthermore, experiment studies have shown that methane absorption by organic rich shales reduces the recovery factor in shale and tight formations. Wang et al. [122] presented a numerical model to analyze the impact of gas adsorption on the recovery factor at the field scale. In this study, the Langmuir extended model for multi-component adsorption was used. The study showed that the adsorption is higher in reservoirs with lower permeability due to nanopore confinement effects. Korobitsyn et al. [123] presented an integrated approach of modeling, planning, and conducting a multistage hydraulic fracturing operation. The study introduced the injection of fluids of different viscosity at different stages of the hydraulic fracturing process. Simulations based on the proposed method effectively increased the initial production of the wells as compared to the traditional approaches. Erofeev et al. [124] presented a numerical model that can be coupled with a hydrodynamic simulator for the optimization of the thermal enhanced oil recovery method from West Siberia shale reservoirs. The module allows the simulation of permeability variation during the thermal treatment of shale reservoirs. Further analysis showed the injection of supercritical water could increase the oil recovery factor from 2% to 5% after 5 years.

3.2. Pore-Scale Digital Rock Physics

Pore-scale simulations have become increasingly popular in the oil and gas industry due to the coupled improvement in both imaging techniques and simulation tools [125]. The digital analogues of pore networks are generated from reconstructed images generated by either Computer Tomography (CT) scans or a Focused Ion Beam–Scanning Electron Microscope (FIB-SEM). Furthermore, the images are then processed to obtain the final digital pore network. The method of obtaining digital pore networks is known as Digital Image Processing (DIP). The complete workflow of DIP is presented in [126]. While the imaging techniques are reliable for generating pore networks for conventional reservoirs, there is a drawback when applied to unconventional reservoirs, which are mainly characterized by submicron pores. With respect to the modeling of fluid flow, the main methods reported in the literature are Direct Numerical simulations (DNSs), Molecular-Scale simulation-Lattice Boltzmann (LBM), and Pore-Network simulations (PNMs). While the flow equations for the modeling of fluid flow in submicron pores are improving, there is still a drawback in the process of generation of a reliable geometry for the simulation. Currently, both FIB-SEM and CT scans have limitations due to resolution; hence, the pore networks generated do not account for the pores below the set resolution. In the research by Wang et al. [127], the experimentally measured porosity and the porosity from the generated FIB-SEM 3D pore network varied by 3%, which was due to the inability to account for the pores less than 20 nm.

Orlov et al. [128] presented a comparative analysis on the different methods of numerical computation of the absolute permeability on digitally generated pore networks. In this research, two comparative analyses were made. First, three different methods of generating digital pore networks were analyzed (Manual DIP, Cross-Laboratory Control DIP, and Automated DIP). Secondly, different numerical methods of computing the relative permeability were applied (PNM, LB, and direct methods). The results show that when different numerical methods are applied on the same DIP, the standard deviation is approximately 0.71 mD. Higher variation was reported for the PNM as compared to the direct simulation methods. This is because the direct simulation methods generate robust geometry from the actual digital pore network, while the PNM is based on approximation of the pores. Furthermore, when the same calculation method (DiMP), but different DIP, is applied, the standard deviation is approximately 1.41 mD. While the choice of numerical method is important, the analyses have shown that the method of construction of the digital pore network and segmentation is more important to obtain an accurate result. Zhao et al. [128,129] made a similar comparison of multi-scale pore-scale modeling. The difference between the comparison of Orlov et al. [128] and Zhao et al. [129] is that Orlov et al. [122] consisted of single-phase flow to compute the absolute permeability of the digital rock samples, while Zhao et al. [129] considered multiphase flow. Regardless, Zhao et al. [129] reached a similar conclusion on the robustness of the direct simulation methods and their ability to capture the detail in the corner and the flow on sharp edges. In contrast, the pore network models lack this robustness. However, the pore network models are computationally efficient as compared to the direct methods. This leads to the tradeoff between the choice of which method to implement.

To eliminate the disadvantages of individual methods, most researchers have implemented multi-scale modeling techniques by coupling the different methods of simulations. Multi-scale digital rock imaging and modeling methods are applied to highly heterogeneous rocks to account for the contribution of the submicron pores to the flow process. A multi-scale approach by Soulaine et al. [130] based on the micro-continuum approach accounts for bulk flow and the mechanism that controls the fluid flow in nanopores (slip flow, adsorption, surface diffusion). Lesinigo et al. [131] presented a multi-scale approach of computing the flow in fractures and matrix by coupling of the Darcy and Brinkman equations. The common multi-scale couplings for shales and tight oil reserves presented in the literature include PNM-LBM [132,133] and Volume of fluid (VOF)-LBM (Chen et al. [134]). To overcome the problem with non-resolved pores in shale and tight oil reservoirs, Ebadi et al. [126] presented a workflow of accounting for the sub-resolved pores in the computation of porosity and permeability from digital rock models obtained from West Siberian cores. The methodology consists of obtaining digital images from the core samples. Furthermore, down-sampling is applied, and the porosity of an effective resolution of 0 µm per voxel (µm/vox) is determined from the approximated trendline of the porosity data. Finally, a bias is computed and used as a correction coefficient for the permeability data. This method improved the absolute permeability predictions when compared to experimental data. In addition, this did not require image scans of different resolution and multi-scale simulation.

You and Lee [135] presented a method of characterizing the impact of accessory minerals in shales (pyrite)–fluid interaction on the overall fluid flow in porous media. You and Lee [134] concluded that the pyrite oxidation does not significantly impact the morphology of the pores. A direct numerical modeling of the deviation of fluid flow from Darcy Law due to osmotic effect in submicron pores of tight and shale formations was presented by Dorhjie et al. [136]. The results show the possibility of deviation of the fluid flow from the Darcy law for pores with a smaller radius below 1000 nm. In addition, the results show that the higher the chemical potential gradient created due to reserves osmosis in the submicron pores, the higher the deviation from Darcy law. Ebadi et al. [137] provided a novel workflow to overcome the unresolved submicron pores in micro-CT images. The process involved generating Xenon enhanced micro-CT images. Subtracting binary classic micro-CT images from the generated Xenon enhanced images enables the resolution of the submicron pores. These images are further processed to extract the porosity, and subsequently, the permeability maps are further used in the computation of the permeability of the porous medium. This allows the contribution of submicron pores to be added to the flow computation and increases the accuracy of results when compared to experimental data.

3.3. Data-Driven Modeling Approaches

Artificial intelligence has gained wide application in various disciplines across the petroleum industry. The application of machine learning and artificial intelligence involves several stages: data collection, data preparation, choosing the machine learning model, training, and testing of the model and validation [138]. Different researchers have applied artificial intelligence methods to predict geological [139], petrophysical [140], and geomechanical properties [141] and production and enhanced oil recovery methods’ efficiency for shale and tight reservoirs [142,143]. A systematic review of the application of artificial intelligence by Syed et al. [144] has shown an exponential increase in the publication of artificial intelligence applications to shale and tight reservoirs. The application of data-driven approaches for the study of West Siberian shales was dedicated mostly to the characterization of geological and petrophysical properties of tight oil reservoirs. With respect to improved and enhanced oil recovery methods, most of the research was dedicated to optimization of hydraulic fracturing techniques [145,146,147,148]. A summary of the pore-scale modeling techniques is presented in

Table 5. Summary of pore-scale modeling of tight/shale reservoirs.

Investigation	Formation	Method	Findings
Digital rock imaging technique and flow modeling Ebadi et al. [126]	Achimov	3D digital rock model with binary of only connected pores (single spatial resolution imaging)	Simplified workflow of improving the computation of the permeability of tight reservoirs.
Pore-scale absolute permeability of tight oil formation Orlov et al. [128]	Achimov	Direct simulation, LBM 3 & PNM 2	It is more vital to execute accurate and precise image processing and segmentation than to use complex computation approaches.
Digital rock imaging technique and flow modeling Kang et al. [149]	Berea sandstone	3D digital rock model with binary of only connected pores (double spatial resolution imaging)	A pore segmentation algorithm for dealing with gray-level pore space while preserving pore connection.
Multi-scale PNM 2 Jiang et al. [150]	-	Two-scale and three-scale PNM 2 construction from micro-CT measurements from different scales	Network integration with spatial correlation of fine-scale network components leads to considerably different relative permeabilities, as shown by applications to rock CT images, as compared to integration with uniformly distributed (uncorrelated) fine-scale pores.
Multi-scale flow modeling Carrillo et al. [151]	-	Micro-continuum model based on modification to the Navier–Stokes equation	A model that only requires a single momentum conservation equation, eliminating the need for multiple meshes, distinct solvers, or complicated interfacial conditions.
Multi-scale flow modeling Guo et al. [152]	-	Micro-continuum gas flow model for transport in organic rich shales (NSB 1)	Conventional pressure-dependent apparent permeability may not accurately depict the transport characteristics of organic rich shale. The findings demonstrate that surface diffusion and non-Darcy effects are significant at low gas pressure (1 MPa), but that these processes are insignificant at high pressure (50 MPa).
Digital rock imaging technique and flow modeling Wang et al. [116]	Sichuan	Pore-scale flow simulation mode	Pore-scale simulation of gas transport with vigorous incorporation of nano-scale flow mechanisms: gas adsorption, diffusion, and surface diffusion.

¹ NSB—Navier–Stokes–Brinkman; ² PNM—Pore Network Model; ³ LBM—Lattice Boltzmann Method.

.

Morozov et al. [142] presented a comprehensive global framework for hydraulic fracturing optimization using field data. The data used in the study were obtained from 5000 wells drilled in West Siberia between 2013 and 2019. For the optimization of hydraulic fracturing, the 92 initial features of the field and wellbore were reduced to 35 features after feature analysis. Among the different machine learning models applied, the decision tree models outperformed the other models in the prediction of the cumulative oil production. Their findings showed the ten most important feature optimizations of hydraulic fracturing. Furthermore, Duplyakov et al. [143] utilized the data of Morozov et al. [142] to identify the parameters for optimization of hydraulic fracturing by an inverse approach. The optimization methods used in this study showed that increasing the proppant mass and reducing the average proppant concentration were vital to maximize the well productivity. Meshalkin et al. [140] presented an automated recognition of lithotypes with thermal core profiling for West Siberian shales based on spatial Fourier transform of high-resolution thermal conductivity profiles. It was demonstrated that the issue of automated rock typing can be resolved with simply high-resolution thermal conductivity data measured parallel and perpendicular to the rock bedding plane. Syed et al. [153] presented a machine learning approach to predict the gas production of shale reservoirs. The data used in the study include geological data, petrophysical data, well completion data, and hydraulic fracturing data. Several machine learning methods were applied, namely, the Gaussian Process, Support Vector Machine, K-mean Clustering, and Artificial and Recurrent Neural Network. From this study, all the developed models were able to effectively predict gas production from the shale formation from which the data were collected and used in the model. Janiga et al. [118] presented a data-driven approach to find the optimal huff-n-puff design through the stochastic, population-based, particle swarm optimization (PSO) method. The proposed methodology includes first the conduction of a huff-n-puff laboratory experiment. Secondly, the results of the laboratory measurements were used to develop a numerical model. The model was validated, and the experimental data and results from the experiment and numerical model were used as input data for the genetic programming model and further used to optimize the huff-n-puff design. A summary of data-driven modeling approaches is presented in

Table 6. Summary of ML application to the exploration of tight and shale reservoirs.

ML Model	Predicted Parameter	Data Points	Formation	Results
NMGM-ARIMA 1 & ARIMA-ANN 2 [154]	Gas production	60	-	Both models produced satisfactory results, with the ARIMA-ANN outperforming the NMGM-ARIMA.
SVM 3 & ANN 4 [155]	Hydrocarbon production	144	Eagle ford, Bakken & Niobrara	It was discovered that RSM and LSSVM had greater oil recovery prediction capabilities than ANN. In addition, LSSVM predicts the gas–oil ratio with the maximum degree of precision.
SVM 3 [156]	Sweet spots	-	Shale reservoir in China	It forecasts numerous features for sweet spots in reservoirs, allowing for an objective assessment of shale gas potential.
LSTM 5 [157]	Gas production	data from 332 shale gas wells	Durvenay & Montney	The suggested approach can be used with conventional wells; however, because of high density drilling and poor decline curve analysis performance, it is more suitable for unconventional wells.
ANN 4 & SVM 3 [158]	Loss circulation	1120 data points from 385 wells	Iranian origin	Out of all the models tested, the SVM with 18 variables had the highest accuracy (accuracy of 0.92 and 0.91 for training and test model, respectively).
SVM 3 & Prevalent classifiers [159]	Sweet spot	Data from 73 wells	-	The maximum agreement rate of 83.37% is produced by XGBoost if the well-log interpreted parameters are omitted from the original data set. During training, GBDT reduces complexity by over 70% compared to SVMs.
LR 6 & ANN 4 & Gradient-boosting decision trees & Extra trees [160]	Gas production	Data from 573 horizontal wells	Duvernay	The results show that the major factors that contribute to shale production for the given formation are the total fluid injection, total proppant mass, well TVD, permeability, porosity, gas saturation, number of stages, formation pressure, horizontal length, distance to fault, and formation thickness.
GA 7 & DE 8 & PSO 9 [161]	Hydraulic fracturing placement	-	-	By adjusting the control vector corresponding to the number of wellbores, HF spacing, fracture half-length, and numerous HF stages, DE and PSO approaches displayed improved performance and objective function improvement. However, the results of the GA examples revealed that this method was unable to discover the optimal values for the decision variables and became stuck in several local optima, resulting in convergence issues.

Autoregressive Integrated Moving Average Model ¹; Autoregressive Integrated Moving Average Model-Artificial neural network ²; Support Vector Machine ³; Artificial neural network ⁴; Long short-term memory ⁵; Logistic regression ⁶; Genetic algorithm ⁷; Differential evolution ⁸; Particle Swarm optimization ⁹.

.

4. Field-Scale Shale and Tight Oil Recovery

One of the greatest projects was conducted from 2017 to 2021, revealing an industry-scale experiment at the shale technology test site in West Siberia. The project was aimed at a comprehensive study of the Bazhenov shale oil formation, assessment of key production and economic drivers, and development of a high-tech oil service market. Following the specific design of the experiment, the company drilled more than 30 horizontal wells and performed more than 620 stages of hydraulic fracturing, while pumping about 70,000 tons of proppant and 650,000 m³ of fracturing fluid. The most technologically advanced solution became horizontal wells with a cemented liner 1500 m long; a pump down plug-and-perf system with dissolvable frac plugs; an engineered completion of 30 stages, with four perforation clusters per stage; and a proppant mass of 150 tons per stage of clean fracturing fluid, with a flow rate of up to 16 m³/min. The average cumulative well production for 180 days for this group of wells reached up to 89,000 bbl, which is comparable to North American analogues with a longer length of wells and greater number of stages. As a result of the experiment, the development cost for Bazhenov shale was reduced by 72%—from 57 to 16 USD/bbl [162]. One of the most significant contributions to the result was the implementation of a unique research program, including covering the entire site with wide-azimuth 3D seismic and controlled source electromagnetic surveys, and acquiring a large set of spatially distributed well data (core, logs, PVT, etc.). This program made it possible to create a digital twin of the most complex reservoir system and move to a new level in the field development management [163]. The progress achieved allowed for starting the commercial development of shale oil as early as 2023. For the full-scale multi-basin production, the company conducts exploration studies in the promising areas in West Siberia (Bazhenov formation) and the Volga-Ural region (Domanik formation) [164,165]. It does not stop there and strives to realize its technological potential as efficiently as possible: in 2023, the program foresees the drilling of 2000 m long wells and an increase in pumping rate up to 18–20 m³/min; R&D projects in the field of waterless fracturing and EOR are approaching the development stage; for 2024, as part of the import substitution program, rotary-steerable systems and a hydraulic fracturing fleet are planned to be tested on the field.

Another case study is a set of pilot projects on the Bazhenov formation at the Salym and Priobskoye fields. By the end of 2020, ten horizontal wells with ball-n-drop completion were drilled. An industrial company managed to minimize complications in the process of drilling in abnormally high-pressure and -temperature conditions. The average length of the horizontal section was 970 m, the maximum 1400 m. Nine horizontal wells with MSHF having an average initial rate up to 55 t/day were put into production. The average proppant mass per stage was 72 tons. As a result of the pilot project, the map of prospects for the Bazhenov formation has been validated, and a drilling program for more than 20 horizontal wells has been launched for the period 2021–2023 [166].

Sredne-Nazym shale field is also a great example of HF application: in the period 2020–2021, an oil company significantly increased the drilling of horizontal wells—21 wells against 8 wells in 2018–2019. Wells in 2018 were characterized by a length of up to 1000 m; an openhole completion system with controllable frac ports, from 6 to 10 zones; and a standard hydraulic fracturing design with 30 tons of proppant per stage. In 2020, the company switched to drilling longer wells with up to 1700 m length, applying its own completion technology-sliding sleeves activated by a special CT-deployed wrench. Hydraulic fracturing was carried out on low-viscosity hybrid fluids at rates up to 8 m³/min, with 60 tons of proppant per stage. The average cumulative well production for 180 days reached up to 69,000 barrels of oil [167].

The effective temperature for thermal cracking of solid organic matter in shales is higher than the one needed for viscous oil. Furthermore, shales have a substantially higher heat conductivity (hence, greater heat loss) than that of a conventional formation [109] caused by low porosity, making it harder to reach the high temperature required for kerogen transformation. This brings thermal EOR methods up to a whole new level of field development and well construction. On the one hand, according to lab studies and numerical simulation, the expected efficiency from thermal EOR is the highest among other EOR methods for shale oil due to kerogen transformations into additional oil. At the same time, the required high temperatures lead to various difficulties in preparations for the well treatment.

The main risks during hot-water injection are associated with well materials and heat losses. Most of the standard well materials are not stable enough to withstand the temperatures of hot water (>350 °C) and the corrosion caused by it [110,111]. Even the temperature of produced fluids, which might exceed 150 °C, is a problem for some operational parts, e.g., the electric pump. Moreover, the well materials should not only be heat-resistant, but also should prevent heat losses, which can be achieved by using thermal insulation [168,169,170]. Otherwise, the heat transfer to the destination would be challenging. However, even the insulated well will not preserve all the heat, and, in some cases, achieving 350 °C in deep reservoirs requires injection of a large amount of vapor overheated up to 500–600 °C from the surface under high pressure. The vapor generator with such parameters does not yet exist on the market. Additionally, the well design, its incline number of fractures, their shapes, and the development scenario should provide heat distribution within the reservoir away from the well. For example, a long horizontal part of the well with many fractures might not allow the heated fluid to distribute evenly and maintain the temperature along the well, losing heat in the first few fractures closest to the vertical part. It should be remembered that the temperature decrease leads to the distribution of “cold” water into the reservoir, which is highly undesirable, as mentioned above. Precise preliminary mathematical and hydrodynamic calculations are required to minimize heat losses and optimize the operational parameters. Finally, although water is commonly used for heat transfer due to its availability, there are specific requirements for water purification [171,172] to secure the well materials. All the described problems are converged towards the main obstacle—significant increase of the treatment preparation costs. Until the challenges are scientifically and economically solved, hot-water injection may never be investigated in a shale field.

Air (or oxygen) injection in shales reveals not much positive state-of-the-art, the base of field tests is still extremely poor, and no commercial projects exist worldwide. The risks of air injection are mainly caused by corrosion [109], which can be prevented by suitable well completion, and explosiveness, which can be avoided by careful selection of the injection regime and development scenario using numerical simulation and laboratory tests. On the other hand, due to the inexpensive cost and widespread availability of air, it is anticipated that this EOR technique is more appropriate for the economics of field development.

Sheng [17] described a few successful cases of air injection in shale-like carbonate reservoirs characterized by permeability 10–20 mD, porosity 11–19%, significant depth, low density and viscosity of oil, and high reservoir temperature. However, the described reservoir properties were far from those of real shales. An overview on a very few thermal EOR field tests for oil shales was presented by Kang et al. [173]. All the performed pilot tests can be divided into classical and alternative methods. Classical methods include convection (thermal fluid injection) and combustion retorting technologies. Alternative methods are conductive (electrical heating) and radiant (radio frequency and microwave) retorting technologies. Each technology demonstrates promising results of the field experience. However, none of the methods have yet been deemed suitable for widespread commercial distribution. The report by Johnson et al. [174] notes that in situ kerogen conversion processes can only be economically viable in reservoirs characterized by sufficient permeability or in reservoirs where such permeability is artificially created by hydraulic fracturing. The well type is also considered to play a key role in the ISC process—ignition initiation requires high pressure of the injected oxygen, which introduces difficulty for performing it in a long horizontal well with many fracturing stages. The most famous pilot field case for thermal gas treatment in a shale oil reservoir is the project at the Sredne-Nazym field, West Siberia [175]. The authors report the potentially successful development of thermal gas treatment technology at Bazhenov shale. The described technology integrates thermal and gas EOR and includes the injection of air and water into the reservoir. For more than 10 years, the company has been studying a way to heat the Bazhenov formation by pumping an air–water mixture with high-performing compressors. First, five wells in the pilot field had been operating for a few years in a depletion mode before the thermal gas treatment started. During the last years, the oil rate dropped from 45 to 5 t/day, and the reservoir pressure dramatically decreased from 300 to 150 atm. At this point, the recovery coefficient was 4%, which indicated the low efficiency of the depletion mode for BF development. Several stages of air and water injection were performed during the following 5 years, which led to additional gas production of up to 80% and oil recovery as a result of the reservoir pressure increase back up to 250 atm and kerogen transformation. Alteration of the composition and decreases in the density and viscosity were observed for the newly produced oil. In sum, 30% of additional oil was recovered during the pilot project, according to the reported data. However, the details of injection parameters, as well as air–water interaction aspects, were not revealed in this study. The pilot phase continues, as the profitability of the technology has not yet been confirmed [176].

5. Conclusions

This study reviewed the progress in the global exploration of shales, with much emphasis on the progress made for the exploration of West Siberian shale and tight oil reservoirs. The advances in gas EOR, thermal EOR, numerical modeling techniques, and field pilots were discussed.

From the review, thermal EOR methods, particularly in situ oil thermal synthesis technologies, like HPAI, combustion, and supercritical-water injection, show promise for recovering shale and tight oil fields, in West Siberia in particular. However, implementing these methods is challenging due to the complexity, high costs, and extensive experimental phases required, including precise determination of kinetics, thermal rock properties, and fluid behavior. Deviations in the recovery factor from field pilots and modeling data have made in situ combustion less attractive, despite its reported advantages.

Each shale and tight oil field has unique characteristics, necessitating a specialized approach for exploration and EOR. There is no universally applicable EOR scenario for all reservoirs; hence, an individualized approach with detailed reservoir characterization, lab testing, numerical simulations, and field trials is essential to develop economically viable and successful projects.

To bridge the gap between lab-recorded recovery factors and pilot studies, improved techniques like the 3D laboratory combustion chamber are crucial. However, specialized equipment and thermal insulation capable of reaching high temperatures are needed. The existing literature provides valuable statistical evidence to enhance this technology.

With respect to gas injection, supercritical CO₂ has gained popularity for the exploitation of North American shales; there seems to be no studies of such technology for West Siberian shale present in the literature. It is recommended that the investigation and implementation of such technology could yield a triple benefit of converting kerogen to liquid hydrocarbon, improving the mobility of liquid hydrocarbons, and potentially leading to geo-sequestration of CO₂.