Pumping well paraffin removal period calculation method and device based on indicator diagram and electronic equipment

Abstract

The invention relates to the technical field of oil well paraffin removal analysis, in particular to a pumping well paraffin removal period calculation method and device based on an indicator diagram and electronic equipment. Inputting 15-dimensional data of the indicator diagram into the paraffin precipitation prediction model to determine a paraffin precipitation degree prediction value, and correcting the paraffin precipitation degree prediction value to obtain a paraffin removal period prediction value; and determining the next paraffin removal date according to the date difference value between the maximum wave crest point and the second maximum wave crest point in the real-time load difference data curve as a paraffin removal period calculation value. According to the method, the paraffin precipitation prediction model is obtained by training the 15-dimensional analysis data of the indicator diagram, and the paraffin removal period prediction value obtained through the paraffin precipitation prediction model and correction is more accurate; and for the purpose of improving the economic benefit, when the predicted value of the paraffin removal period is approached, the next paraffin removal date is further determined by using the real-time load difference data, so that the next paraffin removal date meets the working condition requirements, the production stability and safety are ensured, and the economic benefit can be improved.

Description

Technical Field

[0001] The invention relates to the technical field of oil well wax removal analysis, in particular to a method, a device and an electronic device for calculating the wax removal period of an oil pumping well based on a dynamometer diagram. Background Art

[0002] In the production management of oil wells in oil fields, the main method of crude oil extraction in oil fields is to use pumping units to extract crude oil. However, most of the oil extracted from oil fields is high-wax oil. During the process of oil extraction, due to external factors such as pressure and temperature in the environment and internal factors such as wax content and impurities in the oil, wax crystals are constantly precipitated from the oil, facing serious wax deposition. Finally, wax deposition occurs on the inner wall of equipment such as oil pipelines, resulting in a smaller diameter of the oil well pipeline, a continuous decrease in oil output during oil well extraction, and blockage of the oil well wellbore, which in turn increases the load on the oil well pumping rod. Oil wells and other equipment are difficult to operate and maintain; when the wax deposition of oil wells is serious, it will cause production failures such as oil well wax sticking, well shutdown, oil pipe rupture, and pumping rod equipment breakage, which will directly lead to the shutdown of oil wells for maintenance and cause production failures of oil wells. Therefore, predicting and calculating the wax deposition and wax removal cycle can provide early warning for staff, which is of great significance for avoiding production failures and improving the normal production efficiency of oil wells.

[0003] There are many factors that affect wax deposition in oil wells, and the information involved is very extensive, including data from many fields such as reservoir geology, production dynamics, oil production engineering, and well repair operations. Various parameters have complex correlations with the calculation of the wax removal cycle of oil wells. At present, the main method used by major oil fields is the indicator diagram analysis method. The indicator diagram is a key reference for the normal operation of the pumping unit and can reflect multiple pumping unit operation indicators. The technicians check the operation of the oil well facilities on time, measure a large amount of relevant data, and transfer these data to experts in fault diagnosis. The experts analyze and then determine whether the oil well has wax deposition. The analysis of the indicator diagram mainly calculates the wax deposition cycle from the load difference change and the overall area change of the diagram. However, there are a large number of indirect errors in the data generated by the indicator diagram, which leads to the distortion of the indicator diagram. Analyzing the indicator diagram only from a few dimensions will result in inaccurate analysis results. Since each oil field has a large number of oil wells, this oil well wax removal diagnosis method costs a lot of manpower, material resources, and financial resources, and has certain limitations.

[0004] At present, the wax deposition mechanism modeling, dynamometer image analysis method and machine learning method are mainly used in the market. Among them, the wax deposition mechanism modeling mainly uses fluid mechanics theory, SK equation and other methods to establish mathematical models for molecular diffusion, shear dispersion, Brownian diffusion and gravity sedimentation in the wax deposition process. Since the wax deposition process is very complicated, the wax deposition mechanism modeling is difficult to apply in practice; the dynamometer image analysis method uses image processing technology. The dynamometer is a key reference for the normal operation of the pumping unit and can reflect multiple pumping unit operation indicators. The technicians check the operation of the oil well facilities on time, measure a large amount of relevant data, and transfer these data to experts in fault diagnosis. The experts analyze and determine whether the oil well has wax deposition. The analysis of the dynamometer mainly calculates the wax deposition cycle from the changes in load difference and the overall area of ​​the dynamometer. However, there are a large number of indirect errors in the data used to generate the dynamometer diagram, which leads to its distortion. Analyzing the dynamometer diagram from only a few dimensions will result in inaccurate analysis results. Since each oil field has a large number of oil wells, this oil well wax cleaning diagnosis method consumes a considerable amount of manpower, material resources, and financial resources, and has certain limitations. Machine learning technology has relatively good adaptability and the ability to self-learn, but this type of method focuses on exploring the correlation between things, and the analysis accuracy is slightly insufficient. Summary of the invention

[0005] The present invention provides a method, device and electronic equipment for calculating the wax removal period of an oil pumping well based on an indicator diagram, which overcomes the shortcomings of the above-mentioned prior art and can effectively solve the problems of large indirect errors and inaccurate calculation of the wax removal period in the existing method for calculating the wax removal period of an oil pumping well directly using the indicator diagram analysis data of a few dimensions.

[0006] One of the technical solutions of the present invention is achieved by the following measures: A method for calculating the wax removal cycle of a pumping well based on a dynamometer diagram, comprising:

[0007] Obtain the dynamometer diagram of the pumping well to be analyzed, analyze the dynamometer diagram in n dimensions, and obtain n-dimensional data of the dynamometer diagram, wherein the n-dimensional data of the dynamometer diagram includes maximum load, minimum load, load difference, slope of load-increasing line, slope of unloading line, effective stroke, ABC curvature, BCE curvature, CEA curvature, EAB curvature, dynamometer diagram area, lower left area area, upper left area area, upper right area area, and lower right area area;

[0008] Inputting the n-dimensional data of the dynamometer diagram into the wax deposition prediction model, determining the wax deposition degree prediction value, determining the wax deposition cycle in combination with the wax deposition degree prediction value, and correcting the wax deposition cycle to obtain the wax removal cycle prediction value, wherein the wax deposition prediction model is trained using multiple sets of training data, each set of training data in the multiple sets of training data includes: the n-dimensional data of the dynamometer diagram and identification information corresponding to the wax deposition degree;

[0009] In response to the current date being within the wax cleaning actuarial threshold interval, the date difference between the maximum peak point and the second largest peak point in the real-time load difference data curve is obtained as the wax cleaning cycle calculation value, and the date corresponding to the second largest peak point is shifted back by one wax cleaning cycle calculation value as the next wax cleaning date, wherein the wax cleaning actuarial threshold interval is set according to the wax formation cycle prediction value, and the current date is the difference between the previous wax cleaning date and today's date.

[0010] The following are further optimizations and / or improvements to the above technical solutions:

[0011] The above-mentioned correction of the wax deposition cycle prediction value to obtain the wax removal cycle prediction value includes:

[0012] Analyze the influence of real-time oil parameters and real-time water parameters on the wax removal cycle, and make a correction to the wax deposition cycle prediction value based on the influence value;

[0013] σ=G÷T

[0014] Among them, σ is the influence value; G is the product parameter, x i is the real-time full oil analysis parameter, z j is the real-time water full analysis parameter, a is the slope of the scatter plot of 80 combined product parameters, x is the mean of the real-time full oil analysis parameter, z is the mean of the real-time water full analysis parameter; T is the predicted value of the wax formation cycle;

[0015] The wax-clearing cycle prediction value is corrected twice by using the extended wax-clearing cycle coefficient to obtain the wax-clearing cycle prediction value;

[0016] N 延长 =kN

[0017] Among them, N 延长 is the predicted value of wax cleaning cycle; k is the coefficient of extending wax cleaning cycle; N is the predicted value of wax deposition cycle.

[0018] The construction process of the above wax deposition prediction model includes:

[0019] The trend of the dynamometer diagram change in multiple waxing cycles in history is backtested as a data sample set. Each sample data includes the trend of the dynamometer diagram change and the identification information of the corresponding waxing cycle. The dynamometer diagram change trend dynamometer diagram is the n-dimensional data of the dynamometer diagram. The n-dimensional data includes the maximum load, the minimum load, the load difference, the slope of the load-increasing line, the slope of the unloading line, the effective stroke, the ABC curvature, the BCE curvature, the CEA curvature, the EAB curvature, the dynamometer diagram area, the area of ​​the lower left area, the area of ​​the upper left area, the area of ​​the upper right area, and the area of ​​the lower right area.

[0020] After preprocessing the data sample set, divide it into a training sample set and a test sample set in proportion;

[0021] The long short-term memory network (LSTM) model is trained using the training sample set to obtain a wax deposition prediction model.

[0022] The wax deposition prediction model is tested using the test sample set in combination with the loss function, and the parameters of the long short-term memory network (LSTM) model are adjusted in reverse until the optimal wax deposition prediction model is output.

[0023] In response to the current date being within the wax cleaning actuarial threshold interval, the date difference between the maximum peak point and the second largest peak point in the real-time load difference data curve is obtained as the wax cleaning cycle calculation value, and the date corresponding to the second largest peak point is shifted back by one wax cleaning cycle calculation value as the next wax cleaning date, including:

[0024] Pre-set the wax removal actuarial threshold interval to determine whether the current date is within the wax removal actuarial threshold interval, wherein the wax removal actuarial threshold interval is set according to the wax formation cycle prediction value, and the current date is the difference between the last wax removal date and the current date;

[0025] In response to the current date being within the wax removal actuarial threshold interval, the real-time dynamometer diagram is parsed in n dimensions to obtain a load difference data curve and filter it;

[0026] Perform peak analysis on the filtered load difference data curve to obtain a peak sequence arranged in descending order from large to small, wherein the peak is a point value in the load difference data curve that is greater than the first three values ​​and greater than the last three values;

[0027] The difference between the maximum value and the second maximum value in the peak sequence is used as the calculation value of the wax cleaning cycle;

[0028] The date corresponding to the second largest value is moved back one wax clearing cycle calculation value to become the next wax clearing date.

[0029] The above wax removal date is further determined, including:

[0030] After obtaining the next wax cleaning date, determine whether the next wax cleaning date is less than the current date;

[0031] In response to "yes", the date of the next wax cleaning is moved back by one wax cleaning cycle calculation value, and the next wax cleaning date is updated;

[0032] Repeat the above steps until the next wax cleaning date is not less than the current date, and output the final next wax cleaning date.

[0033] The above also includes, in response to the current date being within the wax cleaning actuarial threshold range, obtaining the real-time load difference data curve and determining whether the load difference data variation range is less than the set variation threshold, and if so, not performing the next wax cleaning date analysis.

[0034] The second technical solution of the present invention is achieved by the following measures: a device for calculating the wax removal cycle of a pumping well based on a dynamometer diagram, comprising:

[0035] A data acquisition unit is used to acquire the dynamometer diagram of the oil well to be analyzed, analyze the dynamometer diagram in n dimensions, and obtain n-dimensional data of the dynamometer diagram, wherein the n-dimensional data of the dynamometer diagram includes maximum load, minimum load, load difference, slope of load-increasing line, slope of unloading line, effective stroke, ABC curvature, BCE curvature, CEA curvature, EAB curvature, dynamometer diagram area, lower left area area, upper left area area, upper right area area, and lower right area area;

[0036] The first analysis unit inputs the n-dimensional data of the dynamometer diagram into a wax deposition prediction model, determines a wax deposition degree prediction value, determines a wax deposition cycle in combination with the wax deposition degree prediction value, and corrects the wax deposition cycle to obtain a wax removal cycle prediction value, wherein the wax deposition prediction model is trained using multiple sets of training data, and each set of training data in the multiple sets of training data includes: the n-dimensional data of the dynamometer diagram and identification information corresponding to the wax deposition degree;

[0037] The second analysis unit, in response to the current date being within the wax cleaning actuarial threshold interval, obtains the date difference between the maximum peak point and the second largest peak point in the real-time load difference data curve as the wax cleaning cycle calculation value, and the date corresponding to the second largest peak point is shifted back by one wax cleaning cycle calculation value as the next wax cleaning date, wherein the wax cleaning actuarial threshold interval is set according to the wax formation cycle prediction value, and the current date is the difference between the previous wax cleaning date and the current date.

[0038] The following are further optimizations and / or improvements to the above technical solutions:

[0039] The first analysis unit comprises:

[0040] The model prediction module inputs the n-dimensional data of the dynamometer diagram into the wax deposition prediction model to determine the wax deposition cycle prediction value, wherein the wax deposition prediction model is trained using multiple sets of training data, each set of training data in the multiple sets of training data includes: the n-dimensional data of the dynamometer diagram and the identification information of the corresponding wax deposition cycle;

[0041] Correction modules, including:

[0042] The first correction submodule analyzes the influence of the real-time oil full parameters and the real-time water full parameters on the wax removal cycle, and corrects the wax deposition cycle prediction value based on the influence degree value;

[0043] σ=G÷T

[0044] Among them, σ is the influence value; G is the product parameter, x iis the real-time full oil analysis parameter, z j is the real-time water full analysis parameter, a is the slope of the scatter plot of 80 combined product parameters, x is the mean of the real-time full oil analysis parameter, z is the mean of the real-time water full analysis parameter; T is the predicted value of the wax formation cycle;

[0045] The second correction submodule uses the extended wax removal cycle coefficient to perform a second correction on the wax deposition cycle prediction value after the first correction to obtain the wax removal cycle prediction value;

[0046] N 延长 =kN

[0047] Among them, N 延长 is the predicted value of wax cleaning cycle; k is the coefficient of extending wax cleaning cycle; N is the predicted value of wax deposition cycle.

[0048] The above also includes a third analysis unit, including:

[0049] After obtaining the next wax cleaning date, determine whether the next wax cleaning date is less than the current date;

[0050] In response to "yes", the date of the next wax cleaning is moved back by one wax cleaning cycle calculation value, and the next wax cleaning date is updated;

[0051] Repeat the above steps until the next wax cleaning date is not less than the current date, and output the final next wax cleaning date.

[0052] The present invention expands the dynamometer diagram analysis method from two analysis dimensions of load difference and graph area to n analysis dimensions, and analyzes the dynamometer diagram in n dimensions, which can more accurately reflect the process of the oil well from normal liquid supply, slight wax deposition, wax deposition (wax influence), to severe wax deposition, and can accurately reflect the process of the oil well from normal liquid supply, slight wax deposition, wax deposition (wax influence), to severe wax deposition, thereby providing technical support for the prediction of the wax removal cycle of the oil well; further, the present invention uses the n-dimensional analysis data of the dynamometer diagram to train a wax deposition prediction model, so that the wax removal cycle prediction value obtained after the wax deposition prediction model and correction is more accurate; because the wax removal cycle of the oil well should be enlarged as much as possible without the occurrence of wax jam accidents, so as to reduce the wax removal cost and improve the economic benefit, the present invention uses the real-time load difference data to further determine the next wax removal date when it is close to the wax removal cycle prediction value, so that the next wax removal date meets the working condition requirements, ensures production stability and safety, and improves the economic benefit. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Attached Figure 1 The present invention is a schematic flow chart of a method.

[0054] Attached Figure 2 It is a schematic diagram of 15-dimensional analysis of the dynamometer diagram in the present invention.

[0055] Attached Figure 3 It is a schematic diagram of multi-dimensional decomposition of the dynamometer diagram in the present invention.

[0056] Attached Figure 4 The figure is a schematic flow chart of the method for constructing a wax prediction model in the present invention.

[0057] Attached Figure 5 The figure is a flow chart of the method for determining the wax removal time in the present invention.

[0058] Attached Figure 6 The present invention is another method flow chart.

[0059] Attached Figure 7 The figure is a schematic diagram of the structure of a device of the present invention.

[0060] Attached Figure 8 This is a schematic diagram of another device structure of the present invention. DETAILED DESCRIPTION

[0061] The present invention is not limited by the following embodiments, and specific implementation methods can be determined based on the technical solution of the present invention and actual conditions.

[0062] The present invention will be further described below in conjunction with embodiments and drawings:

[0063] Embodiment 1: As attached Figure 1 As shown, the embodiment of the present invention discloses a method for calculating the wax removal cycle of a pumping well based on a dynamometer diagram, comprising:

[0064] Step S110, obtaining the dynamometer diagram of the pumping well to be analyzed, parsing the dynamometer diagram in n dimensions, and obtaining n-dimensional data of the dynamometer diagram, wherein the n-dimensional data of the dynamometer diagram includes maximum load, minimum load, load difference, slope of load-increasing line, slope of unloading line, effective stroke, ABC curvature, BCE curvature, CEA curvature, EAB curvature, dynamometer diagram area, lower left area area, upper left area area, upper right area area, and lower right area area;

[0065] Step S120, inputting the n-dimensional data of the dynamometer diagram into the wax deposition prediction model, determining the wax deposition degree prediction value, determining the wax deposition cycle in combination with the wax deposition degree prediction value, and correcting it to obtain the wax removal cycle prediction value, wherein the wax deposition prediction model is trained using multiple sets of training data, and each set of training data in the multiple sets of training data includes: the n-dimensional data of the dynamometer diagram and identification information corresponding to the wax deposition degree;

[0066] Step S130, in response to the current date being within the wax cleaning actuarial threshold interval, the date difference between the maximum peak point and the second largest peak point in the real-time load difference data curve is obtained as the wax cleaning cycle calculation value, and the date corresponding to the second largest peak point is shifted back by one wax cleaning cycle calculation value as the next wax cleaning date, wherein the wax cleaning actuarial threshold interval is set according to the wax formation cycle prediction value, and the current date is the difference between the previous wax cleaning date and the current date.

[0067] The invention discloses a method for calculating the wax removal cycle of a pumping well based on an indicator diagram, which expands the indicator diagram analysis method from two analysis dimensions of load difference and graphic area to n analysis dimensions, analyzes the indicator diagram in n dimensions, and can more accurately reflect the process of the oil well from normal liquid supply, slight wax deposition, wax deposition (wax influence), and severe wax deposition, and can accurately reflect the process of the oil well from normal liquid supply, slight wax deposition, wax deposition (wax influence), and severe wax deposition, thereby providing technical support for the prediction of the wax removal cycle of the oil well; further, the invention uses the n-dimensional analysis data of the indicator diagram to train a wax deposition prediction model, so that the wax removal cycle prediction value obtained after the wax deposition prediction model and correction is more accurate; because the wax removal cycle of the oil well should be enlarged as much as possible without the occurrence of wax jam accidents, so as to reduce the wax removal cost and improve the economic benefit, the invention uses the real-time load difference data to further determine the next wax removal date when it is close to the wax removal cycle prediction value, so that the next wax removal date meets the working condition requirements, ensures production stability and safety, and improves the economic benefit.

[0068] Embodiment 2: The embodiment of the present invention discloses a method for calculating the wax removal cycle of a pumping well based on a dynamometer diagram, comprising:

[0069] Step S210, obtain the dynamometer diagram of the pumping well to be analyzed, analyze the dynamometer diagram in n dimensions, and obtain n-dimensional data of the dynamometer diagram, wherein the n-dimensional data of the dynamometer diagram includes maximum load, minimum load, load difference, slope of loading line, slope of unloading line, effective stroke, ABC curvature, BCE curvature, CEA curvature, EAB curvature, dynamometer diagram area, lower left area area, upper left area area, upper right area area, and lower right area area. Specifically, n dimensions need to be selected according to the actual needs, and 15 dimensions are selected for analysis, wherein the schematic diagram of 15-dimensional analysis of the dynamometer diagram is as shown in the attached figure. Figure 2 As shown in the figure, the multi-dimensional decomposition diagram of the indicator diagram is shown in the attached figure. Figure 3 The 15 dimensions are shown in Table 1.

[0070] Table 1 15 dimension tables

[0071]

[0072]

[0073] Step S220, input the n-dimensional data of the dynamometer diagram into the wax deposition prediction model, determine the predicted value of the wax deposition degree, determine the wax deposition cycle in combination with the predicted value of the wax deposition degree, and correct it to obtain the predicted value of the wax removal cycle, wherein the wax deposition prediction model is trained using multiple sets of training data, and each set of training data in the multiple sets of training data includes: the n-dimensional data of the dynamometer diagram and identification information corresponding to the wax deposition degree.

[0074] The above-mentioned n-dimensional data of the dynamometer diagram is input into the wax deposition prediction model to determine the wax deposition cycle prediction value. The construction process of the wax deposition prediction model is as shown in the attached figure. Figure 4 As shown, including:

[0075] Step S221, backtesting the dynamometer diagram change trend within multiple waxing cycles in history as a data sample set, wherein each sample data includes the dynamometer diagram change trend within a waxing cycle and the identification information of the corresponding waxing cycle, the dynamometer diagram change trend within a waxing cycle is the n-dimensional data change trend obtained by performing n-dimensional analysis on the dynamometer diagram within a waxing cycle, and the n-dimensional data includes the maximum load, the minimum load, the load difference, the slope of the load-increasing line, the slope of the unloading line, the effective stroke, the ABC curvature, the BCE curvature, the CEA curvature, the EAB curvature, the dynamometer diagram area, the lower left area area, the upper left area area, the upper right area area, and the lower right area area;

[0076] Step S222, after preprocessing the data sample set, divide it into a training sample set and a test sample set in proportion; the preprocessing here includes removing outliers and standardizing the data (which may be normalization processing);

[0077] Step S223, training a long short-term memory network (LSTM) model using the training sample set to obtain a wax deposition prediction model;

[0078] LSTM neurons can mine the characteristics and laws of wax deposition data from historical information (trends of changes in the dynamometer diagram within multiple wax deposition cycles in history). Using LSTM neurons to construct a neural network helps solve the problem of nonlinear time series prediction. As a complex time series affected by multiple factors, operating condition prediction can be reasonably predicted using LSTM neural networks, thereby providing data support for operating condition warnings. Therefore, the long short-term memory network LSTM model of the present invention can be composed of three layers of LSTM units and one fully connected layer.

[0079] Step S224, the wax deposition prediction model is tested using the test sample set in combination with the loss function, and the parameters of the long short-term memory network LSTM model are adjusted in reverse until the optimal wax deposition prediction model is output. Here, the wax deposition prediction model is tested using the test sample set in combination with the loss function to verify the applicability of the model. When the loss function result meets the requirements, the corresponding wax deposition prediction model is the optimal model.

[0080] In the above steps, the predicted wax deposition cycle value is corrected to obtain the predicted paraffin removal cycle value, including:

[0081] (1) Analyze the influence degree value of real-time total oil parameters and real-time total water parameters on the paraffin removal cycle, and perform a primary correction on the predicted wax deposition cycle value based on the influence degree value;

[0082] σ = G ÷ T

[0083] Among them, σ is the influence degree value; G is the product parameter, x i is the real-time total oil analysis parameter, z j is the real-time total water analysis parameter, a is the slope of the scatter plot of 80 combined product parameters, x is the mean value of the real-time total oil analysis parameter, z is the mean value of the real-time total water analysis parameter; T is the predicted wax deposition cycle value;

[0084] The construction process of the above product parameter model is as follows:

[0085] (a) Relationship between paraffin removal cycle and liquid production per day

[0086] Through the least squares theory, regression analysis is carried out using paraffin removal cycle data and liquid production per day data to establish the relationship between paraffin removal cycle and oil production per day.

[0087] The basic idea is as follows:

[0088] The least squares method is the most commonly used method to solve the curve fitting problem. Let:

[0089]

[0090] Among them, is a set of linearly independent functions selected in advance, a k is the undetermined coefficient (k = 1, 2,..., m, m < n), and the fitting criterion is to make y i =(i = 1, 2,..., n) and f(x i ) the sum of the squares of the distances δ i the smallest, which is called the least squares criterion.

[0091] (b) Plot the scatter plot of total oil analysis parameters related to crude oil

[0092] The total oil analysis parameters related to crude oil are: crude oil density, acidity, freezing point, wax content, initial boiling point. Making a scatter plot of the paraffin removal cycle and the total oil analysis parameters can reflect the influence degree of these total oil analysis parameters on the paraffin removal cycle. Based on the oil production per day, a calculation formula related to the paraffin removal cycle and crude oil is established. The slope of the scatter plot of the paraffin removal cycle and the total oil analysis parameters is multiplied by the "difference between the specific parameter value and the average value", and divided by the value range of the paraffin removal cycle to reflect the influence degree of a certain parameter on the paraffin removal cycle.

[0093]

[0094] Where y is the wax removal cycle, x is the oil analysis parameter related to crude oil, k is a constant, and the slope of the scatter plot is a i , a i and k are the model parameters obtained by linear regression of y and x, and k is the intercept.

[0095] (c) Establish the relationship between the wax cleaning cycle and the daily water production

[0096] Using the same method as in (a), establish the relationship between the wax cleaning cycle and the daily water production.

[0097] (d) Draw a scatter plot of the oil full analysis parameters related to produced water

[0098] The oil full analysis parameters related to produced water are: pH value, bicarbonate content, chloride content, sulfate content, calcium content, magnesium content, Na_K content, total hardness, temporary hardness, negative hardness, mineralization, water density, first salt, water grade, first alkali, second alkali. The scatter plot of the wax removal cycle and the water full analysis parameters can reflect the influence of these water full analysis parameters on the wax removal cycle. Based on the daily water production, a calculation formula related to the wax removal cycle and produced water is established. The slope of the scatter plot of the wax removal cycle and the water full analysis parameters is multiplied by the "difference between the specific parameter value and the average value" and divided by the value range of the wax removal cycle to reflect the influence of a certain parameter on the wax removal cycle. If all specific parameters are averaged, the wax removal cycle is only related to the daily water production.

[0099]

[0100] Where y is the wax cleaning cycle, x is the oil full analysis parameter related to produced water, k is a constant, and the slope of the scatter plot is formula b. b and k are model parameters obtained by linear regression of y and x, and k is the intercept.

[0101] (e) Draw the interactive parameter scatter plot of oil full analysis parameters and water full analysis parameters

[0102] The oil full analysis parameters related to crude oil and the oil full analysis parameters related to produced water are multiplied in pairs to obtain 80 combinations, and a scatter plot of the wax removal cycle and these product parameters is made to reflect the degree of influence of these product parameters on the wax removal cycle. The slope of the scatter plot of the wax removal cycle and these product parameters is multiplied by the "difference between the specific parameter value and the average value" and divided by the value range of the wax removal cycle to reflect the degree of influence of these product parameters on the wax removal cycle.

[0103]

[0104] Where y is the wax cleaning cycle, x is i is the total oil analysis parameter value, z j is the water full analysis parameter value, a is the slope of the scatter plot of 80 combined product parameters.

[0105] (2) using the extended wax removal cycle coefficient to perform a secondary correction on the wax deposition cycle prediction value after the primary correction to obtain the wax removal cycle prediction value;

[0106] N 延长 =kN

[0107] Among them, N 延长 is the predicted value of wax cleaning cycle; k is the coefficient of extending wax cleaning cycle; N is the predicted value of wax deposition cycle.

[0108] Step S230, in response to the current date being within the wax cleaning actuarial threshold interval, the date difference between the maximum peak point and the second largest peak point in the real-time load difference data curve is obtained as the wax cleaning cycle calculation value, and the date corresponding to the second largest peak point is shifted back by one wax cleaning cycle calculation value as the next wax cleaning date, wherein the wax cleaning actuarial threshold interval is set according to the wax formation cycle prediction value, and the current date is the difference between the previous wax cleaning date and the current date.

[0109] As attached Figure 5 As shown, the above step S230 includes:

[0110] Step S231, pre-setting the wax removal actuarial threshold interval, determining whether the current date is within the wax removal actuarial threshold interval, wherein the wax removal actuarial threshold interval is set according to the wax formation cycle prediction value, and the current date is the difference between the last wax removal date and the current date;

[0111] Step S232, in response to the current date being within the wax removal actuarial threshold interval, the real-time dynamometer diagram is analyzed in n dimensions to obtain a load difference data curve and filter it;

[0112] The filtering method here can adopt the median average filtering algorithm, specifically:

[0113] The load difference data of the dynamometer diagram is packaged into an array [load difference, date] to form a two-dimensional sampling point array P [[load difference 0, date 0], [load difference 1, date 1], [load difference 2, date 2], ... [load difference n, date n]], and then the filter sampling length (N) is set. In the present invention, it is set to 7 (because 7 is close to the threshold of severe wax deposition or wax jam). The larger the value, the smoother the filtered curve is, but the more serious the distortion is. On the contrary, the smaller the distortion is, but the smoothness of the filtered curve is lower. The initial state is all empty.

[0114] When the source index is equal to 0, no filtering is performed and it is directly added to the result sequence (R[]). At this time, R[0] = P[0], P[0] is added to the front end of N[], and N[0] = P[0].

[0115] ...When the source index is greater than 0 and less than 4, the current load difference is first added to the sampling point sequence (N[]) in sequence (first-in-first-out principle), and then the arithmetic mean before the current sampling point is calculated and added to the result sequence.

[0116] P1=AVG(N0,N1);

[0117] P2=AVG(N0,N1,N2);

[0118] P3=AVG(N0,N1,N2,N3);

[0119] When the source index is greater than 4 and less than N, the current load difference is first added to the sampling sequence (N[]) in sequence, and then N[] is sorted, the maximum and minimum values ​​are excluded, and the arithmetic mean of the remaining sequence is calculated. The calculation result is added to the result sequence, and the result is:

[0120] (1) S[] = SORT(N0, N1, N2, N3, N4) - sorting sequence;

[0121] (2) R4 = AVG (S[] - N max -N min )-The arithmetic mean of the sequence after sorting and removing the maximum and minimum values.

[0122] When the source index is greater than or equal to N, the new sampling point data is added to the sampling point sequence N[], the first point in the sampling point sequence is squeezed out, and N[] is then kept at a length of 7. Then the sequence in N[] is also operated in step (2).

[0123] After all sampling points are calculated, a filter curve data sequence is finally obtained. Since the median arithmetic mean filter has a phase lag of N / 2, the final filter curve should be shifted forward by N / 2 lengths. After the filtering is completed, the cycle judgment begins.

[0124] Step S233, performing peak analysis on the filtered load difference data curve to obtain a peak sequence arranged in descending order from large to small, wherein a peak is a point value in the load difference data curve that is greater than the first three values ​​and greater than the last three values;

[0125] Step S234, taking the difference between the maximum value and the second largest value in the peak sequence as the wax cleaning cycle calculation value;

[0126] Step S235, the date corresponding to the second largest value is shifted back by one wax cleaning cycle calculation value as the next wax cleaning date.

[0127] Embodiment 3: As attached Figure 6 As shown, the embodiment of the present invention discloses a method for calculating the wax removal cycle of a pumping well based on a dynamometer diagram, comprising:

[0128] Step S310, obtaining the dynamometer diagram of the pumping well to be analyzed, parsing the dynamometer diagram in n dimensions, and obtaining n-dimensional data of the dynamometer diagram, wherein the n-dimensional data of the dynamometer diagram includes maximum load, minimum load, load difference, slope of load-increasing line, slope of unloading line, effective stroke, ABC curvature, BCE curvature, CEA curvature, EAB curvature, dynamometer diagram area, lower left area area, upper left area area, upper right area area, and lower right area area;

[0129] Step S320, inputting the n-dimensional data of the dynamometer diagram into the wax deposition prediction model, determining the wax deposition degree prediction value, determining the wax deposition cycle in combination with the wax deposition degree prediction value, and correcting it to obtain the wax removal cycle prediction value, wherein the wax deposition prediction model is trained using multiple sets of training data, and each set of training data in the multiple sets of training data includes: the n-dimensional data of the dynamometer diagram and identification information corresponding to the wax deposition degree;

[0130] Step S330, in response to the current date being within the wax clearing actuarial threshold interval, the date difference between the maximum peak point and the second largest peak point in the real-time load difference data curve is obtained as the wax clearing cycle calculation value, and the date corresponding to the second largest peak point is shifted back by one wax clearing cycle calculation value as the next wax clearing date, wherein the wax clearing actuarial threshold interval is set according to the wax formation cycle prediction value, and the current date is the difference between the previous wax clearing date and the current date;

[0131] Step S340, re-determining the wax removal date, including:

[0132] (1) After obtaining the next wax cleaning date, determine whether the next wax cleaning date is less than the current date;

[0133] (2) In response to "yes", the date of the next wax cleaning is shifted back by one wax cleaning cycle calculation value, and the next wax cleaning date is updated;

[0134] (3) Repeat the above steps until the next wax cleaning date is not less than the current date, and output the final next wax cleaning date.

[0135] Embodiment 4: As attached Figure 7 As shown, the embodiment of the present invention discloses a device for calculating the wax removal cycle of an oil pumping well based on a dynamometer diagram, comprising:

[0136] A data acquisition unit is used to acquire the dynamometer diagram of the oil well to be analyzed, analyze the dynamometer diagram in n dimensions, and obtain n-dimensional data of the dynamometer diagram, wherein the n-dimensional data of the dynamometer diagram includes maximum load, minimum load, load difference, slope of load-increasing line, slope of unloading line, effective stroke, ABC curvature, BCE curvature, CEA curvature, EAB curvature, dynamometer diagram area, lower left area area, upper left area area, upper right area area, and lower right area area;

[0137] The first analysis unit inputs the n-dimensional data of the dynamometer diagram into a wax deposition prediction model, determines a wax deposition degree prediction value, determines a wax deposition cycle in combination with the wax deposition degree prediction value, and corrects the wax deposition cycle to obtain a wax removal cycle prediction value, wherein the wax deposition prediction model is trained using multiple sets of training data, and each set of training data in the multiple sets of training data includes: the n-dimensional data of the dynamometer diagram and identification information corresponding to the wax deposition degree;

[0138] The second analysis unit, in response to the current date being within the wax cleaning actuarial threshold interval, obtains the date difference between the maximum peak point and the second largest peak point in the real-time load difference data curve as the wax cleaning cycle calculation value, and the date corresponding to the second largest peak point is shifted back by one wax cleaning cycle calculation value as the next wax cleaning date, wherein the wax cleaning actuarial threshold interval is set according to the wax formation cycle prediction value, and the current date is the difference between the previous wax cleaning date and the current date.

[0139] Embodiment 5: As attached Figure 8 As shown, the embodiment of the present invention discloses a device for calculating the wax removal cycle of an oil pumping well based on a dynamometer diagram, comprising:

[0140] A data acquisition unit is used to acquire the dynamometer diagram of the oil well to be analyzed, analyze the dynamometer diagram in n dimensions, and obtain n-dimensional data of the dynamometer diagram, wherein the n-dimensional data of the dynamometer diagram includes maximum load, minimum load, load difference, slope of load-increasing line, slope of unloading line, effective stroke, ABC curvature, BCE curvature, CEA curvature, EAB curvature, dynamometer diagram area, lower left area area, upper left area area, upper right area area, and lower right area area;

[0141] The first analysis unit inputs the n-dimensional data of the dynamometer diagram into a wax deposition prediction model, determines a wax deposition degree prediction value, determines a wax deposition cycle in combination with the wax deposition degree prediction value, and corrects the wax deposition cycle to obtain a wax removal cycle prediction value, wherein the wax deposition prediction model is trained using multiple sets of training data, and each set of training data in the multiple sets of training data includes: the n-dimensional data of the dynamometer diagram and identification information corresponding to the wax deposition degree;

[0142] The first analysis unit comprises:

[0143] The model prediction module inputs the n-dimensional data of the dynamometer diagram into the wax deposition prediction model to determine the wax deposition cycle prediction value, wherein the wax deposition prediction model is trained using multiple sets of training data, each set of training data in the multiple sets of training data includes: the n-dimensional data of the dynamometer diagram and the identification information of the corresponding wax deposition cycle;

[0144] Correction modules, including:

[0145] The first correction submodule analyzes the influence of the real-time oil full parameters and the real-time water full parameters on the wax removal cycle, and corrects the wax deposition cycle prediction value based on the influence degree value;

[0146] σ=G÷T

[0147] Among them, σ is the influence value; G is the product parameter, x i is the real-time full oil analysis parameter, z j is the real-time water full analysis parameter, a is the slope of the scatter plot of 80 combined product parameters, x is the mean of the real-time full oil analysis parameter, z is the mean of the real-time water full analysis parameter; T is the predicted value of the wax formation cycle;

[0148] The second correction submodule uses the extended wax removal cycle coefficient to perform a second correction on the wax deposition cycle prediction value after the first correction to obtain the wax removal cycle prediction value;

[0149] N 延长 =kN

[0150] Among them, N 延长 is the predicted value of wax cleaning cycle; k is the coefficient of extending wax cleaning cycle; N is the predicted value of wax deposition cycle.

[0151] The second analysis unit, in response to the current date being within the wax clearing actuarial threshold interval, obtains the date difference between the maximum peak point and the second largest peak point in the real-time load difference data curve as the wax clearing cycle calculation value, and the date corresponding to the second largest peak point is shifted back by one wax clearing cycle calculation value as the next wax clearing date, wherein the wax clearing actuarial threshold interval is set according to the wax formation cycle prediction value, and the current date is the difference between the previous wax clearing date and the current date;

[0152] The third analysis unit includes:

[0153] After obtaining the next wax cleaning date, determine whether the next wax cleaning date is less than the current date;

[0154] In response to "yes", the date of the next wax cleaning is moved back by one wax cleaning cycle calculation value, and the next wax cleaning date is updated;

[0155] Repeat the above steps until the next wax cleaning date is not less than the current date, and output the final next wax cleaning date.

[0156] Embodiment 5: The embodiment of the present invention discloses an electronic device, including a processor and a memory, wherein the memory stores a computer program, and the computer program is loaded and executed by the processor to implement a method for calculating the wax removal cycle of an oil well based on a dynamometer diagram.

[0157] The processor may be a central processing unit (CPU), a general purpose processor, a digital signal processor (DSP), an ASIC, an FPGA or other programmable logic device, a transistor logic device, a hardware component or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. It may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like. The memory may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory, a mobile hard disk, a magnetic disk or an optical disk.

[0158] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0159] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0160] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0161] The above technical features constitute the best embodiment of the present invention, which has strong adaptability and best implementation effect. Non-essential technical features can be added or reduced according to actual needs to meet the requirements of different situations.

Claims

1. A method for calculating the wax removal cycle of a pumping well based on a dynamometer diagram, characterized in that: include: Obtain the dynamometer diagram of the pumping well to be analyzed, analyze the dynamometer diagram in n dimensions, and obtain n-dimensional data of the dynamometer diagram, wherein the n-dimensional data of the dynamometer diagram includes maximum load, minimum load, load difference, slope of load-increasing line, slope of unloading line, effective stroke, ABC curvature, BCE curvature, CEA curvature, EAB curvature, dynamometer diagram area, lower left area area, upper left area area, upper right area area, and lower right area area; Inputting the n-dimensional data of the dynamometer diagram into the wax deposition prediction model, determining the wax deposition degree prediction value, determining the wax deposition cycle in combination with the wax deposition degree prediction value, and correcting the wax deposition cycle to obtain the wax removal cycle prediction value, wherein the wax deposition prediction model is trained using multiple sets of training data, each set of training data in the multiple sets of training data includes: the n-dimensional data of the dynamometer diagram and identification information corresponding to the wax deposition degree; In response to the current date being within the wax cleaning actuarial threshold interval, the date difference between the maximum peak point and the second largest peak point in the real-time load difference data curve is obtained as the wax cleaning cycle calculation value, and the date corresponding to the second largest peak point is shifted back by one wax cleaning cycle calculation value as the next wax cleaning date, wherein the wax cleaning actuarial threshold interval is set according to the wax formation cycle prediction value, and the current date is the difference between the previous wax cleaning date and today's date.

2. The method for calculating the wax removal period of a pumping well based on a dynamometer diagram according to claim 1, characterized in that: The wax deposition cycle prediction value is corrected to obtain the wax removal cycle prediction value, including: Analyze the influence of real-time oil parameters and real-time water parameters on the wax removal cycle, and make a correction to the wax deposition cycle prediction value based on the influence value; σ=G÷T Among them, σ is the influence value; G is the product parameter, x i is the real-time full oil analysis parameter, z j is the real-time water full analysis parameter, a is the slope of the scatter plot of 80 combined product parameters, x is the mean of the real-time full oil analysis parameter, z is the mean of the real-time water full analysis parameter; T is the predicted value of the wax formation cycle; The wax-clearing cycle prediction value is corrected twice by using the extended wax-clearing cycle coefficient to obtain the wax-clearing cycle prediction value; N 延长 =kN Among them, N 延长 is the predicted value of wax cleaning cycle; k is the coefficient of extending wax cleaning cycle; N is the predicted value of wax deposition cycle.

3. The method for calculating the wax removal period of a pumping well based on a dynamometer diagram according to claim 1 or 2, characterized in that: The construction process of the wax deposition prediction model includes: The trend of the dynamometer diagram change in multiple waxing cycles in history is taken as a data sample set, wherein each sample data includes the trend of the dynamometer diagram change in a waxing cycle and the identification information of the corresponding waxing cycle. The trend of the dynamometer diagram change in a waxing cycle is the trend of the n-dimensional data change obtained by performing n-dimensional analysis on the dynamometer diagram in a waxing cycle. The n-dimensional data includes the maximum load, the minimum load, the load difference, the slope of the load-increasing line, the slope of the unloading line, the effective stroke, the ABC curvature, the BCE curvature, the CEA curvature, the EAB curvature, the dynamometer diagram area, the area of ​​the lower left area, the area of ​​the upper left area, the area of ​​the upper right area, and the area of ​​the lower right area. After preprocessing the data sample set, divide it into a training sample set and a test sample set in proportion; The long short-term memory network (LSTM) model is trained using the training sample set to obtain a wax deposition prediction model. The wax deposition prediction model is tested using the test sample set in combination with the loss function, and the parameters of the long short-term memory network (LSTM) model are adjusted in reverse until the optimal wax deposition prediction model is output.

4. The method for calculating the wax removal period of a pumping well based on a dynamometer diagram according to claim 1 or 2, characterized in that: In response to the current date being within the wax cleaning actuarial threshold interval, obtaining the date difference between the maximum peak point and the second largest peak point in the real-time load difference data curve as the wax cleaning cycle calculation value, and the date corresponding to the second largest peak point is shifted back by one wax cleaning cycle calculation value as the next wax cleaning date, including: Pre-set the wax removal actuarial threshold interval to determine whether the current date is within the wax removal actuarial threshold interval, wherein the wax removal actuarial threshold interval is set according to the wax formation cycle prediction value, and the current date is the difference between the last wax removal date and the current date; In response to the current date being within the wax removal actuarial threshold interval, the real-time dynamometer diagram is parsed in n dimensions to obtain a load difference data curve and filter it; Perform peak analysis on the filtered load difference data curve to obtain a peak sequence arranged in descending order from large to small, wherein the peak is a point value in the load difference data curve that is greater than the first three values ​​and greater than the last three values; The difference between the maximum value and the second maximum value in the peak sequence is used as the calculation value of the wax cleaning cycle; The date corresponding to the second largest value is moved back one wax clearing cycle calculation value to become the next wax clearing date.

5. The method for calculating the wax removal period of a pumping well based on a dynamometer diagram according to claim 4, characterized in that: It also includes the re-determination of the wax removal date, including: After obtaining the next wax cleaning date, determine whether the next wax cleaning date is less than the current date; In response to "yes", the date of the next wax cleaning is moved back by one wax cleaning cycle calculation value, and the next wax cleaning date is updated; Repeat the above steps until the next wax cleaning date is not less than the current date, and output the final next wax cleaning date.

6. The method for calculating the wax removal period of a pumping well based on a dynamometer diagram according to any one of claims 1 to 4, characterized in that: It also includes, in response to the current date being within the wax cleaning actuarial threshold range, obtaining the real-time load difference data curve and determining whether the change range of the load difference data is less than the set change threshold, and if so, not performing the next wax cleaning date analysis.

7. A device for calculating the wax removal period of a pumping well based on a dynamometer diagram using the method as claimed in any one of claims 1 to 6, characterized in that: include: A data acquisition unit is used to acquire the dynamometer diagram of the oil well to be analyzed, analyze the dynamometer diagram in n dimensions, and obtain n-dimensional data of the dynamometer diagram, wherein the n-dimensional data of the dynamometer diagram includes maximum load, minimum load, load difference, slope of load-increasing line, slope of unloading line, effective stroke, ABC curvature, BCE curvature, CEA curvature, EAB curvature, dynamometer diagram area, lower left area area, upper left area area, upper right area area, and lower right area area; The first analysis unit inputs the n-dimensional data of the dynamometer diagram into a wax deposition prediction model, determines a wax deposition degree prediction value, determines a wax deposition cycle in combination with the wax deposition degree prediction value, and corrects the wax deposition cycle to obtain a wax removal cycle prediction value, wherein the wax deposition prediction model is trained using multiple sets of training data, and each set of training data in the multiple sets of training data includes: the n-dimensional data of the dynamometer diagram and identification information corresponding to the wax deposition degree; The second analysis unit, in response to the current date being within the wax cleaning actuarial threshold interval, obtains the date difference between the maximum peak point and the second largest peak point in the real-time load difference data curve as the wax cleaning cycle calculation value, and the date corresponding to the second largest peak point is shifted back by one wax cleaning cycle calculation value as the next wax cleaning date, wherein the wax cleaning actuarial threshold interval is set according to the wax formation cycle prediction value, and the current date is the difference between the previous wax cleaning date and the current date.

8. The device for calculating the wax removal period of a pumping well based on the dynamometer diagram according to claim 7, characterized in that: The first analysis unit comprises: The model prediction module inputs the n-dimensional data of the dynamometer diagram into the wax deposition prediction model to determine the wax deposition cycle prediction value, wherein the wax deposition prediction model is trained using multiple sets of training data, each set of training data in the multiple sets of training data includes: the n-dimensional data of the dynamometer diagram and the identification information of the corresponding wax deposition cycle; Correction modules, including: The first correction submodule analyzes the influence of the real-time oil full parameters and the real-time water full parameters on the wax removal cycle, and corrects the wax deposition cycle prediction value based on the influence degree value; σ=G÷T Among them, σ is the influence value; G is the product parameter, x i is the real-time full oil analysis parameter, z j is the real-time water full analysis parameter, a is the slope of the scatter plot of 80 combined product parameters, x is the mean of the real-time full oil analysis parameter, z is the mean of the real-time water full analysis parameter; T is the predicted value of the wax formation cycle; The second correction submodule uses the extended wax removal cycle coefficient to perform a second correction on the wax deposition cycle prediction value after the first correction to obtain the wax removal cycle prediction value; N 延长 =kN Among them, N 延长 is the predicted value of wax cleaning cycle; k is the coefficient of extending wax cleaning cycle; N is the predicted value of wax deposition cycle.

9. The device for calculating the wax removal period of a pumping well based on a dynamometer diagram according to claim 7 or 8, characterized in that: Also included is a third analysis unit, including: After obtaining the next wax cleaning date, determine whether the next wax cleaning date is less than the current date; In response to "yes", the date of the next wax cleaning is moved back by one wax cleaning cycle calculation value, and the next wax cleaning date is updated; Repeat the above steps until the next wax cleaning date is not less than the current date, and output the final next wax cleaning date.

10. An electronic device, characterized in that: The method comprises a processor and a memory, wherein a computer program is stored in the memory, and the computer program is loaded and executed by the processor to implement the method for calculating the wax removal cycle of an oil pumping well based on the dynamometer diagram as claimed in any one of claims 1 to 6.

Images