Self-adaptive drilling tool jamming prediction method and device, electronic equipment and storage medium

Abstract

The invention relates to a self-adaptive drilling tool jamming prediction method and device, electronic equipment and a storage medium, and belongs to the technical field of drilling tool jamming prediction.The self-adaptive drilling tool jamming prediction method comprises the steps that drilling tool jamming historical data are obtained, and an initial drilling tool jamming prediction model is trained based on the drilling tool jamming historical data; the method comprises the steps of obtaining an initial drilling tool jamming prediction model, optimizing parameters of the initial drilling tool jamming prediction model by adopting a self-adaptive particle swarm optimization algorithm to obtain a first drilling tool jamming prediction model, training the first drilling tool jamming prediction model based on newly-added drilling tool jamming data when newly-added drilling tool jamming data is received, and obtaining a completely-trained drilling tool jamming prediction model. And inputting the to-be-predicted jamming data into the completely trained jamming prediction model to obtain a jamming prediction result, thereby improving the accuracy of jamming prediction.

Description

Technical Field

[0001] The present invention relates to the technical field of stuck pipe prediction, and in particular to an adaptive stuck pipe prediction method, device, electronic equipment and storage medium. Background Art

[0002] Stuck drill is one of the most common and complex accidents in drilling projects. Its occurrence not only prolongs the project cycle and increases costs, but also may cause serious safety hazards. According to statistics, stuck drill accidents account for about 40% to 30% of drilling accidents, and the financial losses caused by them account for more than 30% of non-production costs. Stuck drill accidents may cause production interruption, equipment damage, environmental pollution, and even casualties; if not handled properly, it may also cause complicated underground conditions such as well wall collapse and well leakage. In severe cases, it may even lead to the scrapping and abandonment of the wellbore, causing huge economic losses.

[0003] In recent years, with the rise of big data and machine learning technologies, machine learning theory has been used to analyze massive amounts of logging data. However, most of these methods rely on controlled variable methods for experiments or on experience and drilling theory formula derivation, which is not only time-consuming and labor-intensive, but also difficult to accurately predict stuck drill situations when considering multiple drilling parameters. Summary of the Invention

[0004] In view of this, it is necessary to provide an adaptive stuck pipe prediction method, device, electronic device and storage medium to solve the technical problem of low accuracy in predicting stuck pipe accidents when considering multiple drilling parameters.

[0005] In order to solve the above problems, in a first aspect, the present invention provides an adaptive stuck pipe prediction method, comprising: Acquiring historical stuck pipe data, training an initial stuck pipe prediction model based on the historical stuck pipe data, and optimizing parameters of the initial stuck pipe prediction model using an adaptive particle swarm optimization algorithm to obtain a first stuck pipe prediction model; When new stuck pipe data is received, training the first stuck pipe prediction model based on the new stuck pipe data to obtain a fully trained stuck pipe prediction model; The stuck pipe data to be predicted is input into the well-trained stuck pipe prediction model to obtain a stuck pipe prediction result.

[0006] In a possible implementation, obtaining historical data of stuck drill includes: Obtain historical data of stuck drills, and perform dimensionality reduction on the historical data using principal component analysis, wherein the historical data of stuck drills includes bit pressure, drilling speed, pump pressure, torque, mechanical drilling speed, and stuck drill mark data.

[0007] In a possible implementation, the principal component analysis method is used to reduce the dimension of the stuck drill historical data, including: The stuck drill historical data is standardized, and a covariance matrix of the stuck drill historical data is calculated. The eigenvalues ​​and eigenvectors of the covariance matrix are calculated. After sorting the eigenvalues, a principal component matrix is ​​constructed. The stuck drill historical data is reduced in dimension based on the principal component matrix.

[0008] In one possible implementation, the covariance matrix is: , in, is the covariance matrix, is the sample size, is the mean vector, For the samples, is transposed.

[0009] In one possible implementation, the parameters of the stuck pipe prediction model include kernel function parameters and penalty parameters; and the step of optimizing the parameters of the initial stuck pipe prediction model using an adaptive particle swarm optimization algorithm includes: A particle swarm of an adaptive particle swarm optimization algorithm is constructed based on the parameters of the initial stuck pipe prediction model, the particle swarm of the adaptive particle swarm optimization algorithm is initialized, and the particle velocity and particle position are obtained; Determine the fitness value of the particle swarm based on the objective function; Iteratively updating the particle position and particle velocity based on the fitness value, determining whether a maximum number of iterations is reached or a convergence condition is satisfied, and obtaining an individual optimal solution for the particle and a global optimal solution for the swarm of particles when the maximum number of iterations is reached or the convergence condition is satisfied; When the maximum number of iterations is not reached or the convergence condition is not met, the current individual optimal position and the global optimal position of the particle swarm are obtained, and the evolution factor is calculated based on the current individual optimal position and the global optimal position of the particle swarm. Based on the evolution factor, the particles are classified using fuzzy logic, and the inertia weight and the acceleration coefficient of the adaptive particle swarm optimization algorithm are adaptively adjusted. The particle position and the particle velocity are updated based on the adjusted inertia weight, the acceleration coefficient, the current individual optimal position and the global optimal position of the particle swarm to obtain the individual optimal solution of the particle and the global optimal solution of the swarm particles; Optimal parameters of an initial stuck pipe prediction model are determined based on the individual optimal solution and the global optimal solution, and the initial stuck pipe prediction model is adjusted based on the optimal parameters of the initial stuck pipe prediction model to obtain a first stuck pipe prediction model.

[0010] In a possible implementation, the evolution factor is: , in, For the evolution factor, is the average distance of the global best particle, is the maximum distance, is the minimum distance; The inertia weight is: , in, is the inertia weight, 、 are the maximum and minimum values ​​of the inertia weight, is the current iteration number, is the maximum number of iterations; The particle velocity is updated as: , in, is the particle velocity, 、 is the learning factor, 、 is a random number, is the particle’s own best historical position, is the best historical position of the particle swarm, is the particle's current position.

[0011] In one possible implementation, the first stuck pipe prediction model includes a first online support vector machine model; and when new stuck pipe data is received, the first stuck pipe prediction model is trained based on the new stuck pipe data to obtain a fully trained stuck pipe prediction model, including: Setting parameters of the first online support vector machine model; Initializing a support vector set and parameters of a first online support vector machine model, wherein newly added stuck drill data is received through the support vector machine; determining whether the support vector machine receives new stuck drill data, and when the support vector machine receives the new stuck drill data, updating the first online support vector machine model, and updating the parameters and support vector set of the first online support vector machine model based on the new stuck drill data; The first online support vector machine model is updated based on the updated parameters of the first online support vector machine model, and the newly added stuck pipe data is predicted based on the updated first online support vector machine model to obtain a stuck pipe prediction result.

[0012] In a second aspect, the present invention further provides an adaptive stuck pipe prediction device, comprising: a model optimization module, configured to obtain historical stuck pipe data, train an initial stuck pipe prediction model based on the historical stuck pipe data, and optimize parameters of the initial stuck pipe prediction model using an adaptive particle swarm optimization algorithm to obtain a first stuck pipe prediction model; a model training module, configured to, when receiving newly added stuck pipe data, train the first stuck pipe prediction model based on the newly added stuck pipe data to obtain a fully trained stuck pipe prediction model; The stuck pipe prediction module is used to input the stuck pipe data to be predicted into the trained stuck pipe prediction model to obtain a stuck pipe prediction result.

[0013] In a third aspect, the present invention further provides an electronic device, comprising: a processor and a memory; The memory stores a computer-readable program executable by the processor; When the processor executes the computer-readable program, the steps in the adaptive stuck pipe prediction method described above are implemented.

[0014] In a fourth aspect, the present invention further provides a computer-readable storage medium for storing computer-readable programs or instructions, which, when executed by a processor, can implement the steps of the adaptive drill stuck prediction method described in any one of the above-mentioned method items.

[0015] The beneficial effects of the present invention are as follows: historical stuck drill data is obtained, various drilling parameter characteristics are comprehensively considered, and the consistency and applicability of the data are ensured; an initial stuck drill prediction model is trained based on the historical stuck drill data, and the parameters of the initial stuck drill prediction model are optimized using an adaptive particle swarm optimization algorithm to obtain a first stuck drill prediction model; when new stuck drill data is received, the first stuck drill prediction model is trained based on the new stuck drill data to obtain a fully trained stuck drill prediction model; the stuck drill data to be predicted is input into the fully trained stuck drill prediction model to obtain a stuck drill prediction result; through parameter adaptive optimization and incremental learning of the stuck drill prediction model, full use is made of massive stuck drill data, the data is fully mined, and the accuracy of stuck drill prediction is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For the technical personnel of the present invention, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 A flow chart of an embodiment of the adaptive stuck pipe prediction method provided by the present invention; Figure 2 A schematic structural diagram of an embodiment of the adaptive stuck pipe prediction device provided by the present invention; Figure 3 This is a schematic structural diagram of an embodiment of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0018] The preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, and are not used to limit the scope of the present invention.

[0019] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0020] Before presenting the embodiments, the following terms are explained.

[0021] Principal Component Analysis (PCA) is a classic unsupervised dimensionality reduction technique that maps high-dimensional data to a low-dimensional space through linear transformation while retaining the main information of the data. It is widely used in data compression, feature extraction, and visualization.

[0022] The present invention discloses an adaptive stuck pipe prediction method, device, electronic device, and storage medium that can be used in a computer. The method, device, or computer-readable storage medium involved in the present invention can be integrated with the above-mentioned device or can be relatively independent.

[0023] A specific embodiment of the present invention discloses an adaptive stuck pipe prediction method, which can be executed by a computer, specifically by one or more processors of the computer. Figure 1 As shown in Figure 2, the adaptive stuck pipe prediction method includes: S101. Obtain historical stuck pipe data, train an initial stuck pipe prediction model based on the historical stuck pipe data, and optimize the parameters of the initial stuck pipe prediction model using an adaptive particle swarm optimization algorithm to obtain a first stuck pipe prediction model. It should be noted that the principal component analysis method was used to reduce the dimensionality of the historical stuck pipe data, thereby reducing data redundancy, lowering computational complexity, and retaining key information. The adaptive particle swarm optimization algorithm was used to optimize the parameters of the stuck pipe prediction model, thereby improving the parameter accuracy of the stuck pipe prediction model, reducing the complexity of hyperparameter tuning, and improving the accuracy of the stuck pipe prediction model.

[0024] S102: When newly added stuck pipe data is received, training the first stuck pipe prediction model based on the newly added stuck pipe data to obtain a fully trained stuck pipe prediction model; It should be noted that the incremental learning of the stuck pipe prediction model fully utilizes the massive amount of stuck pipe data, fully mines the data, and improves the accuracy of stuck pipe prediction; S103: Input the stuck pipe data to be predicted into a well-trained stuck pipe prediction model to obtain a stuck pipe prediction result.

[0025] It should be noted that accurate prediction of stuck pipe accidents was achieved through the online support vector machine model.

[0026] In some embodiments, in step S101, historical data of stuck drill is obtained and historical data related to stuck drill is collected. The historical data of stuck drill includes drilling pressure, drilling speed, pump pressure, torque, mechanical drilling speed, and stuck drill mark data. Combined with the stuck drill mark data (normal, no stuck drill is marked as 0, and stuck drill is marked as 1), missing value processing, outlier detection, and standardization are performed on the data to ensure data consistency and applicability. The stuck drill historical data is used as the original data for establishing a stuck drill prediction model. The stuck drill historical data is reduced in dimension using principal component analysis (PCA). The stuck drill historical data is standardized and a covariance matrix of the stuck drill historical data is calculated. The eigenvalues ​​and eigenvectors of the covariance matrix are calculated. After sorting the eigenvalues, a principal component matrix is ​​constructed. The stuck drill historical data is reduced in dimension based on the principal component matrix to complete the dimensionality reduction of the stuck drill historical data. Specifically, the high-dimensional data in the historical stuck drill data is standardized so that the mean of the data is 0 and the variance is 1. The covariance matrix of the standardized data is calculated. The covariance matrix is: , in, is the covariance matrix, is the sample size, is the mean vector, For the samples, is transposed; Then, the eigenvalues ​​and eigenvectors of the covariance matrix are calculated, and the first K principal components with a cumulative variance contribution rate of 95% are selected to construct the principal component matrix. The principal component matrix is ​​used to reduce the dimension of the stuck drill historical data. The calculation formula is: , in, is the historical data of stuck drill after dimensionality reduction, The historical data of stuck drill, is the principal component matrix.

[0027] The initial stuck pipe prediction model is trained based on the historical data of stuck pipe, and the parameters of the initial stuck pipe prediction model are optimized using the adaptive particle swarm optimization algorithm to obtain the first stuck pipe prediction model. The initial stuck pipe prediction model includes an online support vector machine model (Online-SVM), and the parameters of the initial stuck pipe prediction model include kernel function parameters ( ) and penalty parameter (C); construct an adaptive particle swarm optimization (APSO) particle swarm based on the parameters of the initial stuck drill prediction model, initialize the particle swarm of the adaptive particle swarm optimization algorithm to obtain particle velocity and particle position; randomly initialize the position and velocity of each particle in the particle swarm, where the particle position represents the potential solution to the problem, and the particle velocity determines the movement direction and step size of the particle in the solution space; determine the fitness value of the particle swarm based on the objective function, whose objective function is: , Calculate the fitness of each particle. According to the objective function of the current problem, calculate the fitness value of each particle's current position. The fitness value is used to measure the quality of the particle position. Iteratively update the particle position and particle velocity based on the fitness value, compare the fitness value of the current particle with the historical optimal fitness value of the current particle, and select the best as the individual optimal fitness value; then find the particle with the best fitness value among all particles, and use its position and fitness value as the global optimal position and local optimal fitness value; determine whether the maximum number of iterations has been reached or the convergence condition has been met. When the maximum number of iterations has been reached or the convergence condition has been met, obtain the individual optimal solution of the particle and the global optimal solution of the swarm particles; when the maximum number of iterations has not been reached or the convergence condition has not been met, obtain the current individual optimal position and the global optimal position of the particle swarm, and calculate the evolution factor based on the current individual optimal position and the global optimal position of the particle swarm, that is, at the current position of the particle, use the Euclidean metric to calculate the average distance from each particle to all other particles, and express the average distance of the global optimal particle as , compare the average distance of all particles and determine the maximum distance and minimum distance , its evolution factor is: , in, is the evolution factor, the value of the evolution factor is between 0 and 1, is the average distance of the global best particle, is the maximum distance, is the minimum distance; Based on the evolution factor, fuzzy logic is used to classify particles into different states, such as exploration state, utilization state, convergence state, and exit state, and the inertia weight of the adaptive particle swarm optimization algorithm is and the acceleration factor ( and ) is adaptively adjusted, and its inertia weight is: , in, is the inertia weight, 、 are the maximum and minimum values ​​of the inertia weight, is the current iteration number, is the maximum number of iterations; Based on the adjusted inertia weight, acceleration coefficient, current individual optimal position of the particle swarm and global optimal position, the particle position and particle velocity are updated to obtain the individual optimal solution of the particle and the global optimal solution of the swarm particles. According to the updated parameters and the individual optimal position and global optimal position of the particle, the position and velocity update formula are used to update the position and velocity of each particle. It should be noted that the updated velocity should be limited to prevent the velocity from being too large, which will cause the particle search range to be too large. The particle velocity update is: , in, is the particle velocity, 、 is the learning factor, 、 is a random number, is the particle’s own best historical position, is the best historical position of the particle swarm, is the current position of the particle, and its particle position is updated as follows: , APSO is used for iterative optimization to continuously approach the optimal penalty parameter C and kernel function parameter , determining optimal parameters of an initial stuck pipe prediction model based on the individual optimal solution and the global optimal solution, and adjusting the stuck pipe prediction model based on the optimal parameters of the initial stuck pipe prediction model to obtain a first stuck pipe prediction model.

[0028] In some embodiments, in step S102, when newly added stuck drill data is received, a first stuck drill prediction model is trained based on the newly added stuck drill data to obtain a fully trained stuck drill prediction model; the stuck drill history data is divided into a training set, a test set, and a validation set, and the initial stuck drill prediction model is trained through the training set, and after obtaining the first stuck drill prediction model under the optimized hyperparameters, when the first stuck drill prediction model receives newly added stuck drill data, the first stuck drill prediction model is trained based on the newly added stuck drill data, and the first stuck drill prediction model is a first online support vector machine model; the parameters of the first online support vector machine model are set; the support vector set and the parameters of the first online support vector machine model are initialized, wherein the newly added stuck drill data is received through the support vector machine; it is determined whether the support vector machine receives the newly added stuck drill data, and when the support vector machine receives the newly added stuck drill data, the training set is updated based on the newly added stuck drill data; based on the newly added stuck drill data The stuck drill data is used to train the first online support vector machine model, and the parameters and support vector set of the first online support vector machine model are updated; the first online support vector machine model is updated based on the updated parameters of the first online support vector machine model, and the newly added stuck drill data is predicted based on the updated first online support vector machine model to obtain the stuck drill prediction result, specifically, the online-SVM model is initialized, the online-SVM model parameters are set, the kernel function type (such as linear kernel, radial basis kernel, etc.) used by the online-SVM model is determined, as well as other related parameters such as the penalty parameter C, which will affect the performance and learning effect of the online-SVM model; an empty support vector set and online-SVM model parameters are initialized, an empty set is created for storing support vectors, and the weight vector of the online-SVM model is initialized. and bias b, in the initial stage, the weight vector The and bias are usually set to default values ​​or zero values; start receiving new stuck drill data, the online-SVM enters the state of waiting to receive the new stuck drill data, determines whether there is new stuck drill data, checks whether the new stuck drill data is received, and ends if there is no new stuck drill data, and outputs the final online-SVM model, that is, the first online support vector machine model is the final online-SVM model. If there is new stuck drill data, the current training set is updated by adding the new stuck drill data, and the updated training set is used, that is, the new stuck drill data is used to retrain the online-SVM model. During the process, the parameters of the online-SVM model are adjusted according to the newly added stuck drill data to minimize the classification or regression error and predict the category or value of the newly added stuck drill data. After updating the online-SVM model parameters, the updated online-SVM model is used to predict the newly added stuck drill data. After obtaining the category or predicted value, the model returns to the state of waiting for receiving the newly added stuck drill data and continues to repeat the above process to realize online learning. When no new stuck drill data arrives, the final trained online-SVM model is output. The final trained online-SVM model is used to predict new unknown samples.

[0029] Every fixed time window (e.g. 5 minutes), the Online-SVM model is trained with the reduced-dimensional data, and the Online-SVM model is incrementally updated to use the new data points. Is it near the decision boundary of the Online-SVM model? If the data is within the boundary, it means that the confidence is high and the new data is Adding support vector sets, out-of-bounds data means low confidence and can be ignored to reduce computational overhead.

[0030] In some embodiments, in step S103, the stuck pipe data to be predicted is input into a well-trained stuck pipe prediction model to obtain a stuck pipe prediction result, and the prediction result is calculated as follows: , in, To predict the results, is the new sample data, is the bias, is the transpose of the weight vector, when When the predicted stuck pipe risk is high, , the prediction is normal.

[0031] In summary, the adaptive stuck drill prediction method provided by the present invention obtains historical stuck drill data, trains an initial stuck drill prediction model based on the historical stuck drill data, and uses an adaptive particle swarm optimization algorithm to optimize the parameters of the initial stuck drill prediction model to obtain a first stuck drill prediction model. When new stuck drill data is received, the first stuck drill prediction model is trained based on the new stuck drill data to obtain a fully trained stuck drill prediction model. The stuck drill data to be predicted is input into the fully trained stuck drill prediction model to obtain a stuck drill prediction result, thereby improving the accuracy of the stuck drill prediction.

[0032] In order to better implement the adaptive stuck pipe prediction method in the embodiment of the present invention, based on the adaptive stuck pipe prediction method, correspondingly, Figure 2 As shown, an embodiment of the present invention further provides an adaptive stuck pipe prediction device, and the adaptive stuck pipe prediction device 200 includes: A model optimization module 201 is configured to obtain historical stuck pipe data, train an initial stuck pipe prediction model based on the historical stuck pipe data, and optimize the parameters of the initial stuck pipe prediction model using an adaptive particle swarm optimization algorithm to obtain a first stuck pipe prediction model. A model training module 202 is configured to train a first stuck pipe prediction model based on newly received stuck pipe data to obtain a fully trained stuck pipe prediction model; The stuck pipe prediction module 203 is used to input the stuck pipe data to be predicted into the well-trained stuck pipe prediction model to obtain a stuck pipe prediction result.

[0033] like Figure 3 As shown, the present invention also provides an electronic device 300 , which can be a computing device such as a mobile terminal, a desktop computer, a notebook, a palmtop computer, or a server. The electronic device 300 includes a processor 301 , a memory 302 , and a display 303 . Figure 3 Only some of the components of the electronic device 300 are shown, but it should be understood that implementation of all of the shown components is not required, and more or fewer components may be implemented instead.

[0034] In some embodiments, the memory 302 may be an internal storage unit of the electronic device 300, such as the hard drive or memory of the electronic device 300. In other embodiments, the memory 302 may also be an external storage device of the electronic device 300, such as a plug-in hard drive, a Smart Media Card (SMC), a Secure Digital (SD) card, a flash memory card, etc. Furthermore, the memory 302 may include both the internal storage unit of the electronic device 300 and an external storage device. The memory 302 is used to store application software installed on the electronic device 300 and various data, such as program code installed on the electronic device 300. The memory 302 may also be used to temporarily store data that has been output or is about to be output. In one embodiment, the memory 302 stores an adaptive drill stuck prediction program, which can be executed by the processor 301 to implement the adaptive drill stuck prediction method of various embodiments of the present invention.

[0035] In some embodiments, the processor 301 may be a central processing unit (CPU), a microprocessor, or other data processing chip, configured to execute program codes or process data stored in the memory 302 , such as an adaptive stuck pipe prediction method.

[0036] In some embodiments, display 303 can be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 303 is used to display identification information for the adaptive stuck drill prediction program and to display a visual user interface. Components 301-303 of electronic device 300 communicate with each other via a system bus.

[0037] In some embodiments, when the processor 301 executes the adaptive stuck pipe prediction program in the memory 302, the steps of the adaptive stuck pipe prediction method described in the above embodiments are implemented. Since the adaptive stuck pipe prediction method has been described in detail above, it will not be repeated here.

[0038] Accordingly, the present invention also provides a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the programs or instructions are executed by a processor, the steps or functions of the adaptive drill stuck prediction method provided in the above-mentioned method embodiments can be implemented.

[0039] Those skilled in the art will appreciate that all or part of the process steps of the above-described embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, such as a magnetic disk, an optical disk, a read-only memory, or a random access memory.

[0040] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or replacements that can be easily conceived by any technician familiar with the technical neighbors within the technical scope disclosed by the present invention should be covered by the scope of protection of the present invention.

Claims

1. An adaptive stuck pipe prediction method, characterized in that: include: Acquiring historical stuck pipe data, training an initial stuck pipe prediction model based on the historical stuck pipe data, and optimizing parameters of the initial stuck pipe prediction model using an adaptive particle swarm optimization algorithm to obtain a first stuck pipe prediction model; When new stuck pipe data is received, training the first stuck pipe prediction model based on the new stuck pipe data to obtain a fully trained stuck pipe prediction model; The stuck pipe data to be predicted is input into the well-trained stuck pipe prediction model to obtain a stuck pipe prediction result.

2. The adaptive stuck pipe prediction method according to claim 1, characterized in that: The obtaining of stuck drill historical data includes: Obtain historical data of stuck drills, and perform dimensionality reduction on the historical data using principal component analysis, wherein the historical data of stuck drills includes bit pressure, drilling speed, pump pressure, torque, mechanical drilling speed, and stuck drill mark data.

3. The adaptive stuck pipe prediction method according to claim 2, characterized in that: The principal component analysis method is used to reduce the dimension of the stuck drill historical data, including: The stuck drill historical data is standardized, and a covariance matrix of the stuck drill historical data is calculated. The eigenvalues ​​and eigenvectors of the covariance matrix are calculated. After sorting the eigenvalues, a principal component matrix is ​​constructed. The stuck drill historical data is reduced in dimension based on the principal component matrix.

4. The adaptive stuck pipe prediction method according to claim 3, characterized in that: The covariance matrix is: ， in, is the covariance matrix, is the sample size, is the mean vector, For the samples, is transposed.

5. The adaptive stuck pipe prediction method according to claim 3, characterized in that: The parameters of the stuck pipe prediction model include kernel function parameters and penalty parameters; the adaptive particle swarm optimization algorithm is used to optimize the parameters of the initial stuck pipe prediction model, including: A particle swarm of an adaptive particle swarm optimization algorithm is constructed based on the parameters of the initial stuck pipe prediction model, the particle swarm of the adaptive particle swarm optimization algorithm is initialized, and the particle velocity and particle position are obtained; Determine the fitness value of the particle swarm based on the objective function; Iteratively updating the particle position and particle velocity based on the fitness value, determining whether a maximum number of iterations is reached or a convergence condition is satisfied, and obtaining an individual optimal solution for the particle and a global optimal solution for the swarm of particles when the maximum number of iterations is reached or the convergence condition is satisfied; When the maximum number of iterations is not reached or the convergence condition is not met, the current individual optimal position and the global optimal position of the particle swarm are obtained, and the evolution factor is calculated based on the current individual optimal position and the global optimal position of the particle swarm. Based on the evolution factor, the particles are classified using fuzzy logic, and the inertia weight and the acceleration coefficient of the adaptive particle swarm optimization algorithm are adaptively adjusted. The particle position and the particle velocity are updated based on the adjusted inertia weight, the acceleration coefficient, the current individual optimal position and the global optimal position of the particle swarm to obtain the individual optimal solution of the particle and the global optimal solution of the swarm particles; Optimal parameters of an initial stuck pipe prediction model are determined based on the individual optimal solution and the global optimal solution, and the initial stuck pipe prediction model is adjusted based on the optimal parameters of the initial stuck pipe prediction model to obtain a first stuck pipe prediction model.

6. The adaptive stuck pipe prediction method according to claim 5, characterized in that: The evolution factors are: ， in, For the evolution factor, is the average distance of the global best particle, is the maximum distance, is the minimum distance; The inertia weight is: ， in, is the inertia weight, 、 are the maximum and minimum values ​​of the inertia weight, is the current iteration number, is the maximum number of iterations; The particle velocity is updated as: ， in, is the particle velocity, 、 is the learning factor, 、 is a random number, is the particle’s own best historical position, is the best historical position of the particle swarm, is the particle's current position.

7. The adaptive stuck pipe prediction method according to claim 5, characterized in that: The first stuck pipe prediction model includes a first online support vector machine model; when new stuck pipe data is received, the first stuck pipe prediction model is trained based on the new stuck pipe data to obtain a fully trained stuck pipe prediction model, including: Setting parameters of the first online support vector machine model; Initializing a support vector set and parameters of a first online support vector machine model, wherein newly added stuck drill data is received through the support vector machine; determining whether the support vector machine receives new stuck drill data; and when the support vector machine receives the new stuck drill data, training the first online support vector machine model based on the new stuck drill data, and updating parameters and a support vector set of the first online support vector machine model; The first online support vector machine model is updated based on the updated parameters of the first online support vector machine model, and the newly added stuck pipe data is predicted based on the updated first online support vector machine model to obtain a stuck pipe prediction result.

8. An adaptive stuck pipe prediction device, characterized in that: include: a model optimization module, configured to obtain historical stuck pipe data, train an initial stuck pipe prediction model based on the historical stuck pipe data, and optimize parameters of the initial stuck pipe prediction model using an adaptive particle swarm optimization algorithm to obtain a first stuck pipe prediction model; a model training module, configured to, when receiving newly added stuck pipe data, train the first stuck pipe prediction model based on the newly added stuck pipe data to obtain a fully trained stuck pipe prediction model; The stuck pipe prediction module is used to input the stuck pipe data to be predicted into the trained stuck pipe prediction model to obtain a stuck pipe prediction result.

9. An electronic device, characterized in that: including memory and processor; The memory stores a computer-readable program executable by the processor; When the processor executes the computer-readable program, the processor implements the steps of the adaptive stuck pipe prediction method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps of the adaptive stuck pipe prediction method according to any one of claims 1 to 7.

Images