Horizontal drilling and hydraulic fracturing can increase the contact area of the reservoir and fractures, and thus making oil production from the tight reservoir efficient and economic. According to the U.S. Energy Information Administration (EIA), 54% of the U.S oil production came from the tight reservoirs in 2017^[1]. In contrast, tight oil only accounted for less than 1% of the total oil production in China^[2,3], because of complicated geological conditions and poorer fracturing effect of tight oil reservoirs in China. Although the single well production rate of tight oil reservoirs can be enhanced by increasing the horizontal section length, reducing spacing between stages, and increasing dosage of the fracturing fluid and proppants, the ultimate recovery of this kind of reservoir is still less than 10%^[4,5,6]. Therefore, more advanced techniques are needed to maximize the oil recovery from these reservoirs with hydraulic fracturing. In this work, through analyzing the challenges in fracturing of tight oil reservoirs, integrated techniques of high effective fracture-network stimulation and EOR are proposed.

In tight reservoirs, oil accumulates in pore throats of rocks with permeability less than 0.1×10^-3 μm^2[7], so oil stored in the rock matrix far from the created fractures is hard to be recovered by natural depletion. Previous studies on over 1000 horizontal wells in Barnett and Marcellus gas fields indicated that for a fracture spacing of 15 m, the interference of two neighboring fractures would occur after 4-5 years fracturing, the gas from the matrix in the middle of two parallel fractures can flow to the fractures and be recovered^[8,9]. For oil with higher viscosity, the interference time would be longer unless the fracture spacing is reduced further. Therefore, the first challenge in improving fracturing effect is to create an dense and effective fracture network that enlarges the drainage area and ensures a high fracture conductivity.

Although different clusters and stages of fractures in tight oil and gas wells are typically designed to have similar scales during hydraulic fracturing, distributed temperature sensing (DTS) and production logging results often show that different clusters of fractures develop quite differently. In most cases, approximately 70% of the production comes from 20% of the fractures^[10,11]. Propagation of fractures can change the stress field around, and in turn limits the propagation of neighboring fractures, this interference is known as the “stress shadow” effect^[12]. This effect can make the fractures divert or even two neighboring fractures merge gradually during propagation^[13,14] when arrangement of fractures becomes denser for larger drainage area, resulting in the reduction of the stimulated volume by fracture network. Moreover, this effect can make the fractures different in width, which affects the filling of the proppants^[15].

Furthermore, calculation based on the material balance of the injected fracturing fluid, the wellbore volume and the total volume of designed fractures shows that more than 90% of the fracturing fluid flows into the induced unpropped fractures^[16]. The width of this type of fractures is typically in microns to sub-microns order, too small for conventional proppants to get in to form effective support, so they can’t provide sufficient conductivity for oil flow^[17].

Another characteristic of the tight oil reservoir is that the oil production rate declines sharply in the early stage. For most tight oil horizontal wells in the U.S., the oil production rate falls to 30% of the initial peak rate after one year of production^[18]. During hydraulic fracturing, a significant amount of fracturing fluid is injected into the reservoir to form artificial fracture network^[19,20]. If the fracturing fluid can imbibe into the tight rock and replace the oil, the sweeping volume of the fracturing fluid and the producing degree of oil would be enhanced and the oil production decline would be slowed down effectively. Therefore, the second challenge in the tight oil fracturing is to enhance the imbibition of the fracturing fluid.

In the tight oil reservoir, asphaltene in the crude oil can absorb with the rock surface, changing the original water-wet minerals such as quartz and clay to oil-wet^[21]. Meanwhile, the polar compounds in oil, such as naphthenic acids, can adsorb on the carbonate minerals such as calcite and dolomite, also making them oil-wet^[22]. Since the small pores and pore throats in the tight rock are oil-wet with strong capillary force, which hinders the fracturing fluid flowing into the pores to displace the oil. To alter the rock wettability and enhance the imbibition effect of fracturing fluid, the surfactant can be added into fracturing fluid. This idea initially came from the chemical flooding of fractured oil-wet carbonate reservoir^[23,24]. At present, the similar types of surfactants were chosen and tested in the tight reservoirs^[25,26]. But unlike the conventional fractured carbonate reservoir, tight reservoirs have large specific surface area of matrix pores. This requires a large dosage of surfactant to compensate its adsorption loss on the rock surface, which makes this method uneconomic for the tight oil reservoirs. Moreover, adding surfactants would generate ultra-low interfacial tension (IFT) between oil and water, which would slow down the imbibition rate with capillary force, thus limiting the application of this method in oil fields^[27,28]. For efficient and economic imbibition and replacement, the chosen surfactant needs to have 3 features: (1) can alter the rock from oil-wet to water-wet, (2) has minimal adsorption loss on the rock surface, and (3) can maintain certain oil-water IFT for driving the imbibition.

To overcome the two challenges, based on the previous research results, integrated hydraulic fracturing techniques to enhance oil recovery has been proposed to maximize oil recovery from tight reservoirs. Refracturing with temporary plugging agents is proposed to allow fractures within each stage to propagate uniformly, thus enlarging the contact area with the reservoir rock. Meanwhile, using micro-sized proppants in the pad fluid is proposed to support narrower and deeper micro-fractures. A new surfactant system, liquid nanofluid (LNF), has been developed as a fracturing fluid additive to enhance the imbibition of the fracturing fluid, thus to supplement the reservoir energy and maximize the oil production rate.

A series of temporary plugging agents of different shapes and sizes have been developed for different temperatures and pressures. They are suitable for reservoirs with a temperature up to 200 °C and a pressure up to 140 MPa where the conventional mechanical plugs do not work reliably^[29]. These agents can degrade naturally at the reservoir temperature, and the degradation time can be adjusted by tailoring the relative molecular weights of the polymers for synthesizing the agents. After degradation, they can flow back with the fracturing fluid and oil, leaving no solid residue and thus not causing additional formation damage to the reservoir^[30,31]. Moreover, the degraded solution of temporary plugging agent is acidic, and thus can also help to mitigate the near-wellbore formation damage caused by drilling and completion fluids^[32]. The agents can be made into fibers, particles, and balls with diameters up to 15 mm to meet different operation requirements. By optimizing the combinations, dosage and injection order of temporary plugging agents, the plugging and diversion of previously formed fractures and the perforation holes during the refracturing can be realized.

Through a visual fracture plugging simulation apparatus in the laboratory, the synergy effect of fibers and different-sized particles during the plugging process can be observed. As shown in Fig. 1, when the fibers and particles are injected at the same time, the fibers would attach on rough walls of the fracture first to form a “net”, then the “net” would capture particles injected later, plugging the fracture at last^[33]. Meanwhile, the dynamic plugging tests were conducted on a modified fracture-conductivity evaluation system, where the plugging efficiency of the agents inside a rough fracture could be obtained by measuring the pressure difference between upper and lower parts of fractures under simulated confining pressure and formation pressure. The reservoir rock can be used as experiment sample, or its replica from 3D scanning and printing can be used in multiple comparative experiments on fractures with identical wall roughness but different widths (Fig. 2). Once the plugging was formed, the relationship of plugging locations, wall roughness, initial fracture width and different combinations of agents were analyzed through CT scanning. The experimental results indicate that the particles determine the plugging location within the created fracture, while fibers determine both the plugging efficiency and the plugging strength (Fig. 3). Specifically, the sizes of the temporary plugging particles need to be tailored according to the width of created fracture, larger-sized particles are needed for wider fractures, while small-sized particles are needed for tortuous fractures. The experimental results also show that injecting fiber plugging agents first can increase the plugging efficiency and reduce the amount of agents compared with injecting fiber and particle agents together^[33]. When the reservoir is poorer in fracability, a higher concentration of fibers is needed to enhance the plugging strength of the temporary blocking section in the fracture to meet the higher fracturing pressure needed to open a new fracture^[34,35].

After studying the plugging efficiency in a single fracture, hydraulic fracturing and refracturing tests were further carried out in larger outcrop of 300 mm×300 mm×300 mm under tri-axial stress conditions. The experimental results show the combination of fibers and particles can realize temporary plugging effectively, inducing the initiation of secondary fracture diverting from the directions of the previously formed fracture^[36]. To study the fracture propagation after the temporary plugging in the field scale, a piece of numerical simula-tion software was developed using the cohesive zone model based on the extended finite element method (XFEM-based CZM)^[37]. This software can simulate the formation process of temporary plugging section inside or at the mouth of the fracture with different combinations of the temporary plugging agents at different proportions, and the role of the temporary plugging section in triggering the following fracture. The simulated results were compared with outcrop experiment results under tri-axis stress. The software can be used to analyze the impact of rock mechanic parameters and injection program of temporary plugging agents on the multi-cluster fracturing effect, and to optimize the fracturing and temporary plugging scheme. Simulation results show that by optimizing the formula and injection program of temporary plugging agents, the plugging of fracture at the fracture mouth or the fracture diverting inside the fracture can be realized^[37,38,39]. Therefore, it is possible to use this technology to do multi- layer fracturing in vertical wells and multi-cluster fracturing in one stage of horizontal well.

Supported by the experimental and simulation results, the temporary plugging agents have been applied in the refracturing of 26 horizontal wells in the Santanghu basin. On one hand, all these wells had severe production decline after producing for over two years, so they needed refracturing to restore the economic production rate; on the other hand, it is difficult to apply mechanical sealing in the horizontal well, the production increment after refracturing can directly reflect the effect of multi-cluster fracturing in a stage with temporary plugging agents. After optimizing the formula and injection program of temporary plugging agents, refracturing in horizontal sections over 250 m long has been completed, realizing over 2 times of diverting in each stage. Compared with the first time fracturing, the oil production of these wells after refracturing enhanced by 80% on average.

The horizontal section of Well S-1 is at a depth of around 2300 m, where the reservoir has an average porosity of 18% and average permeability of 0.4×10^-3 μm². At the reservoir temperature of 70 °C, the oil has a viscosity of around 140 mPa•s. During the initial fracturing of this well, the 521 m long horizontal section was divided into 6 stages, with 4 clusters per stage. It was fractured through the perf-and-plug operation. During the refracturing, Stages 1-3 and Stages 4-6 in the first time fracturing were separately by drillable bridge plugs into two stages; eight pumping programs were designed to modify these two stages, detailed steps are: (1) inject the slickwater pad fluid, (2) inject guar gel sand-carrying fluid , (3) inject temporary plugging particles for diverting in stage, (4) inject slickwater pad fluid, (5) inject guar gel sand-carrying fluid, (6) inject temporary plugging particles for diverting in stage, (7) inject slickwater pad fluid, (8) inject guar gel sand-carrying fluid. Table 1 shows the dosage of various fluids, proppants, and temporary plugging particles in the refracturing. Fig. 4 shows the changes of tubing pressure, pumping rate and proppant concentration during refracturing. Pressure increased by 3-10 MPa after the injection of temporary plugging particles, moreover, the injection pressure gradually increased during the injection of pad fluid and sand-carrying fluid in the two stages of refracturing (i.e., tubing pressure increased gradually in the steps 1, 4 and 7, and tubing pressure increased gradually in 2, 5 and 8). These observation results indicate that the injected particles formed plugging within the previously created fractures, thus making the following fracturing fluid stimulate the unstimulated or stimulated regions with low degree of the reservoir^[40]. After this refracturing treatment, Well S-1 had an average oil production of 29 t/d in the first month, that is 26% higher than the initial production after the first fracturing.

Furthermore, it is recommended adding micro-size proppant in pad fluid in order to (1) support narrow fractures that conventional proppants cannot get in, especially micron-size induced fractures, and (2) reduce the settling velocity of proppants so the proppant can reach further fractures.

Previous experimental results indicate that the settling velocity of 0.075 mm (200 mesh) quartz sand is 1/10 of that of 0.800 mm/0.425 mm(20/40 mesh) quartz sand and 1/5 of that of 0.425 mm/0.212 mm (40/70 mesh) quartz sand, while the settling velocity of 0.044 mm (325 mesh) quartz sand is 1/44 of that of 0.800 mm/0.425 mm (20/40 mesh) quartz sand and 1/22 of that of 0.425 mm/0.212 mm (40/70 mesh) quartz sand^[41]. Micro-size proppants have been applied in a few tight/shale gas or condensate reservoirs, and the reported production could increase by 30%-50% compared to the wells not using this type of proppants^[42,43]. However, it remains unclear how micro-sized proppants transport and settle in the complex fracture network, including the induced micro-size fractures as well as the branched fractures formed after refracturing. Moreover, it is also worth studying the surface modification of micro-size proppants to enhance the repelling force between the proppant particles to further reduce the settling velocity^[44].

In recent years, the use of solid nanoparticles to enhance oil recovery from the fractured carbonate reservoirs has been studied extensively^[45,46,47]. The study results showed that they could alter the oil-wet carbonate reservoir rock to water-wet, and meanwhile, maintain a moderate oil-water IFT to enhance the imbibition effect^[48]. Although these two properties meet the requirements on fracturing fluid additive for enhancing oil recovery after fracturing, the nano-scale pore throats in tight rocks can cause gathering or plugging of the nanoparticles, and thus resulting in permanent reservoir damage^[49]. In the meantime, ex-situ microemulsions were studied and proposed as fracturing fluid additives to mitigate the water blockage caused by the invaded fluid, thus enhancing the gas production from shale gas or tight gas reservoirs^[50]. In such microemulsion systems, surfactants form thermodynamically stable micelles that can effectively reduce the adsorption loss of surfactants on the rock surface, which enables surfactants to flow deeper and enhance its flowback by reducing the surface tension^[51,52].

Based on the above studies, we have developed a series of liquid nanofluid (LNF) systems to enhance the oil recovery with fracturing fluid flowing into the reservoir. They are well-dispersed microemulsions with a uniform micelle molecule size of around 10 nm, the narrow distribution peaks represent uniform properties (Fig. 5). According to the surface charge of the reservoir rock, the types of surfactants (i.e., anionic, cationic, or non-ionic surfactants) forming LNF can further minimize the adsorption loss of microemulsion on the rock surface. Like the solid nanoparticles, LNF can change the oil-wet rock to water-wet, and maintain a moderate water-oil IFT (e.g., an IFT of 1-3 mN/m between kerosene and water)^[53]. But unlike the solid nanoparticles, results from the pressure transmission tests indicate LNF with flexible structure can deform and squeeze through the nanoscale pore throats in tight rock^[54]. As shown in Fig. 6, when the simulated formation water with and without LNF were injected into the kerosene-saturated tight rock sample with a perme- ability of 0.05×10^-3 μm² at a constant upstream pressure during the pressure transmission test, the downstream pressure of the rock samples reached the equilibrium at the same rate, indicating no extra formation damage was generated by LNF.

To evaluate the effect of LNF on enhancing the oil recovery from the oil-wet tight rock, the imbibition cell tests were conducted using the naphthenic-acid-treated oil-wet limestone core samples^[56] with a permeability of around 0.7×10^-3 μm². In the tests, core samples of 2.54 cm in diameter and 5 cm in length were firstly saturated with oil, then submerged in the simulated fracturing fluid with and without LNF. The total oil recovery rate of the core sample submerged in the liquid without LNF within 50 h was less than 5%; whereas the total oil recovery rate of core sample submerged in the liquid with LNF was over 50% and rising (Fig. 7). The results show that LNF can alter the oil-wet rock to water-wet and trigger water imbibition to displace the oil. Then, the microscopic oil enhancement mechanism of LNF was further evaluated through the low-field nuclear magnetic resonance (NMR) analyzer. From the T₂ spectrum during the CPMG pulse sequence under different time during oil-water displacement, the NMR analyzer can provide the distribution of water and oil in different size pores in tight rock^[57]. Before the scanning of the core sample by NMR, the tight core sample, as ones tested in the imbibition cell, was saturated with the fluorocarbon oil in order to shield the NMR signals from the oil and only allow the water signals to show in the T₂ spectrum. During the experiment, the simulated fracturing fluid with LNF (i.e., 0.1% LNF+2% KCl) was continuously injected into the core sample at a constant rate of 0.05 mL/min. The experiment was continued for 1290 min until no obvious change was detected in the T₂ spectrum. As shown in Fig. 8a, the simulated fracturing fluid with LNF firstly invaded the bigger pores where the capillary pressure was lower; after 360 min in this case, when LNF diffused into the smaller pores and altered their wettability, water started to invade these regions through spontaneous imbibition and displaced the oil (Fig. 8b). The experimental results suggest that the diffusivity and migration speed of LNF can determine the efficiency of LNF in enhancing the oil recovery from tight rock, and shutting-in the well after hydraulic fracturing with LNF can further enhance the oil recovery^[58]. Nevertheless, more study is needed to clarify the mechanism of LNF and to establish the optimization criterion of this system for different types of tight oil reservoirs.

In short, LNF can minimize the adsorption loss of surfactants on the rock surface, making it possible for surfactants to go deep with fracturing fluid into the tight reservoir rocks. For the oil-wet regions of reservoir rock, LNF can alter the wettability of pore throats and pore walls and maintain a certain level of oil-water IFT for enhancing the water imbibition to displace the oil. For the water-wet regions of reservoir rock, the reduction of oil-water IFT by LNF can mitigate the water blockage, enhancing relative oil permeability and oil production rate at the same time^[56,59-60].

There are two main challenges in maximizing the oil production from the tight reservoirs with hydraulic fracturing, which are generating an extensive and effective fracture network and replenishing energy to the matrix for oil-water displacement respectively.

By adding temporary plugging agents in multi-clusters, the new fractures can be diverted to the unstimulated or under- stimulated regions of the reservoir, thus generating a more uniform fracture network. Meanwhile, adding micro-size proppants in the pad fluid can support narrow fractures and fractures far from the wellbore, especially the widespread micro-size induced fractures, to enhance the conductivity of fractures, making it possible to produce the oil within reservoir matrix.

A new surfactant system (i.e., LNF) has been developed which can minimize the adsorption loss of surfactant on the rock surface and thus allow the surfactant to penetrate deeply into the reservoir with the fracturing fluid. The LNF can alter the rock wettability and meanwhile maintain a certain level of oil-water IFT for motivating water imbibition to displace the oil. Using LNF as a fracturing fluid additive is one of the most effective ways to supplement the reservoir energy with the massive injected fracturing fluid and maximize the oil production.