Horizontal well fluid holdup inversion interpretation method and system based on complex impedance method

Abstract

The invention provides a horizontal well fluid holdup inversion interpretation method and system based on a complex impedance method, and relates to the field of fluid holdup inversion, and the method comprises the steps: obtaining complex impedance frequency spectrums of oil-water stratified flow under different holdups, and extracting characteristic parameters reflecting the water holdup; constructing a data set by using the simulation data, and performing fitting research by using an RBF neural network to obtain a relationship between the characteristic parameters and the water holdup; measuring the oil-water two-phase flow by using a fluid holdup measurement system developed based on a complex impedance method to obtain current data of the oil-water two-phase flow; and processing the acquired current data and extracting characteristic values, and performing inversion by using the trained RBF neural network to obtain the fluid holdup. The method has the advantages that the training set is established through simulation to train the neural network model, the oil-water two-phase flow holdup is measured through the fluid holdup measuring system developed on the basis of the complex impedance method, and reliable inversion explanation is provided for measuring the horizontal well fluid holdup through the complex impedance method.

Description

Technical Field

[0001] This application relates to the field of fluid sustainment inversion, and in particular to a method and system for interpreting fluid sustainment inversion in horizontal wells based on the complex impedance method. Background Technology

[0002] Fluids in horizontal wells are typically two-phase (oil-water) or three-phase (oil-gas-water) flows. Furthermore, due to physical effects such as gravity, the multiphase flow distribution characteristics in horizontal wells are more complex than in vertical wells. To gain a deeper understanding of their flow patterns and establish reliable interpretation models for predicting well flow states, optimizing production equipment design, and controlling the oil extraction process, accurate measurement of multiphase flow parameters is essential. Parameters for evaluating fluid flow in horizontal wells include flow pattern, flow rate, flow velocity, and the holdup of each phase. Among these, the holdup of each phase is particularly critical. In horizontal well development, accurate measurement of the holdup of each phase plays a crucial role in comprehensively understanding the downhole fluid distribution. By accurately obtaining the holdup information of each phase, a deeper understanding of the fluid flow patterns within horizontal wells can be achieved, providing an important basis for developing scientific production plans and significantly contributing to efficient oilfield development and enhanced oil recovery.

[0003] With technological advancements, complex impedance testing technology has been widely applied and continuously developed in various industrial fields, playing an increasingly important role, particularly in electroplating, thin film preparation, materials processing, and petroleum development. Compared to other fields, its application in the petroleum industry has lagged behind. However, due to its advantage of rapid and non-destructive testing of the electrical properties of petroleum fluids, complex impedance testing technology is increasingly valuable in all aspects of petroleum exploration, extraction, and processing, becoming an indispensable testing method in modern petroleum production. Therefore, this paper proposes a horizontal well fluid holdup measurement inversion method based on complex impedance, which is of great value for accurately obtaining horizontal well profile information and is significant for improving the productivity of oilfields and shale oil and gas wells and extending the production cycle of oil and gas wells. Summary of the Invention

[0004] The purpose of this invention is to solve the inversion technology problem in measuring the holdup of oil-water two-phase flow using the complex impedance method, and to provide a method and system for inverting and interpreting the fluid holdup of horizontal wells based on the complex impedance method.

[0005] The above-mentioned objective of this application is achieved through the following technical solution: S1: Obtain the complex impedance spectrum of oil-water stratified flow under different water holdup rates and extract characteristic parameters reflecting the water holdup rate; S2: Train the RBF neural network using the dataset constructed from the feature parameters; S3: Measure the oil-water two-phase flow to obtain the current data of the oil-water two-phase flow; S4: Process the current data and extract feature values, then use the trained RBF neural network to invert the feature values ​​to obtain the fluid holdup.

[0006] Optionally, step S1 includes: The complex impedance spectrum of oil-water stratified flow under different holdup rates in a horizontal well was obtained using the finite element simulation software COMSOL. By using the real and imaginary parts of the complex impedance of the pure water phase as a reference quantity, and combining it with the complex impedance spectrum, the real and imaginary parts of the complex impedance of other water holding capacities are compared to obtain dimensionless characteristic parameters that can reflect water holding capacity information. Pure water phase means water holding capacity of 100%.

[0007] Optionally, the step of using the real and imaginary parts of the complex impedance of the pure aqueous phase as a reference quantity, combining it with the complex impedance spectrum, and comparing it with the real and imaginary parts of the complex impedance of other water holding capacities to obtain dimensionless characteristic parameters that can reflect water holding capacity information specifically includes: If we consider the oil-water two-phase flow as being formed by the gradual addition of oil phase to the water phase, then the change in the complex impedance of the water phase is due to the oil phase altering the electrical properties of the water phase. The ratio of the real part to the imaginary part slope of oil-water two-phase flow with different water holdup ratios and 100% water holdup is defined as follows:

[0008] in, The ratio of the real parts. This is the ratio of the slopes of the imaginary parts. For water holding capacity The real part of the time, For water holding capacity The slope of the imaginary part of time, This is the actual value when the water holding capacity is 100%. This is the real value when the water holding capacity is 100%. Received and As a dimensionless feature parameter that reflects holding rate information.

[0009] Optionally, the RBF neural network includes an input layer, a hidden layer, and an output layer.

[0010] Optionally, step S3 includes: Generate a The excitation signal of the pseudo-random sequence; High-precision current sensors and excitation signals are used to collect current data between electrodes.

[0011] Optionally, step S4 includes: The acquired current signal is subjected to frequency domain analysis to obtain its corresponding real and imaginary part spectrum. Feature values ​​are extracted from the spectrograms of the real and imaginary parts; The fluid holdup is obtained by inverting the eigenvalues ​​using an RBF neural network.

[0012] A horizontal well fluid sustainment inversion and interpretation system based on complex impedance method, the system comprising: an acquisition module and a processing module; The acquisition module is used to measure the oil-water two-phase flow and obtain the current data of the oil-water two-phase flow; The processing module is used to simulate and obtain the complex impedance spectrum of oil-water stratified flow under different holdup rates; The processing module is also used to extract characteristic parameters reflecting water holding capacity from the complex impedance spectrum; The processing module is also used to train the RBF neural network using a dataset constructed from feature parameters; The processing module is also used to process the current data and extract feature values, and to use the trained RBF neural network to invert the feature values ​​to obtain the fluid holdup.

[0013] Optionally, the processing module includes: a sensor array, a transmitter unit, a receiver unit, and a host computer; The transmitter unit is used to generate a The excitation signal of the pseudo-random sequence; The receiver unit uses a high-precision current sensor and an excitation signal to collect current data between the electrodes.

[0014] The beneficial effects of the technical solution provided in this application are: This invention leverages the advantages of the complex impedance method, using a horizontal well fluid holdup measurement device developed based on the complex impedance method to collect oil-water two-phase flow data within a pipeline. Then, it utilizes the aforementioned inversion interpretation method to obtain the holdup of each phase of the oil-water two-phase flow within the pipeline. The horizontal well fluid holdup inversion interpretation method based on the complex impedance method first uses the finite element simulation software COMSOL to obtain the complex impedance spectrum of the oil-water stratified flow under different holdup rates and extracts characteristic parameters reflecting the water holdup. Next, it constructs a dataset using simulation data and uses an RBF neural network for fitting studies to obtain the relationship between the characteristic parameters and the water holdup. Then, it uses the fluid holdup measurement system based on the complex impedance method to measure the oil-water two-phase flow, obtains the current data of the oil-water two-phase flow, processes the acquired current data to extract feature values, and uses the trained RBF neural network for inversion to obtain the fluid holdup. This invention establishes a training set through COMSOL simulation to train the neural network model and combines it with the fluid holdup measurement system based on the complex impedance method to realize the measurement of oil-water two-phase flow holdup, providing a reliable inversion interpretation for measuring horizontal well fluid holdup using the complex impedance method. Attached Figure Description

[0015] The present application will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings: Figure 1 This is a block diagram of the overall structure of the pipe three-phase flow holdup measurement system based on the complex resistivity method in this embodiment of the invention. Figure 2 These are the spectrum diagrams of the real part of the complex impedance at different water holding capacities in embodiments of the present invention; Figure 3 These are the spectrum diagrams of the imaginary part of the complex impedance at different water holding capacities in embodiments of the present invention; Figure 4 The water holding capacity and water holding rate in the embodiments of the present invention and Scatter plot; Figure 5 This is the RBF neural network structure in the embodiments of the present invention; Figure 6 This is a regression graph of the neural network training results in an embodiment of the present invention; Figure 7 This is a comparison chart of the RBF neural network fitting results in the embodiments of the present invention; Figure 8 This is a structural diagram of the fluid holdup measurement system based on the complex impedance method in an embodiment of the present invention; Figure 9 This is a time-domain waveform diagram of the water holding capacity measurement result at 90% in an embodiment of the present invention; Figure 10 This is a spectrum of the real and imaginary parts of the water holding capacity measurement results at 90% in this embodiment of the invention; Figure 11 This is a spectrum of the real and imaginary parts of the complex impedance measured under different water holding capacities in an embodiment of the present invention. Detailed Implementation

[0016] To provide a clearer understanding of the technical features, objectives, and effects of this application, the specific embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0017] The embodiments of this application provide a method for inverting and interpreting fluid sustainment in horizontal wells based on the complex impedance method.

[0018] Please refer to Figure 1 , Figure 1 This is a general structural block diagram of a horizontal well fluid holdup inversion interpretation method based on the complex impedance method in this application embodiment, including: S1: Obtain the complex impedance spectrum of oil-water stratified flow under different water holdup rates and extract characteristic parameters reflecting the water holdup rate; S2: Train the RBF neural network using the dataset constructed from the feature parameters; In one embodiment of this application, the feature parameters and the corresponding water holding rate data constitute a dataset, and the relationship between the feature parameters and the water holding rate is obtained by fitting the constructed RBF neural network.

[0019] S3: Measure the oil-water two-phase flow to obtain the current data of the oil-water two-phase flow; S4: Process the current data and extract feature values, then use the trained RBF neural network to invert the feature values ​​to obtain the fluid holdup.

[0020] Step S1 includes: The complex impedance spectrum of oil-water stratified flow under different holdup rates in a horizontal well was obtained using the finite element simulation software COMSOL. By using the real and imaginary parts of the complex impedance of the pure water phase as a reference quantity, and combining it with the complex impedance spectrum, the real and imaginary parts of the complex impedance of other water holding capacities are compared to obtain dimensionless characteristic parameters that can reflect water holding capacity information. Pure water phase means water holding capacity of 100%.

[0021] As one example, the first step is to use COMSOL simulation to obtain the complex impedance spectrum of oil-water stratified flow under different holdup rates. In real-world environments, the varying flow velocities within horizontal wells result in different mixing states for bubbly and mixed stratified flows. During COMSOL forward modeling, these different mixing states cause variations in the complex impedance spectrum, making the simulation data less reliable. In contrast, stratified flows are less affected by flow velocity, exhibit more stable properties, and their simulation data is more reliable.

[0022] During the simulation, a simulation model with a water holding capacity of 10% to 100% was selected, and 91 sets of complex impedance data were obtained at 1% water holding capacity intervals. A portion of the complex impedance spectrum data was selected for comparative analysis. The spectrum diagrams of the real and imaginary parts of the complex impedance under different water holding capacities are shown below. Figure 2 , Figure 3 As shown.

[0023] The specific steps of using the real and imaginary parts of the complex impedance of the pure aqueous phase as a reference quantity, combining it with the complex impedance spectrum, and comparing it with the real and imaginary parts of the complex impedance of other water holding capacities to obtain dimensionless characteristic parameters that reflect water holding capacity information include: If we consider the oil-water two-phase flow as being formed by the gradual addition of oil phase to the water phase, then the change in the complex impedance of the water phase is due to the oil phase altering the electrical properties of the water phase. The ratio of the real part to the imaginary part slope of oil-water two-phase flow with different water holdup ratios and 100% water holdup is defined as follows:

[0024] in, The ratio of the real parts. This is the ratio of the slopes of the imaginary parts. For water holding capacity The real part of the time, For water holding capacity The slope of the imaginary part of time, This is the actual value when the water holding capacity is 100%. This is the real value when the water holding capacity is 100%. Received and As a dimensionless feature parameter that reflects holding rate information.

[0025] As one example, the varying salinity of the water phase in different horizontal wells affects its electrical properties, leading to discrepancies with theoretical data obtained through simulation. Therefore, the real and imaginary slopes cannot be directly used as characteristic parameters. If the oil-water two-phase flow is considered as being formed by the gradual addition of oil phase to the water phase, then the change in the complex impedance of the water phase is due to the oil phase altering the electrical properties of the water phase. Therefore, by using the real and imaginary parts of the complex impedance of the pure water phase (i.e., 100% water holdup) as a reference and comparing them with the real and imaginary parts of the complex impedance of other water holdup ratios, dimensionless parameters can be obtained. This eliminates the influence of varying water phase salinity in different horizontal wells on the complex impedance parameters.

[0026] The numerical values ​​of the complex impedance spectrum are converted using the formula of the ratio of the real part to the ratio of the slope of the imaginary part. and Scatter plot of water holding capacity, and Scatter plot of water holding capacity as shown Figure 4 As shown in the figure. and There is a certain non-linear relationship between it and the water holding capacity. and As a characteristic parameter reflecting water holding capacity.

[0027] As one embodiment, the obtained and As a dimensionless characteristic parameter that reflects yield information, it can eliminate the influence of different water phase salinity in different horizontal wells on the complex impedance parameters.

[0028] Step S2 includes: The RBF neural network includes: an input layer, a hidden layer, and an output layer.

[0029] As one example, an RBF neural network was used to fit the eigenvalues ​​obtained from S1 with the water holding capacity, revealing the relationship between the eigenvalues ​​and the water holding capacity. The RBF neural network utilizes the radial basis functions of the hidden layers to achieve nonlinear transformations of the input data, exhibiting a powerful ability to approximate complex nonlinear relationships. This unique network structure demonstrates significant advantages in handling highly nonlinear systems, providing an effective solution for modeling and predicting complex problems. Furthermore, the training process of the RBF neural network is completed in a single step, eliminating the need for multiple iterations of optimization. This not only significantly improves training efficiency but also avoids overfitting problems that easily occur in complex situations. More importantly, the RBF neural network has a simple structure and few parameters, mainly including the expansion constant and the maximum number of neurons. This concise structure makes it easier to tune and optimize parameters during fitting, while also reducing the complexity of model development. By gradually increasing the number of neurons in the hidden layers, the RBF neural network can dynamically adjust the model complexity, thereby achieving a good balance between fitting accuracy and generalization performance. In summary, the RBF neural network was further used to study the nonlinear characteristics between the eigenvalues ​​and the water holding capacity.

[0030] RBF neural networks are generally three-layer structures, such as Figure 5 As shown. The network structure includes One input node, Hidden layer nodes and There are 1 output node. Among them, This is the input vector of the neural network. To output the weight matrix, The threshold of the output neuron. This is the output of the neural network. This represents the linear activation function of the output neuron. A core feature of the RBF neural network is that its hidden layers employ basis functions based on distance metrics, while using radially symmetric functions as activation functions. This unique structure makes it excellent at handling nonlinear problems.

[0031] The first layer, or input layer, consists of signal source nodes; the second layer is the hidden layer, the number of hidden units being determined by the needs of the problem being described. The transformation function of the hidden units is a Gaussian function, i.e.:

[0032] in, Indicates the The output values ​​of each hidden layer node; Represents the input sample vector; Indicates the The center point of the Gaussian function of each hidden layer node, and the dimension is consistent with the input vector; Indicates the The width of the basis functions of each hidden layer node; This represents the total number of hidden layer nodes, which exhibit radial symmetry and decay.

[0033] The third layer of the network is the output layer, whose main function is to respond to the input pattern. It uses a linear activation function, where the first... The output of each neuron is represented as:

[0034] in, For the weights of the output layer neural network, This is the network threshold.

[0035] The essence of model training lies in establishing a mapping relationship between input features and target output. This invention uses random sampling to divide the dataset into three parts: 64 groups (70%) of samples are used as the training set, to iteratively optimize network weights and thresholds; 13 groups (15%) of samples serve as the validation set, used to determine optimal network structure parameters, such as the number of hidden layer neurons; and the remaining 14 groups (15%) of samples constitute the test set, used to comprehensively evaluate model performance. The key parameters for network training are set as follows: expansion constant 1, maximum number of neurons 20.

[0036] After training, regression analysis was performed on the training set, validation set, test set, and the complete dataset. The analysis results are as follows: Figure 6 As shown in the figure. It can be seen from the graph that the predicted and actual values ​​are basically the same, and the coefficient of determination... All values ​​are greater than 0.99, further demonstrating that the model can predict water holding capacity very well.

[0037] The test set data is input into the neural network model for training. The predicted values ​​output after training are compared with the true values. The comparison results are as follows: Figure 7 As shown in the figure. "o" in the figure represents the actual water holding capacity. "" represents the predicted water holding capacity obtained after training the neural network. Analysis results show that the model's predicted values ​​are highly consistent with the actual observed values. The following performance indicators were obtained through quantitative calculation: the corrected sum of squared errors (SSE) is 0.21811; the root mean square error (RMSE) is 0.12953; and the coefficient of determination (...). The accuracy reached 0.99997. These data fully demonstrate that the neural network model can accurately predict the water holdup distribution of oil-water two-phase flow based on the input feature parameters.

[0038] To verify the fitting effect of the RBF neural network, polynomial fitting was further introduced for result comparison and analysis. The analysis results are shown in Table 1.

[0039] Table 1 Comparison of fitting errors between the two methods

[0040] As shown in the table, the RBF neural network exhibits a significant advantage in fitting performance: its sum of squared errors (SSE) is reduced by 93.7% compared to polynomial fitting, indicating a substantial improvement in prediction accuracy; the coefficient of determination (... The RBF neural network's value is closer to the ideal value of 1, reflecting a better data fit; the root mean square error (RMSE) is also significantly smaller than that of the polynomial fit. These quantitative analysis results fully demonstrate that the RBF neural network has superior fitting performance and can more accurately characterize the complex nonlinear relationship between characteristic parameters and water holdup. Especially when dealing with high-dimensional, nonlinear data, the RBF neural network exhibits stronger modeling capabilities and prediction accuracy, providing a reliable technical means for the accurate prediction of water holdup in oil-water two-phase flow.

[0041] Step S3 includes: Generate a The excitation signal of the pseudo-random sequence; High-precision current sensors and excitation signals are used to collect current data between electrodes.

[0042] As one embodiment, a fluid holdup measurement system based on the complex impedance method is used to measure the current data of an oil-water two-phase flow. This measurement system is as follows: Figure 8 As shown.

[0043] The transmitter system's function is to provide a field source excitation signal to generate a... A pseudo-random sequence signal is used. This signal allows for multi-frequency measurements with a single power supply, effectively solving the inefficiency problem caused by the need for multiple measurements with traditional single-frequency signals. The sensor array consists of multiple pairs of electrode plates, responsible for sending the excitation signal generated by the transmitter system into the horizontal well and receiving the current signal, which is then transmitted to the receiver system for signal acquisition. The receiver system uses a high-precision current sensor to convert the acquired current signal into a voltage signal and processes it. This part is responsible for filtering the acquired signal, converting single-ended to differential signals, etc., ensuring proper circuit connection and improving the measurement system's anti-interference capability. The main control unit on the receiver is responsible for controlling other modules and data communication. Finally, the receiver system transmits the acquired data to the host computer via the communication module.

[0044] In the process of measuring oil-water two-phase flow using a fluid holdup measurement system based on the complex impedance method, the transmitter system is responsible for generating an excitation signal of a specific frequency. The receiver system uses a high-precision current sensor to acquire the current response signal between the electrodes and transmits the acquired data to the ground via a communication circuit.

[0045] Step S4 includes: The acquired current signal is subjected to frequency domain analysis to obtain its corresponding real and imaginary part spectrum. Feature values ​​are extracted from the spectrograms of the real and imaginary parts; The fluid holdup is obtained by inverting the eigenvalues ​​using an RBF neural network.

[0046] As one embodiment, the obtained current data is processed to obtain feature values, and the fluid holdup is obtained by inversion using the RBF neural network trained by S2. Taking the experimental data with a water holdup of 90% as an example, the time-domain waveform of the measurement result with a water holdup of 90% is as follows: Figure 9 As shown.

[0047] Frequency domain analysis was performed on the time-domain signal with a water holding capacity of 90% to obtain its corresponding real and imaginary part spectrum. The real and imaginary part spectrum of the 90% water holding capacity measurement result is shown below. Figure 10 As shown.

[0048] Since the excitation signal is a 27-sequence pseudo-random voltage signal with a fundamental frequency of 8Hz, the real and imaginary amplitude values ​​of the current signal at frequencies of 8Hz, 16Hz, 32Hz, 64Hz, 128Hz, 256Hz, and 512Hz can be obtained from the figure. Therefore, each frequency point of the acquired current signal can be represented in complex form:

[0049] Where I is the acquired current signal, This represents the real part amplitude of the current signal obtained from the spectrum. The value represents the imaginary part of the current signal obtained from the spectrum.

[0050] Since the frequency and amplitude of the emitted excitation signal are known, the voltage signal at each frequency point of the excitation signal can also be expressed as:

[0051] Where U is the excitation voltage signal. This represents the real part amplitude of the voltage signal obtained from the spectrum. This represents the imaginary amplitude of the voltage signal obtained from the spectrum.

[0052] Therefore, the complex impedance at each frequency can be obtained:

[0053] Where Z is the calculated complex impedance. To calculate the real part magnitude of the complex impedance, This is the magnitude of the imaginary part of the calculated complex impedance.

[0054] By plotting the calculated complex impedances at various frequencies, the real and imaginary part spectrum of the complex impedance can be obtained. Following the above method, all experimentally measured data are processed to calculate the corresponding real and imaginary part values ​​of the complex impedance and plot them. The real and imaginary part spectrums of oil-water two-phase flows at different holds are shown below. Figure 11 As shown.

[0055] The corresponding feature parameters are extracted from the measurement results according to the method of extracting feature parameters in S1, and the measurement results are obtained by fitting the trained RBF neural network. The feature parameters, measurement results and measurement error results are shown in Table 2.

[0056] Table 2 Characteristic parameters, measurement results and measurement errors

[0057] As shown in Table 2, the measurement results contain a certain error, possibly due to the influence of the double electron layer and parasitic capacitance. Secondly, the oil and water phases have surface tension; during measurement, some oil and water phases adhere to the inner wall of the sensor, causing the oil-water interface to be a curved line rather than a straight line. This results in fluctuations between the actual and theoretical oil-water holding capacities within the sensor's measurement range. Furthermore, the simulated wellbore could not be perfectly horizontal during the experiment, resulting in a certain inclination angle. This caused discrepancies between the actual and theoretical holding capacities at the sensor array location, further affecting the accuracy of the measurement results. Table 2 shows that the maximum measurement error for water holding capacity is 12%, meeting the corresponding oil-water identification accuracy requirements. This verifies the feasibility of this fluid holding capacity measurement inversion method in horizontal well fluid holding capacity measurement based on the complex impedance method.

[0058] A horizontal well fluid sustainment inversion and interpretation system based on complex impedance method, the system includes: an acquisition module and a processing module; The acquisition module is used to measure the oil-water two-phase flow and obtain the current data of the oil-water two-phase flow; The processing module is used to simulate and obtain the complex impedance spectrum of oil-water stratified flow under different holdup rates; The processing module is also used to extract characteristic parameters reflecting water holding capacity from the complex impedance spectrum; The processing module is also used to train the RBF neural network using a dataset constructed from feature parameters; The processing module is also used to process the current data and extract feature values, and to use the trained RBF neural network to invert the feature values ​​to obtain the fluid holdup.

[0059] The processing module includes: a sensor array, a transmitter unit, a receiver unit, and a host computer. The transmitter unit is used to generate a The excitation signal of the pseudo-random sequence; The receiver unit uses a high-precision current sensor and an excitation signal to collect current data between the electrodes.

[0060] The above are merely exemplary embodiments of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure.

[0061] This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure. The specification and embodiments are to be considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

Claims

1. A method for inverting and interpreting fluid holdup in horizontal wells based on the complex impedance method, characterized in that, The method includes the following steps: S1: Obtain the complex impedance spectrum of oil-water stratified flow under different water holdup rates and extract characteristic parameters reflecting the water holdup rate; S2: Train the RBF neural network using the dataset constructed from the feature parameters; S3: Measure the oil-water two-phase flow to obtain the current data of the oil-water two-phase flow; S4: Process the current data and extract feature values, then use the trained RBF neural network to invert the feature values ​​to obtain the fluid holdup.

2. The horizontal well fluid sustainment inversion interpretation method based on complex impedance method as described in claim 1, characterized in that, Step S1 includes: The complex impedance spectrum of oil-water stratified flow under different holdup rates in a horizontal well was obtained using the finite element simulation software COMSOL. By using the real and imaginary parts of the complex impedance of the pure water phase as a reference quantity, and combining it with the complex impedance spectrum, the real and imaginary parts of the complex impedance of other water holding capacities are compared to obtain dimensionless characteristic parameters that can reflect water holding capacity information. Pure water phase means water holding capacity of 100%.

3. The horizontal well fluid sustainment inversion interpretation method based on complex impedance method as described in claim 2, characterized in that, The specific steps of using the real and imaginary parts of the complex impedance of the pure aqueous phase as a reference quantity, combining it with the complex impedance spectrum, and comparing it with the real and imaginary parts of the complex impedance of other water holding capacities to obtain dimensionless characteristic parameters that reflect water holding capacity information include: If we consider the oil-water two-phase flow as being formed by the gradual addition of oil phase to the water phase, then the change in the complex impedance of the water phase is due to the oil phase altering the electrical properties of the water phase. The ratio of the real part to the imaginary part slope of oil-water two-phase flow with different water holdup ratios and 100% water holdup is defined as follows: in, The ratio of the real parts. This is the ratio of the slopes of the imaginary parts. For water holding capacity The real part of the time, For water holding capacity The slope of the imaginary part of time, This is the actual value when the water holding capacity is 100%. This is the real value when the water holding capacity is 100%. Received and As a dimensionless feature parameter that reflects holding rate information.

4. The horizontal well fluid sustainment inversion interpretation method based on complex impedance method as described in claim 1, characterized in that, Step S2 includes: The RBF neural network includes: an input layer, a hidden layer, and an output layer.

5. The horizontal well fluid sustainment inversion interpretation method based on complex impedance method as described in claim 1, characterized in that, Step S3 includes: Generate a The excitation signal of the pseudo-random sequence; High-precision current sensors and excitation signals are used to collect current data between electrodes.

6. The horizontal well fluid sustainment inversion interpretation method based on complex impedance method as described in claim 1, characterized in that, Step S4 includes: The acquired current signal is subjected to frequency domain analysis to obtain its corresponding real and imaginary part spectrum. Feature values ​​are extracted from the spectrograms of the real and imaginary parts; The fluid holdup is obtained by inverting the eigenvalues ​​using an RBF neural network.

7. A horizontal well fluid sustainment inversion interpretation system based on the complex impedance method, used to implement the horizontal well fluid sustainment inversion interpretation method based on the complex impedance method as described in any one of claims 1-6, characterized in that, The system includes: an acquisition module and a processing module; The acquisition module is used to measure the oil-water two-phase flow and obtain the current data of the oil-water two-phase flow; The processing module is used to simulate and obtain the complex impedance spectrum of oil-water stratified flow under different holdup rates; The processing module is also used to extract characteristic parameters reflecting water holding capacity from the complex impedance spectrum; The processing module is also used to train the RBF neural network using a dataset constructed from feature parameters; The processing module is also used to process the current data and extract feature values, and to use the trained RBF neural network to invert the feature values ​​to obtain the fluid holdup.

8. The horizontal well fluid sustainment inversion interpretation system based on the complex impedance method as described in claim 7, characterized in that, The processing module includes: a sensor array, a transmitter unit, a receiver unit, and a host computer. The transmitter unit is used to generate a The excitation signal of the pseudo-random sequence; The receiver unit uses a high-precision current sensor and an excitation signal to collect current data between the electrodes.

Images