Oil well operation state analysis system and method

Abstract

The invention provides an oil well operation state analysis system and method, and relates to the technical field of oil well monitoring, and the system comprises an acquisition module which collects oil well operation data in real time; the data processing module preprocesses the operation data to obtain input data; the data analysis module constructs a 1D-CNN network model, and optimizes the 1D-CNN network model by using a grey wolf optimization algorithm to obtain an optimal analysis model; inputting the input data into the optimal analysis model to obtain abnormal data; the fault diagnosis module constructs a random forest network model, and optimizes the random forest network model by using an improved South goose algorithm to obtain an optimal diagnosis model; inputting the abnormal data into the optimal diagnosis model to obtain an abnormal type; and the early warning module receives the abnormal data and the abnormal type to generate a report and transmits the report to a user. Abnormal data are diagnosed through the fault diagnosis module, the abnormal type is obtained, and the oil well operation state analysis accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of oil well monitoring, and particularly to an oil well operation status analysis system and method. Background Art

[0002] In the field of oil extraction, the analysis of the operation status of oil wells is a key link to ensure the efficient and safe production of oil fields, and its importance is reflected in many aspects. First of all, the analysis of the operation status of oil wells can ensure the efficient production of oil fields. By real-time monitoring the operation status of oil wells, including key parameters such as the production volume, pressure, and temperature of oil wells, anomalies and potential faults in the operation of oil wells can be detected in a timely manner. Based on these analysis results, production parameters can be optimized, such as adjusting the water injection volume, oil production speed, etc., so as to increase the production volume of oil wells. At the same time, through the analysis of the performance of oil wells, production bottlenecks can be identified, providing a scientific basis for formulating improvement measures and further enhancing the production efficiency of oil fields. Secondly, the analysis of the operation status of oil wells is an important means to ensure the safe production of oil fields. There are many safety risks in the process of oil extraction, such as well blowout, fire, etc. By real-time monitoring the operation status of oil wells, potential safety risks can be predicted and prevented. Once an anomaly or potential fault is detected, corresponding measures can be taken immediately for repair or adjustment, thus avoiding the occurrence of accidents and ensuring the safe production of oil fields. In addition, the analysis of the operation status of oil wells can also improve the management efficiency of oil fields. Traditional oil field management relies on manual inspections and simple data acquisition systems, which have problems such as poor real-time performance and low efficiency.

[0003] Traditional methods rely on manual inspections or single-sensor threshold alarms, and it is difficult to capture weak anomaly signals and corresponding anomaly types under complex working conditions in real time. Summary of the Invention

[0004] The present invention provides an oil well operation status analysis system and method to solve the defect that traditional methods in the prior art rely on manual inspections or single-sensor threshold alarms and it is difficult to capture weak anomaly signals and corresponding anomaly types under complex working conditions in real time.

[0005] On the one hand, the present invention provides an oil well operation status analysis system, which includes: A data acquisition module for real-time acquisition of oil well operation data.

[0006] A data processing module for preprocessing the operation data to obtain input data.

[0007] A data analysis module for constructing a 1D-CNN network model, optimizing the 1D-CNN network model using the Grey Wolf Optimization algorithm to obtain an optimal analysis model. Inputting the input data into the optimal analysis model to obtain anomaly data.

[0008] A fault diagnosis module, which is used to construct a random forest network model, optimize the random forest network model using an improved snow goose algorithm to obtain an optimal diagnosis model, and input the abnormal data into the optimal diagnosis model to obtain the abnormal type.

[0009] An early warning module, which is used to receive the abnormal data and the abnormal type to generate a report and transmit it to the user.

[0010] For an oil well operation status analysis system provided by the present invention, the pre-processing steps for the operation data include: filtering the operation data using a moving average filtering method to obtain filtered data, and normalizing the filtered data using a minimum-maximum normalization method to obtain input data.

[0011] For an oil well operation status analysis system provided by the present invention, the steps for constructing a 1D-CNN network model include: Setting a network architecture including an input layer, a convolutional layer, and an output layer.

[0012] Adding an activation function between the convolutional layer and the output layer, adding a pooling layer between the activation function and the output layer, and adding a batch normalization layer and a fully connected layer between the pooling layer and the output layer to obtain a 1D-CNN network model.

[0013] For an oil well operation status analysis system provided by the present invention, the steps for optimizing the 1D-CNN network model using a grey wolf optimization algorithm include: Taking the F1 score as the fitness function of the network model.

[0014] Initializing the grey wolf population, randomly generating grey wolf individuals, and each grey wolf individual represents a set of hyperparameters of the LSTM network model.

[0015] Calculating the fitness value of each grey wolf individual in the initialized grey wolf population, classifying each grey wolf individual according to the individual fitness value, and updating the positions of the grey wolf individuals in the initialized population using the encircling behavior.

[0016] Updating the positions of the grey wolf individuals in the grey wolf population using the hunting behavior, calculating the fitness value of each grey wolf individual in the grey wolf population. If the output fitness value is higher than the preset fitness threshold, output the optimal hyperparameters. Otherwise, the grey wolf population continues to iterate until the fitness value is higher than the preset fitness threshold or reaches the maximum number of iterations to obtain the optimal hyperparameters.

[0017] For an oil well operation status analysis system provided by the present invention, the classification of each grey wolf individual includes: Calculate the fitness value of each gray wolf individual in the gray wolf population, select the individual with the optimal fitness as the leading wolf, the individual with the second-best fitness as the subordinate wolf, the individual with the third-best fitness as the scouting wolf, and other gray wolf individuals as ordinary wolves.

[0018] According to an oil well operation status analysis system provided by the present invention, the steps of constructing a random forest network model include: Construct multiple decision trees, and set the maximum depth of the decision trees and the minimum sample output required for the classification point to obtain an initial network model.

[0019] Use experimental data to train the initial network to obtain a random forest network model.

[0020] According to an oil well operation status analysis system provided by the present invention, the steps of optimizing the random forest network model using an improved snow goose algorithm include: Initialize the population. Each snow goose individual represents a set of random forest network model hyperparameters.

[0021] Use the mean squared error value as the fitness function of the model.

[0022] Calculate the fitness value of each snow goose individual in the initialized population, and update the initialized population using the leading goose rotation mechanism.

[0023] Calculate the fitness value of each snow goose individual, make the snow goose population enter the exploitation stage, and update the position of each snow goose individual in cooperation with the boundary strategy.

[0024] Calculate the fitness value of each snow goose individual. If the output fitness value is higher than the preset fitness threshold, select the hyperparameter combination corresponding to the position of the individual with the best fitness as the optimal hyperparameters. Otherwise, the snow goose population continues to iterate until the fitness value is higher than the preset fitness threshold or the maximum number of iterations is reached to obtain the optimal hyperparameters.

[0025] According to an oil well operation status analysis system provided by the present invention, the exploitation stage includes: Adjust the position update amplitude according to the distance between the individual and the leading goose according to a preset weight.

[0026] When the population individuals reach the search boundary, pull the individuals back to the feasible region by means of a direction broken line.

[0027] According to an oil well operation status analysis system provided by the present invention, the boundary strategy includes: Calculate the absolute difference between the fitness value of each snow goose individual and the average fitness value of the group. If the absolute difference between the individual fitness value and the average fitness value of the group is higher than the preset threshold, randomly change the moving direction of the individual position.

[0028] On the other hand, the present invention also provides a method for analyzing the operating state of an oil well, which includes: Collecting the operating data of the oil well in real time, and preprocessing the operating data to obtain input data.

[0029] Constructing a 1D-CNN network model, optimizing the 1D-CNN network model using the grey wolf optimization algorithm to obtain an optimal analysis model. Inputting the input data into the optimal analysis model to obtain abnormal data.

[0030] Constructing a random forest network model, optimizing the random forest network model using an improved snow goose algorithm to obtain an optimal diagnosis model. Inputting the abnormal data into the optimal model to obtain the type of abnormality.

[0031] Generating an analysis report based on the abnormal data and the type of abnormality.

[0032] An oil well operating state analysis system and method provided by the present invention, by collecting the operating data of the oil well in real time, collecting the operating data of the oil well in real time, using the method of moving average filtering to filter the operating data to obtain filtered data. Using the minimum-maximum normalization method to normalize the filtered data to obtain input data, and then constructing a 1D-CNN network model, using the grey wolf optimization algorithm to optimize the 1D-CNN network model to obtain an optimal analysis model. Inputting the input data into the optimal analysis model to obtain abnormal data, constructing a random forest network model, using an improved snow goose algorithm to optimize the random forest network model to obtain an optimal diagnosis model. Inputting the abnormal data into the optimal diagnosis model to obtain the type of abnormality. In the prior art, traditional methods rely on manual inspection or single sensor threshold alarm, and it is difficult to capture weak abnormal signals and abnormal situations under complex working conditions in real time, as well as the defect problems corresponding to the types of abnormalities, improving the accuracy of oil well operating state analysis and the speed of finding the type of abnormality. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0034] Figure 1 is a schematic flow chart of an oil well operating state analysis system provided by an embodiment of the present invention; Figure 2 is a schematic installation diagram of an oil well operating state analysis method provided by an embodiment of the present invention; Figure 3It is a schematic diagram of the optimization step process of the improved Snow Goose algorithm of an oil well operation status analysis system provided by an embodiment of the present invention. Detailed implementation manners

[0035] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0036] The following Figures 1 - 3 describes an oil well operation status analysis system and method of the present invention.

[0037] As Figure 1 shown, an oil well operation status analysis system provided by an embodiment of the present invention includes: A data acquisition module for real-time acquisition of oil well operation data.

[0038] In this embodiment, the acquired data covers key parameters such as wellhead pressure, temperature, flow rate, and motor operation status. For each parameter, a corresponding sensor is selected for efficient and accurate data acquisition. For the acquisition of wellhead pressure, a high-precision piezoelectric sensor, with its fast response and high-precision characteristics, can quickly capture the subtle changes in wellhead pressure, providing a solid foundation for subsequent pressure data analysis. This sensor performs excellently in oil well pressure monitoring, being able to provide real-time feedback on pressure fluctuations, which helps to promptly discover potential production problems or safety hazards. The acquisition of wellhead temperature relies on a platinum resistance thermometer, which is known for its good linearity, high temperature resistance performance, and stability, and can provide accurate temperature readings in the complex and changeable working environment of the oil well. Accurate temperature data is crucial for evaluating the production status and thermal efficiency of the oil well, and also helps to prevent equipment failures caused by too high or too low temperatures. In terms of wellhead flow rate acquisition, the application of a multiphase flowmeter significantly improves the efficiency and accuracy of data acquisition. Compared with traditional single-phase flowmeters, a multiphase flowmeter can measure the flow rates of oil, gas, and water three-phase fluids simultaneously, being more suitable for the complex oilfield development environment, and improving the oil well production and efficiency. In addition, for the monitoring of the motor operation status, a Hall effect sensor can be used to acquire the current and voltage data of the motor. The Hall effect sensor has non-contact measurement, high precision, and high reliability.

[0039] A data processing module for preprocessing the operation data to obtain input data.

[0040] The pre - processing steps for the operation data include: filtering the operation data using the moving average filtering method to obtain filtered data, and normalizing the filtered data using the min - max normalization method to obtain input data.

[0041] In this embodiment, the moving average filtering method reduces the influence of random noise by taking the average of adjacent data points. By presetting the window size and continuously updating the average according to the collected data, the filtered data is obtained. This method has the advantages of simple implementation and high computational efficiency. In the processing of oil well operation data, it can significantly improve the data smoothness and reduce the analysis error caused by random noise. Through moving average filtering, the change trend of the oil well operation state can be more clearly revealed, providing more accurate information for subsequent analysis and decision - making. Normalization processing is another key link in data pre - processing, and its purpose is to unify data with different dimensions to the same scale for easy comparison and analysis. The min - max normalization method is a commonly used normalization technique, and its principle is to linearly map the original data between the specified minimum and maximum values. Specifically, based on the filtered data, its minimum and maximum values are calculated, and then each data point is scaled according to the formula so that the range of the scaled data exactly falls within the preset minimum and maximum value intervals. The benefit of this is that it can eliminate the influence of different data dimensions and improve the comparability and analysis efficiency of data. The expression formula of the min - max normalization method is:

[0042] where, is the normalized data, is the original data point, is the minimum value in the original data set, is the maximum value in the original data set.

[0043] The data analysis module is used to construct a 1D - CNN network model, optimize the 1D - CNN network model using the grey wolf optimization algorithm to obtain the optimal analysis model, and input the input data into the optimal analysis model to obtain abnormal data.

[0044] The steps for constructing the 1D - CNN network model include: Setting a network architecture including an input layer, a convolutional layer, and an output layer.

[0045] Adding an activation function between the convolutional layer and the output layer, adding a pooling layer between the activation function and the output layer, and adding a batch normalization layer and a fully - connected layer between the pooling layer and the output layer to obtain a 1D - CNN network model.

[0046] In this embodiment, the input layer should be set to be able to receive the form of a one-dimensional array, and its length corresponds to the time step or feature dimension of each oil well operating state data point. The convolutional layer is the core component of the 1D-CNN, responsible for extracting features from the input data. One or more convolutional layers can also be set. Each convolutional layer needs to define the size of the convolutional kernel such as 3 or 5, the number such as 32 or 64, and the stride. The size of the convolutional kernel determines the range of the input data covered by each convolution operation, the number affects the diversity of features that the network can learn, and the stride controls the interval at which the convolutional kernel slides on the input data. For example, a typical convolutional layer setting may be: the convolutional kernel size is 3, the number is 32, and the stride is 1. The mathematical expression of the activation function is f(x)=max(0,x). The ReLU activation function has the advantages of simple calculation and not easy to disappear of gradients, and has been widely used in convolutional neural networks. After the activation function, a pooling layer is added to reduce the dimension of the data, reduce the amount of calculation, and enhance the translational invariance of the network. Max pooling or average pooling can also be selected. Max pooling selects the maximum value within the pooling window as the output, while average pooling calculates the average value within the pooling window. The pooling layer needs to set the size of the pooling window such as 2 and the stride 2. For example, a typical pooling layer setting may be: the pooling window size is 2 and the stride is 2. To further improve the training speed and stability of the network, a batch normalization layer can be added after the pooling layer. The batch normalization layer normalizes the input data, so that the inputs of each layer have similar distributions, which helps to speed up the training process and reduce overfitting. In the embodiment, the batch normalization layer can be set between the pooling layer and the fully connected layer. Finally, one or more fully connected layers are added after the batch normalization layer, which are used to map the features learned by the network to the output layer. Each neuron in the fully connected layer is connected to all neurons in the previous layer, and the number of neurons in the fully connected layer can be set according to the specific task requirements. For example, if it is a binary classification task of normal or abnormal, then the output layer can be set to a single neuron, and the sigmoid activation function is used to output the probability value.

[0047] The steps of optimizing the 1D-CNN network model using the Grey Wolf Optimization algorithm include: Taking the F1 score as the fitness function of the network model.

[0048] In this embodiment, the expression formula of the F1 score is:

[0049] Among them, is the harmonic mean of the precision and recall, is the precision, is the recall.

[0050] Initialize the gray wolf population and randomly generate gray wolf individuals. Each gray wolf individual represents a set of hyperparameters of the LSTM network model.

[0051] Calculate the fitness value of each gray wolf individual in the initialized gray wolf population, classify each gray wolf individual according to the individual fitness value, and update the position of the gray wolf individuals in the initialized population using the encircling behavior. In this embodiment, the expression formula for the encircling behavior is.

[0052]

[0053] Where, is the distance between the current gray wolf and the leading wolf, is the distance between the current gray wolf and the subordinate wolf, is the distance between the current gray wolf and the scouting wolf, and and are dynamic weight coefficients used to adjust the influence of the leadership wolves on the update of the current gray wolf position, is the current position of the leading wolf, is the current position of the subordinate wolf, is the current position of the scouting wolf.

[0054] The classification of the gray wolf individuals includes: Calculate the fitness value of each gray wolf individual in the gray wolf population, select the individual with the optimal fitness as the leading wolf, the individual with the second-best fitness as the subordinate wolf, the individual with the third-best fitness as the scouting wolf, and other gray wolf individuals as ordinary wolves.

[0055] In this embodiment, the expression formula for the classification of the gray wolf individuals is:

[0056] Where, is the position of the leading wolf, is the subordinate wolf, is the scouting wolf, represents the that obtains the minimum value in the function , that is, the position of the optimal individual in the current population, represents the that obtains the second smallest value in the function , that is, the position of the second-best individual in the current population, represents the that obtains the third smallest value in the function , that is, the position of the third-best individual in the current population.

[0057] Update the positions of the gray wolf individuals in the gray wolf population using hunting behavior, calculate the fitness values of each gray wolf individual in the gray wolf population. If the output fitness value is higher than the preset fitness threshold, output the optimal hyperparameters. Otherwise, the gray wolf population continues to iterate until the fitness value is higher than the preset fitness threshold or the maximum number of iterations is reached, and the optimal hyperparameters are obtained.

[0058] A fault diagnosis module for constructing a random forest network model, optimizing the random forest network model using an improved snow goose algorithm to obtain an optimal diagnosis model. Input the abnormal data into the optimal diagnosis model to obtain the abnormal type.

[0059] The steps of constructing the random forest network model include: Construct multiple decision trees, and set the maximum depth of the decision trees and the minimum sample output required for the classification point to obtain an initial network model.

[0060] The steps of optimizing the random forest network model using the improved snow goose algorithm include: Initialize the population. Each snow goose individual represents a set of random forest network model hyperparameters.

[0061] Use the mean squared error value as the fitness function of the model.

[0062] Calculate the fitness values of each snow goose individual in the initialized population, and update the initialized population using the leading goose rotation mechanism.

[0063] Calculate the fitness value of each snow goose individual, make the snow goose population enter the exploitation stage, and update the position of each snow goose individual in cooperation with the boundary strategy.

[0064] Calculate the fitness value of each snow goose individual. If the output fitness value is higher than the preset fitness threshold, select the hyperparameter combination corresponding to the position of the individual with the best fitness as the optimal hyperparameters. Otherwise, the snow goose population continues to iterate until the fitness value is higher than the preset fitness threshold or the maximum number of iterations is reached, and the optimal hyperparameters are obtained.

[0065] The exploitation stage includes: Adjust the position update amplitude according to the distance between the individual and the leading goose according to the preset weight.

[0066] When the population individuals reach the search boundary, pull the individuals back to the feasible region by means of a direction broken line.

[0067] In this embodiment, the expression formula of the exploitation stage is:

[0068] Among them, represents the th individual at the The position in the next iteration, is the current position of the ith individual, is the dynamic weight,

[0069] The boundary strategy includes: Calculate the absolute difference between the fitness value of each snow goose individual and the average fitness value of the group. If the absolute difference between the individual fitness value and the average fitness value of the group is higher than the preset threshold, randomly change the moving direction of the individual position.

[0070] In this embodiment, the expression formula of the boundary strategy is:

[0071]

[0072] Wherein, represents the ith individual at the th iteration, is the lower bound of the search space to ensure that the position is not lower than this value, the upper bound of the search space to ensure that the position is not higher than this value, is a random number generated between 0 and 1 for randomly resetting the position within the boundary.

[0073] Based on the same general inventive concept, the present invention also protects an oil well operation status analysis method, which includes: Collect oil well operation data in real time, and preprocess the operation data to obtain input data.

[0074] Construct a 1D-CNN network model, optimize the 1D-CNN network model using the gray wolf optimization algorithm to obtain an optimal analysis model. Input the input data into the optimal analysis model to obtain abnormal data.

[0075] Construct a random forest network model, optimize the random forest network model using an improved snow goose algorithm to obtain an optimal diagnosis model. Input the abnormal data into the optimal model to obtain the abnormal type.

[0076] Generate an analysis report according to the abnormal data and the abnormal type.

[0077] Example 1: Data for 24 consecutive hours, with the pressure fluctuation range of 22 - 32 MPa, the temperature stable at 60 - 75 °C, the flow rate suddenly dropping to 80 m³ / d (normally 120 m³ / d), and the peak value of the vibration signal reaching 18 mm / s. The moving average filter is used to smooth the instantaneous noise, and the normalized data are as follows: pressure 0.45 - 0.85, temperature 0.5 - 0.8, flow rate 0.3 - 0.6, vibration 0.6 - 0.9. Optimization steps of the Grey Wolf Optimization Algorithm: Initialization of the grey wolf population: Randomly generate 30 groups of hyperparameters. Leader wolf: Initial F1 = 0.72, learning rate 0.001, convolution kernel 32, neurons 128. Subordinate wolf: F1 = 0.68, learning rate 0.01, convolution kernel 16, neurons 64. Scout wolf: F1 = 0.65, learning rate 0.005, convolution kernel 64, neurons 256. Optimal hyperparameters: After 50 iterations, F1 is improved to 0.85, learning rate 0.001, convolution kernel 32, neurons 128. Optimization of the Improved Snow Goose Optimization Algorithm: Randomly generate 20 groups of hyperparameters, including the number of trees, maximum depth, and minimum sample split number. Initialization of the snow goose population: Randomly generate 20 groups of hyperparameters of the number of trees, maximum depth, and minimum sample split number. Rotation of the leading goose: Initial MSE = 0.15, number of trees = 100, depth = 10, samples = 20. Adjust the position update amplitude according to the distance, and trigger the individual position correction by the boundary strategy. After 80 iterations, MSE drops to 0.08, number of trees = 150, depth = 15, samples = 30. The normalized value of the sudden drop in the flow rate, which is 0.3, triggers the alarm of the optimal analysis model and is input into the optimal diagnosis model for diagnosing as a pump failure.

[0078] An oil well operation state analysis system and method provided by the present invention collect real-time oil well operation data, construct a 1D-CNN network model, optimize the 1D-CNN network model using the Grey Wolf Optimization Algorithm to obtain an optimal analysis model. Input the input data into the optimal analysis model to obtain abnormal data, construct a random forest network model, optimize the random forest network model using the Improved Snow Goose Algorithm to obtain an optimal diagnosis model. Input the abnormal data into the optimal diagnosis model to obtain the abnormal type. In the prior art, traditional methods rely on manual inspections or single-sensor threshold alarms, making it difficult to capture weak abnormal signals and corresponding abnormal situations under complex working conditions in real time, as well as the defect problems of corresponding abnormal types, improving the accuracy of oil well operation state analysis and the speed of finding abnormal types.

[0079] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative work.

[0080] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. And these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An oil well operating state analysis system, characterized in that, The system includes: A data acquisition module for real-time acquisition of oil well operation data; A data processing module for preprocessing the operation data to obtain input data; A data analysis module for constructing a 1D-CNN network model, optimizing the 1D-CNN network model using the Grey Wolf Optimization algorithm to obtain an optimal analysis model; inputting the input data into the optimal analysis model to obtain abnormal data; A fault diagnosis module for constructing a random forest network model, optimizing the random forest network model using an improved snow goose algorithm to obtain an optimal diagnosis model; inputting the abnormal data into the optimal diagnosis model to obtain the abnormal type; An early warning module for receiving the abnormal data and the abnormal type to generate a report and transmitting it to the user.

2. The oil well operating state analysis system according to claim 1, characterized in that The steps for preprocessing the operation data include: filtering the operation data using the moving average filtering method to obtain filtered data; normalizing the filtered data using the min-max normalization method to obtain input data.

3. The oil well operation status analysis system according to claim 1, characterized in that, The steps for constructing the 1D-CNN network model include: Setting a network architecture including an input layer, a convolutional layer, and an output layer; Adding an activation function between the convolutional layer and the output layer; adding a pooling layer between the activation function and the output layer; adding a batch normalization layer and a fully connected layer between the pooling layer and the output layer to obtain a 1D-CNN network model.

4. The oil well operation state analysis system according to claim 1, wherein The steps for optimizing the 1D-CNN network model using the Grey Wolf Optimization algorithm include: Taking the F1 score as the fitness function of the network model; Initializing the Grey Wolf population, randomly generating Grey Wolf individuals, and each Grey Wolf individual represents a set of hyperparameters of the LSTM network model; Calculating the fitness value of each Grey Wolf individual in the initialized Grey Wolf population, classifying each Grey Wolf individual according to the individual fitness value, and updating the positions of the Grey Wolf individuals in the initialized population using the encircling behavior; Updating the positions of the Grey Wolf individuals in the Grey Wolf population using the hunting behavior, calculating the fitness value of each Grey Wolf individual in the Grey Wolf population, if the output fitness value is higher than the preset fitness threshold, then output the optimal hyperparameters, otherwise, the Grey Wolf population continues to iterate until the fitness value is higher than the preset fitness threshold or reaches the maximum number of iterations to obtain the optimal hyperparameters.

5. The oil well operation status analysis system according to claim 4, characterized in that The classification of each Grey Wolf individual includes: Calculating the fitness value of each Grey Wolf individual in the Grey Wolf population, selecting the individual with the optimal fitness as the lead wolf, the individual with the second-best fitness as the subordinate wolf, the individual with the third-best fitness as the scouting wolf, and other Grey Wolf individuals as ordinary wolves.

6. The oil well operation state analysis system according to claim 1, wherein The steps for constructing the random forest network model include: Constructing multiple decision trees and setting the maximum depth of the decision trees and the minimum sample output required for the classification points to obtain an initial network model; Training the initial network using experimental data to obtain a random forest network model.

7. An oil well operation status analysis system according to claim 1, characterized in that, The steps for optimizing the random forest network model using the improved snow goose algorithm include: Initializing the population; each snow goose individual represents a set of random forest network model hyperparameters; Using the mean squared error value as the fitness function of the model; Calculate the fitness value of each snow goose individual in the initial population, and update the initial population using the leading goose rotation mechanism; Calculate the fitness value of each snow goose individual, enable the snow goose population to enter the exploration stage, and update the position of each snow goose individual in combination with the boundary strategy; Calculate the fitness value of each snow goose individual. If the output fitness value is higher than the preset fitness threshold, select the hyperparameter combination corresponding to the position of the individual with the best fitness as the optimal hyperparameter. Otherwise, the snow goose population continues to iterate until the fitness value is higher than the preset fitness threshold or the maximum number of iterations is reached to obtain the optimal hyperparameter.

8. An oil well operation status analysis system according to claim 7, characterized in that, The exploration stage includes: Adjust the position update amplitude according to the distance between the individual and the leading goose according to the preset weight; When the population individuals reach the search boundary, pull the individuals back to the feasible region by means of a direction broken line.

9. An oil well operation status analysis system according to claim 7, characterized in that, The boundary strategy includes: Calculate the absolute difference between the fitness value of each snow goose individual and the average fitness value of the group; if the absolute difference between the individual fitness value and the average fitness value of the group is higher than the preset threshold, randomly change the moving direction of the individual position.

10. A method for analyzing the operating state of an oil well, which uses an oil well operating state analysis system as described in any one of claims 1-9, characterized in that, The method includes: Collect the operation data of the oil well in real time, and preprocess the operation data to obtain the input data; Construct a 1D-CNN network model, optimize the 1D-CNN network model using the grey wolf optimization algorithm to obtain the optimal analysis model; input the input data into the optimal analysis model to obtain the abnormal data; Construct a random forest network model, optimize the random forest network model using the improved snow goose algorithm to obtain the optimal diagnosis model; input the abnormal data into the optimal model to obtain the abnormal type; Generate an analysis report according to the abnormal data and the abnormal type.

Images