Intelligent analysis method and system for oil pumping unit operating conditions

Abstract

The present application provides a method and system for intelligent analysis of the working condition of an oil pump. The method includes collecting the electrical parameter data of the oil pump, drawing the indicator diagram, and using the optimized graph convolutional neural network GCN to extract features and train the knowledge graph of the working condition; at the same time, the accuracy of the working condition fault analysis is improved by using the knowledge distillation algorithm; the real-time working condition data is analyzed, and a graded warning is performed to provide corresponding treatment measures. The present invention can improve the accuracy of real-time working condition monitoring and analysis of the oil pump through the above technology, and ensure production safety.

Description

Technical Field

[0001] The present invention relates to the field of data analysis and algorithm analysis, and in particular to an intelligent analysis method and system for oil pumping unit operating conditions. Background Art

[0002] The oil pump is an important equipment in the oil extraction process. Its working status and performance directly affect the output and benefits of the oil well. The traditional method of analyzing the working condition of the oil pump mainly relies on experience judgment and manual analysis, which has problems such as strong subjectivity and low efficiency. It lacks the characteristics of intelligence and automation and cannot meet the needs of accurate analysis and fault diagnosis under complex working conditions. With the development of the Internet of Things and big data technology, it has become possible to use the electrical parameter data of the oil pump for intelligent analysis of the working condition.

[0003] In the prior art, the dynamometer diagram can be used to directly diagnose the working condition, but this method has defects such as high cost and deformation of the equipment. At the same time, the complex working environment makes the shape of the dynamometer diagram curve changeable, which increases the difficulty of working condition diagnosis. Therefore, it is necessary to find a method with high recognition efficiency and accuracy to identify the dynamometer diagram. Therefore, the prior art has used deep learning to process the dynamometer diagram to ensure recognition accuracy. However, the existing deep learning framework has defects such as uneven feature extraction, a large number of model parameters, and high hardware cost. Summary of the invention

[0004] In order to solve the above problems, the present invention introduces an improved graph convolutional neural network to extract features and learn fault relationships from the knowledge graph of the pumping unit working conditions, and also combines the means of knowledge distillation, while cooperating with the indicator analysis of the dynamometer diagram to propose an intelligent analysis method and system for the pumping unit working conditions. The present invention specifically provides the following technical solutions:

[0005] A method for intelligent analysis of oil pumping unit working conditions, the method comprising the following steps:

[0006] S1 obtains electrical parameter data of the pumping unit, calculates the power curve of the pumping unit based on the electrical parameter data, and draws the power diagram of the pumping unit;

[0007] S2. Preprocessing the data of the indicator diagram, converting the preprocessed data of the indicator diagram into a graph structure, and constructing a working condition map of the pumping unit;

[0008] S3. Using an optimized graph convolutional neural network GCN to extract features and train the working state map of the pumping unit;

[0009] S4. Using the trained GCN to perform real-time intelligent analysis and fault diagnosis on new operating conditions generated by the pumping unit operation;

[0010] S5. Provide graded warnings for the faults and promptly provide corresponding treatment measures based on the warning strategy library.

[0011] Optionally, in S1, electrical parameter data is obtained from a sensor or monitoring device of the oil pump, and the electrical parameter data includes current, voltage, and power factor;

[0012] Draw the power diagram of the pumping unit, which is an analysis image with time as the horizontal axis and power as the vertical axis; wherein the calculation formula of the value P on the vertical axis is as follows:

[0013] ;

[0014] ;

[0015] Among them, m i is the mass of component i; g is the acceleration due to gravity; v i is the velocity of the center of mass of component i; v i The angle between the horizontal direction and the si is the distance between the center of mass of component i and the corresponding revolute pair; is the angular acceleration of the center of mass of component i; w i J is the angular velocity of the center of mass of component i; i is the equivalent moment of inertia of component i; f is the friction coefficient of the rotating pair; F f is the force acting on the revolute pair; r is the radius of the rotating axis; P is the angular velocity difference between the two components of the kinematic pair i; A is the suspension point load; v A is the suspension point speed; W b is the counterweight; V by is the velocity component of the center of mass of the counterweight in the y direction; B is the structural imbalance value; is the transmission efficiency of the belt; is the transmission efficiency of the reduction box; J is the equivalent moment of inertia; J1, J2, and J3 are the moments of inertia of the crank component, connecting rod component, and beam component around the center of mass respectively; W is the instantaneous angular velocity of the crank; , , , w2, w3 are the angular velocities of the motor, shaft III and shaft II in the reduction gearbox, connecting rod, and walking beam respectively; , , , They are the rotational inertia of the motor rotor, shaft III, shaft II, and shaft I components in the reduction box, respectively;

[0016] The calculation formula of the value t on the horizontal axis is as follows:

[0017] ;

[0018] ;

[0019] in, is the crank angle value; w i is the angular velocity.

[0020] Optionally, the S2 specifically includes the following steps:

[0021] S21. Cleaning and processing the data of the dynamometer diagram;

[0022] S22. Using a directed graph to represent the working state map of the oil pump; wherein each node in the directed graph represents a working state, including different speeds, loads, powers, and times; the edges in the directed graph represent the conversion relationship between states, including speed change rate and speed stability; weights are assigned to the nodes and the edges;

[0023] S23. Construct an operating condition map of the oil pump according to the structure of the directed graph and the properties of the nodes and edges; wherein the structure of the directed graph is converted into an adjacency matrix; the operating condition map is used to analyze the performance of the oil pump under different operating conditions, predict fault conditions, and optimize operation strategies.

[0024] Optionally, in S3, the optimized graph convolutional neural network GCN includes a graph structure module, a graph attention module and a fault knowledge structure module;

[0025] In order to preserve the fault graph structure information of the pumping unit, the degree of the fault entity is integrated into the word vector of the fault entity; for each fault entity, the centrality code of the fault entity is added to the word vector, and the embedding vector x of the fault entity is i Defined as:

[0026] ;

[0027] in, is the embedding vector of the i-th fault entity after integrating fault knowledge; It is a vector concatenation operation; and are the in-degree vector and out-degree vector of the i-th fault entity respectively; Unify the word vectors of the head entity and the tail entity into the word vector of the fault entity; Unify the embedding vectors of the head entity and the tail entity into the embedding vector of the fault entity;

[0028] The graph attention module includes node-level attention and semantic-level attention; the final fault entity x after N paths F It is expressed as:

[0029] ;

[0030] in, is the set of all paths; is the number of paths in the knowledge graph of pumping unit fault diagnosis; Φ is the path from fault entity j to fault entity i; is the transpose of the semantic level attention vector; W is the transformation matrix; is the embedding vector of fault entity i; b is the bias; is the activation function; K is the number of times the node-level attention operation is called; is the set of neighbor fault entities of the i-th fault entity; Deep neural networks for computing node-level attention; is the embedding vector of the neighbor fault entity.

[0031] Optionally, the input of the optimized graph convolutional neural network GCN is the fault knowledge graph of the oil pump and the association matrix of the fault causes; and the output is an inferred sequence of the fault causes.

[0032] Optionally, the training further includes:

[0033] A teacher model is constructed and trained using historical failure data of the oil pumping unit;

[0034] Build a student model and train it using the optimized GCN;

[0035] The knowledge of the teacher model is transferred to the student model using a knowledge distillation algorithm.

[0036] Optionally, in S5, the faults are divided into primary faults, secondary faults, and tertiary faults according to the severity and impact scope of the faults; the early warning strategy library contains early warning rules and treatment measures for various faults; according to the operating conditions and fault characteristics of the oil pump, the corresponding early warning rules are matched to determine the type and level of the fault; an alarm is issued to the operator through sound, light, and text prompts, and the corresponding fault information and treatment measures are displayed in the monitoring system.

[0037] Optionally, S1 further includes: judging the working state of the pumping unit according to the characteristics of the power curve; calculating the wax accumulation degree at the wellhead and the balance degree of the pumping unit according to the working state, the historical electrical parameter data of the pumping unit and the curve fluctuation; and issuing an initial warning according to the changes in the wax accumulation degree and the balance degree. The triggering conditions of the initial warning specifically include: when the wax accumulation degree exceeds 10% of the historical average value, and / or when the balance degree exceeds 20% of the historical average value.

[0038] Optionally, the calculation process of the wax deposition degree is as follows: taking the upward current when the oil well is not waxed, ie, the theoretical current, as a reference, the measured upward current is compared with the theoretical current to determine the wax deposition degree of the oil well.

[0039] The calculation process of the balance degree is as follows:

[0040] Determine a first time period corresponding to when the oil pump is in an upstroke and a second time period corresponding to when the oil pump is in a downstroke from the acceleration curve of the oil pump; obtain a maximum current value in the first time period and a maximum current value in the second time period from the current curve of the oil pump; and correspond the ratio of the maximum current value in the first time period to the maximum current value in the second time period as the balance degree of the oil pump.

[0041] The present invention also provides an intelligent analysis system for operating conditions of an oil pumping unit, the system comprising:

[0042] A data acquisition module is used to collect electrical parameter data of the oil pump;

[0043] A dynamometer module, which realizes the dynamometer analysis function of the oil pumping unit through the electrical parameters;

[0044] A neural network module, used for learning and training the knowledge graph of the oil pumping unit operating conditions;

[0045] A knowledge distillation module, used for optimizing the learning knowledge function of the neural network module;

[0046] The fault diagnosis and early warning module is used to analyze the real-time electrical parameters generated by the oil pump, perform fault diagnosis and early warning.

[0047] The present invention has the following beneficial technical effects: The present invention provides an intelligent analysis method and system for the working condition of an oil pumping unit. The present invention converts the working condition data into a graph structure through the indicator analysis and preprocessing of the indicator diagram, which more intuitively reflects the working condition map of the oil pumping unit. This processing method can better retain the fault graph structure information and provide reliable data support for subsequent fault diagnosis. In this regard, the present invention designs an improved graph convolutional neural network GCN, which extracts and trains features for the knowledge map of the working condition of the oil pumping unit, effectively improving the accuracy and efficiency of fault diagnosis; compared with the traditional deep learning framework, the GCN of the present invention has better parameter settings, reduces the model complexity, and reduces the hardware cost; at the same time, the present invention applies knowledge distillation technology to the model training process, and by constructing a teacher model and a student model, the knowledge of the teacher model is effectively transferred to the student model, further improving the learning efficiency and accuracy of the model, which not only helps to optimize the model, but also enhances the generalization ability of the model in practical applications. The present invention also performs graded warnings according to the severity and impact range of the fault, and is equipped with a corresponding warning strategy library. Through various prompt methods such as sound, light, and text, alarms are promptly issued to operators, and fault information and handling measures are displayed in the monitoring system, effectively improving the operator's ability to respond to faults and handling efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0049] Figure 1 A flow chart of a method for intelligent analysis of oil pumping unit operating conditions provided by an embodiment of the present invention.

[0050] Figure 2 A block diagram of an intelligent analysis system for oil pumping unit operating conditions provided by an embodiment of the present invention.

[0051] Figure 3 It is an example of the indicator diagram of the present invention. DETAILED DESCRIPTION

[0052] The embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0053] The following specific examples are used to illustrate the embodiments of the present application. Those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other in the absence of conflict. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work belong to the scope of protection of the present application.

[0054] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein may be embodied in a wide variety of forms, and any labeled structures and / or functions described herein are merely illustrative. Based on the present application, it should be understood by those skilled in the art that an aspect described herein may be implemented independently of any other aspect, and two or more of these aspects may be combined in various ways. For example, any number and aspect set forth herein may be used to implement an apparatus and / or practice a method.

[0055] It should also be noted that the illustrations provided in the following embodiments are only used to schematically illustrate the basic concept of the present application.

[0056] Additionally, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, it will be understood by those skilled in the art that the examples can be practiced without these marked details.

[0057] The purpose of the present invention is to provide a method and system for intelligent analysis of the working condition of an oil pumping unit, aiming to ensure the real-time performance and accuracy of the analysis of the working condition of the oil pumping unit.

[0058] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0059] Reference Figure 1 , showing an intelligent analysis method of the working condition of an oil pumping unit according to an embodiment of the present application, the method comprises the following steps:

[0060] S1 obtains electrical parameter data of the pumping unit, calculates the power curve of the pumping unit based on the electrical parameter data, and draws the power diagram of the pumping unit;

[0061] Optionally, in S1, electrical parameter data is obtained from a sensor or monitoring device of the oil pump, and the electrical parameter data includes current, voltage, and power factor;

[0062] Draw the power diagram of the pumping unit, which is an analysis image with time as the horizontal axis and power as the vertical axis; wherein the calculation formula of the value P on the vertical axis is as follows:

[0063] ;

[0064] ;

[0065] Among them, m i is the mass of component i; g is the acceleration due to gravity; v i is the velocity of the center of mass of component i; v i The angle between the horizontal direction and the si is the distance between the center of mass of component i and the corresponding revolute pair; is the angular acceleration of the center of mass of component i; w i J is the angular velocity of the center of mass of component i; i is the equivalent moment of inertia of component i; f is the friction coefficient of the rotating pair; F f is the force acting on the revolute pair; r is the radius of the rotating axis; P is the angular velocity difference between the two components of the kinematic pair i; A is the suspension point load; v A is the suspension point speed; W b is the counterweight; V by is the velocity component of the center of mass of the counterweight in the y direction; B is the structural imbalance value; is the transmission efficiency of the belt; is the transmission efficiency of the reduction box; J is the equivalent moment of inertia; J1, J2, and J3 are the moments of inertia of the crank component, connecting rod component, and beam component around the center of mass respectively; W is the instantaneous angular velocity of the crank; , , , w2, w3 are the angular velocities of the motor, shaft III and shaft II in the reduction gearbox, connecting rod, and walking beam respectively; , , , They are the rotational inertia of the motor rotor, shaft III, shaft II, and shaft I components in the reduction box, respectively;

[0066] The calculation formula of the value t on the horizontal axis is as follows:

[0067] ;

[0068] ;

[0069] in, is the crank angle value; w i is the angular velocity.

[0070] Optionally, the typical operating conditions of the oil pump include but are not limited to: normal operating conditions, insufficient fluid supply, oil pipe leakage, rod breakage, wax deposition, sand production, good balance, overbalance, and underbalance.

[0071] Optionally, S1 also includes: judging the working state of the oil pump according to the characteristics of the power curve; calculating the wax deposition degree at the wellhead and the balance degree of the oil pump according to the working state, the historical electrical parameter data of the oil pump and the curve fluctuation; and issuing an initial warning according to the changes in the wax deposition degree and the balance degree.

[0072] Optionally, the calculation process of the wax deposition degree is as follows: taking the upward current when the oil well is not waxed, ie, the theoretical current, as a reference, the measured upward current is compared with the theoretical current to determine the wax deposition degree of the oil well.

[0073] Optionally, the calculation process of the balance degree is as follows: determine the first time period corresponding to when the pumping unit is in an upstroke and the second time period corresponding to when the pumping unit is in a downstroke from the acceleration curve of the pumping unit; obtain the maximum current value in the first time period and the maximum current value in the second time period from the current curve of the pumping unit; and correspond the ratio between the maximum current value in the first time period and the maximum current value in the second time period to the balance degree of the pumping unit. Therefore, under normal working conditions, the current ratio should be between 0.8-1.2.

[0074] An increase in the amount of waxed oil may mean that there is a waxing problem inside the pump, which causes the oil flow to be blocked, thus affecting the working efficiency and output of the pump. At the same time, changes in balance may reflect the imbalance of the pump during the upstroke and downstroke. For example, unbalanced movement caused by wear or failure of internal parts of the pump may lead to increased vibration and reduced working efficiency of the pump. By monitoring the changes in waxing and balance, possible faults of the pump can be discovered in time, so that preliminary warnings can be given and corresponding maintenance measures can be taken to avoid more serious failures.

[0075] For example, the triggering conditions of the initial warning specifically include: when the wax accumulation degree exceeds 10% of the historical average value, and / or when the balance degree exceeds 20% of the historical average value.

[0076] Figure 3 Several examples of dynamometer diagrams are given, among which Figure 3 (a) is the indicator diagram under normal working conditions; Figure 3(b) is the indicator diagram under oil pipe leakage; Figure 3 (c) is the indicator diagram of wax deposition in the oil pipe; Figure 3 (d) is the current diagram of the insufficient fluid supply condition. The overall normal condition is similar, showing a lying "8" shape. The maximum current during the upstroke process is close to the maximum current during the downstroke process. Within the current balance range, the pump is in a balanced state as a whole.

[0077] S2. Preprocessing the data of the indicator diagram, converting the preprocessed data of the indicator diagram into a graph structure, and constructing an operating condition map of the oil pumping unit.

[0078] Optionally, the S2 specifically includes the following steps:

[0079] S21. Cleaning and processing the data of the dynamometer diagram;

[0080] S22. Using a directed graph to represent the working state map of the oil pump; wherein each node in the directed graph represents a working state, including different speeds, loads, powers, and times; the edges in the directed graph represent the conversion relationship between states, including speed change rate and speed stability; weights are assigned to the nodes and the edges;

[0081] S23. Construct an operating condition map of the oil pump according to the structure of the directed graph and the properties of the nodes and edges; wherein the structure of the directed graph is converted into an adjacency matrix; the operating condition map is used to analyze the performance of the oil pump under different operating conditions, predict fault conditions, and optimize operation strategies.

[0082] Preferably, the step of data preprocessing includes data cleaning, such as checking and removing abnormal data points caused by sensor failure or other reasons. For example, when electrical parameter data such as current and voltage exceed the upper and lower limits of the normal working range, it can be judged as an abnormal value. For missing data points, use reasonable interpolation or estimation methods to fill in to ensure the continuity of the data. Convert the units, magnitudes, etc. of the electrical parameter data into a unified standard. Convert the original electrical parameter data into a format suitable for model input, such as graph structure data, which requires defining the attributes and relationships of nodes and edges. Assign specific weights to the nodes and edges in the directed graph. For example, according to the fluctuation of the electrical parameter data, the weight of the working state represented by the node can be set; according to the conversion frequency or historical data between different working states, the weight of the conversion relationship represented by the edge can be set. For example, if the current fluctuation of the pumping unit is large within a certain period of time, then when constructing a directed graph, the weight of the corresponding node may be higher, reflecting its importance or degree of influence. Data preprocessing also includes but is not limited to: data dimensionality reduction processing and data smoothing processing.

[0083] S3. Use an optimized graph convolutional neural network (GCN) to perform feature extraction and training on the operating condition map of the oil pump.

[0084] Optionally, in S3, the optimized graph convolutional neural network GCN includes a graph structure module, a graph attention module and a fault knowledge structure module;

[0085] In order to preserve the fault graph structure information of the pumping unit, the degree of the fault entity is integrated into the word vector of the fault entity; for each fault entity, the centrality code of the fault entity is added to the word vector, and the embedding vector x of the fault entity is i Defined as:

[0086] ;

[0087] in, is the embedding vector of the i-th fault entity after integrating fault knowledge; It is a vector concatenation operation; and are the in-degree vector and out-degree vector of the i-th fault entity respectively; Unify the word vectors of the head entity and the tail entity into the word vector of the fault entity; Unify the embedding vectors of the head entity and the tail entity into the embedding vector of the fault entity;

[0088] The graph attention module includes node-level attention and semantic-level attention; the final fault entity x after N paths F It is expressed as:

[0089] ;

[0090] in, is the set of all paths; is the number of paths in the knowledge graph of pumping unit fault diagnosis; Φ is the path from fault entity j to fault entity i; is the transpose of the semantic level attention vector; W is the transformation matrix; is the embedding vector of fault entity i; b is the bias; is the activation function; K is the number of times the node-level attention operation is called; is the set of neighbor fault entities of the i-th fault entity; Deep neural networks for computing node-level attention; is the embedding vector of the neighbor fault entity.

[0091] Optionally, the input of the optimized graph convolutional neural network GCN is the fault knowledge graph of the oil pump and the association matrix of the fault causes; and the output is an inferred sequence of the fault causes.

[0092] Preferably, the training further comprises:

[0093] A teacher model is constructed and trained using the historical failure data of the oil pump; the model can accurately infer the cause of the failure.

[0094] A student model is constructed, and the optimized GCN is used to train the fault knowledge graph of the pumping unit and the association matrix of the fault-cause to obtain a preliminary student model.

[0095] The knowledge of the teacher model is transferred to the student model using a knowledge distillation algorithm.

[0096] Optionally, the teacher model is used to infer a set of training samples to obtain the output of the teacher model. The same training samples are inferred using the student model to obtain the output of the student model. The difference between the output of the teacher model and the output of the student model is calculated as the loss function of knowledge distillation, where the loss function can be a weighted combination of cross entropy loss and mean square error loss. During the training process, the output of the teacher model is used as an additional supervisory signal and optimized together with the output of the student model. Through repeated iterative training, the student model is continuously optimized to gradually approach the performance of the teacher model. Through knowledge distillation, the student model can obtain more knowledge and experience from the teacher model, improving its accuracy and robustness in inferring the cause of the fault.

[0097] S5. Provide graded warnings for the faults and promptly provide corresponding treatment measures based on the warning strategy library.

[0098] Optionally, in S5, the faults are divided into primary faults, secondary faults, and tertiary faults according to the severity and impact scope of the faults; the early warning strategy library contains early warning rules and treatment measures for various faults; according to the operating conditions and fault characteristics of the oil pump, the corresponding early warning rules are matched to determine the type and level of the fault; an alarm is issued to the operator through sound, light, and text prompts, and the corresponding fault information and treatment measures are displayed in the monitoring system.

[0099] Preferably, the early warning can be achieved through but not limited to the following methods: using a high-volume, urgent sound alarm to attract the attention of the operator to indicate the severity of the fault; using flashing, color changing or special light signals to attract the attention of the operator to indicate the occurrence of a fault; displaying obvious fault information and handling measures on the display screen of the monitoring system so that the operator can quickly understand the type and level of the fault and take appropriate measures.

[0100] For example, take the processing method in the underbalanced fault scenario as an example:

[0101] Level 1 Fault (Severe): If the pumping unit has a serious underbalance problem, it may cause severe vibration and damage.

[0102] Send out high-frequency and urgent alarm sounds; flash red warning lights; display large-font warning messages in the monitoring system.

[0103] Stop the operation of the pumping unit immediately and cut off the power supply. Contact maintenance personnel or professional teams as soon as possible, provide fault information and request emergency maintenance. Before waiting for the maintenance personnel to arrive, take necessary safety measures, such as blocking the faulty equipment and clearing the surrounding dangerous objects.

[0104] Level 2 Fault (Moderate): If the pumping unit is underbalanced, you can try simple adjustments and calibration.

[0105] Send out medium-frequency and urgent alarm sounds; flash yellow warning lights; display warning messages in the monitoring system; provide brief handling suggestions and ways to contact maintenance personnel.

[0106] Level 3 Fault (Minor): If the pump has a minor underbalance problem, you can try simple adjustments and calibration.

[0107] Low-frequency and urgent alarm sounds; flashing orange warning lights; displaying warning messages in the monitoring system. Treatment reminder: Provides a way to contact maintenance personnel in an emergency and recommends taking necessary safety measures to prevent further damage or accidents.

[0108] In addition, measures also include ensuring the safety of operators, such as wearing personal protective equipment; checking the balancing device, including checking whether the balancing device of the oil pump, such as the balancing block, balancing shaft, etc., is loose, worn or damaged; according to the inspection results, make corresponding repairs or replace the balancing device; after repairing or replacing the balancing device, perform balancing debugging to ensure that the balancing performance of the oil pump returns to normal; after completing the repair and debugging, perform test operation to ensure the normal operation of the oil pump; record the fault handling process and results, and report to relevant personnel.

[0109] Reference Figure 2 , shows an intelligent analysis system for the working condition of an oil pumping unit according to an embodiment of the present application, the system comprising:

[0110] A data acquisition module is used to collect electrical parameter data of the oil pump;

[0111] A dynamometer module, which realizes the dynamometer analysis function of the oil pumping unit through the electrical parameters;

[0112] A neural network module, used for learning and training the knowledge graph of the oil pumping unit operating conditions;

[0113] A knowledge distillation module, used for optimizing the learning knowledge function of the neural network module;

[0114] The fault diagnosis and early warning module is used to analyze the real-time electrical parameters generated by the oil pump, perform fault diagnosis and early warning.

[0115] Optionally, the system also includes but is not limited to:

[0116] The data storage module is used to upload the collected data to the cloud for storage and analysis.

[0117] The user interface module is used to display analysis results and provide user interaction functions.

[0118] Remote monitoring module, used to remotely monitor the working status of the oil pump.

[0119] The control module is used to control the operating status of the oil pump.

[0120] Communication module, used for data exchange with external devices.

[0121] The above are exemplary embodiments disclosed in the present invention, but it should be noted that various changes and modifications may be made without departing from the scope disclosed in the embodiments of the present invention as defined in the claims. The functions, steps and / or actions of the method claims according to the disclosed embodiments described herein do not need to be performed in any marked order. In addition, although the elements disclosed in the embodiments of the present invention may be described or required in individual form, they may also be understood as multiple unless explicitly limited to the singular.

[0122] In this specification, the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the embodiments described later, the description is relatively simple, and the relevant parts can be referred to the partial description of the previous embodiments.

[0123] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A method for intelligent analysis of oil pumping unit operating conditions, characterized in that: The method comprises the following steps: S1 obtains the electrical parameter data of the pumping unit, calculates the power curve of the pumping unit based on the electrical parameter data, and draws the indicator diagram of the pumping unit; S2. Preprocessing the data of the indicator diagram, converting the preprocessed data of the indicator diagram into a graph structure, and constructing a working condition map of the pumping unit; S3. Using an optimized graph convolutional neural network GCN to extract features and train the working condition map of the pumping unit; S4. Using the trained GCN to perform real-time intelligent analysis and fault diagnosis of new operating conditions generated by the pumping unit operation; S5. Provide graded warnings for the faults and give timely reminders of corresponding treatment measures based on the warning strategy library; In S1, electrical parameter data is obtained from a sensor or monitoring device of the oil pump, wherein the electrical parameter data includes current, voltage, and power factor; Draw the power diagram of the pumping unit, which is an analysis image with time as the horizontal axis and power as the vertical axis; wherein the calculation formula of the value P on the vertical axis is as follows: ； ； Among them, m i is the mass of component i; g is the acceleration due to gravity; v i is the velocity of the center of mass of component i; v i The angle between the horizontal direction and the si is the distance between the center of mass of component i and the corresponding revolute pair; is the angular acceleration of the center of mass of component i; w i J is the angular velocity of the center of mass of component i; i is the equivalent moment of inertia of component i; f is the friction coefficient of the rotating pair; F f is the force acting on the revolute pair; r is the radius of the revolving axis; P is the angular velocity difference between the two components connected by kinematic pair i; A is the suspension point load; v A is the suspension point speed; W b is the balance weight; V by is the velocity component of the center of mass of the counterweight in the y direction; B is the structural imbalance value; is the transmission efficiency of the belt; is the transmission efficiency of the reduction box; J is the equivalent moment of inertia; J1, J2, and J3 are the moments of inertia of the crank component, the connecting rod component, and the beam component around the center of mass respectively; W is the instantaneous angular velocity of the crank; , , , w2, w3 are the angular velocities of the motor, shaft III and shaft II in the reduction gearbox, connecting rod, and walking beam respectively; , , , They are the rotational inertia of the motor rotor, shaft III, shaft II, and shaft I components in the reduction box, respectively; The calculation formula of the value t on the horizontal axis is as follows: ； ； in, is the crank angle value; w i is the angular velocity.

2. The method for intelligent analysis of oil pumping unit operating conditions according to claim 1, characterized in that: The S2 specifically includes the following steps: S21. Cleaning and processing the data of the dynamometer diagram; S22. Using a directed graph to represent the working state map of the oil pump; wherein each node in the directed graph represents a working state, including different speeds, loads, powers, and times; the edges in the directed graph represent the conversion relationship between states, including speed change rate and speed stability; weights are assigned to the nodes and the edges; S23. Construct an operating condition map of the oil pump according to the structure of the directed graph and the properties of the nodes and edges; wherein the structure of the directed graph is converted into an adjacency matrix; the operating condition map is used to analyze the performance of the oil pump under different operating conditions, predict fault conditions, and optimize operation strategies.

3. The method for intelligent analysis of oil pumping unit operating conditions according to claim 2, characterized in that: In S3, the optimized graph convolutional neural network GCN includes a graph structure module, a graph attention module and a fault knowledge structure module; In order to preserve the fault graph structure information of the pumping unit, the degree of the fault entity is integrated into the word vector of the fault entity; for each fault entity, the centrality code of the fault entity is added to the word vector, and the embedding vector x of the fault entity is i Defined as: ； in, is the embedding vector of the i-th fault entity after integrating fault knowledge; It is a vector concatenation operation; and are the in-degree vector and out-degree vector of the i-th fault entity respectively; Unify the word vectors of the head entity and the tail entity into the word vector of the fault entity; Unify the embedding vectors of the head entity and the tail entity into the embedding vector of the fault entity; The graph attention module includes node-level attention and semantic-level attention; the final fault entity x after N paths F It is expressed as: ； in, is the set of all paths; is the number of paths in the knowledge graph of pumping unit fault diagnosis; Φ is the path from fault entity j to fault entity i; is the transpose of the semantic level attention vector; W is the transformation matrix; is the embedding vector of fault entity i; b is the bias; is the activation function; K is the number of times the node-level attention operation is called; is the set of neighbor fault entities of the i-th fault entity; Deep neural networks for computing node-level attention; is the embedding vector of the neighbor fault entity.

4. The method for intelligent analysis of oil pumping unit operating conditions according to claim 3, characterized in that: The input of the optimized graph convolutional neural network GCN is the fault knowledge graph of the oil pump and the association matrix of the fault and the cause; the output is the inferred sequence of the fault cause.

5. The method for intelligent analysis of oil pumping unit operating conditions according to claim 4, characterized in that: The training also includes: A teacher model is constructed and trained using historical failure data of the oil pumping unit; Build a student model and train it using the optimized GCN; The knowledge of the teacher model is transferred to the student model using a knowledge distillation algorithm.

6. The method for intelligent analysis of oil pumping unit operating conditions according to claim 1, characterized in that: In S5, the faults are divided into primary faults, secondary faults, and tertiary faults according to the severity and impact range of the faults; the warning strategy library contains warning rules and treatment measures for various faults; according to the operating conditions and fault characteristics of the oil pump, the corresponding warning rules are matched to determine the type and level of the fault; an alarm is issued to the operator through sound, light, and text prompts, and the corresponding fault information and treatment measures are displayed in the monitoring system.

7. The method for intelligent analysis of oil pumping unit operating conditions according to claim 1, characterized in that: The S1 also includes: judging the working state of the oil pump according to the characteristics of the power curve; calculating the wax accumulation degree at the wellhead and the balance degree of the oil pump according to the working state, the historical electrical parameter data of the oil pump and the curve fluctuation; making an initial warning according to the changes in the wax accumulation degree and the balance degree; the triggering conditions of the initial warning specifically include: when the wax accumulation degree exceeds 10% of the historical average value, and / or when the balance degree exceeds 20% of the historical average value.

8. The method for intelligent analysis of oil pumping unit operating conditions according to claim 7, characterized in that: The calculation process of the wax deposition degree is as follows: taking the upward current when the oil well is not waxed, that is, the theoretical current, as a reference, the measured upward current is compared with it to determine the wax deposition degree of the oil well; The calculation process of the balance degree is as follows: determine the first time period corresponding to the upstroke of the oil pump and the second time period corresponding to the downstroke of the oil pump from the acceleration curve of the oil pump; obtain the maximum current value in the first time period and the maximum current value in the second time period from the current curve of the oil pump; and correspond the ratio between the maximum current value in the first time period and the maximum current value in the second time period as the balance degree of the oil pump.

9. An intelligent analysis system for operating conditions of an oil pumping unit, used to implement an intelligent analysis method for operating conditions of an oil pumping unit according to any one of claims 1 to 8, characterized in that: The system comprises: A data acquisition module is used to collect electrical parameter data of the oil pump; A dynamometer module, which realizes the dynamometer analysis function of the oil pumping unit through the electrical parameters; A neural network module, used for learning and training the knowledge graph of the oil pumping unit operating conditions; A knowledge distillation module, used for optimizing the learning knowledge function of the neural network module; The fault diagnosis and early warning module is used to analyze the real-time electrical parameters generated by the oil pump, perform fault diagnosis and early warning.

Images