Differential pressure jamming prediction method and device based on machine learning, equipment and medium

Abstract

The invention provides a differential pressure jamming prediction method and device based on machine learning, equipment and a medium. According to the method, a machine learning model is trained and established on the basis of historical data of historical wells with the differential pressure drilling tool jamming problem and historical wells without the differential pressure drilling tool jamming problem, so that differential pressure drilling tool jamming condition early warning is carried out by comprehensively considering conditions such as tripping working conditions and abnormal resistance characteristics during drilling activities, and drilling risks are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of oil drilling engineering, and in particular to a method, device, equipment and medium for predicting differential pressure sticking based on machine learning. Background Art

[0002] Differential pressure stuck drill pipe is one of the most common types of stuck drill pipe during drilling. The occurrence of such accidents will seriously affect the normal operation of on-site drilling operations, reduce drilling efficiency, prolong the drilling cycle and cause serious economic losses. Improper handling of differential pressure stuck drill pipe may induce accidents such as collapse stuck drill pipe, broken drill bits, or even buried drill bits and scrapped drill bits. The prediction and evaluation of differential pressure stuck drill pipe can effectively avoid the occurrence of drilling risks, reduce the probability and frequency of drilling engineering risks, and shorten the well construction cycle. Therefore, the accurate prediction of differential pressure stuck drill pipe is of great significance for preventing and reducing the occurrence of stuck drill pipe accidents during drilling.

[0003] Currently, there are few prediction methods for differential pressure drilling, which mainly rely on empirical methods. However, they lack immediacy and are highly subjective based on experience. There is still much room for improvement in the prediction effect. Summary of the invention

[0004] In order to solve the technical problems that the method of predicting differential pressure drill sticking based on experience is not timely enough and involves subjective judgment, and the prediction effect is low in accuracy, the present invention proposes a method, device, equipment and medium for predicting differential pressure drill sticking based on machine learning.

[0005] In a first aspect, an embodiment of the present invention provides a method for predicting differential pressure stuck pipe based on machine learning, comprising:

[0006] Obtain historical data of historical wells with differential pressure sticking problems and historical data of historical wells without differential pressure sticking problems;

[0007] Classifying historical data of historical wells with differential pressure stuck drill problems according to corresponding differential pressure stuck drill problems, wherein the classification of differential pressure stuck drill problems includes differential pressure stuck drill section, near differential pressure stuck drill section and normal section;

[0008] Performing data preprocessing on the historical data of the historical wells without the differential pressure sticking problem and the historical data of the historical wells with the differential pressure sticking problem;

[0009] Based on the pre-processed historical data of historical wells without the problem of differential pressure sticking and the pre-processed historical data of historical wells with the problem of differential pressure sticking, model training is performed to obtain a machine learning model;

[0010] Predicting by using the machine learning model according to the real-time data of the on-site well to obtain a predicted hook load curve;

[0011] Predict differential pressure sticking based on the predicted hook load curve numerical trend and current tripping conditions;

[0012] Early warning is given based on the predicted differential pressure problem.

[0013] In some implementations, predicting the differential pressure sticking problem according to the predicted numerical trend of the hook load curve and the current tripping condition includes:

[0014] During the drilling period, the maximum hook load value in the predicted hook load curve shows an increasing trend in the process of lowering the column for a set number of consecutive times, and the growth rate each time is greater than the preset growth rate threshold value for several consecutive times; or during the drilling period, the maximum hook load value in the predicted hook load curve shows a decreasing trend in the process of lowering the column for a set number of consecutive times, and the decline rate each time is greater than the preset decline rate threshold value for several consecutive times, then the differential pressure stuck drill problem is a differential pressure stuck drill section;

[0015] During the drilling period, the maximum hook load value in the predicted hook load curve shows an increasing trend in the process of lowering the column for a set number of consecutive times, and the growth rate each time is less than the preset growth rate threshold value for several consecutive times; or during the drilling period, the maximum hook load value in the predicted hook load curve shows a decreasing trend in the process of lowering the column for a set number of consecutive times, and the decline rate each time is less than the preset decline rate threshold value for several consecutive times, then the differential pressure stuck drill problem is close to the differential pressure stuck drill section;

[0016] During drilling, if the maximum hook load value in the predicted hook load curve shows a decreasing trend in the process of lowering the column for a consecutive set number of times, or during drilling, if the maximum hook load value in the predicted hook load curve shows an increasing trend in the process of lowering the column for a consecutive set number of times, the pressure difference drill sticking problem is in the normal section.

[0017] In some implementations, the performing of model training based on preprocessed historical data of historical wells without differential pressure sticking problems and preprocessed historical data of historical wells with differential pressure sticking problems to obtain a machine learning model includes:

[0018] Model training and model performance evaluation are performed based on pre-processed historical data of historical wells without differential pressure sticking problems and pre-processed historical data of historical wells with differential pressure sticking problems;

[0019] If the evaluation result does not meet the preset model performance standard, model training and model performance evaluation will continue until the evaluation result meets the preset model performance standard to obtain a machine learning model.

[0020] In some implementations, the data preprocessing includes: performing data cleaning on the historical data of the historical wells without the differential pressure sticking problem and the historical data of the historical wells with the differential pressure sticking problem.

[0021] In some implementations, the historical data of the historical wells without differential pressure sticking problems and the historical data of the historical wells with differential pressure sticking problems include time series, hook loads, drill bit positions, column numbers, and lift-off weights corresponding to the historical wells.

[0022] In some implementations, the step of providing an early warning based on the predicted differential pressure sticking problem includes:

[0023] If the predicted differential pressure sticking problem is a differential pressure sticking section, a first-level warning is issued;

[0024] If the predicted differential pressure sticking problem is close to the differential pressure sticking section, a second-level warning will be issued.

[0025] In some implementations, the performing of model training and model performance evaluation based on preprocessed historical data of historical wells without differential pressure sticking problems and preprocessed historical data of historical wells with differential pressure sticking problems includes:

[0026] Divide the preprocessed historical data of the historical wells without the differential pressure sticking problem and the preprocessed historical data of the historical wells with the differential pressure sticking problem into a training set, a validation set and a test set according to a preset ratio;

[0027] Normalizing the data in the training set, validation set, and test set;

[0028] Perform model training based on the normalized training set, and evaluate model performance and adjust model hyperparameters based on the normalized validation set during model training;

[0029] The performance and generalization ability of the trained model are evaluated based on the normalized test set to obtain the evaluation results.

[0030] In a second aspect, an embodiment of the present invention provides a differential pressure drill sticking prediction device, comprising:

[0031] An acquisition module, used for acquiring historical data of historical wells with differential pressure sticking problems and historical data of historical wells without differential pressure sticking problems;

[0032] A classification module, for classifying historical data of historical wells with differential pressure stuck drill problems according to corresponding differential pressure stuck drill problems, wherein the classification of the differential pressure stuck drill problem includes differential pressure stuck drill section, near differential pressure stuck drill section and normal section;

[0033] A processing module, used for performing data preprocessing on the historical data of the historical wells without the differential pressure sticking problem and the historical data of the historical wells with the differential pressure sticking problem;

[0034] A model building module, for performing model training based on pre-processed historical data of historical wells without differential pressure sticking problems and pre-processed historical data of historical wells with differential pressure sticking problems to obtain a machine learning model;

[0035] A prediction module, used to predict the real-time data of the field well through the machine learning model to obtain a predicted hook load curve, and predict the differential pressure sticking problem according to the numerical trend of the predicted hook load curve and the current drilling conditions;

[0036] The early warning module is used to issue an early warning based on the predicted pressure difference drill sticking problem.

[0037] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by at least one processor, the method described in the first aspect is implemented.

[0038] In a fourth aspect, an embodiment of the present invention provides an electronic device, comprising a memory and at least one processor, wherein the memory stores a computer program, and when the computer program is executed by the at least one processor, the method described in the first aspect is implemented.

[0039] One or more embodiments of the present invention bring at least the following beneficial effects:

[0040] The present invention classifies the historical data of historical wells with differential pressure stuck drill problems according to differential pressure stuck drill sections, near differential pressure stuck drill sections and normal sections, thereby training a machine learning model based on the historical data of historical wells without differential pressure stuck drill problems and the classified historical data of historical wells with differential pressure stuck drill problems. Thus, the real-time data of the field wells can be transmitted to the trained machine learning model to predict the hook load curve, and based on the numerical trend of the hook load curve and the current tripping and drilling conditions, it is predicted whether the differential pressure stuck drill problem is a differential pressure stuck drill section, a near differential pressure stuck drill section or a normal section, and a corresponding warning is made, thereby effectively avoiding the occurrence of drilling risks, reducing the probability and frequency of drilling engineering risks, and shortening the well construction cycle. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and should not be regarded as limiting the scope.

[0042] Figure 1 A flow chart of a method for predicting differential pressure stuck drill based on machine learning provided in an embodiment of the present invention;

[0043] Figure 2A structural diagram of a pressure differential drilling sticking prediction device based on machine learning provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.

[0045] Embodiment 1

[0046] This embodiment provides a method for predicting differential pressure stuck pipe based on machine learning, such as Figure 1 As shown, including:

[0047] Step S100, obtaining historical data of historical wells with differential pressure sticking problems and historical data of historical wells without differential pressure sticking problems. The historical data of historical wells without differential pressure sticking problems and historical data of historical wells with differential pressure sticking problems include time series, hook load, drill bit position, column number and lifting weight of corresponding historical wells.

[0048] Step S200, classifying historical data of historical wells with differential pressure stuck pipe problems according to corresponding differential pressure stuck pipe problems, wherein the classification of differential pressure stuck pipe problems includes differential pressure stuck pipe section, near differential pressure stuck pipe section and normal section.

[0049] Step S300, preprocessing the historical data of the historical wells without the differential pressure sticking problem and the historical data of the historical wells with the differential pressure sticking problem. The data preprocessing includes data cleaning to eliminate noise points in the data.

[0050] Step S400, based on the pre-processed historical data of historical wells without differential pressure sticking problems and the pre-processed historical data of historical wells with differential pressure sticking problems, model training is performed to obtain a machine learning model.

[0051] In some implementations, step S400 includes:

[0052] Step S410 , performing model training and model performance evaluation based on the pre-processed historical data of historical wells without the differential pressure sticking problem and the historical data of historical wells with the differential pressure sticking problem.

[0053] ,Step S420, if the evaluation result does not meet the preset model performance standard, continue model training and model performance evaluation until the evaluation result meets the preset model performance standard and obtains the machine learning model.

[0054] In some implementations, step S410 includes:

[0055] Step S411, dividing the pre-processed historical data of the historical wells without the differential pressure sticking problem and the historical data of the historical wells with the differential pressure sticking problem into a training set, a validation set and a test set according to a preset ratio;

[0056] Step S412, normalizing the data in the training set, validation set, and test set;

[0057] Step S413, perform model training based on the normalized training set, and perform model performance evaluation based on the normalized validation set during the model training process and adjust the model hyperparameters. Specifically, adjust the model hyperparameters based on the model performance evaluation results and relevant petroleum engineering knowledge. The model hyperparameters include the number of hidden layers, the number of neurons in each layer, the learning rate, etc.

[0058] Step S414, evaluating the performance and generalization ability of the trained model based on the normalized test set to obtain an evaluation result.

[0059] Step S500, predicting by the machine learning model according to the real-time data of the on-site well, obtaining a predicted hook load curve, and predicting the differential pressure sticking problem according to the numerical trend of the predicted hook load curve and the current drilling conditions.

[0060] The method of predicting the differential pressure sticking problem according to the numerical trend of the predicted hook load curve and the current drilling conditions includes:

[0061] During the drilling period, the maximum hook load value in the predicted hook load curve shows an increasing trend in the process of lowering the column for a set number of consecutive times, and the growth rate each time is greater than the preset growth rate threshold value for several consecutive times; or during the drilling period, the maximum hook load value in the predicted hook load curve shows a decreasing trend in the process of lowering the column for a set number of consecutive times, and the decline rate each time is greater than the preset decline rate threshold value for several consecutive times, then the differential pressure stuck drill problem is a differential pressure stuck drill section;

[0062] During the drilling period, the maximum hook load value in the predicted hook load curve shows an increasing trend in the process of lowering the column for a set number of consecutive times, and the growth rate each time is less than the preset growth rate threshold value for several consecutive times; or during the drilling period, the maximum hook load value in the predicted hook load curve shows a decreasing trend in the process of lowering the column for a set number of consecutive times, and the decline rate each time is less than the preset decline rate threshold value for several consecutive times, then the differential pressure stuck drill problem is close to the differential pressure stuck drill section;

[0063] During drilling, if the maximum hook load value in the predicted hook load curve shows a decreasing trend in the process of lowering the column for a consecutive set number of times, or during drilling, if the maximum hook load value in the predicted hook load curve shows an increasing trend in the process of lowering the column for a consecutive set number of times, the pressure difference drill sticking problem is in the normal section.

[0064] Step S600, issuing an early warning based on the predicted differential pressure drill sticking problem.

[0065] In some implementations, an early warning is issued based on the predicted differential pressure pipe sticking problem, including:

[0066] If the predicted differential pressure sticking problem is a differential pressure sticking section, a first-level warning is issued;

[0067] If the predicted differential pressure sticking problem is close to the differential pressure sticking section, a second-level warning will be issued.

[0068] The method provided in this embodiment is to sort out and classify the historical data of historical wells with differential pressure stuck drill problems into differential pressure stuck drill sections, near differential pressure stuck drill sections and normal sections, so as to train a machine learning model based on the historical data of historical wells without differential pressure stuck drill problems and the sorted and classified historical data of historical wells with differential pressure stuck drill problems. Thus, the hook load curve can be predicted by passing the real-time data of the field wells into the trained machine learning model, and based on the numerical trend of the predicted hook load curve and the current tripping and drilling conditions, it can be predicted whether the differential pressure stuck drill problem is a differential pressure stuck drill section, a near differential pressure stuck drill section or a normal section, and a corresponding warning can be issued.

[0069] In actual application, the method provided in this embodiment is used to predict differential pressure sticking during the construction process of key wells in an oilfield exploration block. The machine learning model in this method can predict differential pressure sticking three hours in advance when the hook load does not change significantly, which can effectively avoid the occurrence of drilling situations, reduce the probability and frequency of drilling engineering risks, and shorten the well construction cycle.

[0070] Embodiment 2

[0071] This embodiment provides a pressure differential drill bit sticking prediction device, such as Figure 2 As shown, including:

[0072] The acquisition module 100 is used to acquire historical data of historical wells with differential pressure sticking problems and historical data of historical wells without differential pressure sticking problems. The historical data of historical wells without differential pressure sticking problems and historical data of historical wells with differential pressure sticking problems include time series, hook load, drill bit position, column number and lifting weight of corresponding historical wells.

[0073] The classification module 200 is used to classify the historical data of the historical wells with the differential pressure stuck pipe problem according to the corresponding differential pressure stuck pipe problem, wherein the classification of the differential pressure stuck pipe problem includes the differential pressure stuck pipe section, the near differential pressure stuck pipe section and the normal section.

[0074] The processing module 300 is used to perform data preprocessing on the historical data of the historical wells without the differential pressure sticking problem and the historical data of the historical wells with the differential pressure sticking problem. The data preprocessing includes data cleaning to eliminate noise points in the data.

[0075] The model building module 400 is used to perform model training based on the pre-processed historical data of historical wells without differential pressure sticking problems and the pre-processed historical data of historical wells with differential pressure sticking problems to obtain a machine learning model. Specifically, it includes: performing model training and model performance evaluation based on the pre-processed historical data of historical wells without differential pressure sticking problems and the pre-processed historical data of historical wells with differential pressure sticking problems; if the evaluation result does not meet the preset model performance standard, continuing model training and model performance evaluation until the evaluation result meets the preset model performance standard to obtain a machine learning model.

[0076] Among them, the model training and model performance evaluation based on the pre-processed historical data of historical wells without pressure differential sticking problems and the pre-processed historical data of historical wells with pressure differential sticking problems include: dividing the pre-processed historical data of historical wells without pressure differential sticking problems and the pre-processed historical data of historical wells with pressure differential sticking problems into training sets, validation sets and test sets according to preset proportions; normalizing the data in the training sets, validation sets and test sets; training the model based on the normalized training set, and performing model performance evaluation and adjusting model hyperparameters based on the normalized validation set during the model training process; and evaluating the performance and generalization ability of the trained model based on the normalized test set to obtain an evaluation result.

[0077] The prediction module 500 is used to predict the real-time data of the field well through the machine learning model to obtain the predicted hook load curve, and predict the pressure difference pipe sticking problem according to the numerical trend of the predicted hook load curve and the current drilling condition. The prediction of the pressure difference pipe sticking problem according to the numerical trend of the predicted hook load curve and the current drilling condition includes:

[0078] During the drilling period, the maximum hook load value in the predicted hook load curve shows an increasing trend in the process of lowering the column for a set number of consecutive times, and the growth rate each time is greater than the preset growth rate threshold value for several consecutive times; or during the drilling period, the maximum hook load value in the predicted hook load curve shows a decreasing trend in the process of lowering the column for a set number of consecutive times, and the decline rate each time is greater than the preset decline rate threshold value for several consecutive times, then the differential pressure stuck drill problem is a differential pressure stuck drill section;

[0079] During the drilling period, the maximum hook load value in the predicted hook load curve shows an increasing trend in the process of lowering the column for a set number of consecutive times, and the growth rate each time is less than the preset growth rate threshold value for several consecutive times; or during the drilling period, the maximum hook load value in the predicted hook load curve shows a decreasing trend in the process of lowering the column for a set number of consecutive times, and the decline rate each time is less than the preset decline rate threshold value for several consecutive times, then the differential pressure stuck drill problem is close to the differential pressure stuck drill section;

[0080] During drilling, if the maximum hook load value in the predicted hook load curve shows a decreasing trend in the process of lowering the column for a consecutive set number of times, or during drilling, if the maximum hook load value in the predicted hook load curve shows an increasing trend in the process of lowering the column for a consecutive set number of times, the pressure difference drill sticking problem is in the normal section.

[0081] The early warning module 600 is used to issue an early warning based on the predicted pressure difference pipe sticking problem.

[0082] In some implementations, an early warning is issued based on the predicted differential pressure pipe sticking problem, including:

[0083] If the predicted differential pressure sticking problem is a differential pressure sticking section, a first-level warning is issued;

[0084] If the predicted differential pressure sticking problem is close to the differential pressure sticking section, a second-level warning will be issued.

[0085] The device provided in this embodiment classifies the historical data of historical wells with differential pressure stuck drill problems into differential pressure stuck drill sections, near differential pressure stuck drill sections, and normal sections, thereby training a machine learning model based on the historical data of historical wells without differential pressure stuck drill problems and the classified historical data of historical wells with differential pressure stuck drill problems. Thus, the hook load curve can be predicted by transferring the real-time data of the field wells into the trained machine learning model, and based on the numerical trend of the predicted hook load curve and the current tripping and drilling conditions, it can be predicted whether the differential pressure stuck drill problem is a differential pressure stuck drill section, a near differential pressure stuck drill section, or a normal section, and a corresponding warning can be issued.

[0086] Embodiment 3

[0087] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by at least one processor, the pressure differential drill sticking prediction method based on machine learning of the aforementioned embodiment is implemented.

[0088] Among them, the computer-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, disk or optical disk.

[0089] Embodiment 4

[0090] This embodiment provides an electronic device, including a memory and at least one processor, wherein a computer program is stored in the memory, and when the computer program is executed by the at least one processor, the method of the above-mentioned embodiment is implemented.

[0091] The processor may be an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a controller, a microcontroller unit (MCU), a microprocessor or other electronic components to implement the differential pressure stuck drill prediction method based on machine learning in the above embodiment.

[0092] Embodiment 5

[0093] This embodiment provides a computer program product. When the computer program product is run on a processor, it executes the pressure differential pipe sticking prediction method based on machine learning of the aforementioned embodiment.

[0094] In practical applications, the computer program product can be implemented and run in electronic devices.

[0095] In the several embodiments provided in the embodiments of the present invention, it should be understood that the disclosed apparatus and method can also be implemented in other ways. The apparatus and method embodiments described above are merely illustrative.

[0096] It should be noted that, in this article, the terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. The terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0097] Although the embodiments disclosed in the present invention are as above, the above contents are only embodiments adopted for facilitating the understanding of the present invention and are not intended to limit the present invention. Any technician in the technical field to which the present invention belongs can make any modifications and changes in the form and details of the implementation without departing from the spirit and scope disclosed in the present invention, but the patent protection scope of the present invention shall still be subject to the scope defined in the attached claims.

Claims

1. A method for predicting differential pressure stuck pipe based on machine learning, characterized in that: include: Obtain historical data of historical wells with differential pressure sticking problems and historical data of historical wells without differential pressure sticking problems; Classifying historical data of historical wells with differential pressure stuck drill problems according to corresponding differential pressure stuck drill problems, wherein the classification of differential pressure stuck drill problems includes differential pressure stuck drill section, near differential pressure stuck drill section and normal section; Performing data preprocessing on the historical data of the historical wells without the differential pressure sticking problem and the historical data of the historical wells with the differential pressure sticking problem; Based on the pre-processed historical data of historical wells without the problem of differential pressure sticking and the pre-processed historical data of historical wells with the problem of differential pressure sticking, model training is performed to obtain a machine learning model; Predicting by using the machine learning model according to the real-time data of the on-site well to obtain a predicted hook load curve; Predict differential pressure sticking based on the predicted hook load curve numerical trend and current tripping conditions; Early warning is given based on the predicted differential pressure problem.

2. The differential pressure pipe sticking prediction method according to claim 1, characterized in that: The method of predicting the differential pressure sticking problem based on the numerical trend of the predicted hook load curve and the current tripping and drilling conditions includes: During the drilling period, the maximum hook load value in the predicted hook load curve shows an increasing trend in the process of lowering the column for a set number of consecutive times, and the growth rate each time is greater than the preset growth rate threshold value for several consecutive times; or during the drilling period, the maximum hook load value in the predicted hook load curve shows a decreasing trend in the process of lowering the column for a set number of consecutive times, and the decline rate each time is greater than the preset decline rate threshold value for several consecutive times, then the differential pressure stuck drill problem is a differential pressure stuck drill section; During the drilling period, the maximum hook load value in the predicted hook load curve shows an increasing trend in the process of lowering the column for a set number of consecutive times, and the growth rate each time is less than the preset growth rate threshold value for several consecutive times; or during the drilling period, the maximum hook load value in the predicted hook load curve shows a decreasing trend in the process of lowering the column for a set number of consecutive times, and the decline rate each time is less than the preset decline rate threshold value for several consecutive times, then the differential pressure stuck drill problem is close to the differential pressure stuck drill section; During drilling, if the maximum hook load value in the predicted hook load curve shows a decreasing trend in the process of lowering the column for a consecutive set number of times, or during drilling, if the maximum hook load value in the predicted hook load curve shows an increasing trend in the process of lowering the column for a consecutive set number of times, the pressure difference drill sticking problem is in the normal section.

3. The differential pressure pipe sticking prediction method according to claim 1, characterized in that: The method of performing model training based on the pre-processed historical data of historical wells without the problem of differential pressure sticking and the pre-processed historical data of historical wells with the problem of differential pressure sticking to obtain a machine learning model includes: Model training and model performance evaluation are performed based on pre-processed historical data of historical wells without differential pressure sticking problems and pre-processed historical data of historical wells with differential pressure sticking problems; If the evaluation result does not meet the preset model performance standard, model training and model performance evaluation will continue until the evaluation result meets the preset model performance standard to obtain a machine learning model.

4. The differential pressure pipe sticking prediction method according to claim 1, characterized in that: The data preprocessing includes: cleaning the historical data of the historical wells without the differential pressure sticking problem and the historical data of the historical wells with the differential pressure sticking problem.

5. The differential pressure pipe sticking prediction method according to claim 1, characterized in that: The historical data of the historical wells without the pressure differential sticking problem and the historical data of the historical wells with the pressure differential sticking problem include the time series, hook load, drill bit position, column number and lifting weight of the corresponding historical wells.

6. The differential pressure pipe sticking prediction method according to claim 1, characterized in that: The early warning of the drill bit sticking problem according to the predicted differential pressure includes: If the predicted differential pressure sticking problem is a differential pressure sticking section, a first-level warning is issued; If the predicted differential pressure sticking problem is close to the differential pressure sticking section, a second-level warning will be issued.

7. The differential pressure pipe sticking prediction method according to claim 3, characterized in that: The model training and model performance evaluation based on the pre-processed historical data of historical wells without the differential pressure sticking problem and the pre-processed historical data of historical wells with the differential pressure sticking problem include: Divide the preprocessed historical data of the historical wells without the differential pressure sticking problem and the preprocessed historical data of the historical wells with the differential pressure sticking problem into a training set, a validation set and a test set according to a preset ratio; Normalizing the data in the training set, validation set, and test set; Perform model training based on the normalized training set, and evaluate model performance and adjust model hyperparameters based on the normalized validation set during model training; The performance and generalization ability of the trained model are evaluated based on the normalized test set to obtain the evaluation results.

8. A pressure differential drill sticking prediction device, characterized in that: include: An acquisition module, used for acquiring historical data of historical wells with differential pressure sticking problems and historical data of historical wells without differential pressure sticking problems; A classification module, for classifying historical data of historical wells with differential pressure stuck drill problems according to corresponding differential pressure stuck drill problems, wherein the classification of the differential pressure stuck drill problem includes differential pressure stuck drill section, near differential pressure stuck drill section and normal section; A processing module, used for performing data preprocessing on the historical data of the historical wells without the differential pressure sticking problem and the historical data of the historical wells with the differential pressure sticking problem; A model building module, for performing model training based on pre-processed historical data of historical wells without differential pressure sticking problems and pre-processed historical data of historical wells with differential pressure sticking problems to obtain a machine learning model; A prediction module, used to predict the real-time data of the field well through the machine learning model to obtain a predicted hook load curve, and predict the differential pressure sticking problem according to the numerical trend of the predicted hook load curve and the current drilling conditions; The early warning module is used to issue an early warning based on the predicted pressure difference drill sticking problem.

9. An electronic device, characterized in that: The method comprises a memory and at least one processor, wherein a computer program is stored in the memory, and when the computer program is executed by the at least one processor, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by at least one processor, the method according to any one of claims 1 to 7 is implemented.

Images