Pumping unit fault detection method, device, equipment, medium and product

Abstract

The invention discloses an oil pumping unit fault detection method and device, equipment, a medium and a product, and relates to the field of fault detection.The method comprises the steps that a to-be-detected indicator diagram of an oil pumping unit is obtained; performing feature extraction and classification on the indicator diagram to be detected by adopting a fault diagnosis model, and determining the working condition type of the oil pumping unit; the working condition type is a normal working condition or various fault working conditions; wherein the fault diagnosis model is obtained by training an improved Inception V3 network by adopting a training sample set in advance; according to the improved Inception V3 network, an attention mechanism is added behind a full connection layer of the Inception V3 network; the training sample set comprises a plurality of sample indicator diagrams and a working condition type label of each sample indicator diagram. The accuracy and efficiency of fault detection of the oil pumping unit are improved, and the method is suitable for fault detection under complex working conditions.

Description

Technical Field

[0001] The present application relates to the field of fault detection, and in particular to a method, device, equipment, medium and product for detecting oil pump faults based on an Inception V3 network. Background Art

[0002] With the increasing shortage of oil resources and the increasing difficulty of oil extraction, the stability and reliability of oil extraction equipment have become particularly important. As a key equipment for oil extraction, the operating status of the oil pumping unit directly affects the oil production efficiency. Traditionally, the dynamometer diagram is used to analyze the working status of the oil pumping unit, but manual interpretation is subjective and has errors. Summary of the invention

[0003] The purpose of this application is to provide a method, device, equipment, medium and product for detecting oil pump faults, which can improve the accuracy and efficiency of oil pump fault detection and is suitable for fault detection under complex working conditions.

[0004] To achieve the above objectives, this application provides the following solutions: In a first aspect, the present application provides a method for detecting a pumping unit fault, comprising: Obtain the indicator diagram of the oil pumping unit to be tested; The fault diagnosis model is used to extract and classify the characteristics of the dynamometer diagram to be detected, and the working condition type of the oil pumping unit is determined; the working condition type is a normal working condition or various fault working conditions; Among them, the fault diagnosis model is obtained by pre-training the improved Inception V3 network with a training sample set; the improved Inception V3 network is obtained by adding an attention mechanism after the fully connected layer of the Inception V3 network; the training sample set includes multiple sample dynamometer diagrams and the operating condition type label of each sample dynamometer diagram.

[0005] In a second aspect, the present application provides a pumping unit fault detection device comprising: The indicator diagram acquisition module is used to obtain the indicator diagram of the pumping unit to be tested; A classification module is used to extract and classify the characteristics of the dynamometer diagram to be detected by using a fault diagnosis model, and determine the working condition type of the oil pump; the working condition type is a normal working condition or various fault working conditions; Among them, the fault diagnosis model is obtained by pre-training the improved Inception V3 network with a training sample set; the improved Inception V3 network is obtained by adding an attention mechanism after the fully connected layer of the Inception V3 network; the training sample set includes multiple sample dynamometer diagrams and the operating condition type label of each sample dynamometer diagram.

[0006] In a third aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned method for detecting oil pump faults.

[0007] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which implements the above-mentioned method for detecting oil pump faults when executed by a processor.

[0008] In a fifth aspect, the present application provides a computer program product, including a computer program, which implements the above-mentioned method for detecting oil pump faults when executed by a processor.

[0009] According to the specific embodiments provided in this application, this application has the following technical effects: The present application provides a method, device, equipment, medium and product for detecting oil pump faults, which adopts an improved Inception V3 network for detecting oil pump faults. The Inception V3 network has powerful feature extraction capabilities and excellent generalization performance. After adding the attention mechanism, the improved Inception V3 network can automatically focus on important features and is more suitable for processing complex tasks such as compound fault detection. At the same time, it reduces manual intervention, reduces the risk of misjudgment, improves the accuracy of identifying oil pump operating conditions, and can also judge compound operating conditions on the basis of fault detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, 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.

[0011] Figure 1 This is an application environment diagram of a method for detecting oil pump faults in one embodiment of the present application.

[0012] Figure 2 An overall flow chart of a method for detecting oil pump faults provided in one embodiment of the present application.

[0013] Figure 3 A detailed flow chart of a method for detecting oil pump faults provided in one embodiment of the present application.

[0014] Figure 4 Schematic diagram of the structure of the improved Inception V3 network in one embodiment of the present application.

[0015] Figure 5 This is a power diagram under normal working conditions in an embodiment of the present application.

[0016] Figure 6 This is a power diagram under insufficient liquid supply conditions in one embodiment of the present application.

[0017] Figure 7 This is a power diagram under gas-affected working conditions in one embodiment of the present application.

[0018] Figure 8 This is a performance diagram of an oil well under sand production conditions in one embodiment of the present application.

[0019] Fig. 9 A schematic diagram of functional modules of a pumping unit fault detection device provided in one embodiment of the present application.

[0020] Fig.10 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0022] With the development of deep learning technology, it has shown strong capabilities in the fields of image processing and pattern recognition. Through intelligent processing of the indicator diagram by deep learning algorithms, various operating states of the pumping unit can be automatically identified and classified, including normal operating conditions and fault conditions. This method not only improves the accuracy and efficiency of pumping unit fault detection, but also reduces manual intervention and reduces the risk of misjudgment. This application combines the indicator diagram processing of deep learning technology to provide an intelligent solution for pumping unit compound fault detection based on single fault detection.

[0023] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0024] The pumping unit fault detection method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be set up separately, or it can be integrated on the server 104, or it can be placed on the cloud or other servers. The terminal 102 can send the indicator diagram to be detected to the server 104. After the server 104 receives the indicator diagram to be detected, it uses the fault diagnosis model to extract and classify the features of the indicator diagram to be detected to determine the operating condition type of the oil pump. The server 104 can feedback the operating condition type of the oil pump to the terminal 102. In addition, in some embodiments, the oil pump fault detection method can also be implemented by the server 104 or the terminal 102 alone.

[0025] The terminal 102 may be, but is not limited to, various desktop computers, laptop computers, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart vehicle-mounted devices, etc. The portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers, or may be a cloud server.

[0026] In an exemplary embodiment, Figure 2 and Figure 3 As shown, a method for detecting oil pump failure is provided. The method is executed by a computer device, and can be executed by a computer device such as a terminal or a server alone, or by a terminal and a server together. In the embodiment of the present application, the method is applied to Figure 1 The server 104 in the example is used for explanation, and the steps include the following steps 201 to 202.

[0027] Step 201, obtaining the dynamometer diagram of the oil pump to be tested.

[0028] In an exemplary embodiment, step 201 includes the following steps 301 to 303 .

[0029] Step 301, obtaining the operation data of the oil pump, wherein the operation data includes the load, displacement, well number and time of multiple sampling points.

[0030] Step 302, the operation data is subjected to denoising and normalization processing in sequence to obtain pre-processed operation data. Specifically, invalid, erroneous and repeated information in the operation data is removed, and high-frequency noise is removed using filtering technology, and then normalization processing is performed to make the processed operation data meet the input requirements of the fault diagnosis model. Among them, the normalization processing adopts the minimum-maximum normalization method to scale the data to between 0 and 1 to ensure that the fault diagnosis model can correctly interpret the input data.

[0031] Step 303: Draw the dynamometer diagram to be tested based on the pre-processed operation data, with displacement as the abscissa and load as the ordinate, and further adjust the dynamometer diagram to be tested to a space size suitable for the input of the fault diagnosis model.

[0032] Step 202: Use the fault diagnosis model to extract and classify the characteristics of the dynamometer diagram to be detected, and determine the working condition type of the oil pumping unit. The working condition type is a normal working condition or various fault working conditions.

[0033] The fault diagnosis model is obtained by pre-training the improved Inception V3 network using a training sample set. The improved Inception V3 network is obtained by adding an attention mechanism after the fully connected layer of the Inception V3 network. The training sample set includes multiple sample dynamometer diagrams and the operating condition type label of each sample dynamometer diagram. The Inception V3 network is the third version of the deep learning network (Inception).

[0034] This application adds an attention mechanism to the commonly used Inception V3 network to help the fault diagnosis model automatically focus on important features. Since there are too many Inception modules in the Inception V3 network, adding an attention mechanism after each Inception module is too much work, so this application only adds an attention mechanism after the fully connected layer to focus on feature extraction.

[0035] In an exemplary embodiment, the fault diagnosis model includes a feature extractor, a fully connected layer, an attention mechanism layer, a similarity calculation layer and an output layer.

[0036] The improved Inception V3 network mainly consists of two main parts: the backbone network and the classification network. The backbone network is responsible for extracting high-level features of the image, and the classification network is responsible for classifying and outputting these feature results. The backbone network adopts a multi-scale convolution design, which can simultaneously capture image features at different scales, so that the improved Inception V3 network has better generalization ability when processing images of different sizes. This application uses the backbone network as a feature extractor. The classification network mainly includes a fully connected layer, an attention mechanism layer, a similarity calculation layer, and an output layer.

[0037] In an exemplary embodiment, Figure 4As shown in the figure, the structure of the improved Inception V3 network is: input → 3×3 convolution layer → ReLU activation function → 3×3 convolution layer → ReLU activation function → 3×3 convolution layer → ReLU activation function → maximum pooling layer → 1×1 convolution layer → 2×2 convolution layer → ... multiple Inception modules ... → global average pooling layer → fully connected layer → ReLU activation function → → attention mechanism layer → similarity calculation layer → output layer. Among them, each Inception module includes 1×1 convolution, 3×3 convolution, 5×5 convolution and maximum pooling layer.

[0038] The dynamometer diagram to be tested is sent to the fault diagnosis model, triggering the forward propagation mechanism. During the forward propagation process, the fault diagnosis model sequentially passes through multiple convolutional layers, pooling layers, and Inception modules to gradually extract high-level features of the dynamometer diagram to be tested. These features are continuously refined and extracted in the fault diagnosis model until they reach the fully connected layer.

[0039] Step 202 includes the following steps 401 to 405 .

[0040] Step 401: extract features from the dynamometer diagram to be detected by the feature extractor to obtain a high-dimensional feature vector.

[0041] The core of the feature extractor consists of multiple convolutional layers and multiple Inception modules.

[0042] The Inception module combines a variety of convolution kernels, including 1×1, 3×3, and 5×5, to extract features at different levels. Multiple outputs can be obtained, and then multiple outputs are combined and transferred to the next Inception module. The combination of convolution kernels in each Inception module is different. Each convolution layer extracts local features in the input image through convolution operations. ReLU is added as an activation function after each convolution layer.

[0043] In order to reduce the dimension and computational complexity of the feature map, multiple pooling layers are embedded in the feature extractor, including the maximum pooling layer and the average pooling layer. The pooling layer reduces the size of the feature map through downsampling operations while retaining important information to prevent overfitting.

[0044] Step 402: Map the high-dimensional feature vector to a low-dimensional space through the fully connected layer to obtain a low-dimensional feature vector.

[0045] After the multi-layer convolution and global average pooling operations of multiple Inception modules, the dynamometer diagram to be tested is flattened and input into the fully connected layer. Inception V3 sets 2 fully connected layers. The number of nodes in the first fully connected layer is set to 1024, which is used to map high-dimensional features to low-dimensional space. The number of nodes in the second fully connected layer is the same as the number of categories of the classification task.

[0046] Step 403: learn the correlation between the low-dimensional feature vector and the working condition type through the attention mechanism layer, assign a weight to each feature, and obtain an attention feature vector.

[0047] The attention mechanism layer is introduced to dynamically assign weights to each feature in the low-dimensional feature vector by learning the correlation with the working condition type. The attention weight reflects the influence of each feature on the final judgment.

[0048] Step 404, calculating the similarity between the attention feature vector and each sample attention feature vector in the feature vector library through the similarity calculation layer. The feature vector library pre-stores a plurality of sample attention feature vectors and a working condition type label of each sample attention feature vector. The sample attention feature vector is obtained based on a sample dynamometer diagram.

[0049] Step 405: Determine the operating condition type of the oil pumping unit through the output layer according to the similarity between the attention feature vector and each sample attention feature vector in the feature vector library.

[0050] The output layer outputs the similarity ratio between the attention feature vector and the attention feature vector of each sample through the Softmax function. Then a judgment program is added to output the final working condition type through the threshold judgment.

[0051] Specifically, it is determined whether the maximum value of the similarity is greater than a set threshold. If so, the operating condition type label of the sample attention feature vector corresponding to the maximum value of the similarity is used as the operating condition type of the pumping unit. If not, the operating condition type label of the sample attention feature vector corresponding to the maximum value and the second maximum value of the similarity is used as the operating condition type of the pumping unit. Among them, a reasonable setting threshold is determined through multiple experiments.

[0052] In an exemplary embodiment, the fault conditions include fixed valve leakage, floating valve leakage, dual valve leakage, insufficient fluid supply, sand production in the oil well, gas impact, gas lock, pump rod breakage and well stuck. Figures 5 to 8 The indicator diagrams under normal operating conditions and some fault conditions are shown.

[0053] In an exemplary embodiment, the training process of the fault diagnosis model includes the following steps 501 to 505.

[0054] Step 501, obtaining multiple groups of sample operation data of the oil pump under different operating conditions.

[0055] In this application, 6000 sets of operating data representing the load and displacement of the pumping unit suspension point were collected from the field, including data under normal working conditions and various fault conditions, to ensure the diversity and representativeness of the data to cover a sufficient number of working conditions. Each set of sample operating data is stored in an Excel table, and each Excel table stores the displacement, load, well number and time of no less than 200 sampling points.

[0056] Step 502 , denoising and normalizing each group of sample operation data are performed in sequence to obtain multiple groups of pre-processed sample operation data.

[0057] Specifically, the sample running data is denoised to eliminate the influence of noise on feature extraction, and then normalized to make different features have the same scale.

[0058] Step 503 : for any set of preprocessed sample operation data, draw a sample dynamometer diagram according to the preprocessed sample operation data, and determine the operating condition type label of the sample dynamometer diagram according to the operating condition type of the preprocessed sample operation data.

[0059] First, use Python to read the data in the Excel table and draw the indicator diagram with displacement as the horizontal axis and load as the vertical axis. Further remove the invalid, erroneous and repeated information in the sample operation data.

[0060] Among them, filtering technology is used to remove high-frequency noise in the sample operation data. All processed sample dynamometer diagrams are normalized and limited to a 1×1 rectangle to avoid interference from other factors. The minimum-maximum normalization method is used to scale the data in the sample dynamometer diagram to between 0 and 1. Min-max, also known as deviation standardization, is a linear transformation of the original data so that the result falls into the [0,1] interval, where max is the maximum value of the sample operation data and min is the minimum value of the sample operation data. Then the displacement data of the sample dynamometer diagram is obtained. x and load data y : ; ; in, n is the number of sampling points in the sample run data, x i For the sample run data i The displacement of the sampling points, y i For the sample run data i The load at each sampling point,j is the sampling point number, ranging from 1 to n .

[0061] The spatial size of the normalized sample dynamometer diagram is further adjusted to adapt to the input conditions of the improved Inception V3 network.

[0062] Step 504: perform data enhancement on each sample dynamometer diagram to obtain a training sample set.

[0063] Since the collected sample operation data is limited, data enhancement is used to increase the number of dynamometer diagrams. Each sample dynamometer diagram after preprocessing is expanded by 30-degree image rotation, 45-degree image rotation, horizontal flipping, and image translation, so that the original dynamometer diagram data is expanded four times for subsequent model training.

[0064] Step 505: Train the improved Inception V3 network according to the training sample set to obtain a fault diagnosis model. The fault diagnosis model is the trained improved Inception V3 network.

[0065] The training sample set is divided into training set and test set according to a certain ratio (7:3). The sample dynamometer diagrams are placed in different folders according to the working condition type, and are named as normal working condition, fixed valve leakage, floating valve leakage, double valve leakage, insufficient fluid supply, sand production in oil wells, gas influence, gas lock, pump rod breakage, and well stuck. The folder naming directly reflects the single working condition type of its dynamometer diagram.

[0066] The training sample set is used to optimize the hyperparameters of the improved Inception V3 network. Specifically, a grid search strategy is used to adjust hyperparameters such as learning rate, batch size, and weight decay within a reasonable range. Through multiple experiments and verifications, a set of optimal hyperparameter combinations is found, which enables the network to show good performance on both the training set and the verification set.

[0067] During the training process, the cross-validation method is used to further divide the training set into multiple subsets, and the stability and generalization ability of the network are evaluated through multiple training and validation. At the same time, the back-propagation algorithm and gradient descent optimizer are used to update the parameters of the network so that the loss function of the network gradually decreases. The performance indicators of the network, such as accuracy, recall rate, and F1 score, are monitored to ensure that the network can accurately identify the operating condition type in the indicator diagram. When the performance of the network on the validation set no longer improves, the early stopping method is used to avoid overfitting.

[0068] Furthermore, after the training is completed, the sample dynamometer diagram of the known working condition is input into the trained improved Inception V3 network, and the attention feature vector of each sample dynamometer diagram is extracted through the forward propagation process of the network, that is, the sample attention feature vector is obtained, and the sample attention feature vector and the corresponding working condition type label are stored in the vector feature library. The vector feature library can be implemented using a database or file system to facilitate subsequent retrieval and matching.

[0069] This application uses an improved Inception V3 network to detect compound faults of oil pumps, which has the following advantages: First, the Inception V3 network has powerful feature extraction capabilities and excellent generalization performance. The improved Inception V3 network with the addition of an attention mechanism can automatically focus on important features and is more suitable for processing complex tasks such as compound fault detection. At the same time, it reduces manual intervention, reduces the risk of misjudgment, improves the accuracy of working condition identification, and can also judge compound working conditions based on fault detection.

[0070] The oil pump fault detection method provided in this application can be applied to oil production sites.

[0071] Based on the same inventive concept, the embodiment of the present application also provides a pumping unit fault detection device for implementing the pumping unit fault detection method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the pumping unit fault detection device provided below can refer to the limitations of the pumping unit fault detection method above, and will not be repeated here.

[0072] In an exemplary embodiment, Fig. 9 As shown, a pumping unit fault detection device is provided, including: a dynamometer diagram acquisition module 901 and a classification module 902 .

[0073] The dynamometer diagram acquisition module 901 is used to acquire the dynamometer diagram to be tested of the oil pumping unit.

[0074] The classification module 902 is used to extract features and classify the dynamometer diagram to be detected using a fault diagnosis model to determine the operating condition type of the pumping unit, which is a normal operating condition or various fault operating conditions.

[0075] The fault diagnosis model is obtained by pre-training the improved Inception V3 network using a training sample set. The improved Inception V3 network is obtained by adding an attention mechanism after the fully connected layer of the Inception V3 network. The training sample set includes multiple sample dynamometer diagrams and the operating condition type label of each sample dynamometer diagram.

[0076] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Fig.10 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store fault diagnosis models. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for detecting oil pump faults is implemented.

[0077] Those skilled in the art will understand that Fig.10 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0078] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0079] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0080] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0081] In this application, all actions to obtain signals, information or data are carried out in compliance with the relevant data protection laws and policies of the country where they are located and with the authorization given by the owner of the corresponding device.

[0082] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0083] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0084] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0085] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for detecting a pumping unit fault, characterized in that: The pumping unit fault detection method comprises: Obtain the indicator diagram of the oil pumping unit to be tested; The fault diagnosis model is used to extract and classify the characteristics of the dynamometer diagram to be detected, and the working condition type of the oil pumping unit is determined; the working condition type is a normal working condition or various fault working conditions; Among them, the fault diagnosis model is obtained by pre-training the improved Inception V3 network with a training sample set; the improved Inception V3 network is obtained by adding an attention mechanism after the fully connected layer of the Inception V3 network; the training sample set includes multiple sample dynamometer diagrams and the operating condition type label of each sample dynamometer diagram.

2. The method for detecting oil pump failure according to claim 1, characterized in that: Obtain the dynamometer diagram of the pumping unit to be tested, including: Acquire the operation data of the oil pump; the operation data includes the load, displacement, well number and time of multiple sampling points; Performing denoising and normalization processing on the operation data in sequence to obtain preprocessed operation data; According to the pre-processed operation data, a dynamometer diagram to be detected is drawn with displacement as the abscissa and load as the ordinate.

3. The method for detecting oil pump failure according to claim 1, characterized in that: The fault diagnosis model includes a feature extractor, a fully connected layer, an attention mechanism layer, a similarity calculation layer and an output layer; The fault diagnosis model is used to extract and classify the features of the dynamometer diagram to be detected, and the working condition type of the pumping unit is determined, specifically including: Extracting features from the dynamometer diagram to be detected by the feature extractor to obtain a high-dimensional feature vector; Mapping the high-dimensional feature vector to a low-dimensional space through the fully connected layer to obtain a low-dimensional feature vector; The attention mechanism layer learns the correlation between the low-dimensional feature vector and the working condition type, assigns a weight to each feature, and obtains an attention feature vector; The similarity between the attention feature vector and each sample attention feature vector in the feature vector library is calculated by the similarity calculation layer; the feature vector library pre-stores a plurality of sample attention feature vectors and a working condition type label of each sample attention feature vector; the sample attention feature vector is obtained based on a sample indicator diagram; The operating condition type of the oil pump is determined through the output layer according to the similarity between the attention feature vector and each sample attention feature vector in the feature vector library.

4. The method for detecting oil pump failure according to claim 3, characterized in that: Determining the working condition type of the oil pumping unit through the output layer according to the similarity between the attention feature vector and each sample attention feature vector in the feature vector library, specifically includes: Determine whether the maximum value of similarity is greater than a set threshold; If yes, the operating condition type label of the sample attention feature vector corresponding to the maximum similarity value is used as the operating condition type of the pumping unit; If not, the operating condition type labels of the sample attention feature vectors corresponding to the maximum and second maximum values ​​of similarity are taken as the operating condition type of the pumping unit.

5. The method for detecting oil pump failure according to claim 1, characterized in that: The fault conditions include fixed valve leakage, floating valve leakage, simultaneous double valve leakage, insufficient fluid supply, sand production in the oil well, gas impact, gas lock, sucker rod breakage and well stuck.

6. The method for detecting oil pump failure according to claim 1, characterized in that: The training process of the fault diagnosis model includes: Obtain multiple groups of sample operation data of the pumping unit under different operating conditions; De-noising and normalizing are performed on each group of sample operation data in turn to obtain multiple groups of pre-processed sample operation data; For any set of pre-processed sample operation data, draw a sample dynamometer diagram according to the pre-processed sample operation data, and determine the operating condition type label of the sample dynamometer diagram according to the operating condition type of the pre-processed sample operation data; Perform data enhancement on each sample dynamometer diagram to obtain a training sample set; According to the training sample set, the improved Inception V3 network is trained to obtain a fault diagnosis model; the fault diagnosis model is the trained improved Inception V3 network.

7. A pumping unit fault detection device, applied to the pumping unit fault detection method according to any one of claims 1 to 6, characterized in that: The pumping unit fault detection device comprises: The indicator diagram acquisition module is used to obtain the indicator diagram of the pumping unit to be tested; A classification module is used to extract and classify the characteristics of the dynamometer diagram to be detected by using a fault diagnosis model, and determine the working condition type of the oil pump; the working condition type is a normal working condition or various fault working conditions; Among them, the fault diagnosis model is obtained by pre-training the improved Inception V3 network with a training sample set; the improved Inception V3 network is obtained by adding an attention mechanism after the fully connected layer of the Inception V3 network; the training sample set includes multiple sample dynamometer diagrams and the operating condition type label of each sample dynamometer diagram.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for detecting oil pump faults according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for detecting oil pump failure according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for detecting oil pump failure according to any one of claims 1 to 6 is implemented.

Images