Method and device for diagnosing working condition of rod-pumped well

Abstract

The invention relates to the technical field of mechanical oil extraction, and discloses a rod-pumped well working condition diagnosis method and device, and the method comprises the following steps: S01, obtaining target data; step S02, preprocessing the target data obtained in the step S01; s03, judging whether working condition argument needs to be carried out or not; step S04, analyzing the indicator diagram of the rod-pumped well based on the constructed convolutional neural network so as to perform fault diagnosis on the working condition of the rod-pumped well and obtain the working condition fault type of the rod-pumped well; s05, outputting a working condition argument result of the rod-pumped well; according to the method, the step S03 and the step S04 are set, the pump efficiency is analyzed, the working condition dissection of the rod-pumped well is recognized in advance, the dissection efficiency is improved, the characteristics of the indicator diagrams are extracted by constructing the convolutional neural network, the characteristics of the different indicator diagrams are recognized, and the working condition fault category of the rod-pumped well is accurately obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of mechanical oil production, and more specifically to a method and device for diagnosing the working condition of an oil pumping well. Background Art

[0002] A pumping unit, also known as a pumping well, is a type of machine equipment for oil extraction, commonly known as a "kowtow machine"; a pumping unit is the most important lifting equipment in a rod pumping system; depending on whether there is a walking beam, it can be divided into a walking beam pumping unit and a non-walking beam pumping unit; with the development of the oil industry, the importance of pumping wells in the production process is becoming increasingly higher; pumping wells are prone to operating problems during the production process, such as thick oil, wax deposition, wellbore failure, etc.; if these problems are not discovered and handled in a timely manner, they will affect the normal production of the oil well and even lead to production accidents, so real-time monitoring and diagnosis of the operating conditions of the pumping wells is very necessary.

[0003] Most methods for diagnosing the working condition of pumping wells use dynamometer diagrams for analysis. By recording the production parameters of the pumping wells, the closed relationship curve generated according to the change of the suspension point load with the suspension point displacement in a complete reciprocating motion process is the dynamometer diagram. When different fault types occur, the normal production state is affected, resulting in different work performance, and the corresponding closed curves of the dynamometer diagrams have their own characteristics. However, due to the complexity of the working environment of the pumping wells, there are many types of faults, so it is difficult to effectively analyze the working conditions based on the dynamometer diagram. Traditionally, the diagnosis of the working condition of pumping wells is mostly performed by analyzing the dynamometer diagram through manual recognition methods to perform fault diagnosis. This requires a lot of on-site work experience and prior knowledge of oilfield experts, has high technical requirements for staff, and has low diagnostic efficiency. Therefore, a method for diagnosing the working condition of pumping wells is needed, which can effectively and accurately analyze the dynamometer diagram, so as to accurately obtain the working condition of the pumping well. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method and device for diagnosing the working condition of a pumping well to solve the problems existing in the above-mentioned background technology.

[0005] The present invention provides the following technical solution: a method for diagnosing the working condition of a pumping well, comprising the following steps: Step S01: Acquire target data; Step S02: preprocessing the target data acquired in step S01; Step S03: Determine whether it is necessary to perform a working condition judgment: Analyze and calculate the actual pump efficiency and theoretical pump efficiency of the pumping well, and calculate the pump efficiency difference index. If the pump efficiency difference index exceeds the set threshold, it is determined that it is necessary to perform a working condition judgment, and step S04 is executed; otherwise, it is determined that it is not necessary to perform a working condition judgment, and step S05 is executed; Step S04: Analyze the dynamometer diagram of the pumping well based on the constructed convolutional neural network, thereby diagnosing the working condition of the pumping well and obtaining the working condition fault type of the pumping well; Step S05: outputting the operating condition judgment result of the pumping well.

[0006] Preferably, the formulas for analyzing and calculating the actual pump efficiency and theoretical pump efficiency of the oil pumping well in step S03 are respectively expressed as: , where η B is the actual pumping efficiency of the pumping well, Q 实际 is the actual production of the oil well, Q 理论 is the theoretical production of the oil well; , where η 理 is the theoretical pump efficiency; η 冲 is the stroke loss influence coefficient, that is, the pump efficiency when the elastic expansion and contraction of the tubing and the sucker rod produces stroke loss; β is the filling degree of the pump, η 漏 is the pump leakage coefficient, that is, the influence coefficient of pump leakage on pump efficiency; η 气 The coefficient of displacement affected by liquid shrinkage, that is, the coefficient of the volume shrinkage of crude oil degassing on the pump efficiency; like , where C η is the pump efficiency difference index, when C η The value satisfies C η >YU, it is determined that the working condition judgment is required; when C η The value satisfies C η ≤YU, it is determined that no operating condition judgment is required; YU is a set threshold, and YU≤30%.

[0007] Preferably, the calculation formula of the stroke loss influence coefficient is: , where S is the stroke length of the polished rod at the suspension point, λ j is the stroke loss caused by static load, λ g The stroke loss is caused by inertial load; the stroke loss refers to the difference between the suspension point stroke and the actual stroke of the plunger.

[0008] Preferably, the calculation formula for the stroke loss caused by the static load is: , where Δp is the pressure difference between the discharge port and the suction port of the oil well pump; L is the depth of the oil well pump; f p is the cross-sectional area of ​​the oil well pump plunger; E is the elastic modulus of the material used in the oil well pump; f r is the average cross-sectional area of ​​the sucker rod string; f t is the average cross-sectional area of ​​the oil pipe; The calculation formula for the stroke loss caused by the inertial load is: , where μ is the inertia factor coefficient, , where ω is the angular velocity of the pumping unit crank, L is the depth of the pump, and a is the propagation speed of the sound wave in the sucker rod string, and a=5100.

[0009] Preferably, the calculation formula of the filling degree β of the pump is: , where X y1 is the first displacement factor affected by free gas, X y2 The second displacement factor affected by the clearance volume; , where f w is the water content of the oil layer, R is the original oil-gas ratio, R' is the dissolved oil-gas ratio under the pump suction pressure, B m is the gas volume coefficient at the submerged pressure; , where S x is the anti-stroke distance length, and S is the stroke length of the bare rod at the suspension point.

[0010] Preferably, the calculation formula of the leakage coefficient of the pump is: , where Q 漏 is the leakage between the oil well pump plunger and the pump barrel, Q 内 The flow rate of crude oil above the plunger of the oil pump; , where D is the diameter of the pump barrel, δ is the gap between the pump plunger and the pump barrel, Δp is the pressure difference between the discharge port and the suction port of the pump, γ is the dynamic viscosity of the gap liquid, and L' is the effective plunger length, that is, the length of the sealing contact part of the pump plunger; V p is the plunger stroke volume of the oil well pump; , where P 理 is the theoretical displacement of the oil well pump, , where S is the stroke length of the bare rod at the suspension point, and n is the number of pumping unit strokes.

[0011] Preferably, the calculation formula for the displacement coefficient affected by liquid shrinkage is: , where B0 is the volume coefficient of crude oil under sinking pressure, R is the original oil-gas ratio, and f w is the water content of the oil layer.

[0012] A pumping well working condition diagnosis device, comprising a data acquisition module, a data preprocessing module, a working condition judgment module, a working condition fault judgment module and a data output module; The data acquisition module is used to collect target data, including but not limited to the dynamic liquid level, stroke, stroke frequency, wellhead back pressure, wellhead casing pressure, wellhead pump plunger cross-sectional area, wellhead pump depth and ground dynamometer diagram of the pumping well; The data preprocessing module is used to preprocess the target data to obtain target data that can be directly used; the preprocessing operation includes filtering and normalizing the target data; The working condition judgment module is used to analyze and calculate the actual pump efficiency and theoretical pump efficiency of the pumping well, and calculate the pump efficiency difference index, so as to determine whether it is necessary to perform working condition judgment on the pumping well; The working condition fault diagnosis module analyzes the dynamometer diagram of the pumping well based on the constructed convolutional neural network, thereby diagnosing the working condition of the pumping well and obtaining the working condition fault type of the pumping well; The data output module is used to output the working condition judgment result of the oil pumping well.

[0013] Technical effects and advantages of the present invention: The present invention is provided with step S03 and step S04, which is conducive to analyzing and calculating the actual pump efficiency and theoretical pump efficiency of the oil pumping well, and calculating the pump efficiency difference index, so as to determine whether it is necessary to perform a working condition diagnosis on the oil pumping well. If a working condition diagnosis is required, the indicator diagram of the oil pumping well is analyzed based on the constructed convolutional neural network, so as to diagnose the working condition of the oil pumping well and obtain the working condition fault type of the oil pumping well; by analyzing the pump efficiency, the working condition diagnosis of the oil pumping well is identified in advance, and the diagnosis efficiency is improved; by constructing a convolutional neural network to extract features of the indicator diagram, the features of different indicator diagrams are identified, and the working condition fault category of the oil pumping well is accurately obtained, thereby improving the analysis and diagnosis efficiency and accuracy of the indicator diagram, without the need to analyze the indicator diagram through an artificial recognition method, and improving the accuracy of fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 The present invention is a flow chart of the method for diagnosing the working condition of a pumping well.

[0015] Figure 2 It is a structural diagram of the pumping well working condition diagnosis device of the present invention. DETAILED DESCRIPTION

[0016] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the present invention. In addition, the forms of the various structures recorded in the following embodiments are merely illustrative. The method and device for diagnosing the working conditions of a pumping well involved in the present invention are not limited to the various structures recorded in the following embodiments. All other embodiments obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention.

[0017] like Figure 1 As shown, the present invention provides a method for diagnosing the working condition of a pumping well, comprising the following steps: Step S01: Obtain target data: obtain data such as the dynamic liquid level, stroke, stroke frequency, wellhead back pressure, wellhead casing pressure, wellhead pump plunger cross-sectional area, wellhead pump depth, and ground indicator diagram of the well; Step S02: preprocessing the target data acquired in step S01: the preprocessing operation includes filtering and normalizing the target data, wherein the filtering can eliminate noise interference, and the normalization can improve the comparability of the data; Step S03: Determine whether it is necessary to perform a working condition diagnosis: Analyze and calculate the actual pump efficiency and theoretical pump efficiency of the pumping well, and calculate the pump efficiency difference index. If the pump efficiency difference index exceeds the set threshold, it is determined that a working condition diagnosis is required, and step S04 is executed; otherwise, it is determined that no working condition diagnosis is required, and step S05 is executed; if the pump efficiency difference index exceeds the set threshold, it means that there is a working condition failure in the pumping well, which affects the actual pump efficiency, resulting in a large difference between the actual pump efficiency and the theoretical pump efficiency. The main factors affecting the pump efficiency include stroke loss of the rod string, gas and insufficiency, leakage and volume coefficient, etc., so it is necessary to further diagnose the working condition failure of the pumping well; Step S04: Analyze the indicator diagram of the pumping well based on the constructed convolutional neural network, so as to diagnose the working condition of the pumping well and obtain the working condition fault type of the pumping well; the indicator diagram when the pumping well fails is different from the indicator diagram of the normal working condition, so the indicator diagram is feature extracted and analyzed to obtain whether the working condition of the pumping well is abnormal, and identify the working condition fault type corresponding to the indicator diagram through the convolutional neural network; Step S05: output the operating condition diagnosis result of the pumping well; if it is determined in step S03 that the operating condition diagnosis is not required, the output operating condition diagnosis result is "normal"; if it is determined in step S03 that the operating condition diagnosis is required, the output operating condition diagnosis result is "abnormal", and the operating condition fault type output in step S04 is output at the same time.

[0018] In this embodiment, it should be specifically explained that the formulas for analyzing and calculating the actual pump efficiency and theoretical pump efficiency of the oil pumping well in step S03 are respectively expressed as: , where η B is the actual pumping efficiency of the pumping well, Q 实际 is the actual production of the oil well, Q 理论 is the theoretical production of the oil well; , where η 理 is the theoretical pump efficiency; η 冲 is the stroke loss influence coefficient, that is, the pump efficiency when the elastic expansion and contraction of the tubing and the sucker rod produces stroke loss; β is the filling degree of the pump, η 漏 is the pump leakage coefficient, that is, the influence coefficient of pump leakage on pump efficiency; η气 The coefficient of displacement affected by liquid shrinkage, that is, the coefficient of the volume shrinkage of crude oil degassing on the pump efficiency; like , where C η is the pump efficiency difference index, when C η The value satisfies C η >YU, it is determined that the pump efficiency difference index exceeds the set threshold value, and it is necessary to make an operating condition judgment; at this time, the actual pump efficiency is too different from the theoretical pump efficiency, which exceeds the set difference range, indicating that the operating condition of the pumping well has a fault, resulting in the efficiency and capacity of the pumping unit being insufficient to reach the expected pump efficiency level, so it is necessary to further judge the operating condition of the pumping unit; at this time, step S04 is executed; when C η The value satisfies C η ≤YU, it is judged that the pump efficiency difference index does not exceed the set threshold value, and no working condition judgment is required, which is a normal value; at this time, the actual pump efficiency is slightly different from the theoretical pump efficiency and does not exceed the set difference range. The part that does not reach the theoretical pump efficiency is a normal loss phenomenon; The YU is a set threshold value, satisfying YU≤30%. The specific value of YU can be set by technical personnel in the field according to actual conditions. This embodiment does not specifically limit the specific value of YU.

[0019] In this embodiment, it should be specifically explained that the calculation formula of the stroke loss influence coefficient is: , where S is the stroke length of the polished rod at the suspension point, λ j is the stroke loss caused by static load, λ g The stroke loss caused by inertial load; the stroke loss refers to the difference between the suspension point stroke and the actual stroke of the plunger; where: , where Δp is the pressure difference between the discharge port and the suction port of the oil well pump; L is the depth of the oil well pump; f p is the cross-sectional area of ​​the oil well pump plunger; E is the elastic modulus of the material used for the oil well pump. In this embodiment, steel is preferred, so E can be the elastic modulus of steel; f r is the average cross-sectional area of ​​the sucker rod string; f t is the average cross-sectional area of ​​the oil pipe; , where μ is the inertia factor coefficient, , where ω is the angular velocity of the pumping unit crank, L is the depth of the pump, and a is the propagation speed of the sound wave in the sucker rod string, and a=5100.

[0020] In this embodiment, it should be specifically explained that the calculation formula of the filling degree β of the pump is: , where X y1 is the first displacement factor affected by free gas, Xy2 The second displacement factor affected by the clearance volume; , where f w is the water content of the oil layer, R is the original oil-gas ratio, R' is the dissolved oil-gas ratio under the pump suction pressure, B m is the gas volume coefficient at the submerged pressure; , where S x is the anti-stroke distance length, and S is the stroke length of the bare rod at the suspension point.

[0021] In this embodiment, it should be specifically stated that the calculation formula of the leakage coefficient of the pump is: , where Q 漏 is the leakage between the oil well pump plunger and the pump barrel, Q 内 The flow rate of crude oil above the plunger of the oil pump; , where D is the diameter of the pump barrel, δ is the gap between the pump plunger and the pump barrel, Δp is the pressure difference between the discharge port and the suction port of the pump, γ is the dynamic viscosity of the gap liquid, and L' is the effective plunger length, that is, the length of the sealing contact part of the pump plunger; V p is the plunger stroke volume of the oil well pump; , where P 理 is the theoretical displacement of the oil well pump, , where S is the stroke length of the bare rod at the suspension point, and n is the number of pumping unit strokes.

[0022] In this embodiment, it should be specifically explained that the calculation formula of the displacement coefficient affected by the liquid shrinkage is: , where B0 is the volume coefficient of crude oil under sinking pressure, R is the original oil-gas ratio, and f w is the water content of the oil layer.

[0023] In this embodiment, it should be specifically explained that the convolutional neural network in step S04 includes an input layer, a convolution layer, a pooling layer, an activation function, a fully connected layer and an output layer; The input layer is used to input the dynamometer diagram of the pumping well; The convolution layer is used to extract the image features of the indicator diagram; the convolution kernel of adjustable size is slid with a certain step length to perform convolution operation on the input matrix data to extract the feature information of the input matrix data; the convolution kernel gradually performs point convolution operation on the input image matrix data through its matrix window of fixed width and height; the convolution kernel size must be smaller than the size of the image data matrix; the convolution kernel performs convolution operation on the input image data matrix to achieve feature extraction of the image data; The pooling layer is also called the downsampling layer, which compresses the feature image obtained by the convolution layer, reduces the dimension of the feature map, reduces the subsequent parameter calculation amount, and improves the generalization performance of the convolutional neural network. When the feature map data is compressed, the main features implied by the feature map are still retained. The feature map data is compressed, which greatly reduces the network calculation amount and helps to avoid network overfitting. Maximum pooling or average pooling can be used; The activation function adopts the Tanh function, and the formula is expressed as: ; The Tanh function can be evenly mapped in the 1 to -1 space, has negative values, and is symmetric about the origin, so it will not cause the network to be difficult to converge; The fully connected layer is used to classify the working conditions of the pumping well; the fully connected layer is located in the last layer of the convolutional neural network, each neuron is connected to all the output mappings of the previous layer, and the two-dimensional feature matrix obtained after the convolution layer and the pooling layer is processed and calculated into a one-dimensional feature vector by weighted addition; The output layer is used to output the operating fault category of the oil pumping unit.

[0024] like Figure 2 As shown, a pumping well operating condition diagnosis device for using a pumping well operating condition diagnosis method comprises: Data acquisition module: used to collect target data, including but not limited to the dynamic liquid level, stroke, stroke frequency, wellhead back pressure, wellhead casing pressure, wellhead pump plunger cross-sectional area, wellhead pump depth and ground dynamometer diagram of the pumping well; Data preprocessing module: used to preprocess the target data to obtain target data that can be directly used; the preprocessing operation includes filtering and normalizing the target data; Working condition judgment module: used to analyze and calculate the actual pump efficiency and theoretical pump efficiency of the pumping well, and calculate the pump efficiency difference index, so as to determine whether it is necessary to judge the working condition of the pumping well; Working condition fault diagnosis module: Based on the constructed convolutional neural network, the indicator diagram of the pumping well is analyzed to diagnose the working condition of the pumping well and obtain the working condition fault type of the pumping well; Data output module: used to output the working condition diagnosis results of the pumping well; if the working condition diagnosis module determines that it is not necessary to perform working condition diagnosis on the pumping well, the output working condition diagnosis result is "normal"; if the working condition diagnosis module determines that it is necessary to perform working condition diagnosis on the pumping well, the output working condition diagnosis result is "abnormal" and the working condition fault type output by the working condition fault diagnosis module is output at the same time.

[0025] In the present embodiment, it should be specifically explained that the difference between the present embodiment and the prior art lies mainly in that the present embodiment has step S03 and step S04, which is conducive to analyzing and calculating the actual pump efficiency and theoretical pump efficiency of the pumping well, and calculating the pump efficiency difference index, so as to determine whether it is necessary to make an operating condition diagnosis for the pumping well. If an operating condition diagnosis is required, the indicator diagram of the pumping well is analyzed based on the constructed convolutional neural network, so as to diagnose the operating condition of the pumping well and obtain the operating condition fault type of the pumping well; by analyzing the pump efficiency, the operating condition diagnosis of the pumping well is identified in advance, thereby improving the diagnosis efficiency; by constructing a convolutional neural network to extract features of the indicator diagram, the features of different indicator diagrams are identified, and the operating condition fault category of the pumping well is accurately obtained, thereby improving the analysis and diagnosis efficiency and accuracy of the indicator diagram, and there is no need to analyze the indicator diagram through manual recognition methods, thereby improving the accuracy of fault diagnosis.

[0026] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0027] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for diagnosing the working condition of a pumping well, characterized in that: The following steps are involved: Step S01: Acquire target data; Step S02: preprocessing the target data acquired in step S01; Step S03: Determine whether it is necessary to perform a working condition judgment: Analyze and calculate the actual pump efficiency and theoretical pump efficiency of the pumping well, and calculate the pump efficiency difference index. If the pump efficiency difference index exceeds the set threshold, it is determined that it is necessary to perform a working condition judgment, and step S04 is executed; otherwise, it is determined that it is not necessary to perform a working condition judgment, and step S05 is executed; Step S04: Analyze the dynamometer diagram of the pumping well based on the constructed convolutional neural network, thereby diagnosing the working condition of the pumping well and obtaining the working condition fault type of the pumping well; Step S05: outputting the operating condition judgment result of the pumping well.

2. A method for diagnosing the working condition of a pumping well according to claim 1, characterized in that: The formulas for analyzing and calculating the actual pump efficiency and theoretical pump efficiency of the pumping well in step S03 are respectively expressed as: , where η B is the actual pumping efficiency of the well, Q 实际 is the actual production of the oil well, Q 理论 is the theoretical production of the oil well; , where η 理 is the theoretical pump efficiency; η 冲 is the stroke loss influence coefficient, that is, the pump efficiency when the elastic expansion and contraction of the tubing and the sucker rod produces stroke loss; β is the filling degree of the pump, η 漏 is the pump leakage coefficient, that is, the influence coefficient of pump leakage on pump efficiency; η 气 The coefficient of displacement affected by liquid shrinkage, that is, the coefficient of the volume shrinkage of crude oil degassing on the pump efficiency; like , where C η is the pump efficiency difference index, when C η The value satisfies C η >YU, it is determined that the working condition judgment is required; when C η The value satisfies C η ≤YU, it is determined that no operating condition judgment is required; YU is a set threshold, and YU≤30%.

3. A method for diagnosing the working condition of a pumping well according to claim 2, characterized in that: The calculation formula of the stroke loss influence coefficient is: , where S is the stroke length of the polished rod at the suspension point, λ j is the stroke loss caused by static load, λ g The stroke loss is caused by inertial load; the stroke loss refers to the difference between the suspension point stroke and the actual stroke of the plunger.

4. A method for diagnosing the working condition of a pumping well according to claim 3, characterized in that: The calculation formula for the stroke loss caused by the static load is: , where Δp is the pressure difference between the discharge port and the suction port of the oil well pump; L is the depth of the oil well pump; f p is the cross-sectional area of ​​the oil well pump plunger; E is the elastic modulus of the material used in the oil well pump; f r is the average cross-sectional area of ​​the sucker rod string; f t is the average cross-sectional area of ​​the oil pipe; The calculation formula for the stroke loss caused by the inertial load is: , where μ is the inertia factor coefficient, , where ω is the angular velocity of the pumping unit crank, L is the depth of the pump, and a is the propagation speed of the sound wave in the sucker rod string, and a=5100.

5. A method for diagnosing the working condition of a pumping well according to claim 2, characterized in that: The calculation formula for the filling degree β of the pump is: , where X y1 is the first displacement factor affected by free gas, X y2 The second displacement factor affected by the clearance volume; , where f w is the water content of the oil layer, R is the original oil-gas ratio, R' is the dissolved oil-gas ratio under the pump suction pressure, B m is the gas volume coefficient at the sinking pressure; , where S x is the anti-stroke distance length, and S is the stroke length of the bare rod at the suspension point.

6. A method for diagnosing the working condition of a pumping well according to claim 2, characterized in that: The calculation formula of the leakage coefficient of the pump is: , where Q 漏 is the leakage between the oil well pump plunger and the pump barrel, Q 内 The flow rate of crude oil above the plunger of the oil pump; , where D is the diameter of the pump barrel, δ is the gap between the pump plunger and the pump barrel, Δp is the pressure difference between the discharge port and the suction port of the pump, γ is the dynamic viscosity of the gap liquid, and L' is the effective plunger length, that is, the length of the sealing contact part of the pump plunger; V p is the plunger stroke volume of the oil well pump; , where P 理 is the theoretical displacement of the oil well pump, , where S is the stroke length of the bare rod at the suspension point, and n is the number of pumping unit strokes.

7. A method for diagnosing the working condition of a pumping well according to claim 2, characterized in that: The calculation formula of the displacement coefficient affected by liquid shrinkage is: , where B0 is the volume coefficient of crude oil under sinking pressure, R is the original oil-gas ratio, and f w is the water content of the oil layer.

8. A pumping well operating condition diagnosis device, used for using a pumping well operating condition diagnosis method according to any one of claims 1 to 7, characterized in that: It includes a data acquisition module, a data preprocessing module, an operating condition judgment module, an operating condition fault judgment module and a data output module; The data acquisition module is used to collect target data, including but not limited to the dynamic liquid level, stroke, stroke frequency, wellhead back pressure, wellhead casing pressure, wellhead pump plunger cross-sectional area, wellhead pump depth and ground dynamometer diagram of the pumping well; The data preprocessing module is used to preprocess the target data to obtain target data that can be directly used; the preprocessing operation includes filtering and normalizing the target data; The working condition judgment module is used to analyze and calculate the actual pump efficiency and theoretical pump efficiency of the pumping well, and calculate the pump efficiency difference index, so as to determine whether it is necessary to perform working condition judgment on the pumping well; The working condition fault diagnosis module analyzes the dynamometer diagram of the pumping well based on the constructed convolutional neural network, thereby diagnosing the working condition of the pumping well and obtaining the working condition fault type of the pumping well; The data output module is used to output the working condition judgment result of the oil pumping well.

Images