Most oilfields in China are continental ones. These oilfields have reached an average water flooding recovery rate of about 33% in recent years, entering the two-high development stage (high water cut and high recovery ratio). Chemical flooding is usually used to further improve oil recovery of these oilfields^[1]. Based on the successful application of polymer flooding in Daqing Oilfield, NE China, alkaline-surfactant-polymer (ASP) flooding technology has been further developed^[2]. From 2014 to 2017, this technology was promoted, as a result, the blocks developed by ASP had an increase of oil recovery of over 18%. Support technologies of reservoir engineering, oil production engineering, and surface engineering for ASP also have been developed. Sun Longde et al.^[3] pointed out that the accurate characterization of remaining oil was the key to further improve chemical flooding recovery.

Currently, the methods used for oil displacement monitoring include the chemical tracer method^[4-5], surface seismic method^[6⇓-8], cross-well seismic method^[9], surface electromagnetic method^[10-11], cross-well electro-magnetic method, well-ground electromagnetic method, direct current method^[12], electrical logging method, and gravity logging method^[13]. As ASP flooding is mainly applied to the multi-layer and thin-layer reservoirs in Daqing oilfield, which have uneven distribution of remaining oil, fuzzy contrast in seismic attributes, and small resistivity differences, traditional oil displacement monitoring methods cannot reach high detection accuracy for these reservoirs^[8]. In addition, in the oilfield, various pipelines (including power cables) are densely distributed at 2 m depth, causing serious electromagnetic interference and even shielding signals, so traditional electromagnetic methods cannot get reliable data^[10⇓-12]. In order to improve the recovery rate of ASP flooding, we should work out an effective method to monitor the oil displacement process.

He Jishan^[14] proposed the differential wide-field electromagnetic method in 2010. In this study, we used this method (WFEM) to monitor ASP flooding in a block of Daqing Oilfield. All data measurement equipment, including the transmitting electrode, measuring electrode, and measuring cable, was buried underground to avoid interference. Before the ASP flooding, we measured the electromagnetic field and took it as the background field. In the ASP flooding, the electromagnetic field was measured several times. By differential of these obtained electromagnetic fields, we got the WFEM data, from which we got the 3D resistivity model from inversion. Using the logging resistivity data, we calibrated the 3D resistivity model of the test area and the target formation. Finally, the distribution range and edge of the ASP flooding system was extracted from the first-order differential of the 3D resistivity model in the spatial domain.

In oilfield development, water flooding, gas flooding and chemical flooding will cause changes in reservoir physical properties through displacement, correspondingly, reservoir geophysical responses will change over time too. Therefore, time-lapse geophysical detection methods have been proposed to monitor remaining oil^[7]. Traditional time-lapse electromagnetic method needs to acquire electric and magnetic field data, but as the test area has serious electromagnetic interference, the electromagnetic response differences of the reservoir before and after ASP flooding there can’t be acquired. The WFEM method only needs to acquire the electrical field, and it has the advantages such as strong signals, strong anti-interference ability, large detection depth, high efficiency, and low consumption^[14]. In this study, the differential WFEM method was used to monitor the reservoir resistivity variations during the ASP flooding.

Fig. 1 shows a universal electromagnetic method model. The source is on the ground. The coordinate system origin is at the center point of the source. The value of $z$ is positive downward and negative upward. The bottom layer is assumed to be infinitely thick. The relative permittivity and relative permeability of each layer are set as 1.

The wave equation of harmonic electromagnetic field is derived from Maxwell equations, from which the integral expression of the harmonic electromagnetic field generated by the horizontal electric dipole source on the surface of the layered earth model is obtained. For an electric dipole source along the x direction, the horizontal x component x of the electrical field is:

For an N-layer earth medium, ${{R}_{1}}$ and $R_{1}^{*}$ are the functions linking the electrical characteristics of the surface and that of the lower half space (the Earth), which are related to the conductivity of the N-th layer and the conductivity and layer thickness of all the layers above. The specific expression is as follows:

where

${{u}_{j}}=\sqrt{{{\lambda }^{2}}+k_{j}^{2}}$$k_{j}^{2}=-\text{i}\omega {{\mu }_{0}}{{\sigma }_{j}}$${{\sigma }_{j}}\text{=}\frac{1}{{{\rho }_{j}}}$

(j=1, 2, …, N)

After ASP flooding starts, electrical characteristics of the displaced formation will change. In this study, the observation system was fixed to ensure parameters of the system are constant, so the time-domain differential method can be used to accurately measure the changes in reservoir electrical property. We simulated inversely the electrical field or resistivity data collected at two different time, and processed the resistivity data volume by differential to work out the influence of the ASP flooding on the electrical field or resistivity at two different time, and infer the distribution range and front edge of the ASP flooding. As Eq. (3) shows, η_IJ is not related to current source strength, but is related to offset, observation angle, electrical parameters, layer thickness, and frequency.

The time-domain differential WFEM method has the following steps:

(1) Measure the background electrical field before the ASP flooding, and get the 3D resistivity data volume of the working area from inversion.

(2) Measure the electromagnetic field at different time during the ASP flooding, and get the 3D resistivity data from inversion under constraint of log data.

(3) Calculate the 3D resistivity difference between two different time by Eq. (3) and analyze the oil displacement efficacy of ASP flooding.

The WFEM method obtains geoelectric information at different depths by transmitting and receiving signals of different frequencies. The pseudo-random current signals sent at one time contain multiple main frequency components with similar amplitude. Therefore, this method can get the differential time-domain anomaly of 3D resistivity data by only measuring one electrical field component. Compared with the methods that measure the magnetic field, this method has stronger anti-interference ability, smaller calculation error, and higher reliability for resistivity anomaly measurement. Furthermore, through differential processing, the background noise and system error can be reduced.

Performing the spatial domain first-order differential to the obtained resistivity anomalies can effectively separate and highlight the superimposed anomalies in all directions. Then, in the data processing, we adopted the moduli of the vectors in all directions to analyze the resistivity changes, as shown in Eq. (4). The moduli of the vectors reflect the total gradient change of the resistivity at a certain point. By using the spatial domain first-order difference and differential moduli as parameters to monitor the resistivity anomaly in the target area, we can extract the resistivity changes in the target layer caused by ASP flooding effectively, overcoming the low measurement accuracy caused by low background resistivity, small differences and large target depth.

The spatial domain first-order differential of the zero line can reflect the abnormal boundary, so we use it to determine the distribution range and front edge of the ASP flooding in the target layer. By solving the first-order differential, we get the 3D distribution of the anomalous body to examine its scale, shape and trend, and infer the range and boundary of ASP flooding. M₁ of the spatial domain first-order differential was used to infer the front edge of the ASP flooding in this study.

The test area is located in the north of the central depression of the Songliao Basin, China. The target layer is the sub-layer of the Cretaceous Saertu oil reservoir S II sand group (S II-7 - S II-14). The overlying strata are Tertiary and Quaternary, and the underlying strata are Jurassic and Triassic. Due to differences in lithology, the Cretaceous is characterized by lower resistivity, while the overlying and underlying strata have higher resistivity. The “high-low-high” resistivity feature helps the identification and calibration of the target stratum. S II-7 - S II-14 is channel deposit of a large delta plain. On the top of S II-8 is a set of stable and high resistivity sandstone layer, which can be taken as the dividing line between S II-7 and S II-8 and electrical marker to calibrate the inversion results.

We tested the resistivity of 72 core plugs in 3 groups (24 small bedded sandstone core plugs from S II-7 - S II-14 in each group) by forcing current method. Water flooding, polymer flooding and ASP flooding were conducted on the 3 group of core plugs respectively, and the resistivity values of them before and after displacement were measured (Table 1). The test results show that injecting water, polymer and ASP flooding system can all make the resistivity of the target layer reduce. The resistivity after ASP flooding is much lower than that after polymer flooding or water flooding. Low resistivity can indicate the distribution and migration of the ASP flooding system.

We used the JSGY-2 WFEM system developed by Hunan Geosun High-tech Co. Ltd. for data acquisition (Fig. 2). The system has a high-performance computer to realize the real-time acquisition, operation, and processing of multi-channel parallel high-speed data. The transmitting system consists of a portable signal controller and a power cabinet (Fig. 2a). The system has separated strong current system and control system, to ensure the safety of operators and reduce the influence of the strong current system on the control system. The precision clock source and digital signal synthesis technology are used to ensure that the signal frequency is accurate, and the pseudo- random signal controller can transmit multiple frequencies at one time.

Following the principle of operation safety and easy construction, we arranged the power supply and selected the power transformer nearby. The transformer was 380V three-phase four-wire main power, and the power was 120 kV∙A. The power supply voltage was set at 130 V. The transmitting station consisted of a switching power supply (100 kW), an inverter cabinet (100 kW), and a signal controller. The offset was about 4 700 m. Due to the site condition, the distance between two grounding electrodes A and B was 890 m (Fig. 2b), and the maximum transmitting current was 70 A.

In the test area, there are a lot of metal and plastic pipelines at 1.5-2.5 m depth. To avoid their interference, we drilled holes with a depth of 5 m and a diameter of more than 3 cm by high-pressure water drills at 256 measurement stations, and buried the measuring electrodes inside. The distribution of survey grid is shown in Fig. 3. Because the electrode wires need to be buried for about 1 year, we covered the electrode wires with PVC sleeves to prevent oxidation and corrosion. One end of the wire was attached to the ground, and the other end was connected to the electrode. Both ends were sealed and waterproof. The top end of the PVC sleeve was slightly lower than the ground, and the measuring cable was pulled inside the horizontal PVC sleeve buried in a groove on the ground. We put some straws in the groove for water proofing and heat insulation. The distance between measured electrodes M and N was 20 m (Fig. 2b).

In the test, multiple receivers were used for simultaneous measurement at 40 frequencies in the range of 0.0156-8192.000 0 Hz. Based on the skin depth formula^[14], the detection depth was over 3 km, meeting the require-ments of this test. In order to test the efficacy of the buried electrodes, we used two channels of a receiver to collect the apparent resistivity of the electrodes at surface and buried at 5 m depth of one measuring station. The results are shown in Fig. 4. For the electrode at surface, the interference of shallowly buried pipelines is serious, so the apparent resistivity curve fluctuates severely. For the electrode buried at 5 m depth, the interference of shallow buried pipelines is negligible, so the apparent resistivity data is much better in quality (Fig. 4).

On November 20, 2015, one month after the fluid injection, we collected data for the first time. The data collected this time was taken as the baseline. And we collected data four times in total (Table 2).

We processed the data with the data processing module of the WFEM in the integrated system GME_3DI (V6.1), which is a system for 3D inversion and interpretation of gravity, magnetism, and electricity. We established the initial geoelectric model of the survey area using the logging data, regional geological information and formation electrical property data. The resistivity data collected by the WFEM were preprocessed to get the 1D model from inversion, on the basis of which, we obtained the 2D layered medium model from inversion. Next, constrained by the horizons and resistivity ranges from the logging data, we got the 3D model from inversion based on the 2D model.

The resistivity data volumes collected in the first and second times are shown in Fig. 5. According to the resis-tivity model, the test area has 9 electrical layers. The resistivity stratification in the area is relatively simple and in horizontal distribution. Vertically, high resistance layers and low resistance layers appear alternately. The background layer of S Ⅱ sand formation is located in the 9th electrical layer, which is a high resistivity layer. The depth of S Ⅱ-7 - S Ⅱ-14 is 980-1015 m.

Only the resistivity characteristics of the S II-7 layer are illustrated in this paper. The resistivity data of the S II-7 layer collected in the first and third times are shown in Fig. 6. The resistivity data acquired in the first time is used as the background data for the time domain differential of the subsequent acquisition. The target layer has high background resistivity, and its resistivity presents as a near-circular anomaly centering around the injection well, which could be caused by the ASP flooding. In the data acquired in the third time, the resistivity anomaly still exists, and further expanded in range. We can infer the distribution and migration of the ASP flooding system between the two times of acquisition by the differential in time domain and spatial domain.

The first-order tangential differential method was used to extract the information from the 3D resistivity data in the time domain and spatial domain. By using the data acquired in the first time as background data, the time-domain differential was performed to the data collected in the second, third and fourth times to get static maps for dynamic analysis. The residual gradient was used to infer the horizontal distribution characteristics and migration pattern of the ASP flooding system. In this study, the residual gradient was calculated by the difference between data acquired at any time and that at the first time.

Fig. 7 and Fig. 8 show the distribution of the ASP flooding system in S II-7 and S II-8a (the thickest sub-layers in S II-8), two thick oil layers, respectively. Half a month into ASP flooding (the second acquisition), the ASP flooding system in the S II-7 sublayer hadn’t migrated to any producing well. Six months into flooding (the third acquisition), the ASP flooding system had reached three production wells, N4-11-P3039, N4-D20- P3139, and N4-20-SP3039. The displacement fronts in the NE, NW and SW directions were all beyond the boundary of the test area, which may be due to the migration of the ASP flooding systems of other wells outside the monitoring area. The fourth time of data collection was only 14 days after the third time, and the distribution range of the ASP flooding system didn’t expand further significantly.

It can be seen from Fig. 8 that the ASP flooding system hadn’t reached any production well half a month into the flooding. Six months later, the ASP flooding system had reached two production wells, N4-D20-P3039 and N4- D20-P3139, in the east and west directions, respectively. The displacement front on the east side was beyond the boundary of the test area, which may be caused by the migration of the ASP flooding systems in other well groups. The interval between the fourth time and the third time of data collection was only 14 days, and the distribution range of the ASP flooding system expanded slightly.

The block had been flooded with the polymer before the ASP flooding. As Table 3 shows, on December 5, 2015 (the second collection), none of the four production wells produced alkali or surfactant. On May 15, 2016 (the third collection), the concentrations of alkali and surfactant in the four wells increased, indicating that the ASP flooding system had broken through. The production data are consistent with the wide-field resistivity model and the distribution model of the ASP flooding system in the target area. In addition, there was no obvious correlation between water content fluctuation and polymer concentration.

The ASP flooding system has low resistivity, while the reservoir has high resistivity. Therefore, the electromagnetic method can be used for oil displacement monitoring. To avoid the interference or shielding of useful signals, we buried the electrode 5 m deep and reduced the offset. By doing this, the quality of data acquisition, signal-to-noise ratio, and observation accuracy were improved. The collected data truly and effectively reflected the formation information. We inferred the distribution range and edge of the ASP flooding system, which can be used to judge the inter-well connectivity of the ASP flooding system and provide data for optimizing injection and production parameters.

At present, ground excitation and ground receiving mode can hardly realize thin-layer identification, so the migration features of the ASP flooding system in each oil layer can’t be analyzed yet. To improve the resolution, further researches will focus on improving excitation effect, sampling rate, and spatial data density: (1) Use wellbore power supply to excite the electromagnetic field caused by the ASP flooding system. (2) Adopt continuous monitoring mode to improve the data density in the time-domain to reflect migration features of the ASP flooding system dynamically. (3) Densify the excitation signal frequency to improve the longitudinal resolution. Reduce the distance between measuring points to 5 m or 10 m to improve the lateral resolution.

E—current, A;

E_x—the horizontal component of the electric field in the direction of the source, V/m;

h_j—the thickness of layerj (m);

i—imaginary unit;

${{\text{J}}_{0}}\left( \lambda r \right)$, ${{\text{J}}_{1}}\left( \lambda r \right)$—zero-order and first-order Bessel functions with variableλr;

j—the electrical layer number,j=1, 2, …, N;

k_j—the wave number of the j^th electrical layer;

M₁—the modulus of the first-order difference (Ω);

N—quantity of electrical layers;

P_E—electric dipole moment of the dipole source (A·m);

r—space between transmitter and receiver (m);

R₁ and R₁^*—functions relating the electrical characteristics of the surface and the lower half of space (the Earth);

x, y, z—the three directions of the axis;

${{\eta }_{IJ}}$—residual gradient, dimensionless;

θ—the angle between the direction of the dipole source and the midpoint of the source and the vector diameter of the receiving point (°);

λ—integral variable;

μ₀—the permeability of free space (F/m);

ρ—resistivity (Ω·m);

ρ_I, ρ_J—3D resistivity at the I-th and J-th times of acquisition after displacement (Ω·m);

ρ_j—the resistivity of the j-th electrical layer (Ω·m);

σ_j—the conductivity of the j-th electrical layer (S/m);

ϕ—measurement range in angle (°);

ω—angular frequency (rad/s).