An automatic matching method of oil and gas production decline curve, electronic equipment and storage medium

Abstract

The present application belongs to the technical field of electric digital data processing, and particularly relates to an automatic matching method of oil and gas production decline curve, an electronic device and a storage medium. The method utilizes the mathematical characteristics of double logarithmic space to convert the complex curve matching problem into a linear translation optimization problem, and realizes the automatic alignment of data points in combination with an interpolation algorithm, so that the single well analysis time is shortened from 1-2 hours of the traditional method to within 1 minute, and the efficiency is improved by two orders of magnitude. A numerical optimization algorithm is adopted to replace artificial visual matching, an optimal translation parameter is automatically calculated by establishing a least square optimization model based on centroid initialization, and the deviation caused by human factors is eliminated, so that the parameter inversion accuracy is greatly improved. A coincidence quantitative evaluation mechanism is introduced, the matching result is objectively evaluated by setting a scientific threshold, the problem of lacking quality evaluation standard in the traditional method is solved, and the high-precision and high-efficiency automatic matching of oil and gas well production data and theoretical curve is realized.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of digital data processing, and particularly relates to an automatic matching method for oil and gas production decline curve, an electronic device and a storage medium. BACKGROUND

[0002] In the development of oil and gas fields, production decline analysis is a key technology for evaluating the development effect of single wells and oil fields. The main existing analysis method is manual matching method. Engineers use double logarithmic coordinate paper or professional software such as Harmony, OFM, etc. to perform graphical superposition, visual comparison, and manually adjust the position of the theoretical curve to match the production data. The specific process is as follows:

[0003] The production data is logarithmically transformed, and the stable production section is usually selected as the reference point. The theoretical curve transparency film or software layer is manually translated to match different production stages. After matching, the visual consistency of the whole curve is checked, the key point error is manually calculated, and the error is verified. If the error is verified, the result is output. If not, the reference point is reselected or adjusted, and the matching is performed again. The verification is repeated until it is passed.

[0004] As a traditional standard method for production decline analysis, the manual matching method has significant technical defects. The highly dependent on the subjective experience of engineers leads to large differences in matching results, with a parameter deviation of 15%-30%. The analysis efficiency is low, and it takes 1-2 hours for a single well, and the time increases linearly with the amount of data. In the operation, 2%-5% of cumulative errors are generated in the process of logarithmic transformation and visual alignment. Only a local optimal solution can be obtained, and global optimization and matching quality quantitative evaluation cannot be realized, which seriously restricts the reliability and timeliness of large-scale data analysis.

[0005] There are also some methods that attempt to overcome the defects of manual matching or fitting. For example, Chinese patent document CN113935253A discloses a shale gas well empirical production decline model fitting method based on data weighting. Based on the decline characteristics of the production history data decline stage, the outlier factor detection algorithm is used to identify the abnormal points in the production history data, the exponential smoothing method is used to correct the abnormal values, the Euclidean distance is combined to reasonably assign the production history data fitting weight, and the weighted least squares method is used in the process of solving the empirical production decline model parameters. The obtained fitting precision is higher, and the prediction result is more reliable. However, it is designed for shale gas wells, relies on the definition of the decline stage and the Euclidean distance weighting, and the application scenario is relatively limited. It emphasizes the prediction accuracy, but does not show the quantitative curve matching degree, resulting in the lack of intuitive evaluation standard, and the local matching quality of the fitted curve and the actual production data cannot be directly measured. Only global error indicators are used, which may cover up local deviations. Moreover, this method requires multiple steps of outlier detection, smoothing and filling, and weight calculation, and has high computational complexity.

[0006] As disclosed in Chinese patent document CN110610288A, an oil and gas well production data intelligent system analysis method is disclosed, the data analysis method comprises a variable production pressure data interpretation method, a production decline analysis method and a single well water drive curve analysis method, the production data of the oil and gas well production is first subjected to data preprocessing to obtain more reliable production data, since the daily production data of the single well usually has great volatility, the production data is subjected to one-time spline noise reduction interpolation to reduce the fluctuation of the production data, so that the overall trend is more obvious, and the data after the noise reduction interpolation is segmented. However, although automation is emphasized, the segmented fitting still needs to preset a correlation coefficient threshold, and the processing of abnormal values relies on artificial experience judgment, and the secondary fitting method causes time-consuming calculation due to large parameter search space. SUMMARY

[0007] In view of the above problems, the present application provides an automatic matching method for oil and gas production decline curve, which realizes high-precision and high-efficiency automatic matching of oil and gas well production data and theoretical curve by organically combining double logarithmic space transformation, least square optimization and coincidence quantification evaluation technologies.

[0008] The present application also discloses an electronic device for implementing the above method.

[0009] The present application also discloses a machine-readable storage medium for implementing the above method.

[0010] In order to achieve the above purpose, the present application adopts the following technical solutions:

[0011] An automatic matching method for oil and gas production decline curve, comprising the following steps:

[0012] S101, data preprocessing: performing the same logarithmic conversion on the time-production data of the actual production data and all curves in the theoretical curve library to obtain the logarithmic time-logarithmic production data of the actual production data, the logarithmic time-logarithmic production data of the theoretical curve, the actual production data curve after logarithmic conversion and the theoretical curve after logarithmic conversion, respectively;

[0013] S102, selecting the logarithmic time-logarithmic production data of a theoretical curve in step S101, calculating the centroid offset (dx, dy) between the logarithmic time-logarithmic production data of the actual production data, and taking the centroid offset (dx, dy) as an initial translation vector;

[0014] S103, translating the actual production data curve after logarithmic conversion according to the initial translation vector obtained in step S102 and calculating the residual error of the logarithmic production of the theoretical curve after logarithmic conversion selected in step S102, and minimizing the residual error square sum based on the least square method to obtain an optimal translation vector;

[0015] S104. After shifting the actual production data curve obtained in step S101 through logarithmic transformation according to the optimal translation vector obtained in step S103, count the number of points of overlap between it and the theoretical curve obtained in step S102 through logarithmic transformation, and calculate the degree of overlap.

[0016] S105. Traverse all logarithmically transformed theoretical curves in the theoretical curve library, repeating steps S102-S104. Calculate the overlap between the logarithmically transformed actual production data curve and each logarithmically transformed theoretical curve. Select the logarithmically transformed theoretical curve with the highest overlap as the optimal matching curve, and output the optimal matching curve and key parameters: drain radius r. eD Decreasing exponent b;

[0017] S106. Plot the optimal matching curve output in step S105 in a double logarithmic coordinate system.

[0018] Preferably, in step S102, the theoretical curve initially selected is r. eD The minimum theoretical curve, from r eD The minimum theoretical curve is then traversed.

[0019] Preferably, the centroid offset (dx,dy) in step S102 is the difference between the centroid of the logarithmic time-logarithmic output data of the theoretical curve and the centroid of the logarithmic time-logarithmic output data of the actual production data.

[0020] More preferably, the centroid mentioned in step S102 is calculated according to formulas (1) and (2), as follows:

[0021] , (1)

[0022] , (2)

[0023] Among them, formula (1) , The theoretical curve after logarithmic transformation l1 The coordinates of the centroid of the point on the curve in a logarithmic coordinate system, where N is the number of time points for the theoretical curve. For dimensionless time, For dimensionless output; in formula (2), l2 This is the actual production data curve after logarithmic transformation. , Actual production data curve after logarithmic transformation l2 The coordinates of the centroid of the point on the logarithmic coordinate system are given, where n is the number of time points in the actual production data. It is the number of production days at the i-th time point. is the daily production at the i-th time point, and i in formulas (1) and (2) represents the i-th time point.

[0024] Preferably, the step S103 calculates the residual of the logarithmic production of the shifted logarithmically converted actual production data curve relative to the logarithmically converted theoretical curve selected in the step S102, specifically, the theoretical production value corresponding to the time point of the shifted logarithmically converted actual production data curve is calculated on the logarithmically converted theoretical curve by linear interpolation, and then the difference between the logarithmic production of the shifted logarithmically converted actual production data curve and the theoretical production value of the logarithmically converted theoretical curve is calculated as the residual.

[0025] Preferably, the step S103 minimizes the residual sum of squares based on the least square method to obtain the optimal shift vector, specifically, the optimization function for solving the nonlinear least square problem in the SciPy library is used to minimize the residual sum of squares to obtain the optimal shift vector.

[0026] Preferably, the step S104 specifically calculates the vertical coordinate value of the logarithmically converted theoretical curve corresponding to the horizontal coordinate of the shifted logarithmically converted actual production data curve within the intersection range of the horizontal coordinates of the logarithmically converted theoretical curve and the shifted logarithmically converted actual production data curve by linear interpolation, calculates the absolute value of the difference between the vertical coordinates of the shifted logarithmically converted actual production data curve and the logarithmically converted theoretical curve after interpolation, and if the absolute value of the difference is less than a threshold value, it is considered that the point is coincident, the coincident points meeting the condition are counted, and the coincidence degree is calculated.

[0027] Further preferably, the threshold value is 0.01-0.1.

[0028] Further preferably, the coincidence degree is the proportion of the number of points meeting the difference between the actual production and the theoretical production, and the coincidence degree = the number of coincident points meeting the condition / the number of time points of the logarithmically converted actual production data.

[0029] Preferably, the step S106 specifically uses matplotlib to draw a double logarithmic curve in a double logarithmic coordinate system, superimposes the actual production data points in the form of scattered points and the optimal matching curve displayed in the form of a continuous curve, and labels the coincidence degree.

[0030] The application also provides an application of the automatic matching method for oil and gas well production dynamic analysis, productivity prediction and recoverable reserves evaluation, which realizes rapid and accurate inversion of reservoir parameters by automatically matching the actual production data and the theoretical curve.

[0031] In another aspect of the application, an electronic device is also provided, comprising:

[0032] at least one processor; and

[0033] A memory storing instructions that, when executed by the at least one processor, cause the at least one processor to perform the method for automatically matching oil and gas production decline curves as described above.

[0034] In another aspect of the present application, there is also provided a machine-readable storage medium storing executable instructions that, when executed, cause the machine to perform the method for automatically matching oil and gas production decline curves as described above.

[0035] Compared with the prior art, the present application has the following beneficial effects:

[0036] (1) The present application addresses the core problems of strong experience dependence, low efficiency, error accumulation and lack of objective standards in the prior art, and builds an automated solution: by utilizing the mathematical properties of the double logarithmic space, the complex curve matching problem is transformed into a linear translation optimization problem, and the interpolation algorithm is used to realize the automatic alignment of data points, so that the single well analysis time is shortened from 1-2 hours in the traditional method to less than 1 minute, with an efficiency improvement of two orders of magnitude; numerical optimization algorithm is used instead of manual visual matching, by establishing a least squares optimization model based on centroid initialization, the optimal translation parameters are automatically calculated, and the deviation caused by human factors is eliminated, so that the parameter inversion accuracy is greatly improved; the coincidence quantitative evaluation mechanism is introduced, and the matching results are objectively evaluated by setting a scientific threshold, solving the problem of lack of quality evaluation standard in the traditional method.

[0037] (2) The present application transforms the complex nonlinear decline relationship into a simple linear translation optimization problem through innovative double logarithmic transformation linearization, greatly reducing the difficulty of solving. The method uses centroid initialization to accelerate convergence, and uses the centroid offset of the two curves as the initial guess, which reduces the number of iterations by more than 50% compared with random initialization; the dynamic interpolation target function is used to align the data points in real time, solving the problem of time point mismatch; combined with thresholding coincidence evaluation and intelligent search of the whole curve library, the optimal matching result is automatically selected, the parameter inversion accuracy is improved to within ±3%, and the subjective deviation of manual selection is eliminated. Compared with the traditional manual matching, the present application realizes end-to-end automated analysis, and the single well analysis time is shortened from 1-2 hours to less than 1 minute, with an efficiency improvement of 120 times. Standardized result output ensures data comparability, and visual diagnosis intuitively displays the matching effect. This method not only is simple to operate and can be completed by non-professionals with one key, but also significantly reduces resource consumption: 5 tons of paper can be saved per year, storage space requirement is reduced by 99%, and at the same time, the working environment of engineers is improved, 90% of physical labor and occupational disease risk is reduced, realizing the dual breakthroughs of environmental protection and efficiency.

[0038] (3) Compared with manual matching, the method of the present invention has the advantages of significantly improved analysis accuracy, significantly improved analysis efficiency, simpler and more standardized operation, and energy and resource savings: manual matching relies on experience, and the analysis results of different engineers for the same well can vary by more than ±20%. Through automatic optimization algorithms and overlap quantification, the parameter inversion accuracy is improved to within ±3%. Traditional manual matching takes 1-2 hours per well, including data processing, curve overlay, parameter adjustment, etc. Automated analysis takes less than 60 seconds per well with an Intel i7 processor and Python environment, improving efficiency by 120 times. Manual matching requires professional engineers to operate, with a training period of ≥6 months, and the results are greatly affected by human factors. The present invention can automatically output results by inputting the original data with one-click analysis, and can be operated by non-professionals. Manual analysis requires printing a large number of curve charts, averaging 10 sheets per well, and requires the assistance of a high-performance workstation. The paperless operation of the present invention reduces paper consumption by about 5 tons per year based on 1000 wells; it also improves environmental protection and labor intensity. Manual matching leads to engineers working at their desks for long periods of time, resulting in a high risk of occupational diseases. Paper reports typically require approximately 1 m³ of storage space per 100 wells. This invention reduces manual labor by 90%, eliminating repetitive operations such as curve overlay and manual adjustments. Fully electronic storage achieves a data compression ratio of 1:100, reducing storage space requirements by 99%. Attached Figure Description

[0039] Figure 1 This is a flowchart illustrating the method of the present invention;

[0040] Figure 2 This is a graph showing the actual output over time, based on actual production data from Embodiment 1 of the present invention.

[0041] Figure 3 This is an image showing the automatic matching results of Embodiment 1 of the present invention;

[0042] Figure 4 This is a time diagram of the automatic matching method in Embodiment 1 of the present invention;

[0043] Figure 5 This is the theoretical curve chart used in the manual matching method of Embodiment 1 of the present invention;

[0044] Figure 6 This is a theoretical curve chart used in the automatic matching method of Embodiment 1 of the present invention. Detailed Implementation

[0045] Example 1

[0046] This invention provides an automatic matching method for oil and gas production decline curves. Figure 1 This is a flowchart illustrating the method of the present invention, with reference to... Figure 1 It can be seen that the method includes the following steps:

[0047] S101: data preprocessing: the same logarithmic conversion is performed on the time-yield data of the actual production data and all curves in the theoretical curve library, respectively, to obtain the logarithmic time-logarithmic yield data of the actual production data, the logarithmic time-logarithmic yield data of the theoretical curve, the logarithmically converted actual production data curve and the logarithmically converted theoretical curve;

[0048] Specifically, the A well production history data table is read from the Excel file, the time (d)-yield (m3 / d) data is extracted, which is converted into logarithmic coordinates, the same logarithmic conversion is performed on all curves in the theoretical curve library, and the curve matching problem is converted into a linear translation optimization problem.

[0049] Figure 2 The actual production data is the actual yield-time change graph.

[0050] S102: initial translation vector: the logarithmic time-logarithmic yield data of a theoretical curve in step S101 is selected, the centroid offset (dx, dy) between the logarithmic time-logarithmic yield data of the actual production data is calculated, and the centroid offset (dx, dy) is used as the initial translation vector, wherein dx and dy represent the translation amount required to move the centroid of the actual production data to the centroid of the theoretical curve data;

[0051] Specifically, the first selected theoretical curve is r eD the smallest theoretical curve, starting from r eD the smallest theoretical curve starts to traverse.

[0052] Specifically, the centroid, i.e. the mean coordinate, of the logarithmically converted actual production data and the logarithmically converted theoretical curve is calculated, the centroid represents the "average position" of the data points of the curve in the double logarithmic coordinate, and the centroid offset (dx, dy) is taken as the initial translation vector, which makes the centroid of the logarithmically converted actual production data and the centroid of the logarithmically converted theoretical curve initially coincide. Specifically, the logarithmically converted actual production data curve is defined as l2, the logarithmically converted theoretical curve is defined as l1, the centroid of the logarithmically converted actual production data curve l2 and the logarithmically converted theoretical curve l1 is calculated by the following formula, the initial translation vector is the centroid of the logarithm time-yield data of the theoretical curve minus the centroid of the logarithm time-yield data of the actual production data, if the two curves match completely, there should be centerxl1 = centerxl2 and centeryl1 =centeryl2, i.e. dx=dy=0. When there is an offset, dx and dy represent the translation amount required to move the centroid of the logarithmically converted actual production data to the centroid of the logarithmically converted theoretical curve. Taking the centroid offset (dx, dy) as the initial translation vector, compared with random initialization, the number of iterations is reduced by 60%, and the probability of avoiding local optimal solution is improved to more than 95%.

[0053] , (1)

[0054] , (2)

[0055] wherein formula (1), 、 is the centroid coordinate of the point on the logarithmically converted theoretical curve l1 in the double logarithmic coordinate system, N is the number of time points of the theoretical curve, is the dimensionless time, is the dimensionless yield; in formula (2), l2 is the logarithmically converted actual production data curve, , is the centroid coordinate of the point on the logarithmically converted actual production data curve l2 in the double logarithmic coordinate system, n is the number of time points of the actual production data, is the production day of the i-th time point, is the daily yield of the i-th time point, i in formula (1) and formula (2) represents the i-th time point.

[0056] S103, least square optimization of translation parameters: according to step S102, an initial translation vector is obtained, the actual production data curve after logarithmic conversion is translated, and the residual of the logarithmic production thereof and the logarithmic converted theoretical curve selected in step S102 is calculated, and the least square method is used to minimize the residual sum of squares to obtain the optimal translation vector;

[0057] The translation vector (dx, dy) is optimized by the least square method to minimize the matching error of the actual production data curve after translation in the logarithmic coordinate and the theoretical curve. The translation vector is applied to the logarithmically converted actual production data curve to obtain the translated curve. The theoretical production value corresponding to the time point of the translated curve is calculated on the logarithmically converted theoretical curve by linear interpolation, and the difference between the interpolated theoretical value and the translated actual production is the logarithmic production difference. The optimization function for solving the nonlinear least square problem in the SciPy library is used to minimize the residual sum of squares to obtain the optimal translation vector.

[0058] Specifically, the matching error of the translated logarithmically converted actual production data curve and the theoretical curve is calculated. After the logarithmically converted actual production data curve is translated, the theoretical production value corresponding to the time point of the translated logarithmically converted l2 curve is calculated on the logarithmically converted theoretical curve l1 by linear interpolation. The logarithmic production difference between the translated logarithmically converted actual production data curve and the interpolated point is calculated as the residual.

[0059] Specifically, the current translation amount is applied to the logarithmically converted actual production data curve l2, and the translation is applied to each point of the logarithmically converted actual production data curve l2 to move the logarithmically converted actual production data curve l2 to the currently guessed optimal position.

[0060] (3)

[0061] wherein, is the translated logarithmically converted actual production data curve l2, is the nth time point of the actual production data, is the nth production value of the actual production data, dx is the translation amount in the time direction, and dy is the translation amount in the production direction.

[0062] For each time point of the translated logarithmically converted actual production data curve , find the two adjacent time points and in the theoretical curve, and then linearly interpolate to calculate the corresponding theoretical production value. Calculate the residual, which is the difference between the translated actual production and the interpolated production of the logarithmically converted theoretical curve. Minimize the residual sum of squares to best match the logarithmically converted actual production data to the logarithmically converted theoretical curve.

[0063] S104, calculate the matching degree: after the log-transformed actual production data curve obtained in step S101 is translated by the optimal translation vector obtained in step S103, the number of coincident points with the log-transformed theoretical curve selected in step S102 is counted, and the degree of coincidence is calculated;

[0064] After translation by the optimal translation vector (dx, dy) obtained by optimization, the corresponding vertical coordinate value of the log-transformed theoretical curve under the horizontal coordinate value of the translated log-transformed actual production data curve is calculated by linear interpolation in the horizontal coordinate intersection range of the log-transformed theoretical curve l1 and the translated log-transformed actual production data curve l2, that is, the theoretical yield, and the absolute value of the vertical coordinate difference between the interpolated log-transformed theoretical curve l1 and the translated log-transformed actual production data curve l2 is compared. If the absolute value of the difference is less than the threshold value, it is considered that the point is coincident. The coincident points that meet the condition are counted, and the degree of coincidence is calculated.

[0065] The threshold value is 0.1, and the degree of coincidence is the proportion of the number of points that meet the condition to the number of time points of the log-transformed actual production data. Coincidence degree = number of coincident points that meet the condition / number of time points of log-transformed actual production data.

[0066] S105, automatic search for the best match in the whole curve library: traverse all the log-transformed theoretical curves in the theoretical curve library, that is, different combinations of r eD and b, repeat steps S102-S104, calculate the degree of coincidence of the log-transformed actual production data curve with each log-transformed theoretical curve, select the curve with the highest degree of coincidence as the optimal matching curve, and output the optimal matching curve and the key parameters: drainage radius r eD , decline index b;

[0067] Specifically, the steps S102-S104 are repeated in a loop: the translation vector that makes the log-transformed actual production data curve l2 coincide with the log-transformed l1 after translation is calculated by an optimization algorithm. Then, the vertical coordinate value of the log-transformed theoretical curve l1 corresponding to the horizontal coordinate value of the log-transformed actual production data curve l2 is calculated by interpolation, and the proportion of the number of points with a vertical coordinate difference less than the threshold value is calculated as the degree of coincidence of the current l1. The result is updated, and if the degree of coincidence of the current curve is higher than the historical best value, the optimal matching curve, the translation vector and the degree of coincidence are updated. The result is returned, and after the traversal is completed, the theoretical curve with the highest degree of coincidence is returned. The optimal theoretical matching curve and the key parameters: r eD drainage radius, b decline index are output.

[0068] S106, visualization and result output: draw the optimal matching curve output in step S105 in a double logarithmic coordinate system.

[0069] In the double logarithmic coordinate system, a double logarithmic curve is drawn using matplotlib, actual production data points in the form of scatter points and the optimal matching curve in the form of a continuous curve are superimposed, and the coincidence degree is marked, as shown in Figure 3 .

[0070] The single-well analysis time of the method is 11.4 s, and the specific running time result is shown in Figure 4 , according to the same scene of example 1, the single-well analysis is carried out by using the traditional method, and the approximate time range, operation steps and time cost are shown in table 1 according to experience.

[0071] Table 1 Operation steps and time cost

[0072]

[0073] The artificial matching method has strict requirements for the values of r eD and b, and must be consistent with the discrete values preset by the theoretical chart, the theoretical curve chart used is shown in Figure 5 , the number of theoretical curves is relatively small, and the matching error is large; the method can break through the chart limitation, output continuous values, the theoretical curve chart used is shown in Figure 6 , the number of curve combinations of r eD and b is more, the number of theoretical curves is much larger than that of the artificial matching method, the error is effectively reduced, and the precision is improved.

[0074] Example 2,

[0075] The embodiment also provides an electronic device, comprising:

[0076] at least one processor; and

[0077] a memory, the memory stores instructions, when the instructions are executed by the at least one processor, the at least one processor executes the automatic matching method of the oil and gas production decline curve as described above.

[0078] In the embodiment, the electronic device can include but is not limited to: personal computer, server computer, workstation, desktop computer, laptop computer, notebook computer, mobile computing device, smart phone, tablet computer, cellular phone, personal digital assistant (PDA), handheld device, messaging device, wearable computing device, consumer electronic device, etc.

[0079] Example 3,

[0080] The embodiment also provides a machine readable storage medium, which stores executable instructions, the instructions when executed make the machine execute the automatic matching method of the oil and gas production decline curve as described above.

[0081] In particular, a system or apparatus can be provided with a readable storage medium on which software program codes implementing the functions of any of the above embodiments are stored, and a computer or processor of the system or apparatus is caused to read out and execute the instructions stored in the readable storage medium.

[0082] In this case, the program codes read from the readable medium can themselves implement the functions of any of the above embodiments, and thus the machine readable codes and the readable storage medium storing the machine readable codes constitute a part of the present specification.

[0083] Embodiments of the readable storage medium include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROMs, CD-Rs, CD-RWs, DVD-ROMs, DVD-RAMs, DVD-RWs, DVD-RWs), magnetic tapes, non-volatile memory cards and ROMs. Alternatively, the program codes can be downloaded from a server computer or the cloud over a communication network.

[0084] Obviously, the above embodiments of the present application are merely examples for clearly illustrating the technical solutions of the present application, and are not intended to limit the specific embodiments of the present application. Any modification, equivalent replacement and improvement, etc. made within the spirit and principle of the claims of the present application shall be included in the protection scope of the claims of the present application.

Claims

1. A method for automatic matching of oil and gas production decline curves, characterized by, The method comprises the following steps: S101, data preprocessing: performing the same logarithmic conversion on the time-yield data of the actual production data and all curves in the theoretical curve library to obtain the logarithmic time-logarithmic yield data of the actual production data, the logarithmic time-logarithmic yield data of the theoretical curve, the logarithmically converted actual production data curve and the logarithmically converted theoretical curve, respectively; S102, selecting the logarithmic time-logarithmic yield data of a theoretical curve in step S101, calculating the centroid offset (dx, dy) between the logarithmic time-logarithmic yield data of the actual production data, and taking the centroid offset (dx, dy) as an initial translation vector; S103, according to the initial translation vector obtained in step S102, translating the logarithmically converted actual production data curve and calculating the residual error of the logarithmic yield between the logarithmically converted actual production data curve and the logarithmically converted theoretical curve selected in step S102, and minimizing the residual error square sum based on the least square method to obtain an optimal translation vector; S104, after the logarithmically converted actual production data curve obtained in step S101 is translated by the optimal translation vector obtained in step S103, counting the number of coincident points of the logarithmically converted actual production data curve and the logarithmically converted theoretical curve selected in step S102, and calculating the coincidence degree; S105, traverse all the logarithmically converted theoretical curves in the theoretical curve library, repeat steps S102-S104, calculate the coincidence degree of the logarithmically converted actual production data curve and each logarithmically converted theoretical curve, select the logarithmically converted theoretical curve with the highest coincidence degree as the optimal matching curve, and output the optimal matching curve and key parameters: drainage radius r eD , decline exponent b; S106, drawing the optimal matching curve output in step S105 in a double logarithmic coordinate system.

2. The automatic matching method according to claim 1, characterized in that, The centroid offset (dx, dy) in step S102 is the difference between the centroid of the logarithmic time-logarithmic yield data of the theoretical curve and the centroid of the logarithmic time-logarithmic yield data of the actual production data.

3. The automatic matching method according to claim 2, characterized in that, The centroid in step S102 is calculated according to formulas (1) and (2) as follows: ， （1） ， （2） wherein formula (1), , is the centroid coordinate of the points on the logarithmically converted theoretical curve l1 in the double logarithmic coordinate system, N is the number of time points of the theoretical curve, is the dimensionless time, is the dimensionless production; in formula (2), l2 is the logarithmically converted actual production data curve, , is the centroid coordinate of the points on the logarithmically converted actual production data curve l2 in the double logarithmic coordinate system, n is the number of time points of the actual production data, is the production days of the i th time point, is the daily production of the i th time point, i in formula (1) and formula (2) represents the i th time point.

4. The automatic matching method according to claim 1, characterized in that, The residual error of the logarithmic yield between the logarithmically converted actual production data curve and the logarithmically converted theoretical curve selected in step S102 in step S103 is calculated by linear interpolation on the logarithmically converted theoretical curve to calculate the theoretical yield value corresponding to the time point of the logarithmically converted actual production data curve after translation, and then calculating the difference between the logarithmic yield of the logarithmically converted actual production data curve after translation and the theoretical yield value of the logarithmically converted theoretical curve as the residual error.

5. The automatic matching method according to claim 1, characterized in that, Step S104 is specifically: in the intersection range of the horizontal coordinates of the logarithmically converted theoretical curve and the logarithmically converted actual production data curve after translation, the vertical coordinate value of the logarithmically converted theoretical curve corresponding to the horizontal coordinate of the logarithmically converted actual production data curve after translation is calculated by linear interpolation, and the absolute value of the difference between the vertical coordinates of the logarithmically converted theoretical curve after interpolation and the logarithmically converted actual production data curve after translation is compared. If the absolute value of the difference is less than a threshold value, it is considered that the point is coincident, the coincident points satisfying the condition are counted, and the coincidence degree is calculated.

6. The automatic matching method according to claim 5, characterized in that, The threshold value is 0.01-0.

1.

7. The automatic matching method according to claim 5, characterized in that, The coincidence degree is the proportion of the number of points satisfying the difference between the actual yield and the theoretical yield, and the coincidence degree = the number of coincident points satisfying the condition / the number of time points of the logarithmically converted actual production data.

8. The automatic matching method according to any one of claims 1 to 7, characterized in that, The automatic matching method is used for oil and gas well production dynamic analysis, productivity prediction and recoverable reserves evaluation.

9. An electronic device, comprising: The electronic device comprises: at least one processor; and a memory storing instructions that, when executed by the at least one processor, cause the at least one processor to perform the method of automatically matching production decline curves as claimed in any one of claims 1 to 7.

10. A machine-readable storage medium having stored thereon executable instructions, the method comprising: the instructions, when executed, cause the machine to perform the method of automatically matching production decline curves as claimed in any one of claims 1 to 7.

Images