Wellbore fluid velocity determination methods, apparatuses, and computing devices based on fiber optic logging

Abstract

The application discloses a wellbore fluid velocity determination method and device based on optical fiber logging and a computing device. The method comprises the following steps: converting distributed optical fiber acoustic wave monitoring logging data into an energy density spectrum in a frequency-wave number domain; extracting candidate feature points related to acoustic velocity from the energy density spectrum by using a pre-trained one-class support vector machine; wherein the hyperparameters of the one-class support vector machine are determined in advance based on a particle swarm algorithm; filtering noise points in the candidate feature points to extract target feature points from the candidate feature points; and calculating the wellbore fluid velocity based on the target feature points. The present scheme can realize accurate calculation of the wellbore fluid velocity and improve the determination efficiency of the wellbore fluid velocity.

Description

Technical Field

[0001] This application relates to the field of data processing technology, specifically to a method, apparatus, and computing device for determining wellbore fluid velocity based on fiber optic logging. Background Technology

[0002] Distributed Acoustic Sensing (DAS) logging is a fiber optic logging technology that uses fiber optic cables laid in the wellbore to convert downhole acoustic signals into quantifiable optical signals using the Rayleigh scattering mechanism. It can be applied to scenarios such as wellbore fluid velocity analysis.

[0003] In current technologies for determining wellbore fluid velocity based on DAS data, the slopes of the ascending and descending waves are typically determined manually to obtain the fluid velocity. This method is inefficient and lacks sufficient accuracy in determining wellbore fluid velocity. Summary of the Invention

[0004] In view of the above problems, this application is made in order to provide a method, apparatus and computing device for determining wellbore fluid velocity based on fiber optic logging that overcomes or at least partially solves the above problems.

[0005] According to a first aspect of this application, a method for determining wellbore fluid velocity based on fiber optic logging is provided, comprising:

[0006] Distributed fiber optic acoustic monitoring logging data is converted into energy density spectra in the frequency-wavenumber domain.

[0007] Candidate feature points related to sound speed are extracted from the energy density spectrum using a pre-trained single-class support vector machine; wherein the hyperparameters of the single-class support vector machine are pre-determined based on the particle swarm optimization algorithm.

[0008] Noise points are filtered out from the candidate feature points to extract the target feature points from the candidate feature points;

[0009] The wellbore fluid velocity is calculated based on the target feature points.

[0010] In one optional implementation, the conversion of distributed fiber optic acoustic monitoring logging data into an energy density spectrum in the frequency-wavenumber domain includes:

[0011] The distributed fiber optic acoustic monitoring logging data is segmented to generate multiple data segments;

[0012] For any data slice, perform a two-dimensional fast Fourier transform on the data slice to generate a candidate frequency-wavenumber domain spectrum;

[0013] The zero frequency and zero wavenumber components in the candidate frequency-wavenumber domain spectrum are moved to the center of the candidate frequency-wavenumber domain spectrum to generate the energy density spectrum of the frequency-wavenumber domain.

[0014] In one optional implementation, the hyperparameters of the single-class support vector machine are determined as follows:

[0015] Particle initialization;

[0016] For any given particle, calculate the fitness of applying that particle to a single-class support vector machine;

[0017] Update the global optimal position based on the fitness;

[0018] Determine whether the iteration termination condition is met; if yes, determine the hyperparameters based on the current global optimal position; if no, update the particle velocity and particle position, and execute the step of calculating the fitness of any particle for a single class of support vector machines.

[0019] In one optional implementation, filtering noise points from the candidate feature points to extract the target feature points includes:

[0020] Noise points in the candidate feature points are filtered out using preset feature distribution rules in order to extract the target feature points from the candidate feature points;

[0021] And / or, using a pre-trained support vector machine to extract target feature points from the candidate feature points.

[0022] In one optional implementation, calculating the wellbore fluid velocity based on the target feature point includes:

[0023] Filter out the first target feature points with a wavenumber greater than 0 and the second target feature points with a wavenumber less than 0;

[0024] The upward sound velocity is obtained by fitting the first target feature point, and the downward sound velocity is obtained by fitting the second target feature point;

[0025] The wellbore fluid velocity is calculated based on the upward and downward sound velocities.

[0026] In one optional implementation, fitting the first target feature point to obtain the upward sound velocity and fitting the second target feature point to obtain the downward sound velocity includes:

[0027] Using a pre-trained linear regression model, the upward sound velocity is obtained by fitting the first target feature point, and the downward sound velocity is obtained by fitting the second target feature point;

[0028] The intercept of the linear regression model is 0.

[0029] According to a second aspect of this application, a wellbore fluid velocity determination device based on fiber optic logging is provided, comprising:

[0030] The conversion module is used to convert distributed fiber optic acoustic monitoring logging data into an energy density spectrum in the frequency-wavenumber domain.

[0031] An extraction module is used to extract candidate feature points related to sound speed from the energy density spectrum using a pre-trained single-class support vector machine; wherein the hyperparameters of the single-class support vector machine are pre-determined based on a particle swarm optimization algorithm.

[0032] An enhancement module is used to filter out noise points in the candidate feature points in order to extract the target feature points from the candidate feature points;

[0033] The calculation module is used to calculate the wellbore fluid velocity based on the target feature points.

[0034] In one optional implementation, the conversion module is used to: perform fragmentation processing on the distributed fiber optic acoustic monitoring logging data to generate multiple data fragments;

[0035] For any data slice, perform a two-dimensional fast Fourier transform on the data slice to generate a candidate frequency-wavenumber domain spectrum;

[0036] The zero frequency and zero wavenumber components in the candidate frequency-wavenumber domain spectrum are moved to the center of the candidate frequency-wavenumber domain spectrum to generate the energy density spectrum of the frequency-wavenumber domain.

[0037] In one alternative implementation, the extraction module is used to:

[0038] Particle initialization;

[0039] For any given particle, calculate the fitness of applying that particle to a single-class support vector machine;

[0040] Update the global optimal position based on the fitness;

[0041] Determine whether the iteration termination condition is met; if yes, determine the hyperparameters based on the current global optimal position; if no, update the particle velocity and particle position, and execute the step of calculating the fitness of any particle for a single class of support vector machines.

[0042] In one optional implementation, the enhancement module is used to: filter noise points in the candidate feature points using a preset feature distribution rule, so as to extract the target feature points from the candidate feature points;

[0043] And / or, using a pre-trained support vector machine to extract target feature points from the candidate feature points.

[0044] In one optional implementation, the calculation module is used to: filter out a first target feature point with a wavenumber greater than 0 and a second target feature point with a wavenumber less than 0;

[0045] The upward sound velocity is obtained by fitting the first target feature point, and the downward sound velocity is obtained by fitting the second target feature point;

[0046] The wellbore fluid velocity is calculated based on the upward and downward sound velocities.

[0047] In one optional implementation, the calculation module is used to: use a pre-trained linear regression model to fit the first target feature point to obtain the upward sound speed, and to fit the second target feature point to obtain the downward sound speed;

[0048] The intercept of the linear regression model is 0.

[0049] According to a third aspect of this application, a computing device is provided, comprising: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus;

[0050] The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the above-described method for determining wellbore fluid velocity based on fiber optic logging.

[0051] According to a fourth aspect of this application, a computer storage medium is provided, wherein at least one executable instruction is stored in the storage medium, the executable instruction causing a processor to perform the operation corresponding to the above-described method for determining wellbore fluid velocity based on fiber optic logging.

[0052] According to a fifth aspect of this application, a computer program product is provided, comprising at least one executable instruction that causes a processor to perform the operations corresponding to the above-described method for determining wellbore fluid velocity based on fiber optic logging.

[0053] The method, apparatus, computing device, computer storage medium, and computer program product for determining wellbore fluid velocity based on fiber optic logging provided in this application utilize a single-class support vector machine to extract candidate feature points correlated with sound velocity from the energy density spectrum in the frequency-wavenumber domain. Furthermore, feature enhancement is performed on these candidate feature points to obtain pure target feature points correlated with sound velocity. Finally, the wellbore fluid velocity is calculated based on these target feature points. This approach enables accurate calculation of wellbore fluid velocity and improves the efficiency of wellbore fluid velocity determination.

[0054] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0055] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0056] Figure 1 A schematic flowchart of a wellbore fluid velocity determination method based on fiber optic logging provided in an embodiment of this application is shown.

[0057] Figure 2 A schematic flowchart of a frequency-wavenumber domain energy density spectrum generation method provided in an embodiment of this application is shown;

[0058] Figure 3 A flowchart illustrating a method for determining hyperparameters of a single-class support vector machine according to an embodiment of this application is shown.

[0059] Figure 4 A flowchart illustrating a method for calculating wellbore fluid velocity based on target feature points, provided in an embodiment of this application, is shown.

[0060] Figure 5 This illustration shows a structural schematic diagram of a wellbore fluid velocity determination device based on fiber optic logging, provided in an embodiment of this application.

[0061] Figure 6 A schematic diagram of the structure of a computing device provided in an embodiment of this application is shown. Detailed Implementation

[0062] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0063] Figure 1 The diagram shows a flowchart of a wellbore fluid velocity determination method based on fiber optic logging provided in an embodiment of this application.

[0064] Specifically, such as Figure 1 As shown, the method includes the following steps:

[0065] S101 converts distributed fiber optic acoustic monitoring logging data into an energy density spectrum in the frequency-wavenumber domain.

[0066] This application embodiment uses distributed fiber optic acoustic monitoring logging data to determine wellbore fluid velocity. Distributed fiber optic acoustic monitoring logging data is typically time-depth domain data. This step obtains the frequency-wavenumber domain energy density spectrum by preprocessing the distributed fiber optic acoustic monitoring logging data. The frequency-wavenumber domain energy density spectrum is a spectrum in which the horizontal and vertical axes are wavenumber and frequency, respectively, and the grayscale or color of the spectrum represents the energy density (i.e., power density). Furthermore, the zero frequency and zero wavenumber components of the spectrum in the converted energy density spectrum are located at the center of the spectrum.

[0067] In one alternative implementation, specifically, it can be adopted Figure 2 The steps shown are used to generate the energy density spectrum in the frequency-wavenumber domain, thereby improving the efficiency of subsequent wellbore fluid velocity determination:

[0068] S1011 performs fragmentation processing on distributed fiber optic acoustic monitoring logging data to generate multiple data fragments.

[0069] In actual implementation, since the amount of data in distributed fiber optic acoustic monitoring logging data is usually large, in order to improve data processing efficiency, this embodiment can divide the distributed fiber optic acoustic monitoring logging data into multiple data segments according to the data acquisition time dimension. For example, the data collected by the distributed fiber optic acoustic monitoring logging equipment every N seconds (such as 15 seconds, 30 seconds, 60 seconds, etc.) can be used as a data segment, and each data segment can be stored and processed subsequently.

[0070] S1012, for any data slice, perform a two-dimensional fast Fourier transform on the data slice to generate a candidate frequency-wavenumber domain spectrum.

[0071] For each data slice, a two-dimensional fast Fourier transform (2D FFT) is performed on the data slice to convert the original time-depth domain data into frequency-wavenumber domain data. After the data slice undergoes the two-dimensional fast Fourier transform, candidate frequency-wavenumber domain spectra are obtained.

[0072] S1013, move the zero frequency and zero wavenumber components in the candidate frequency-wavenumber domain spectrum to the center of the candidate frequency-wavenumber domain spectrum to generate the energy density spectrum in the frequency-wavenumber domain.

[0073] The zero-frequency and zero-wavenumber components in the candidate frequency-wavenumber domain spectrum are usually located at the four corners of the spectrum. Therefore, these components are moved from the corners to the center of the spectrum to center the low-frequency, low-wavenumber components, facilitating subsequent analysis. After the movement, the energy density corresponding to any coordinate (wavenumber, frequency) is calculated and represented in grayscale or a corresponding color, thus obtaining the frequency-wavenumber domain energy density spectrum. The generated frequency-wavenumber domain energy density spectrum will appear as two bright, sloping lines radiating outwards from the center, resembling a "V".

[0074] S102 uses a pre-trained single-class support vector machine to extract candidate feature points related to sound speed from the energy density spectrum; wherein, the hyperparameters of the single-class support vector machine are pre-determined based on the particle swarm optimization algorithm.

[0075] The frequency-wavenumber domain energy density spectrum obtained in step S101 contains both sound speed-related feature points and a large number of background data points. The energy density of sound speed-related feature points is high, while the energy density of background data points is low. Therefore, this embodiment of the application pre-trains a One-Class Support Vector Machine (OCSVM) based on a machine learning algorithm. The OCSVM can build a model using only "normal samples" (single-class data) to detect "abnormal samples." For example, a large number of background data point samples (i.e., normal samples) can be used to train the OCSVM so that it can detect sound speed-related feature points (i.e., abnormal samples). Furthermore, since the energy density of sound speed-related feature points is higher than that of background data points, the trained OCSVM can be used to extract sound speed-related feature points from the energy density spectrum; these proposed feature points are called candidate feature points.

[0076] In a single-class support vector machine (SVM), the optimal hyperplane is found in a high-dimensional feature space, such that the vast majority of data points are divided into two sides (corresponding to background data points) and outliers (corresponding to candidate feature points) on the other side, maximizing the margin between the hyperplane and the training sample points. This hyperplane is the decision boundary of the single-class SVM.

[0077] Specifically, the optimization of a single-class support vector machine can be shown in Equation 1:

[0078]

[0079] in, Represents the normal vector of the hyperplane; This represents a slack variable (used to handle situations where the i-th data point may not meet the boundary condition). b This represents the offset of the decision boundary from the origin (the intercept of the hyperplane). The hyperparameters are between 0 and 1 (used to control the proportion of outliers); n represents the total number of data points. This represents the feature mapping function.

[0080] Using the Lagrangian function to transform the optimization problem shown in Equation 1 into a dual problem, we can obtain the following Equation 2:

[0081]

[0082] in, Represent the Lagrange function; and It represents the Lagrange multiplier.

[0083] By using the Lagrange function about , b、 By setting all partial derivatives to zero, the dual problem can be derived, resulting in the following formula 3:

[0084]

[0085] in, This represents the kernel function of a single-class support vector machine. ;x i and x j These represent two data points in the dataset. This application does not limit the specific type of the kernel function; for example, a linear kernel function can be used.

[0086] This shows that the hyperparameter v is a key hyperparameter in a single-class support vector machine, affecting its sensitivity and generalization ability.

[0087] In existing single-class support vector machines, hyperparameters are typically determined based on human experience; however, this approach results in insufficient detection accuracy. To improve the detection accuracy of candidate feature points and thus the accuracy of wellbore fluid velocity determination, this application employs Particle Swarm Optimization (PSO) to pre-determine the hyperparameters in the single-class support vector machine. PSO utilizes the interactions between individuals within a swarm to find the optimal solution. Each particle in PSO represents a possible solution, and particles update their position and velocity based on individual and swarm experience to gradually approach the optimal solution.

[0088] In one alternative implementation, specifically, it can be adopted Figure 3 The steps shown are for determining the hyperparameters in a single-class support vector machine:

[0089] S1021, Particle initialization.

[0090] In the initial state, a predetermined number (e.g., 20-50) of particles are generated within the domain (0-1) of the hyperparameter v. Each particle has a corresponding position and velocity. The particle position represents a candidate solution for the hyperparameter v, and the particle velocity represents the trend and magnitude of change of the candidate solution, i.e., the update step size and update direction during iteration. Furthermore, coefficients such as inertia weight and acceleration constant in the particle swarm optimization algorithm can be further set.

[0091] In one alternative implementation, the initial particle position and particle velocity can be randomly generated, thereby simplifying the execution process of this implementation.

[0092] In one optional implementation, the energy density spectrum (referred to as historical spectrum) in the historical frequency-wavenumber domain can be pre-acquired. The proportion of sound velocity-related feature points in each historical spectrum is statistically analyzed to generate a proportion range. The particle positions in the initial state are determined based on this proportion range. For example, a first proportion of particle positions can be located within this proportion range, and a second proportion of particle positions can be located outside this proportion range. The first proportion is greater than the second proportion, meaning that a first number of particle positions are randomly selected within the proportion range, and a second number of particle positions are randomly selected outside the proportion range. The first number / total number of particles = the first proportion, and the second number / total number of particles = the second proportion. Using this method, the particle positions in the initial state can be concentrated within the proportion range of historically statistically analyzed sound velocity-related feature points. The hyperparameter is the upper bound of the proportion of points judged as outliers, and also the lower bound of the proportion of support vectors. Therefore, the optimal hyperparameter will appear within this proportion range with a higher probability. Thus, this implementation method can accelerate the subsequent iteration process, reduce the number of iterations, and improve the overall execution efficiency.

[0093] S1022, For any particle, calculate the fitness of applying that particle to a single-class support vector machine.

[0094] Fitness is used to evaluate the performance of a single-class support vector machine (SVM) with its hyperparameter v set to that particle position; in other words, it evaluates the quality of the particle. A higher fitness indicates that the model performs better when the hyperparameter v of the single-class SVM is equal to that particle position.

[0095] In the specific implementation process, the hyperparameter v of the single-class support vector machine is set to equal the particle position. The single-class support vector machine is trained using a preset number of training samples to obtain the output result. Based on the output result, the fitness of the particle applied to the single-class support vector machine is generated. For example, the fitness can be generated based on the clustering index of the output result, such as the Davies-Bouldin Index (DBI); or based on the ratio of inter-class variance (one class is feature data points, and the other is background data points) to intra-class variance; or a combination of clustering index and the above ratio can be used to generate the fitness.

[0096] S1023, Update the global optimal position based on fitness.

[0097] The global optimal position is the position of the particle with the highest current fitness. After obtaining the particle fitness, the position of the particle with the highest current fitness is selected as the updated value of the global optimal position. Whenever a new particle position is generated, its fitness is compared with the fitness of the current global optimal position. If the fitness of the new particle position is higher than that of the current global optimal position, the global optimal position is updated, and the new particle position is used as the updated value of the global optimal position. If the fitness of the new particle position is not higher than that of the current global optimal position, the global optimal position is not updated.

[0098] S1024, determine whether the iteration termination condition is met; if yes, execute S1025; if no, execute S1026.

[0099] The iteration termination condition can be reaching the maximum number of iterations, or the convergence of the difference between the global optimal positions of adjacent N iterations, etc.

[0100] S1025, determine the hyperparameters based on the current global optimal position.

[0101] If the iteration termination condition is met, the hyperparameters are determined based on the current global optimal position, i.e., the final hyperparameter value is equal to the global optimal position at the end of the iteration.

[0102] S1026, update particle velocity and particle position.

[0103] If the iteration termination condition is not met, update the particle velocity and particle position, and execute step S1022 for the next iteration.

[0104] Optionally, particle velocity can be updated according to Formula 4, and particle position can be updated according to Formula 5:

[0105]

[0106]

[0107] in, This represents the position of particle i at time t; This indicates the position of particle i at time t+1; This represents the particle velocity of particle i at time t+1; This represents the particle velocity of particle i at time t; This represents the best position currently found by particle i (i.e., the position of the particle with the highest fitness). Indicates the globally optimal position; w p c1 and c2 represent inertial weights, which control the influence of velocity at the next moment; c1 and c2 represent acceleration coefficients, indicating the degree to which the particle is influenced by global and individual experience; r1 and r2 are random numbers between 0 and 1.

[0108] After obtaining the hyperparameters, the single-class support vector machine containing the hyperparameters is trained using sample data to obtain a trained single-class support vector machine. The single-class support vector machine can identify candidate feature points with high energy density. For example, in the output of the single-class support vector machine, 1 represents a candidate feature point and 0 represents a background data point.

[0109] Step S103: Filter out noise points in the candidate feature points to extract the target feature points from the candidate feature points.

[0110] Step S102 allows for the preliminary identification of high-energy-density feature points, i.e., preliminary identification of sound velocity-related feature points. To further improve the accuracy of wellbore fluid velocity determination, feature enhancement is performed in this step, i.e., filtering out noise points among candidate feature points to extract target feature points from the candidate feature points.

[0111] In one optional implementation, a preset feature distribution rule can be used to filter noise points among candidate feature points in order to extract target feature points from the candidate feature points. Specifically, the preset feature distribution rule is generated by analyzing the distribution pattern of sound velocity-related feature points in a large number of historical frequency-wavenumber domain energy density spectra. For example, the distribution pattern of sound velocity-related feature points is as follows: the feature points are distributed in a "V" shape, with the two sides of the "V" located in the positive and negative wavenumber regions respectively. The two sides of the "V" are almost straight lines, with a certain continuity and width. The preset feature distribution rule obtained from this can be: removing candidate feature points in the region near zero wavenumber (within a preset distance range); removing completely isolated points that are not connected to the main feature region (e.g., points whose distance from the cluster center is greater than a preset threshold after clustering candidate feature points, or points whose data points in the cluster are less than a preset number), and so on.

[0112] In one optional implementation, a pre-trained Support Vector Machine (SVC) is used to extract target feature points from candidate feature points. Specifically, unlike a single-class support vector machine, the SVC uses supervised training, while the single-class support vector machine uses unsupervised training. This allows the SVC to be trained using sample data (such as background data points or sound velocity-related data points) and sample annotations. The trained SVC can then reclassify candidate feature points, identifying noise points and target feature points.

[0113] Step S104: Calculate the wellbore fluid velocity based on the target feature points.

[0114] By implementing steps S101-S103 above, pure sound velocity-related feature points (i.e., target feature points) can be extracted, that is, the coordinates of the target feature points in the frequency-wavenumber domain can be obtained. Then, the wellbore fluid velocity can be obtained by fitting the target feature points.

[0115] In one alternative implementation, reference can be made to Figure 4 The steps shown calculate the wellbore fluid velocity based on the target feature points:

[0116] S1041, filter out the first target feature point with a wave number greater than 0 and the second target feature point with a wave number less than 0.

[0117] Using a wave number of 0 as the dividing line, the target feature point can be divided into the first target feature point and the second target feature point. The wave number of the first target feature point is positive, corresponding to the upward wave; the wave number of the second target feature point is negative, corresponding to the downward wave.

[0118] S1042, the upward sound speed is obtained by fitting the first target feature point, and the downward sound speed is obtained by fitting the second target feature point.

[0119] In the frequency-wavenumber domain, the speed of sound is related to the slope of the straight line formed by the target feature points. Therefore, a first straight line passing through the origin is fitted to the first target feature point, and a second straight line passing through the origin is fitted to the second target feature point. The absolute value of the slope of the first straight line is the upward speed of sound, and the absolute value of the slope of the second straight line is the downward speed of sound.

[0120] In one optional implementation, to improve the calculation accuracy of uplink / downlink sound speeds, a linear regression (LR) model is pre-trained. The intercept of the linear regression model is 0. The pre-trained linear regression model is used to fit the first target feature point to obtain the uplink sound speed, and to fit the second target feature point to obtain the downlink sound speed. The training method of the linear regression model is not limited in this embodiment.

[0121] S1043, calculate the fluid velocity in the wellbore based on the upward and downward sound velocities.

[0122] When sound waves propagate upwards or downwards in a stationary fluid, their speed is affected by the fluid flow. The upward speed of sound is the speed at which a sound wave propagates upwards from below the fluid, while the downward speed of sound is the speed at which a sound wave propagates downwards from above the fluid. Using the Doppler effect, the wellbore fluid velocity can be calculated based on Equation 6:

[0123]

[0124] in, Represents the wellbore fluid velocity; Indicates the speed of sound as it travels upwards; This indicates the speed of sound in the downward direction.

[0125] As shown in Table 1, the wellbore fluid velocity calculated using this method (corresponding to the calculated flow velocity in Table 1) is very close to the actual measured flow velocity (corresponding to the recorded flow velocity in Table 1). Therefore, this method can accurately determine the wellbore fluid velocity.

[0126]

[0127] Therefore, the wellbore fluid velocity determination method based on fiber optic logging provided in this application utilizes a single-class support vector machine to extract candidate feature points correlated with sound velocity from the energy density spectrum in the frequency-wavenumber domain. Furthermore, feature enhancement is performed on these candidate feature points to obtain pure target feature points correlated with sound velocity. Finally, the wellbore fluid velocity is calculated based on these target feature points. This approach enables accurate calculation of wellbore fluid velocity and improves the efficiency of wellbore fluid velocity determination.

[0128] Figure 5 This illustration shows a schematic diagram of a wellbore fluid velocity determination device based on fiber optic logging, according to an embodiment of this application. Figure 5 As shown, the device 500 includes: a conversion module 510, an extraction module 520, an enhancement module 530, and a calculation module 540.

[0129] The conversion module 510 is used to convert distributed fiber optic acoustic monitoring logging data into an energy density spectrum in the frequency-wavenumber domain.

[0130] Extraction module 520 is used to extract candidate feature points related to sound speed from the energy density spectrum using a pre-trained single-class support vector machine; wherein, the hyperparameters of the single-class support vector machine are pre-determined based on particle swarm optimization algorithm;

[0131] Enhancement module 530 is used to filter noise points in the candidate feature points in order to extract the target feature points from the candidate feature points;

[0132] The calculation module 540 is used to calculate the wellbore fluid velocity based on the target feature points.

[0133] In one optional implementation, the conversion module 510 is used to: perform fragmentation processing on the distributed fiber optic acoustic monitoring logging data to generate multiple data fragments;

[0134] For any data slice, perform a two-dimensional fast Fourier transform on the data slice to generate a candidate frequency-wavenumber domain spectrum;

[0135] The zero frequency and zero wavenumber components in the candidate frequency-wavenumber domain spectrum are moved to the center of the candidate frequency-wavenumber domain spectrum to generate the energy density spectrum of the frequency-wavenumber domain.

[0136] In one alternative implementation, the extraction module 520 is used for: particle initialization;

[0137] For any given particle, calculate the fitness of applying that particle to a single-class support vector machine;

[0138] Update the global optimal position based on the fitness;

[0139] Determine whether the iteration termination condition is met; if yes, determine the hyperparameters based on the current global optimal position; if no, update the particle velocity and particle position, and execute the step of calculating the fitness of any particle for a single class of support vector machines.

[0140] In an optional implementation, the enhancement module 530 is used to: filter noise points in the candidate feature points using a preset feature distribution rule, so as to extract the target feature points from the candidate feature points;

[0141] And / or, using a pre-trained support vector machine to extract target feature points from the candidate feature points.

[0142] In one optional implementation, the calculation module 540 is used to: filter out a first target feature point with a wavenumber greater than 0 and a second target feature point with a wavenumber less than 0;

[0143] The upward sound velocity is obtained by fitting the first target feature point, and the downward sound velocity is obtained by fitting the second target feature point;

[0144] The wellbore fluid velocity is calculated based on the upward and downward sound velocities.

[0145] In one optional implementation, the calculation module 540 is used to: use a pre-trained linear regression model to fit the first target feature point to obtain the upward sound speed, and to fit the second target feature point to obtain the downward sound speed;

[0146] The intercept of the linear regression model is 0.

[0147] Therefore, the wellbore fluid velocity determination device based on fiber optic logging provided in this application utilizes a single-class support vector machine to extract candidate feature points correlated with sound velocity from the energy density spectrum in the frequency-wavenumber domain. Furthermore, it performs feature enhancement on these candidate feature points to obtain pure target feature points correlated with sound velocity, and finally calculates the wellbore fluid velocity based on these target feature points. This approach enables accurate calculation of wellbore fluid velocity and improves the efficiency of wellbore fluid velocity determination.

[0148] This application provides a non-volatile computer storage medium storing at least one executable instruction or computer program that enables a processor to perform the operation corresponding to the wellbore fluid velocity determination method based on fiber optic logging in any of the above method embodiments.

[0149] This application provides a computer program product, which includes at least one executable instruction or computer program that enables a processor to perform the operation corresponding to the wellbore fluid velocity determination method based on fiber optic logging in any of the above method embodiments.

[0150] Figure 6 The diagram shows a structural schematic of a computing device provided in an embodiment of this application. The specific embodiments of this application do not limit the specific implementation of the computing device.

[0151] like Figure 6 As shown, the computing device may include: a processor 602, a communication interface 604, a memory 606, and a communication bus 608.

[0152] The processor 602, communication interface 604, and memory 606 communicate with each other via communication bus 608. Communication interface 604 is used to communicate with other network elements such as clients or other servers. Processor 602 executes program 610, specifically performing the relevant steps in the above-described embodiment of the method for determining wellbore fluid velocity based on fiber optic logging for computing devices.

[0153] Specifically, program 610 may include program code that includes computer operation instructions.

[0154] The processor 602 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application. The computing device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0155] Memory 606 is used to store program 610. Memory 606 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device. Program 610 can specifically be used to cause processor 602 to perform the operations described in the method embodiments above.

[0156] In summary, based on the computing device, computer storage medium, and computer program product provided in the embodiments of this application, a single-class support vector machine is used to extract candidate feature points related to sound velocity from the energy density spectrum in the frequency-wavenumber domain. Furthermore, feature enhancement is performed on these candidate feature points to obtain pure target feature points related to sound velocity. Finally, the wellbore fluid velocity is calculated based on these target feature points. This solution enables accurate calculation of wellbore fluid velocity and improves the efficiency of wellbore fluid velocity determination.

[0157] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. The required structure for constructing such systems is apparent from the above description. Furthermore, the embodiments of this application are not directed to any particular programming language. It should be understood that the content of this application described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of this application.

[0158] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of this application may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0159] Similarly, it should be understood that, in order to streamline this application and aid in understanding one or more of the various inventive aspects, features of the embodiments of this application are sometimes grouped together in a single embodiment, figure, or description thereof in the above description of exemplary embodiments of this application. However, this method of disclosure should not be construed as reflecting an intention that the claimed application requires more features than expressly recited in each claim. Rather, as reflected in the claims, the inventive aspect lies in fewer than all features of a single foregoing disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into that detailed description, wherein each claim itself is a separate embodiment of this application.

[0160] Those skilled in the art will understand that modules in the device of the embodiments can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or device so disclosed. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0161] Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are intended to be within the scope of this application and form different embodiments. For example, in the following claims, any of the claimed embodiments can be used in any combination.

[0162] The various component embodiments of this application can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some or all of the components according to the embodiments of this application. This application can also be implemented as a device or apparatus program (e.g., a computer program and computer program product) for performing part or all of the methods described herein. Such an implementation of this application can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0163] It should be noted that the above embodiments are illustrative of this application and not restrictive, and those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. This application can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be construed as limiting the order of execution.

Claims

1. A method for determining wellbore fluid velocity based on fiber optic logging, the method comprising: The method comprises: converting distributed optical fiber acoustic monitoring well data into an energy density spectrum in a frequency-wave number domain; extracting candidate feature points related to acoustic velocity from the energy density spectrum using a pre-trained one-class support vector machine, wherein the hyperparameters of the one-class support vector machine are pre-determined based on a particle swarm algorithm; filtering noise points in the candidate feature points to extract target feature points from the candidate feature points; screening out first target feature points with a wave number greater than 0 and second target feature points with a wave number less than 0; fitting the first target feature points to obtain an uplink acoustic velocity using a pre-trained linear regression model, and fitting the second target feature points to obtain a downlink acoustic velocity; wherein the intercept of the linear regression model is 0; and calculating a wellbore fluid velocity based on the uplink acoustic velocity and the downlink acoustic velocity.

2. The method of claim 1, wherein, The conversion of the distributed optical fiber acoustic monitoring well data into the energy density spectrum in the frequency-wave number domain comprises: performing slice processing on the distributed optical fiber acoustic monitoring well data to generate a plurality of data slices; for any data slice, performing two-dimensional fast Fourier transform on the data slice to generate a candidate frequency-wave number domain spectrum; moving zero-frequency and zero-wave number components in the candidate frequency-wave number domain spectrum to the center of the candidate frequency-wave number domain spectrum to generate the energy density spectrum in the frequency-wave number domain.

3. The method of claim 1, wherein, The hyperparameters of the one-class support vector machine are determined as follows: particle initialization; for any particle, calculating the fitness of the particle applied to the one-class support vector machine; updating the global optimal position according to the fitness; determining whether the iteration termination condition is met; if yes, determining the hyperparameters according to the current global optimal position; if no, updating the particle velocity and the particle position, and performing the step of calculating the fitness of the particle applied to the one-class support vector machine.

4. The method according to any one of claims 1 to 3, characterized in that, The filtering of noise points in the candidate feature points to extract target feature points from the candidate feature points comprises: filtering noise points in the candidate feature points to extract target feature points from the candidate feature points using a pre-set feature distribution rule; and / or, extracting target feature points from the candidate feature points using a pre-trained support vector machine.

5. A wellbore fluid velocity determination apparatus based on fiber optic logging, characterized by, The method comprises: a conversion module configured to convert distributed optical fiber acoustic monitoring well data into an energy density spectrum in a frequency-wave number domain; an extraction module configured to extract candidate feature points related to acoustic velocity from the energy density spectrum using a pre-trained one-class support vector machine, wherein the hyperparameters of the one-class support vector machine are pre-determined based on a particle swarm algorithm; an enhancement module configured to filter noise points in the candidate feature points to extract target feature points from the candidate feature points. A computing module is configured to screen out a first target feature point with a wave number greater than 0 and a second target feature point with a wave number less than 0; a linear regression model trained in advance is used to fit the first target feature point to obtain an upgoing acoustic velocity and to fit the second target feature point to obtain a downgoing acoustic velocity; wherein the intercept of the linear regression model is 0; and a wellbore fluid velocity is calculated according to the upgoing acoustic velocity and the downgoing acoustic velocity of the target feature point.

6. A computing device, comprising: Comprise: a processor, a memory, a communication interface and a communication bus, the processor, the memory and the communication interface complete the communication among each other through the communication bus; The memory is used to store at least one executable instruction, and the executable instruction makes the processor execute the operation corresponding to the wellbore fluid velocity determination method based on optical fiber logging in any one of claims 1-4.

7. A computer storage medium, characterized in that The storage medium has at least one executable instruction stored therein, and the executable instruction makes the processor execute the operation corresponding to the wellbore fluid velocity determination method based on optical fiber logging in any one of claims 1-4.

8. A computer program product, characterised in that, Comprise at least one executable instruction, and the executable instruction makes the processor execute the operation corresponding to the wellbore fluid velocity determination method based on optical fiber logging in any one of claims 1-4.

Images