Introduction

The concept of Smart Cities has become an increasingly common trend in technology-based projects (De Paz et al. 2016). Balancing the environment and natural resources is a practical and responsible way to ensure that the environment and resources on the planet are conserved appropriately for the next generation. Geological sequestration is a good means of reducing anthropogenic atmospheric emissions of CO2 that is immediately available and technologically feasible. Three kinds of geological settings have been recognized as major potential CO2 sinks: deep saline-filled sedimentary formations, depleted oil and natural gas reservoirs, and un-mineable coal seams (Bradshaw et al. 2007; Brown et al. 2014; Gislason and Oelkers 2014; Abidoye et al. 2015; Du et al. 2016). However, the geological sequestration in saline aquifers is considered a most viable option as it seems to have very large carbon storage potential (Bachu 2008; Zahid et al. 2011; Zheng et al. 2013). Parts of the reasons for this include the stability, capacity, and ubiquity of these aquifers (Bachu 2015). Stable sedimentary basins are necessary for dependable sequestration, and such basins are found in most continents with estimated capacities of around 1000–100,000 gigatonnes of CO2 (Zahid et al. 2011).

One of the primary concerns in CO2 sequestration is the safety and long term of CO2 immobilization. There are mainly four storage mechanisms for immobilizing CO2 in a porous medium: structural trapping, residual trapping, dissolution trapping, and mineral trapping (Xu et al. 2005; Bachu 2008; Zhang et al. 2009; Hidalgo et al. 2013; Martinez-Landa et al. 2013). When CO2 is injected into the subsurface, it is first trapped by the first two storage mechanisms (Bachu 2008). During this period, the CO2 existing as a free phase in the reservoir pore space would not dissolve quickly in groundwater, which would cause CO2 transportation and pressure buildup in the reservoir and increase the risk of CO2 leakage (Yang et al. 2014, 2015; Meng et al. 2015).

Studying and modeling CO2 sequestration in geological formations need a clear understanding of multiphase flow characteristics and its behavior in porous media. Briefly, CO2 was injected into a formation at high flow rate through an injection well. The supercritical CO2 fluid flowed into the relatively more permeable regions surrounding the well and displaced native formation water under strong pressure gradients (e.g., brine) (Birkholzer et al. 2015). The CO2 injected at the bottom of the storage formation migrates upward rapidly by buoyancy forces because the density of the supercritical CO2 phase is lower than that of the aqueous phase. A small fraction of CO2 gas is trapped in the porous rock as a residual gas after injection, and most of the free CO2 gas accumulates below the caprock, forms a CO2 plume, and is transported far away (Zhang et al. 2009).The spreading and migration of mobile CO2 could increase the risk of CO2 leakage into shallower formations through fractures, outcrops, or abandoned wells (Lewicki et al. 2007; Lu et al. 2010). Moreover, the increased pressures caused by CO2 injection in the storage formations could induce geomechanical alteration of the reservoirs and their surroundings, for example, creating new fractures or reactivating larger faults (Rutqvist 2012). These changes taking place in the caprock or overburden could result in new leakage pathways for brine and/or CO2. If such leakage cannot be properly assessed, geological sequestration of CO2 might lead to undesirable environmental and safety consequences that may ultimately prevent future deployment of geological sequestration of CO2 (Jung et al. 2015). Therefore, flow barriers, such as faults, increase induced pressure considerably; for reservoirs with such features, careful site characterization and operational planning will be required for large storage projects (Chadwick et al. 2009).

Many researchers have significantly improved the knowledge base and addressed many of the technical gaps in CO2 storage. A large body of research has been devoted to identify and verify the main processes that control CO2 migration, trapping, and containment in deep saline aquifers. However, less attention has been paid to the effects of pressure buildup associated with CO2 injection (Birkholzer et al. 2015).

In this study, four sets of reservoirs suitable for CO2 storage in deep saline aquifers in Dongying Sag with different conditions were selected and analyzed. CO2 plume behavior and pressure buildup in different reservoirs, which are extremely important to CO2 storage safely, were analyzed using numerical simulation by TOUGHREACT. In this study, an attempt is made to provide an important basis for carrying out the actual CO2 injection project in the future.

Geological and hydrogeological setting

Geological background

The Dongying Sag is a sub-tectonic unit lying in the southeastern part of the Jiyang Depression of the Bohai Bay Basin (Cao et al. 2014). It is a Mesozoic–Cenozoic half graben-like basin that is bounded to the east by the Qingtuozi Bulge, to the south by the Luxi Uplift and Guangrao Bulge, to the west by the Gaoqing Fault, and to the north by the Chenjiazhuang–Binxian Bulge (Fig. 1). The Dongying Sag covers an area of 5700–5850 km2 with an east–west axis of 90 km and a north–south axis of 65 km (Cao et al. 2014; Wang et al. 2016). The tectonic evolution of the Dongying Sag can be subdivided into three stages: the rifting, faulting, and depression stages. During the rifting period, the Dongying Sag was affected by intensive structural uplift with soils developed, mainly in alluvial fans and deltas. In the faulting stage, sediment deposition was effected by tectonic activity, paleoclimatic changes, and sediment supply. With declining tectonic activity, lakes shrank gradually, forming monospecific facies dominated by fluvial facies in the depression stage (Li and Li 2016). The sag is filled with Cenozoic sediments, which are formations from the Paleogene, Neogene, and Quaternary periods. The formations from the Paleogene period are the Kongdian (Ek), Shahejie (Es), and Dongying (Ed); the formations from the Neogene period are the Guantao (Ng) and Minghuazhen (Nm); and the formation from the Quaternary period is the Pingyuan (Qp) (Guo et al. 2012; Zhang et al. 2004, 2010).

Fig. 1
figure 1

Location of the Dongying Depression (a Position and structure of Bohai Bay Basin, (I) Jiyang subbasin, (II) Huanghua subbasin, (III) Bozhong subbasin, (IV) Zhezhong subbasin, (V) Liaohe subbasin, (VI) Dongpu subbasin; b Structural units of Dongying Depression; c Cross section (A–A′) showing the various tectono-structural zones and key stratigraphic intervals within the Dongying Depression) (Modified from Guo et al. 2012)

Full size image

Hydrogeological background

Palaeohydrogeology and hydrogeochemistry of water-bearing rock series in Dongying Sag are divided into different periods through the combination of hydrogeological cycle with hydrogeological period. Atmospheric water infiltrates from the edge of the depression to the center forming a centripetal flow. The mudstone-compacted water migrates from the center of the depression to the surrounding forming a centrifugal flow. The groundwater drains to southern slope, northern fault zone, and the central fault zone (Yang 1985). Mudstone-compacted water is an important source of sedimentary groundwater and controls the hydrodynamic field formation and evolution of the Dongying Sag (Xie and Wang 1998). Lateral movement of the compacted water in the Tertiary Shahejie Formations is extremely slow, and such a low flow rate is essentially close to still water. The formation water salinity is generally 15,700–26,817 mg/L, and the hydrochemical types of groundwater in Shahejie Formations are mainly CaCl2 type, and secondly NaHCO3 type.

Numerical approaches

Numerical tool

Following Xu et al. (2006, 2011), we conducted the numerical simulations by TOUGHREACT, a program being applied to the study of non-isothermal multiphase reactive geochemical transport. TOUGHREACT was produced on the base of TOUGH2 V2—the multiphase fluid and heat flow code developed by Pruess et al. (1999, 2004), with the introduction of reactive geochemistry (Zhang et al. 2009; Xu et al. 2010). Please see TOUGHREACT User’s Guide for the specific information (Xu et al. 2012).

Model setup

Hydrogeological conceptual model

The test results of 154 core and water samples from an exploration well in the study area showed the Shahejie Formation consists of interbedded layers of sand rocks and mudstones, with a suitable thickness and good stability of each layer. Among the Shahejie Formation layers, Es2, the upper and middle Es3 layers with a burial depth of 1464–3102 m, there are four sets of reservoirs suitable for CO2 storage, shown in Fig. 2. The reservoirs are sand rocks, with appropriate depth, thickness, porosity and permeability. The groundwater in these aquifers is saline water, with a salinity of 15,700–26,817 mg/L. The specific parameters for the four reservoirs are summarized in Table 1. Above each reservoir, there is a layer of comparatively impermeable mudstone as caprock to prevent CO2 from leaking out of the reservoirs. The specific parameters for the four caprocks are summarized in Table 2. Besides, above the Shahejie Formation, there are some thin mudstones as buffer layers in the Minghuazhen and Guantao Formations to ensure CO2 storage in deep saline aquifers.

Fig. 2
figure 2

Stratigraphic column of the site for CO2 storage

Full size image
Table 1 Properties of the four reservoirs for CO2 geological storage
Full size table
Table 2 Properties of the four caprocks for CO2 geological storage
Full size table

Because of the pure lithology and stable deposition of the study area, the reservoirs were assumed to be horizontal homogeneous sandstone aquifers, the thicknesses of which were stable. Considering the vertical pressure during the development of strata, the reservoirs were assumed to be anisotropic, and the ratio of vertical to horizontal permeability is 0.1. Two layers of mudstones, respectively, overlie and underlie the sandstone reservoir. CO2 is injected into the reservoir from a well, and the CO2 outlet is located in the half bottom portion of the well. Therefore, a generalized hydrogeological conceptual model of taking these features into account is of the form shown in Fig. 3.

Fig. 3
figure 3

Schematic diagram for CO2 storage in deep saline aquifer

Full size image

Boundary and meshing

Based on the above-mentioned hydrogeological conceptual model, a simple two-dimensional (R–Z, 2D) radial model was used to simulate the CO2 plume migration and pressure buildup during and after injecting CO2 into the reservoirs. According to the area and the shape of Dongying Depression, the simulated range is generalized to a circle with a radius of 77,500 m. For each reservoir, 101 radial grid elements with spacing increasing gradually away from the injection well were used in the horizontal direction. Different numbers of vertical grid elements but with a constant spacing of 10 m were used in the vertical direction with reference to the specific thickness of the four reservoirs. The injection well is a flow boundary with a constant CO2 injection rate of 5 kg/s 10 m at half of the well, which is located on the Z-axis. Upper and lower mudstone layers are impermeable boundaries. Lateral is an infinite boundary, which was reflected in the model by giving the outer grid element a large volume of 1030 m3.

Simulation parameters

In the model, three types of parameters were input, including hydrogeological conditions, initial minerals, and water chemical composition of the reservoirs. Hydrogeological conditions of the four reservoirs were different due to their different depths, shown in Table 3. The values of thicknesses, porosity, and permeability are assigned with reference to the core test data. The values of temperature are assigned according to the average depth for the specific reservoir with a geothermal gradient of 3.6 °C/100 m and the average annual surface temperature of 14 °C (Yang 1984). The values of pressure are assigned according to the average depth for the specific reservoir with a usual pressure gradient of 1 MPa/100 m. Values of compressibility, residual water saturation, and residual gas saturation are assigned according to the properties of sandstone (Zhou et al. 2008; Birkholzer et al. 2009; Zhang et al. 2009; Shevalier et al. 2011; Xu et al. 2011; Beni et al. 2012). The initial mineral composition and water chemical composition of the four reservoirs were the same, because they were all sandstone reservoirs of the Shahejie Formation (As we do not discuss chemical reactions in this paper, the specific data are not shown here).

Table 3 Hydrogeological parameters used in the simulations
Full size table

Results and discussion

Though the hydrogeochemical reaction among CO2, rock, and brine was considered in the simulations, we discuss here mainly the CO2 plume migration and the pressure buildup by CO2 injection, which are vitally important to safe CO2 storage in deep saline reservoirs. And these two processes are mainly controlled by the physical properties of reservoirs including depth (which determines the reservoir temperature and pressure), thickness, porosity, and permeability. The simulation period up to 1000 years includes an injection period of 10 years and an after injection period of 990 years.

CO2 migration in different reservoirs

During the CO2 injection period, CO2 displaces the formation water in the pore volumes, and the displaced water flows into the surroundings. The injected CO2 moves upward by buoyancy and simultaneously moves laterally by injection pressure. Thus, a drainage zone was formed and spread out in the shape of a CO2 plume in the four reservoirs (Fig. 4). Especially, in the vicinity of the injection well, the saturation of the formation water was zero. A larger drainage zone of less porosity and permeability was formed in the reservoir; for example, after injecting CO2 for 10 years, the range of the dry out zone in the fourth reservoir was the largest (Fig. 4p). This is mainly because the rock pore volume of a reservoir with smaller permeability and porosity can be quickly filled with CO2. In other words, injecting the same volume of CO2, the distribution of the CO2 plume in a smaller porosity reservoir was wider.

Fig. 4
figure 4

Spatial distribution of CO2 gas saturation during the CO2 injection period for the four reservoirs

Full size image

The radial distance of CO2 migration at the bottom of the fourth reservoir (Fig. 4d, h, l, p) was obviously farther than that in the other three reservoirs. This is because CO2 injection can cause a rock pore pressure increase, the rock pore pressure increased more in reservoirs with a smaller porosity. Moreover, as the ratio of vertical permeability to horizontal permeability was set to 0.1, CO2 was easier to move in the horizontal direction in the reservoirs with a small porosity. Thus, when the reservoir pore pressure increases, radial distance of CO2 migration at the bottom in a reservoir with a smaller permeability was larger than that in a reservoir with a greater permeability.

In the early period of CO2 injection, in the reservoirs with similar porosities and permeabilities, the radial distance of CO2 migration in the upper part of a thin reservoir was greater than that in the upper part of a thick reservoir. In Fig. 4, it is seen that after injecting CO2 for 1 year, the radial distance of CO2 migration was about 200 m in the upper part of the first reservoir (Fig. 4a), while CO2 just migrated up to the top of the second reservoir (Fig. 4b). This is mainly because in the thin reservoir, CO2 was first vertically transported to the top of the reservoir and then horizontally migrated in the radial direction; however, in the thick reservoir, CO2 should migrate upward over a long distance and take more time to reach the top of reservoir.

In the CO2 injection period of 10 years, the CO2 plume front migrated approximately 1200, 1200, 1000, and 1000 m downward in the four reservoirs.

After injection, the buoyant CO2 will further spread upward and migrate laterally. As no new CO2 is injected into the reservoirs, the early injected CO2 gradually began to dissolve in the formation water (Fig. 5). In the first and second reservoirs, most of the injected CO2 dissolved after the CO2 injection period of 20 years (Fig. 5a, b). However, a fair portion of the CO2 still existed as a gas phase in the third and fourth reservoirs (Fig. 5c, d), especially in the fourth reservoir. This is because CO2 could migrate easily and have a large contact area with brine in reservoirs with a great porosity and permeability, which promoted the dissolution of CO2 in saline water, and CO2 could not dissolve quickly in the reservoirs with a small porosity and permeability. After a CO2 injection period of 1000 years, almost all the CO2 dissolved in the four reservoirs (Fig. 5m–p).

Fig. 5
figure 5

Spatial distribution of CO2 gas saturation after the CO2 injection period for the four reservoirs

Full size image

Pressure buildup in different reservoirs

Before CO2 injection, the pressure at the same depth is the same for each reservoir (Fig. 6a–d). Because of different depths and thicknesses of the four reservoirs, the pressure ranges and the contour legend levels in the figures are also different for the four reservoirs simulated.

Fig. 6
figure 6

Spatial distribution of the pressure buildup during the CO2 injection period

Full size image

When injecting CO2 into saline aquifers, the pressure fields for the four reservoirs changed. At the beginning, the pressure near the injection well increased immediately. With increased injection time, the pressure buildup zone was extended. Since the original reservoir is completely filled with groundwater, CO2 injection resulted in an instantaneous pressure increase near the injection well, and the influence zone spread far away from the location of the well. The pressure increased in magnitude close to injection well was greater than away from it (shown in Fig. 6e–t).

In the reservoirs with similar porosities and permeabilities, the radial range of pressure buildup in a thick reservoir was smaller than that in a thin reservoir. As shown in Fig. 6e, f, the thickness of the first reservoir is only half that of the second reservoir, and the zone of pressure buildup in the first reservoir was larger than that in the second reservoir. This is mainly because the CO2 injected into reservoir first moves upward by buoyancy. A thick reservoir has a high capacity to relieve the increased pressure in a vertical direction. A thin reservoir was limited by thickness, and CO2 should move more in a radial direction to relieve the increased pressure.

The pressure buildup by injecting CO2 was mostly affected by reservoir porosity and permeability. The pressure increased in magnitude, and the spread range was larger in a reservoir with small porosity and permeability than that in a reservoir with large porosity and permeability. As shown in Fig. 6h, l, p, t, the pressure increased in magnitude and the distribution range was larger in the fourth reservoir than the other three reservoirs. This is because groundwater flows slowly in a reservoir with a small porosity and permeability, and during CO2 injection, the pore pressure increased drastically and formed a large pressure gradient.

The distance of pressure propagation was nearly up to the boundary after a CO2 injection of 10 years in the four reservoirs (Fig. 6q–t), especially in the third and fourth reservoirs, greater than the distance of CO2 plume migration, which indicated that the range of pressure increase caused by CO2 injection was far away from the injection well and much wider than the range of CO2 storage location in the reservoir.

After CO2 injection, the pressure in the four reservoirs began to recover (Fig. 7). Pressure recovery can be explained in two ways: On the one hand, the increased pressure transferred far away (lateral was defined as an infinite boundary in the model) and on the other hand, a large amount of CO2 dissolved in the reservoir water. After a CO2 injection period of 1000 years, the pressures in the four reservoirs recovered to each one’s original levels.

Fig. 7
figure 7

Spatial distribution of the pressure dissipation after the CO2 injection period

Full size image

In order to further evaluate the safety of storage CO2 in the four reservoirs, the pressure changes in different locations in each reservoir and at different times during the CO2 injection period were analyzed. As shown in Fig. 8 and Table 4, pressure changes in different locations of the four reservoirs reflected the same pattern with injection time.

Fig. 8
figure 8

Changes in the pressure in different locations of the four reservoirs with injection time [where (21.32, − 5) represents the location 21.32 m away from injection well, and 5 m below the top of the reservoir, and so forth for others]

Full size image
Table 4 Pressure changes in different position in the four reservoirs
Full size table

For the four reservoirs, the pressure increment tends to decrease away from the injection well in a radial direction. This is because the thickness of the CO2 plume gradually decreases from the injection well to a location far away from the well. And the undissolved CO2 in the reservoir can greatly increase the pore pressure.

Close to the injection well, the pressure increment in the upper layer was smaller than that in the lower layer. At the location about 200 m away from the injection well, the pressure increment in the upper layer was greater than that in the lower layer. And at the location about 2000 m away from the injection well, the pressure increment in the upper layer was as large as that in the lower layer. This phenomenon can be explained by assuming that near to injection well the CO2 was continuously injected into the lower half of the well, and the CO2 cannot dissolve or move upward quickly, and this causes the pressure increment in lower layer to be greater than that in the upper layer; at the location not far away from the injection well, the CO2 gradually dissolved or moved upward, which leads to the pressure increment in upper layer being greater than that in the lower layer at the location about 200 m away from the injection well; and at the location far away from the injection well, there was no undissolved CO2 and the pressure increment was mainly caused by a lateral pressure transmission, therefore, the pressure increment in the upper layer was as great as that in the lower layer.

For the same location in different reservoirs, the pressure increment in the lower reservoir was larger than that in the upper reservoir. This is because the porosity and permeability of the lower reservoir were smaller than those of the upper reservoir. As the amount of CO2 injected into the unit thickness reservoir was equal, the CO2 injected into the lower reservoir with a smaller porosity and permeability could increase the pore pressure greatly.

Sensitivity analysis

In order to analyze the numerical model sensitivity to the parameters of reservoir porosity, permeability, and thickness, based on the parameters of the first reservoir, we, respectively, changed the value of porosity to 0.167, the value of permeability to 115.9 mD, and the value of reservoir thickness to 200 m.

The effect of different parameter values on the CO2 transport is represented in Fig. 9. Porosity and permeability have a significant impact on the CO2 migration. Compared with the CO2 plume distribution of original first reservoir (Fig. 9a, e, i, m), the CO2 moves farther in the upper reservoir when the value of porosity is reduced (Fig. 9b, f, j, n), and the CO2 moves farther in the lower reservoir when the value of permeability is reduced (Fig. 9c, g, k, o). The thickness of reservoir has little effect on the CO2 plume distribution (Fig. 9d, h, l, p).

Fig. 9
figure 9

Effect of reservoir porosity, permeability, and thickness on the CO2 transport (Where n*, k*, and m*, respectively, indicate the value of porosity is equal to 0.167, the value of permeability is equal to 115.9 mD, and the value of reservoir thickness is equal to 200 m)

Full size image

The effect of different parameter values on the reservoir pressure transport is represented in Fig. 10. Permeability has a great influence on reservoir pressure. Compared with the pressure of original first reservoir (Fig. 10a, e, i, m), the pressure increased much when the value of permeability is reduced (Fig. 10c, g, k, o). The porosity and the thickness of reservoir affect the reservoir pressure little (Fig. 9b, f, j, n, d, h, l, p).

Fig. 10
figure 10

Effect of reservoir porosity, permeability, and thickness on the reservoir pressure (The meanings of n*, k*, and m* are the same as Fig. 9.)

Full size image

Limitation discussion

The geological model was simplified to a circle in the plane according to the area and the shape of Dongying Depression, and the outer boundary of the model was given at the edge of Dongying Depression. It did not consider the influence of faults, since there was not enough information to confirm the location and conductivities of the faults. And this study mainly focuses on the effect of reservoir porosity, permeability, and thickness on the CO2 migration and reservoir pressure buildup.

Currently, operational data are not available to fully calibrate the numerical model though laboratory experiment. But the numerical software used in this study has been widely recognized in CO2 geological storage (Xu and Pruess 2001). We will compare the numerical model results to some basic analytical solutions or real field data to further prove that the simulation results are reasonable in the future research.

Conclusions

In the study area of Dongying Sag, there are four sets of reservoirs suitable for CO2 storage. The reservoirs are all made of sandrock, but at different depth, thickness, porosity, and permeability. During the CO2 injection period, a drainage zone was formed and spread out in the shape of a CO2 plume in the four reservoirs. The migration of CO2 plumes was mainly controlled by the porosity and permeability of the reservoirs. For the same injection time and an equal amount of CO2 injection per unit thickness, the radial distance of CO2 migration at the bottom of a reservoir of smaller porosity and permeability was wider than that in a reservoir of greater porosity and permeability. In the early period of CO2 injection, the radial distance of CO2 migration in the upper part of a thin reservoir was greater than that in the upper part of a thick reservoir. In the late period of CO2 injection, the radial distance of CO2 migration in the upper part of a thick reservoir was larger than that in the upper part of a thin reservoir. After CO2 injection period, CO2 gradually moves upward and is transported laterally along the upper parts of the reservoirs, and almost all the CO2 is dissolved in the reservoir water after 1000 years.

Injecting CO2 into saline aquifers caused an increase in the reservoir pressure. The pressure buildup by CO2 injection was mostly affected by reservoir porosity, permeability, and thickness. On the whole, the pressure increment in a reservoir of smaller porosity and permeability was greater than that in a reservoir of a greater porosity and permeability. On the other hand, the pressure increment could be decreased in a thick reservoir. The pressure increment close to the injection well was larger than that far away from the injection well. During the injection period, the pressure increment in different position of the reservoir changed with injection time and was affected by the porosity and permeability of the reservoir. Upon discontinuation of CO2 injection, the reservoir pressure is restored to the original pressure.

Reservoirs with large values of porosity, permeability, and thickness are more suitable for a CO2 geological storage with respect to CO2 transport and pressure buildup in the reservoirs. The monitoring wells for CO2 migration should be sited nearby the injection well (in the range of CO2 plume distribution) and monitored not only during the injection period but also after the injection; the monitoring wells for reservoir pressure should be sited in the whole study area and monitored just during the injection period. In the particular reservoir area like the study area, CO2 migration monitoring wells should be sited at about 20 and 2000 m away from the injection well and monitored for 1000 years as possible; reservoir pressure monitoring wells should be sited at about 20, 2000, 10000 m or far away from the injection well and monitored for 10 years.