1. Introduction
The development of oil deposits is conducted mainly by the injection of water to maintain reservoir pressure. If oil is produced without water injection, then there will be a decrease in the reservoir pressure. The decrease in the reservoir pressure causes decreases in the permeability and oil recovery [
1,
2]. Managing oil production from reservoirs to maximize the future economic return of an asset is an important issue in petroleum engineering [
3]. Water injection contributes to an increase in oil recovery but also leads to an increase in the well water cut. At high value of the layer-by-layer heterogeneity of the reservoir and oil viscosity, the probability of the uneven movement of water along the reservoirs and premature watering of production wells increases [
4,
5,
6].
In the presence of clays in the geological section of the oil reservoir during water injection, a decrease in the permeability is possible due to kaolinite particles blocking the pore channels. In this case, the permeability can decrease to 64% of the initial value [
7]. Owing to the incompatibility of the injected and formation water during flooding, salts can be deposited inside the pores of the reservoir rock to reduce the effective permeability, which ultimately affects the fluid flow and thus leads to a decrease in the overall productivity [
8]. In addition, a decrease in permeability can occur due to the detachment, migration, and capture of small particles in small-diameter pore channels [
9].
The higher the viscosity of the oil is, the stronger the dynamic heterogeneity will be as the displacement process proceeds and the greater the difference in the distribution of flow rates will be in different layers, which will lead to the formation and growth of viscous fingers. The rock wettability and pore-space structure play an important role in determining the dynamics of oil–water flow at the pore scale and affect the final water velocity during waterflooding [
10].
Taking into account the properties of the field as well as individual productive zones makes it possible to increase the efficiency of waterflooding [
11]. The main causes of the water flooding of wells include cone formation and movement of the water front through layers with high permeability [
12,
13,
14]. The effect of water injection into the reservoir depends on the permeability. The effect of water injection in low-permeability reservoirs is very different from the effect in medium- and high-permeability reservoirs [
15]. Therefore, when developing fields, it is important to study the patterns of waterflooding front movement in formations with low permeability [
16]. An anisotropic relative permeability not only influences the breakthrough time and waterflood coverage but also has a significant impact on the overall water cut [
17].
In oil fields, watering of production wells in the most permeable reservoirs is one of the main problems [
18]. To prevent premature watering and increase the production of an oil reservoir, injection of polymers and cyclic flooding are conducted, although gel-forming technologies have been noted for having the highest manufacturability [
19]. Therefore, forecasting of the water migration time is an urgent task.
The change in the water cut in wells is often studied based on the production of recoverable oil reserves of various viscosities [
20,
21,
22]. The known parameter dependencies allow the prediction of the water cut at various stages of reservoir development [
23,
24]. Oil recovery factors depend on the reservoir permeability, dynamic viscosity of the reservoir oil, production levels, and compensation of fluid production by water injection [
25]. However, the dependence of the water cut on the oil recovery is applicable for the field as a whole, but it poorly predicts the time at which water will completely fill the oil reservoir. The stimulation of oil production using software allows for the comprehensive assessment of the dynamics of oil and water production by wells and provides forecasts for the near future. Reservoir models are built based on various types of information, including laboratory data on fluids, the results of processing the geophysical characteristics, and well tests. If the reservoir area is poorly covered by wells, then the model will not be accurate. In this case, to predict the production indicators, it is possible to use the equations obtained by processing information from analogous fields.
Along with geological factors [
26], the water cut is affected by the development system and well operation modes [
27,
28,
29,
30]. With an increase in the water cut, the current oil production often decreases, and the deposition of salts in wells and gathering systems increases [
31,
32].
In this work, modeling in a hydrodynamic simulator of a section of an oil reservoir with injection and production wells was conducted to estimate the time of water migration between wells. The issue of rapid water movement is especially relevant for oil fields with high viscosities and permeabilities. The model predicted permeabilities of up to 500 × 10−3 µm2 and oil viscosities of up to 100 MPa·s. The dependencies of the water migration time on the reservoir permeability, bottom hole pressure, and oil viscosity were obtained.
Information was collected on fields where a reservoir pressure maintenance system was being created. The actual water migration time from the injection well to the production well was determined. A comparison was made between the actual values of the irrigation time and the model data. Real time is less.
Because the creation of a reservoir model that takes into account all possible natural factors is quite laborious and requires a large amount of information, the use of statistical regression models allows the time of water migration between wells to be quickly predicted with less labor and activity planning in a timely manner.
2. Simulation of Water Movement in Reservoir
Simulations of the movement of water in the reservoir will help to highlight the most significant factors that affect the time it takes for water to move from an injection well to a production well [
33]. To assess the estimated water migration time, a model of a reservoir area with production and injection wells was created in the Tempest MORE hydrodynamic simulator (Roxar). In the simulator, the flow of oil (o), water (w), and gas (g) in a porous medium obey Darcy’s law and the equations of continuity of the flow of a weakly compressible multiphase fluid and a Newtonian flow.
The continuity equations for water, oil, and gas include the formation permeability k. The permeability changes its value with changes in the reservoir pressure. The following calculation algorithm is proposed:
- -
the pressure in the reservoir and the saturation of the reservoir with oil, water, and gas are determined by solving the material balance equation;
- -
the effective pressure is calculated;
- -
the fluid flow between cells is determined.
The model consisted of three layers with a thicknesses of 1.5 m. The perforation intervals—the entire thickness of the layers. The distance between the wells was 500 m. The layers were homogeneous. The cell thickness in the model was 1.5 m, and the depth and width were both 25 m. The production and injection of water in the wells were controlled through the values of the bottomhole pressure (
Table 1). The calculated initial data is shown in
Table 1,
Table 2,
Table 3 and
Table 4.
During the simulation, the value of one of the parameters was changed step by step, and the remaining parameters were not changed. The time of water movement from the injection well to the production was measured. The speed and time of water movement at specified distances were estimated.
Figure 1 shows an image of the computational hydrodynamic model.
For the effective application of water cut reduction technology, it is necessary to predict the time of water movement through reservoir sections with different permeabilities. When modeling, calculations were performed for the movement of water from an injection well to a production well, depending on the properties of the reservoir oil, reservoir permeability, and bottomhole pressure. The position of the water front in the reservoir was noted by changing the water saturation of the model cells.
The dependence of the velocity of the displacement front on the permeability of the interlayer and the dynamic viscosity of the oil from the simulations is shown in
Figure 2. The values in
Figure 2 were obtained by determining the time of movement of water from an injection well over a distance of 400 m at various values of the reservoir permeability and oil viscosity. The average speed of the injected water at a permeability of 450 × 10
−3 µm
2 and a viscosity of 30 mPa·s was 57 m/year.
The average velocities of the movement of the water–oil front with the increase in the permeability increased significantly, while the travel time decreased. The speed of water movement increased as the mobility coefficient increased (
Figure 3). The duration of the movement of the displacement front at a certain distance was influenced by the mobility factor (the ratio of the permeability and dynamic viscosity of the reservoir oil) (
Figure 4).
The results of the model calculations showed that when water approached a production well, its velocity increased significantly because the filtration area decreased. The mobility factor (k/µo) had the greatest influence on the movement of water. At a distance of 100 m in the layers with a mobility factor of less than 2 μm2/Pa s, the timing of the advance of the displacement front from the injection well to the production well increased sharply. For mobility factors greater than 2 μm2/Pa·s, the time was less than 8–9 years.
3. Water Cut in Fields
Research of the dynamics of the water migration time of the Tournaisian reservoirs of the Nozhovskaya group of fields was conducted (
Figure 5). The fields are characterized by a high oil viscosity. The oil reservoirs contained up to 14 oil-saturated permeable layers with a thicknesses from 0.2 to 14.6 m and core permeabilities from 0.00036 to 0.774 µm
2. Intensive watering of production wells was observed, which was associated with a high dynamic viscosity of the reservoir oil as well as the presence of bottom water, oil–water zones, and heterogeneity of the reservoir in terms of the filtration properties. During the development of oil reserves from such reservoirs, an uneven movement of the displacement front occurred, and the likelihood of advanced watering of the production wells increased. The geological parameters of the deposits are shown in
Table 5.
The well was considered to be watered when the water cut was above 80%. The results of the analysis of the field information made it possible to assess the regularity of the water migration time based on the permeability, the distance between the production and injection wells, the bottomhole pressure, and the oil viscosity. Information on the predicted water cut allowed for prompt planning of measures to regulate the oil displacement front by aligning the injectivity and production profiles in the injection and production wells.
A database on the history of the oil and water production by the wells was created (
Table 6). The water migration time was determined according to the history of the production wells. The data were compared in terms of the geological parameters of the reservoir and the distances between the wells. The permeability in the well drainage zone was obtained from the results of well tests.
The results in the table make it possible to find patterns in the water migration time variations with the permeability, the distance between the production and injection wells, the pressure at the bottom of the wells, and the viscosity of the reservoir oil. Information about the predicted water migration time allows the planning of measures to block the reservoirs.
The water migration time was significantly reduced with an increase in the permeability (
Figure 6) and mobility factor (
Figure 7). If the permeability was greater than 0.25 μm
2, the average period of watering did not exceed 6 years. When comparing
Figure 6b,c, it can be noted that with an increase in the difference in the bottomhole pressures between the injection and production wells, the water migration time decreased.
The dependence of the water migration time on the mobility factor was plotted. The data dependencies for real wells were similar in trend to the theoretical dependencies built using the hydrodynamic model (
Figure 8). However, according to the model, the water migration time from the injection well to the production well turned out to be longer. This was because in the model, the pore space was homogeneous, while in the real oil reservoir, the distribution had a complex structure and communication of voids occurred. This conclusion is consistent with previous results [
34], where using an example of sandstone, large differences in the oil production were revealed depending on small changes in the structure of the pore spaces of the rock. It can be concluded that having an accurate understanding of the pore structure in waterflooding is necessary to develop strategies aimed at maximizing the oil recovery.
The average water migration time with a mobility factor of more than 2 μm2/(Pa·s) was 6.3 years (with a distance between the production and injection wells of 400 m). These results may indicate a longer period for the watering of production wells when predicting using geological and hydrodynamic models.
For the mathematical processing of the data in
Table 6, the classical multiple linear regression model was used. It was assumed that the relationship between the parameters could be described by a linear relationship of the form:
where Y is the dependent variable, u is a random component of the model, x
i represents the independent variables, and β
i represents the regression coefficient.
The relationship between the dependent variable T and independent variables k, µ, L, P
inj, and P
bot were determined via regression, where the least squares method was used in the regression analysis. The following equation was obtained to estimate the water migration time (years) at a distance L from the injection well:
where L is the distance from the injection to the production well (m), k is the permeability (μm
2), µ is the dynamic viscosity of the reservoir oil (Pa·s), P
bot is the bottomhole pressure in the production well (MPa), and P
inj is the bottomhole pressure in the injection well (MPa). For Equation (2), the following limits for the initial parameters should be used:
A comparison of the predicted (according to Equation (2)) and actual values is shown in
Figure 9. Good convergence of the predicted and actual values of the water migration time confirmed the possibility of forecasting using Equation (2).
When considering a water migration time of up to 10 years, the predicted and actual values of the water migration time agreed closely. When comparing data from wells that watered after more than 10 years, the discrepancy between the actual and predicted values increased. Wells that flooded 10 years after the start of waterflooding were operating in low permeability formations, so a high accuracy of the water migration time prediction was not required. Equation (2) makes it possible to predict the water migration time within the limits described above. The issue of the water migration time for interlayers with high permeabilities (greater than 0.2 µm2) is especially important. In most of the explored fields, rapid flooding of the wells was observed. As shown by the analysis of the results to identify sources of watering, along with the high viscosity of the reservoir oil, the reasons for the more intensive watering were the advanced penetration of water through the more permeable part (separated layers) of the layered heterogeneous reservoir (56.3%). The fields were developed with low efficiency due to their complex geological and physical structures, i.e., the separation of productive layers into a large number of thin interlayers with a high viscosity of the reservoir oil. Under these conditions, the development of the oil reserves occurred by displacing the oil through the most permeable interlayers with water.
The timely application of polymer or gel compositions makes it possible to increase the oil recovery in oil fields by at least 6% [
35,
36]. The main goal in conducting such work in wells is to reduce the filtration parameters in the washed part of the formation to reduce the water cut of the product. Under field conditions, various compositions are used to form insulating screens in flooded interlayers, which can significantly reduce the permeability of rock. It is planned to use the proposed equation to identify wells with short watering periods for modeling measures using polymer or gel compositions.