Horizontal well pumping bridge plug perforation parameter prediction method, system, equipment and medium

Abstract

The invention belongs to the field of horizontal well multi-section hydraulic fracturing, and particularly discloses a horizontal well pumping bridge plug perforation parameter prediction method, system and device and a medium, and the method comprises the steps that horizontal well related parameters and corresponding site construction actual pumping displacement and cable tension are collected; dividing the well bore track into a plurality of infinitesimal sections, and constructing a training database by taking relevant parameters of the infinitesimal sections as basic learning units; establishing a neural network prediction model, and training and optimizing the neural network prediction model; dividing the well body track of the horizontal well to be predicted into a plurality of infinitesimal sections, taking the relevant parameters of the infinitesimal sections as input, outputting prediction results of the pumping displacement and the cable tension of all the infinitesimal sections, and integrating to generate prediction results of the pumping displacement and the cable tension of the whole well. According to the method, autonomous learning and continuous optimization of the prediction model are realized, the accuracy of simulation and prediction of construction parameters is guaranteed, and the construction success rate is improved. The method is suitable for predicting the pumping displacement and the cable tension.

Description

Technical Field

[0001] This invention belongs to the field of horizontal well multi-stage hydraulic fracturing, specifically a method, system, equipment, and medium for predicting perforation parameters of a horizontal well pumped bridge plug. Background Technology

[0002] With the deepening understanding of unconventional oil and gas reservoirs in my country, pumped bridge plug perforation has become a crucial technology. In recent years, the development of the oil and gas industry and continuous technological advancements have led to an increasing demand for precise control and optimization of pumped bridge plug perforation operations in horizontal wells. Simulating and predicting construction parameters to guide pumped bridge plug perforation operations can ensure accurate control, improve the success rate of construction, and reduce the incidence of engineering accidents.

[0003] Chinese patent CN117390816A, entitled "Method for Establishing a Bridge Plug Pumping Model and Bridge Plug Pumping Method," proposes a method for establishing a bridge plug pumping model and a bridge plug pumping method. This method simplifies the shape of the bridge plug tool string to a cylindrical shape; performs force analysis on the bridge plug tool string to determine the forces acting on it in its direction of movement; and establishes a bridge plug pumping model based on the relationship between the forces acting on the bridge plug tool string and the pumping parameters. The method then substitutes the predetermined parameters into the bridge plug pumping model to obtain reference values ​​for the undetermined parameters, thus providing effective guidance for setting the undetermined parameters during bridge plug pumping. However, this method simplifies the shape of the bridge plug tool string to a cylindrical shape, which may overlook some important practical factors. Furthermore, the established model is relatively fixed, lacking the ability to learn and optimize autonomously, making it difficult to adapt to complex patterns and changes in data, and its generalization ability may be insufficient, thus limiting the model's predictive ability.

[0004] Chinese patent CN114329956A, entitled "A Method, System, and Terminal for Controlling Critical Discharge Rate of Pumped Perforation in Horizontal Wells," proposes a method, system, and terminal for controlling critical discharge rate of pumped perforation in horizontal wells. This method establishes a critical discharge rate model by acquiring horizontal well parameters, perforation gun string parameters, and pumping fluid parameters and performing force balance analysis. It then collects the inclination angle of the perforation gun string and the real-time pumping discharge rate, inputting the inclination angle into the critical discharge rate model to calculate the critical pumping discharge rate. Finally, a control strategy for controlling the perforation gun string discharge rate is generated based on the real-time pumping discharge rate and the critical pumping discharge rate. However, this method establishes a relatively fixed model, lacking the ability to learn and optimize autonomously. It is not easily adaptable to complex patterns and changes in the data, and its generalization ability may be insufficient, thus limiting the model's predictive capability.

[0005] Therefore, in order to ensure the accuracy of the simulated and predicted construction parameters and improve the success rate of construction, it is urgent to establish a method for predicting the perforation parameters of horizontal well pumped bridge plugs with autonomous learning and optimization capabilities to guide the construction of pumped bridge plug perforations. Summary of the Invention

[0006] The purpose of this invention is to provide a method, system, equipment, and medium for predicting perforation parameters of horizontal well pumped bridge plugs, so as to achieve autonomous learning and continuous optimization of the prediction model, ensure the accuracy of simulated and predicted construction parameters, and improve the success rate of construction.

[0007] To achieve the above objectives, the present invention employs the following technical methods:

[0008] A method for predicting perforation parameters of a pumped bridge plug in a horizontal well includes the following steps performed sequentially:

[0009] S1. Collection well basic parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, cable parameters, and corresponding actual pumping displacement and cable tension during on-site construction;

[0010] S2. Divide the well trajectory into multiple micro-segments, and use the basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of the micro-segments, as well as the corresponding actual pumping displacement and cable tension in the field construction, as the basic learning units to build a training database.

[0011] S3. Establish a neural network prediction model, using the basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of the micro-segment in the training database as inputs, and the corresponding actual pumping displacement and cable tension in the field as outputs, to train and optimize the neural network prediction model.

[0012] S4. Divide the well trajectory of the horizontal well to be predicted into multiple micro-segments. Use the basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of each micro-segment as input to the optimized neural network prediction model. The optimized neural network prediction model outputs the prediction results of pumping displacement and cable tension for all micro-segments. Then, integrate the prediction results of all micro-segments to generate the prediction results of pumping displacement and cable tension for the entire well.

[0013] As a limitation: the hyperparameters of the neural network prediction model are optimized using a genetic algorithm. These hyperparameters include the number of layers in the neural network and the number of nodes in each layer. The specific optimization process is as follows:

[0014] Set the range of values ​​for the number of layers and the number of nodes in each layer of the neural network prediction model, and then select multiple sets of hyperparameter data within the range to form an initial population;

[0015] The selected hyperparameter data is converted into binary code. Then, based on the predicted values ​​obtained by the neural network prediction model using the selected hyperparameter data, the fitness of each set of hyperparameter data is evaluated. The sets of hyperparameter data with the highest fitness are retained. New hyperparameter data are generated through crossover and mutation and added to the population to keep the population size the same as the initial population size.

[0016] Determine whether the set number of generations has been reached. If it has, output the hyperparameter data with the highest fitness in the population as the hyperparameter data for the neural network prediction model. If it has not been reached, repeat the previous step until the set number of generations is reached.

[0017] To further define this: the neural network prediction model specifically includes an input layer, hidden layers, and an output layer, with the calculation formula for each node in each layer being:

[0018] Y = WX + B,

[0019] In the formula, Y represents the output of the node, X represents the features of the input node, W represents the weight of the node, and B represents the bias of the node; nonlinearity is introduced through an activation function, specifically the sigmoid function. Represents the predicted value matrix;

[0020] Constructing the loss function for:

[0021]

[0022] In the formula, z represents the predicted value of the j-th sample for the i-th feature. ij Let represent the actual value of the j-th sample on the i-th feature, n represent the number of features, p represent the number of samples, tr represent the trace of the matrix, Z represent the actual value matrix, and T represent the transpose of the matrix.

[0023] To further define this, training and optimizing the neural network prediction model specifically involves: before training begins, initializing the weights and biases in the neural network, propagating the input data forward through the layers of the neural network, calculating the output of each layer until the output of the last layer, calculating the value of the loss function, and continuously updating the weights and biases in the neural network using the gradient descent algorithm until the preset maximum number of iterations is reached, so that the value of the loss function is minimized.

[0024] The update rules for updating the weights and biases in a neural network using the gradient descent algorithm are as follows:

[0025]

[0026]

[0027] Where η is the learning rate. It is a loss function Partial derivatives with respect to the weights, It is a loss function Partial derivative with respect to the bias;

[0028] Calculate according to the chain rule and The principle is and The specific calculation process is as follows:

[0029]

[0030] Get ⊙ indicates that elements at corresponding positions are multiplied;

[0031] Similarly, we can obtain

[0032] As a limitation: basic well parameters include wellbore diameter; well trajectory parameters include vertical depth, sounding depth, inclination angle, and azimuth angle; wellhead sealing parameters include sealing resin viscosity, gap between the baffle and the cable, baffle length, wellhead pressure, pressure of the grease pump above the wellhead, baffle sealing pressure, sealing resin density, coefficient of friction between the cable and the wellbore, cable type, cable tensile strength, fluid density, and fluid viscosity; tubing string parameters include bridge plug length, perforating gun outer diameter, bridge plug weight, bridge plug outer diameter, perforating gun weight, perforating gun length, total weight of the tool string, and total length of the tool string; pumping parameters include cable lowering speed, cable pulling speed, and coefficient of friction between the fracturing bridge plug and the wellbore; cable parameters include cable diameter, cable weight, and cable safety tensile strength.

[0033] This invention also provides a horizontal well pumped bridge plug perforation parameter prediction system, comprising:

[0034] The data collection module is used to collect basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, cable parameters, as well as the corresponding actual pumping displacement and cable tension during on-site construction.

[0035] The training database construction module is used to divide the well trajectory into multiple micro-segments. The basic learning units are the well basic parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of the micro-segments, as well as the corresponding actual pumping displacement and cable tension in the field construction.

[0036] The model building and training module is used to build a neural network prediction model. It takes the basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters and cable parameters of the micro-segment in the training database as inputs, and the corresponding actual pumping displacement and cable tension in the field construction as outputs to train and optimize the neural network prediction model.

[0037] The prediction module is used to divide the well trajectory of the horizontal well to be predicted into multiple micro-segments. The basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of the micro-segments are used as inputs to the optimized neural network prediction model. The optimized neural network prediction model outputs the prediction results of pumping displacement and cable tension.

[0038] The present invention also provides an electronic device, including a memory and a processor. The memory stores a computer program, and the processor calls the computer program in the memory to execute the above-described method for predicting perforation parameters of a horizontal well pumped bridge plug.

[0039] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, is used to implement the above-described method for predicting perforation parameters of a horizontal well pumped bridge plug.

[0040] The beneficial effects achieved by this invention, due to the adoption of the above-described solution, compared with the prior art, are as follows:

[0041] A method, system, equipment, and medium for predicting perforation parameters of a horizontal well pumped bridge plug are disclosed. This method establishes a neural network prediction model, using basic well parameters, wellbore trajectory parameters, wellhead sealing parameters, tubing string parameters, pumping parameters, and cable parameters as inputs, and the corresponding actual pumping displacement and cable tension in the field as outputs. The neural network prediction model is trained and optimized, and then used to predict the pumping displacement and cable tension of the horizontal well pumped bridge plug perforation. This achieves autonomous learning and continuous optimization of the prediction model, resulting in greater overall flexibility and adaptability, ensuring the accuracy of the simulated and predicted construction parameters, effectively guiding horizontal well pumped bridge plug perforation operations, and improving the success rate of construction. The method comprehensively considers the influence of basic well parameters, wellbore trajectory parameters, wellhead sealing parameters, tubing string parameters, pumping parameters, and cable parameters, ensuring the accuracy of the simulated and predicted construction parameters. Furthermore, the use of a genetic algorithm to optimize the hyperparameters of the neural network prediction model improves the model's convergence speed, avoids getting trapped in local optima, and enhances the prediction accuracy of the neural network prediction model.

[0042] This invention is applicable to the prediction of pumping displacement and cable tension. Attached Figure Description

[0043] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0044] Figure 1 This is a flowchart of a method for predicting perforation parameters of a horizontal well pumped bridge plug according to Embodiment 1 of the present invention;

[0045] Figure 2 This is a graph showing the relationship between the pumping capacity of the horizontal well CJHC4-2-6H predicted in Embodiment 1 of the present invention and the well depth, inclination angle, and azimuth angle.

[0046] Figure 3 This is a graph showing the relationship between the predicted cable tension of the horizontal well CJHC4-2-6H and the well depth, inclination angle, and azimuth angle in Embodiment 1 of the present invention.

[0047] Figure 4 This is a structural block diagram of a horizontal well pumped bridge plug perforation parameter prediction system according to Embodiment 2 of the present invention;

[0048] Figure 5 This is a schematic diagram of the structure of the electronic device in Embodiment 3 of the present invention. Detailed Implementation

[0049] The present invention will be further described below with reference to the embodiments. However, those skilled in the art should understand that the present invention is not limited to the following embodiments. Any improvements and equivalent changes made based on the specific embodiments of the present invention are within the scope of protection of the claims of the present invention.

[0050] Example 1: A method for predicting perforation parameters of a horizontal well pumped bridge plug

[0051] A method for predicting perforation parameters of a pumped bridge plug in a horizontal well includes the following steps performed sequentially:

[0052] S1. Collection well basic parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, cable parameters, and corresponding actual pumping displacement and cable tension during on-site construction;

[0053] Among them, the basic parameters of a well include the wellbore diameter;

[0054] Well trajectory parameters include vertical depth, well depth sounding, well inclination angle, and azimuth angle;

[0055] Wellhead sealing parameters include sealing resin viscosity, gap between the flow-blocking pipe and the cable, flow-blocking pipe length, wellhead pressure, grease injection pump pressure above wellhead pressure, flow-blocking pipe sealing pressure, sealing resin density, friction coefficient between the cable and the wellbore, cable type, cable tensile strength, fluid density, and fluid viscosity.

[0056] The parameters of the tubing string include bridge plug length, perforator gun outer diameter, bridge plug weight, bridge plug outer diameter, perforator gun weight, perforator gun length, total weight of the tool string, and total length of the tool string; the total weight of the tool string is the sum of the bridge plug weight and the perforator gun weight, and the total length of the tool string is the sum of the bridge plug length and the perforator gun length.

[0057] Pumping parameters include cable lowering speed, cable pulling speed, and friction coefficient between fracturing bridge plug and wellbore;

[0058] Cable parameters include cable diameter, cable weight, and cable safety tensile strength.

[0059] S2. Divide the well trajectory into multiple micro-segments, and use the basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of the micro-segments, as well as the corresponding actual pumping displacement and cable tension in the field construction, as the basic learning units to build a training database.

[0060] S3. Establish a neural network prediction model, using the basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of the micro-segment in the training database as inputs, and the corresponding actual pumping displacement and cable tension in the field as outputs, to train and optimize the neural network prediction model.

[0061] The specific steps for establishing a neural network prediction model are as follows: The neural network prediction model includes an input layer, hidden layers, and an output layer. The calculation formula for each node in each layer is as follows:

[0062] Y = WX + B,

[0063] In the formula, Y represents the output of the node, X represents the features of the input node, W represents the weight of the node, and B represents the bias of the node; nonlinearity is introduced through an activation function, specifically the sigmoid function. Represents the predicted value matrix;

[0064] Constructing the loss function for:

[0065]

[0066] In the formula, z represents the predicted value of the j-th sample for the i-th feature. ij Let represent the actual value of the j-th sample on the i-th feature, n represent the number of features, p represent the number of samples, tr represent the trace of the matrix, Z represent the actual value matrix, and T represent the transpose of the matrix;

[0067] The hyperparameters of a neural network prediction model are optimized using a genetic algorithm. These hyperparameters include the number of layers in the neural network and the number of nodes in each layer. The specific optimization process is as follows:

[0068] Set the range of values ​​for the number of layers and the number of nodes in each layer of the neural network prediction model, and then select multiple sets of hyperparameter data within the range to form an initial population;

[0069] The selected hyperparameter data is converted into binary code. Then, based on the predicted values ​​obtained by the neural network prediction model using the selected hyperparameter data, the fitness of each set of hyperparameter data is evaluated. The sets of hyperparameter data with the highest fitness are retained. New hyperparameter data are generated through crossover and mutation and added to the population to keep the population size the same as the initial population size.

[0070] Determine whether the set number of generations has been reached. If it has, output the hyperparameter data with the highest fitness in the population as the hyperparameter data for the neural network prediction model. If it has not been reached, repeat the previous step until the set number of generations has been reached.

[0071] The specific steps for training and optimizing a neural network prediction model are as follows: Before training begins, the weights and biases in the neural network are initialized, the input data is propagated forward through the layers of the neural network, the output of each layer is calculated until the output of the last layer is reached, the value of the loss function is calculated, and the weights and biases in the neural network are continuously updated through the gradient descent algorithm until the preset maximum number of iterations is reached, so that the value of the loss function is minimized.

[0072] The update rules for updating the weights and biases in a neural network using the gradient descent algorithm are as follows:

[0073]

[0074] Where η is the learning rate. It is a loss function Partial derivatives with respect to the weights, It is a loss function Partial derivative with respect to the bias;

[0075] Calculate according to the chain rule and The principle is and The specific calculation process is as follows:

[0076]

[0077]

[0078] Get ⊙ indicates that elements at corresponding positions are multiplied;

[0079] Similarly, we can obtain

[0080] S4. Divide the well trajectory of the horizontal well to be predicted into multiple micro-segments. Use the basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of each micro-segment as input to the optimized neural network prediction model. The optimized neural network prediction model outputs the prediction results of pumping displacement and cable tension for all micro-segments. Then, integrate the prediction results of all micro-segments to generate the prediction results of pumping displacement and cable tension for the entire well.

[0081] The prediction method of this embodiment is applied to predict the pumping displacement and cable tension during the pumping bridge plug perforation operation of a horizontal well CJHC4-2-6H in a certain area. The well trajectory of the horizontal well CJHC4-2-6H is divided into multiple micro-segments. The basic well parameters, well trajectory parameters, wellhead sealing parameters, tubing string parameters, pumping parameters, and cable parameters of each micro-segment are used as inputs to the optimized neural network prediction model. The basic well parameters, wellhead sealing parameters, tubing string parameters, pumping parameters, cable parameters, and some well trajectory parameters of the horizontal well CJHC4-2-6H are input to the optimized neural network training model, as shown in Tables 1-5. The optimized neural network prediction model outputs the predicted pumping displacement and cable tension. The relationship between the predicted pumping displacement and well depth, inclination angle, and azimuth angle is as follows: Figure 2 As shown, the predicted cable tension is related to the well depth, well inclination angle, and azimuth angle as follows: Figure 3 As shown. Based on the predicted pumping capacity and cable tension, the horizontal well pumping bridge plug perforation operation is guided, improving the success rate of the operation.

[0082] Table 1. Basic well parameters and cable parameters for horizontal well CJHC4-2-6H

[0083]

[0084]

[0085] Table 2. String parameters for horizontal well CJHC4-2-6H

[0086]

[0087] Table 3 Pumping parameters for horizontal well CJHC4-2-6H

[0088]

[0089] Table 4 Wellhead sealing parameters for horizontal well CJHC4-2-6H

[0090]

[0091] Table 5. Partial wellbore trajectory parameters of horizontal well CJHC4-2-6H

[0092]

[0093]

[0094] Example 2: A Horizontal Well Pumped Bridge Plug Perforation Parameter Prediction System

[0095] A horizontal well pumped bridge plug perforation parameter prediction system, the structural block diagram of which is shown below. Figure 4 As shown, it includes:

[0096] The data collection module is used to collect basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, cable parameters, as well as the corresponding actual pumping displacement and cable tension during on-site construction.

[0097] Among them, the basic parameters of a well include the wellbore diameter;

[0098] Well trajectory parameters include vertical depth, well depth sounding, well inclination angle, and azimuth angle;

[0099] Wellhead sealing parameters include sealing resin viscosity, gap between the flow-blocking pipe and the cable, flow-blocking pipe length, wellhead pressure, grease injection pump pressure above wellhead pressure, flow-blocking pipe sealing pressure, sealing resin density, friction coefficient between the cable and the wellbore, cable type, cable tensile strength, fluid density, and fluid viscosity.

[0100] The parameters of the tubing string include bridge plug length, perforator gun outer diameter, bridge plug weight, bridge plug outer diameter, perforator gun weight, perforator gun length, total weight of the tool string, and total length of the tool string; the total weight of the tool string is the sum of the bridge plug weight and the perforator gun weight, and the total length of the tool string is the sum of the bridge plug length and the perforator gun length.

[0101] Pumping parameters include cable lowering speed, cable pulling speed, and friction coefficient between fracturing bridge plug and wellbore;

[0102] Cable parameters include cable diameter, cable weight, and cable safety tensile strength.

[0103] The training database construction module is used to divide the well trajectory into multiple micro-segments. The basic learning units are the well basic parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of the micro-segments, as well as the corresponding actual pumping displacement and cable tension in the field construction.

[0104] The model building and training module is used to build a neural network prediction model. It takes the basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters and cable parameters of the micro-segment in the training database as inputs, and the corresponding actual pumping displacement and cable tension in the field construction as outputs to train and optimize the neural network prediction model.

[0105] The specific steps for establishing a neural network prediction model are as follows: The neural network prediction model includes an input layer, hidden layers, and an output layer. The calculation formula for each node in each layer is as follows:

[0106] Y = WX + B,

[0107] In the formula, Y represents the output of the node, X represents the features of the input node, W represents the weight of the node, and B represents the bias of the node; nonlinearity is introduced through an activation function, specifically the sigmoid function. Represents the predicted value matrix;

[0108] Constructing the loss function for:

[0109]

[0110] In the formula, z represents the predicted value of the j-th sample for the i-th feature. ij Let represent the actual value of the j-th sample on the i-th feature, n represent the number of features, p represent the number of samples, tr represent the trace of the matrix, Z represent the actual value matrix, and T represent the transpose of the matrix;

[0111] The hyperparameters of a neural network prediction model are optimized using a genetic algorithm. These hyperparameters include the number of layers in the neural network and the number of nodes in each layer. The specific optimization process is as follows:

[0112] Set the range of values ​​for the number of layers and the number of nodes in each layer of the neural network prediction model, and then select multiple sets of hyperparameter data within the range to form an initial population;

[0113] The selected hyperparameter data is converted into binary code. Then, based on the predicted values ​​obtained by the neural network prediction model using the selected hyperparameter data, the fitness of each set of hyperparameter data is evaluated. The sets of hyperparameter data with the highest fitness are retained. New hyperparameter data are generated through crossover and mutation and added to the population to keep the population size the same as the initial population size.

[0114] Determine whether the set number of generations has been reached. If it has, output the hyperparameter data with the highest fitness in the population as the hyperparameter data for the neural network prediction model. If it has not been reached, repeat the previous step until the set number of generations has been reached.

[0115] The specific steps for training and optimizing a neural network prediction model are as follows: Before training begins, the weights and biases in the neural network are initialized, the input data is propagated forward through the layers of the neural network, the output of each layer is calculated until the output of the last layer is reached, the value of the loss function is calculated, and the weights and biases in the neural network are continuously updated through the gradient descent algorithm until the preset maximum number of iterations is reached, so that the value of the loss function is minimized.

[0116] The update rules for updating the weights and biases in a neural network using the gradient descent algorithm are as follows:

[0117]

[0118] Where η is the learning rate. It is a loss function Partial derivatives with respect to the weights, It is a loss function Partial derivative with respect to the bias;

[0119] Calculate according to the chain rule and The principle is and The specific calculation process is as follows:

[0120]

[0121] Get ⊙ indicates that elements at corresponding positions are multiplied;

[0122] Similarly, we can obtain

[0123] The prediction module is used to divide the well trajectory of the horizontal well to be predicted into multiple micro-segments. The basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of the micro-segments are used as inputs to the optimized neural network prediction model. The optimized neural network prediction model outputs the prediction results of pumping displacement and cable tension.

[0124] Example 3: An electronic device

[0125] The electronic device of this embodiment includes a memory and a processor. The memory stores a computer program, and the processor calls the computer program in the memory to execute a method for predicting perforation parameters of a horizontal well pumped bridge plug according to Embodiment 1. Figure 5 This is a schematic diagram of the structure of the electronic device provided in this embodiment. The electronic device can be a terminal device or a server. The terminal device can include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, personal digital assistants (PDAs), portable Android devices (PADs), portable media players (PMPs), and in-vehicle terminals (such as in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 5 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of this embodiment.

[0126] like Figure 5 As shown, electronic devices may include processing units, such as central processing units (CPUs) and graphics processors (GPUs), which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) or loaded from storage devices into random access memory (RAM). RAM also stores various programs and data required for the operation of the electronic device. The processing unit, ROM, and RAM are interconnected via a bus. Input devices, output devices, communication devices, and storage devices are also connected to the bus via I / O interfaces.

[0127] Typically, the following devices can be connected to an I / O interface: input devices such as touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices such as liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices such as magnetic tapes, hard drives, etc.; and communication devices. Communication devices allow electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 5 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown. More or fewer devices may be implemented or have alternatively.

[0128] Example 4: A computer-readable medium

[0129] The computer-readable storage medium of this embodiment stores a computer program, which, when executed by a processor, is used to implement a method for predicting perforation parameters of a horizontal well pumped bridge plug according to Embodiment 1. The computer-readable storage medium of this embodiment may be included in an electronic device; alternatively, it may exist independently and not assembled into an electronic device.

[0130] The computer-readable storage medium of this embodiment may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0131] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this disclosure, and are not intended to limit them. Although this disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this disclosure.

Claims

1. A method for predicting perforation parameters of a pumped bridge plug in a horizontal well, characterized in that, This includes the following steps performed sequentially: S1. Collection well basic parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, cable parameters, and corresponding actual pumping displacement and cable tension during on-site construction; S2. Divide the well trajectory into multiple micro-segments, and use the basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of the micro-segments, as well as the corresponding actual pumping displacement and cable tension in the field construction, as the basic learning units to build a training database. S3. Establish a neural network prediction model, using the basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of the micro-segment in the training database as inputs, and the corresponding actual pumping displacement and cable tension in the field as outputs, to train and optimize the neural network prediction model. S4. Divide the well trajectory of the horizontal well to be predicted into multiple micro-segments. Use the basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of each micro-segment as input to the optimized neural network prediction model. The optimized neural network prediction model outputs the prediction results of pumping displacement and cable tension for all micro-segments. Then, integrate the prediction results of all micro-segments to generate the prediction results of pumping displacement and cable tension for the entire well.

2. The method for predicting perforation parameters of a horizontal well pumped bridge plug according to claim 1, characterized in that, The hyperparameters of a neural network prediction model are optimized using a genetic algorithm. These hyperparameters include the number of layers in the neural network and the number of nodes in each layer. The specific optimization process is as follows: Set the range of values ​​for the number of layers and the number of nodes in each layer of the neural network prediction model, and then select multiple sets of hyperparameter data within the range to form an initial population; The selected hyperparameter data is converted into binary code. Then, based on the predicted values ​​obtained by the neural network prediction model using the selected hyperparameter data, the fitness of each set of hyperparameter data is evaluated. The sets of hyperparameter data with the highest fitness are retained. New hyperparameter data are generated through crossover and mutation and added to the population to keep the population size the same as the initial population size. Determine whether the set number of generations has been reached. If it has, output the hyperparameter data with the highest fitness in the population as the hyperparameter data for the neural network prediction model. If it has not been reached, repeat the previous step until the set number of generations is reached.

3. The method for predicting perforation parameters of a horizontal well pumped bridge plug according to claim 2, characterized in that, The specific steps for establishing a neural network prediction model are as follows: The neural network prediction model includes an input layer, hidden layers, and an output layer. The calculation formula for each node in each layer is as follows: Y = WX + B, In the formula, Y represents the output of the node, X represents the features of the input node, W represents the weight of the node, and B represents the bias of the node; nonlinearity is introduced through an activation function, specifically the sigmoid function. Represents the predicted value matrix; Constructing the loss function for: In the formula, Let z represent the predicted value of the j-th sample for the i-th feature. ij Let represent the actual value of the j-th sample on the i-th feature, n represent the number of features, p represent the number of samples, tr represent the trace of the matrix, Z represent the actual value matrix, and T represent the transpose of the matrix.

4. The method for predicting perforation parameters of a horizontal well pumped bridge plug according to claim 3, characterized in that, The specific steps for training and optimizing a neural network prediction model are as follows: Before training begins, the weights and biases in the neural network are initialized, the input data is propagated forward through the layers of the neural network, the output of each layer is calculated until the output of the last layer is reached, the value of the loss function is calculated, and the weights and biases in the neural network are continuously updated through the gradient descent algorithm until the preset maximum number of iterations is reached, so that the value of the loss function is minimized. The update rules for updating the weights and biases in a neural network using the gradient descent algorithm are as follows: Where η is the learning rate. It is a loss function Partial derivatives with respect to the weights, It is a loss function Partial derivative with respect to the bias; Calculate according to the chain rule and The principle is and The specific calculation process is as follows: Get ⊙ indicates that elements at corresponding positions are multiplied; Similarly, we can obtain 5. A method for predicting perforation parameters of a horizontal well pumped bridge plug according to any one of claims 1-4, characterized in that, Basic well parameters include wellbore diameter; well trajectory parameters include vertical depth, sounding depth, inclination angle, and azimuth angle; wellhead sealing parameters include sealing resin viscosity, gap between the baffle and the cable, baffle length, wellhead pressure, pressure difference between the grease pump and the wellhead, baffle sealing pressure, sealing resin density, coefficient of friction between the cable and the wellbore, cable type, cable tensile strength, fluid density, and fluid viscosity; tubing string parameters include bridge plug length, perforating gun outer diameter, bridge plug weight, bridge plug outer diameter, perforating gun weight, perforating gun length, total weight of the tool string, and total length of the tool string; pumping parameters include cable lowering speed, cable pulling speed, and coefficient of friction between the fracturing bridge plug and the wellbore; cable parameters include cable diameter, cable weight, and cable safety tensile strength.

6. A horizontal well pumped bridge plug perforation parameter prediction system, characterized in that, include: The data collection module is used to collect basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, cable parameters, as well as the corresponding actual pumping displacement and cable tension during on-site construction. The training database construction module is used to divide the well trajectory into multiple micro-segments. The basic learning units are the well basic parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of the micro-segments, as well as the corresponding actual pumping displacement and cable tension in the field construction. The model building and training module is used to build a neural network prediction model. It takes the basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters and cable parameters of the micro-segment in the training database as inputs, and the corresponding actual pumping displacement and cable tension in the field construction as outputs to train and optimize the neural network prediction model. The prediction module is used to divide the well trajectory of the horizontal well to be predicted into multiple micro-segments. The basic well parameters, well trajectory parameters, wellhead sealing parameters, pipe string parameters, pumping parameters, and cable parameters of the micro-segments are used as inputs to the optimized neural network prediction model. The optimized neural network prediction model outputs the prediction results of pumping displacement and cable tension.

7. An electronic device, characterized in that, It includes a memory and a processor. The memory stores a computer program, and the processor calls the computer program in the memory to execute the method for predicting perforation parameters of a horizontal well pumped bridge plug as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, A computer-readable storage medium stores a computer program, which, when executed by a processor, is used to implement the method for predicting perforation parameters of a horizontal well pumped bridge plug as described in any one of claims 1-5.

Images