Method and system for determining casing deformation prevention perforation parameters based on SVM model and numerical simulation

Abstract

The invention belongs to the technical field of casing damage, and particularly relates to a method and system for determining casing deformation prevention perforation parameters based on an SVM model and numerical simulation. The method comprises the steps that a double-perforating-bullet numerical model is established; casing stress is obtained through the double-perforating-bullet numerical model, and the casing stress serves as a data set; based on the data set and an SVM learning algorithm, constructing a pre-trained casing perforation stress calculation model; and utilizing the pre-trained casing perforation stress calculation model to determine casing deformation prevention perforation parameters. The method overcomes the defects that when traditional numerical calculation software is used for calculating the casing stress, input parameters are many, parameter adjustment is complex, the calculation speed is low, post-processing is needed and the like, and has the advantages of being friendly to users, high in calculation efficiency, convenient to use and the like. The effective stress of the casing pipe can be quickly calculated according to perforation parameter data selected by a user, whether the casing pipe fails or not is judged, and the optimal perforation parameter combination is provided.

Description

Technical Field

[0001] This disclosure belongs to the field of casing damage technology, and specifically relates to a method and system for determining anti-casing variable orifice parameters based on SVM model and numerical simulation. Background Technology

[0002] Perforation technology connects the formation at the bottom of the well to the wellbore through perforation, effectively improving the connection channel between the two and increasing the efficiency of hydraulic fracturing operations, thereby significantly improving oil well productivity. This advantage has led to the widespread application of perforated completion technology in the oil and gas extraction field both domestically and internationally. However, perforation technology also has some drawbacks and shortcomings. In perforation engineering, the enormous heat generated by the explosive detonation melts the shaped charge liner, forming a metal jet. Under the detonation force, this jet penetrates the rock formation and wellbore structure. During the perforation penetration stage, the jet inevitably causes damage and deformation to the metal casing, affecting its structural integrity. After casing deformation, the overall stress uniformity of the casing is disrupted, and the deformed area becomes a weak point. Under the action of complex geostress and local high stress, casing deformation further exacerbates the damage to the wellbore structure, reducing the productivity and service life of oil and gas wells. In recent years, in particular, casing deformation problems have become significant in shale gas wells in Fuling, shale oil horizontal wells in Jiyang, and the Tarim Oilfield in my country, leading to increased extraction costs and a sharp decline in production, severely hindering the development of shale gas in the country. Therefore, ensuring the stability of the casing structure during perforation operations is crucial.

[0003] Perforation parameters are a key factor in perforation operations, and their design directly affects the degree of damage and deformation of the perforated casing. Regarding the impact of perforation parameters on the structural integrity of the casing, some scholars have proposed optimization methods. For example, Wang Haiyang et al. proposed a method for predicting formation fracture pressure in perforated wells and optimizing construction parameters. This method adjusts and optimizes construction parameters based on the magnitude of formation fracture pressure at the perforation hole. By predicting the formation fracture pressure of vertical and horizontal wells at any perforation angle, they further optimize the design of perforation parameters and casing type. Hu Zheyu et al. used elasticity and the dichotomy method to determine the formation fracture pressure during sandblasting perforation at different perforation phase angles, proposing the influence law of perforation phase angle on fracture pressure. Li Zhong et al., by determining the maximum production capacity of oilfields corresponding to different perforation process parameters, proposed a method for predicting production capacity based on perforation parameters and determined the optimal perforation parameters within each stage. The above-mentioned perforation parameter design optimization methods are all based on ensuring the production efficiency of oil and gas wells. They ignore the damage to the wellbore integrity caused by the perforation parameter design during perforation operations, which still poses certain safety hazards to the stability of oil and gas wells and is not conducive to the long-term development of oil and gas wells.

[0004] In addition, a series of studies have been conducted by relevant scholars. GEKing conducted experimental research on the external extrusion load on perforated casing, studying the influence of high perforation density on the external extrusion load on the casing, and obtained the extrusion resistance strength coefficient Ktest of perforated casing. Dou Yihua et al. studied the mechanical properties of perforated casing through experiments, and obtained the calculation formulas for the casing hole stress intensity factor and its correction coefficient, as well as the calculation formula for the remaining extrusion resistance strength of perforated casing. Guo Yonggui et al., based on the deep-water oil reservoir in the Gulf of Mexico, conducted high-density perforation tests and used finite element simulation to analyze perforated and unperforated casing, which to some extent explained the failure mode and instability failure mechanism of perforated casing. Liu Xianbo et al. used a double-perforated elastic finite element model to study the influence of perforation parameters such as perforation density and perforation phase angle on the effective stress of the casing, and revealed the stress law at the position between the two holes of the perforated casing. Many scholars have analyzed the impact of perforation parameters on casing integrity by establishing theoretical models. However, these theoretical models make too many assumptions and cannot accurately simulate the complex environment inside the well. While numerical simulation methods for casing stress calculation have obvious advantages, they also have disadvantages such as high professional requirements, many input parameters, complex parameter adjustment, slow calculation speed, and the need for post-processing. Summary of the Invention

[0005] To address the above problems, this disclosure provides a method for determining the parameters of the anti-sleeve variable aperture based on an SVM model and numerical simulation, the method comprising:

[0006] Establish a numerical model for a double-perforation projectile;

[0007] The casing stress was obtained through a numerical model of a double-perforated spring, and the casing stress was used as a dataset.

[0008] Based on the dataset and SVM learning algorithm, a pre-trained casing perforation stress calculation model is constructed.

[0009] The parameters for preventing casing perforation variation were determined using a pre-trained casing perforation stress calculation model.

[0010] Furthermore, the casing stress is obtained through a numerical model of a double-perforation spring, including:

[0011] Determine the combination of geometric parameters, the range of variation of the main control factor parameters, and the step size of variation of the main control factor parameters in the numerical model of the dual-perforation projectile. Perform the following steps for each set of geometric parameters:

[0012] Input the initial values ​​of the main control factor parameters into the dual-perforation numerical model to obtain the casing stress corresponding to the current input value;

[0013] Adjust the values ​​of the main control factor parameters according to the step size of the change, and then input them into the dual-perforation numerical model to obtain the casing stress corresponding to the current input value.

[0014] Furthermore, based on the dataset and the SVM learning algorithm, a pre-trained casing perforation stress calculation model is constructed, including:

[0015] Construct an initial casing perforation stress calculation model;

[0016] By using a dataset and an SVM learning algorithm, with the goal of minimizing the model loss of the initial casing perforation stress calculation model, the model parameters of the initial casing perforation stress calculation model are fine-tuned to obtain a pre-trained casing perforation stress calculation model.

[0017] Furthermore, using a dataset and an SVM learning algorithm, with the goal of minimizing the model loss of the initial casing perforation stress calculation model, the model parameters of the initial casing perforation stress calculation model are fine-tuned to obtain a pre-trained casing perforation stress calculation model, including:

[0018] Obtain the input vector and the desired output, wherein the input vector is a combination of geometric parameters and the desired output is the sleeve stress;

[0019] Input the input vector into the initial casing perforation stress calculation model to obtain the actual output value;

[0020] By comparing the expected output with the actual output, the output deviation is obtained;

[0021] Determine whether the output deviation meets the preset conditions;

[0022] If not, determine whether the current training has reached the training count threshold;

[0023] When the training count threshold is reached, the current training ends, and the pre-trained casing perforation stress calculation model is obtained.

[0024] If the training count threshold is not reached, the hidden layer unit error is calculated, the error gradient is calculated, and the weights of each function in the hidden layer are adjusted.

[0025] Furthermore, the anti-casing perforation parameters are determined using a pre-trained casing perforation stress calculation model, including:

[0026] By inputting the geometric parameters and the main control parameters into the pre-trained casing perforation stress calculation model, the actual casing stress is obtained.

[0027] Determine whether the casing strength has failed based on the Mises yield strength criterion and the actual casing stress, and determine the anti-casing perforation parameters.

[0028] This disclosure also proposes a system for determining the parameters of the anti-sleeve variable aperture based on an SVM model and numerical simulation, the system comprising:

[0029] Establish a module for building a numerical model of a dual-perforation projectile;

[0030] The dataset module is used to obtain the casing stress through the numerical model of the double-perforation spring and to use the casing stress as the dataset.

[0031] The building block is used to construct a pre-trained casing perforation stress calculation model based on the dataset and the SVM learning algorithm.

[0032] The determination module is used to determine the anti-casing perforation parameters using a pre-trained casing perforation stress calculation model.

[0033] Furthermore, the dataset module, used to obtain the casing stress through a dual-perforation spring numerical model, includes:

[0034] The dataset module is used to determine the combination of geometric parameters, the range of variation of the main control factor parameters, and the step size of variation of the main control factor parameters in the numerical model of the dual-perforation projectile. The following steps are performed for each set of geometric parameters:

[0035] Input the initial values ​​of the main control factor parameters into the dual-perforation numerical model to obtain the casing stress corresponding to the current input value;

[0036] Adjust the values ​​of the main control factor parameters according to the step size of the change, and then input them into the dual-perforation numerical model to obtain the casing stress corresponding to the current input value.

[0037] Furthermore, a building block is provided for constructing a pre-trained casing perforation stress calculation model based on the dataset and the SVM learning algorithm, including:

[0038] The building block is used to construct the initial casing perforation stress calculation model;

[0039] By using a dataset and an SVM learning algorithm, with the goal of minimizing the model loss of the initial casing perforation stress calculation model, the model parameters of the initial casing perforation stress calculation model are fine-tuned to obtain a pre-trained casing perforation stress calculation model.

[0040] Furthermore, using a dataset and an SVM learning algorithm, with the goal of minimizing the model loss of the initial casing perforation stress calculation model, the model parameters of the initial casing perforation stress calculation model are fine-tuned to obtain a pre-trained casing perforation stress calculation model, including:

[0041] Obtain the input vector and the desired output, wherein the input vector is a combination of geometric parameters and the desired output is the sleeve stress;

[0042] Input the input vector into the initial casing perforation stress calculation model to obtain the actual output value;

[0043] By comparing the expected output with the actual output, the output deviation is obtained;

[0044] Determine whether the output deviation meets the preset conditions;

[0045] If not, determine whether the current training has reached the training count threshold;

[0046] When the training count threshold is reached, the current training ends, and the pre-trained casing perforation stress calculation model is obtained.

[0047] If the training count threshold is not reached, the hidden layer unit error is calculated, the error gradient is calculated, and the weights of each function in the hidden layer are adjusted.

[0048] Furthermore, a module is defined to determine the anti-casing perforation parameters using a pre-trained casing perforation stress calculation model, including:

[0049] The determining module is used to input the combination of geometric parameters and the main control parameters into the pre-trained casing perforation stress calculation model to obtain the actual casing stress.

[0050] Determine whether the casing strength has failed based on the Mises yield strength criterion and the actual casing stress, and determine the anti-casing perforation parameters.

[0051] This disclosure has the following beneficial effects:

[0052] This disclosure establishes a training dataset for a casing stress calculation model based on simulation results from numerical calculation software under different parameter combinations. It employs the SVM algorithm to extract features from the dataset and establishes a casing stress calculation model, which greatly improves computational efficiency. It overcomes the shortcomings of traditional numerical calculation software in calculating casing stress, such as numerous input parameters, complex parameter adjustments, slow calculation speed, and the need for post-processing. It features user-friendliness, high computational efficiency, and ease of use. It can quickly calculate the effective stress of the casing based on the perforation parameter data selected by the user, determine whether the casing has failed, and provide the optimal perforation parameter combination.

[0053] Other features and advantages of this disclosure will be set forth in the following description and will be apparent in part from the description or may be learned by practicing the disclosure. The objects and other advantages of this disclosure may be realized and obtained by means of the structures pointed out in the description and the accompanying drawings. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1This diagram illustrates a method for determining the parameters of the anti-sleeve variable aperture based on an SVM model and numerical simulation in an embodiment of this disclosure.

[0056] Figure 2 This invention discloses a detailed flowchart of a method for determining the parameters of the anti-sleeve variable aperture based on an SVM model and numerical simulation in an embodiment of the present invention.

[0057] Figure 3 This diagram illustrates a numerical model of a dual-perforation projectile penetrating a casing-cement sheath-rock stratum in an embodiment of this disclosure.

[0058] Figure 4 This diagram shows the effective stress value cloud after the casing double perforation in an embodiment of this disclosure;

[0059] Figure 5 This diagram illustrates the data set generation process in an embodiment of the present disclosure.

[0060] Figure 6 This diagram illustrates the structure of a deep learning model in an embodiment of this disclosure.

[0061] Figure 7 This diagram illustrates the maximum margin hyperplane of the SVM algorithm in an embodiment of this disclosure.

[0062] Figure 8 This diagram illustrates the neural network training process in an embodiment of the present disclosure.

[0063] Figure 9 This diagram illustrates the target analysis points of the casing in an embodiment of this disclosure.

[0064] Figure 10 This diagram illustrates a system for determining the parameters of the anti-sleeve variable aperture based on an SVM model and numerical simulation in an embodiment of this disclosure. Detailed Implementation

[0065] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this disclosure more comprehensive and complete, and to fully convey the concept of the example embodiments to those skilled in the art. The described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a full understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced with one or more of the specific details omitted, or other methods, components, apparatus, steps, etc., can be employed. In other instances, well-known technical solutions are not shown or described in detail to avoid obscuring various aspects of this disclosure.

[0066] Furthermore, the accompanying drawings are merely illustrative of this disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware units or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0067] The flowchart shown in the attached diagram is merely an illustrative example and does not necessarily include all steps. For example, some steps may be broken down, while others may be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0068] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented, for example, in orders other than those illustrated or described herein.

[0069] Furthermore, the terms “comprising” and “having”, and any variations thereof, are intended to cover non-exclusive inclusion, such that a process, method, system, product, or device that includes a series of steps or sub-modules is not necessarily limited to those steps or sub-modules that are explicitly listed, but may include other steps or sub-modules that are not explicitly listed or that are inherent to such process, method, product, or device.

[0070] like Figure 1 As shown, this disclosure proposes a method for determining the parameters of the anti-sleeve variable aperture based on an SVM model and numerical simulation, the method comprising:

[0071] Establish a numerical model for a double-perforation projectile;

[0072] The casing stress was obtained through a numerical model of a double-perforated spring, and the casing stress was used as a dataset.

[0073] Based on the dataset and SVM learning algorithm, a pre-trained casing perforation stress calculation model is constructed.

[0074] The parameters for preventing casing perforation variation were determined using a pre-trained casing perforation stress calculation model.

[0075] The detailed technical solution of this method is as follows: Figure 2 As shown, the steps are as follows:

[0076] Step 1: Use numerical simulation software to establish a numerical model of the double-perforation spring and analyze the stress sensitivity factors of the casing.

[0077] Considering the actual wellbore structure and related perforation engineering parameters, the commercial finite element software ANSYS / LS-DYNA was used for modeling, such as... Figure 3 As shown, the model consists of a perforating projectile, casing, cement sheath, and rock formation. The overall height of the model is 300cm. Parameters such as the casing inner diameter and wall thickness can be selected with reference to the actual wellbore structure. The model adopts a dual-perforating projectile setup method. The perforating projectile model selected in the actual perforation operation should be used as the modeling basis, and the established perforating projectile model and parameters should be consistent with the parameters of that perforating projectile model. During the model building process, considering that the entire perforation model is a symmetrical structure, only a 1 / 4 three-dimensional model of casing-cement sheath-formation needs to be built. At the same time, considering the complexity of the numerical simulation process of perforation operation, in order to improve computational efficiency and ensure computational accuracy, only the air domain is used as the propagation medium of the perforation jet. The model only considers the jet envelope and the location through which the jet passes.

[0078] During the model mesh generation process, considering that the numerical calculation model focuses on analyzing the damage of the casing and the jet development process, and that the formation and movement of the jet are the result of the interaction between the explosive and the shaped charge liner, the mesh generation of both is particularly important. In order to avoid model mesh distortion caused by unreasonable mesh generation, which would seriously affect the calculation results, the air domain and the casing-cement ring part were meshed.

[0079] The numerical simulation of perforation explosion described in this invention employs a traditional point-detonation method. The centers of the two perforating projectiles along the wellbore axis are designated as the detonation points for the two projectiles. Pressure is applied to the outer wall of the rock strata in the model to simulate the pressure of deep formations on the wellbore. Free boundary conditions are set for the upper and lower walls of the formation to simulate infinitely bounded formations. Considering the interference caused by the reflection of explosive detonation waves, the outer boundaries of the air domain are set as non-reflective boundaries. Considering the large deformation of the perforating projectiles and mesh distortion caused by the explosive explosion during perforation, the Arbitrary Lagrange-Euler (ALE) multi-material algorithm is applied to the perforating projectile section (shaped charge liner, explosive, outer shell) and the air domain, allowing the metal jet formed after the explosion to flow freely within the defined ALE mesh. The casing-cement sheath-formation (CCF) section uses the Lagrange algorithm, and a fluid-structure interaction method is used to connect the meshes at the interface between the two meshes. The established dual-perforating projectile numerical model and calculation results are as follows: Figure 3 , 4 As shown.

[0080] Step 2: Calculate the effective stress value of the casing under different controlling factors, such as perforation density, perforation phase angle, casing internal pressure, casing thickness, cement sheath elastic modulus, and rock layer elastic modulus.

[0081] There are many factors affecting casing damage and deformation during the perforation stage, which can be summarized into 13 main controlling factors: perforation density, perforation phase angle, wellbore size, casing internal pressure, casing thickness, casing inner and outer diameters, cement sheath elastic modulus, cement sheath Poisson's ratio, rock formation elastic modulus, rock formation Poisson's ratio, maximum horizontal stress, minimum horizontal stress, and vertical stress. Among these, the combination of wellbore diameter, casing thickness, and casing inner and outer diameters is relatively fixed, and its parameter combinations are generally determined based on the actual engineering situation (see Table 1). The parameter variation range and variation step size of the remaining 10 main controlling factors are determined based on the analysis of on-site drilling and completion operations (see Table 2).

[0082] Table 1

[0083] Wellbore size (mm) Outer diameter of the sleeve (mm) Casing wall thickness (mm) 215.9 127 11.1 / 12.14 215.9 139.7 9.17 / 10.54 / 12.09 / 12.34 / 12.7 215.9 145.6 13.49

[0084] Table 2

[0085] Parameter name Parameter variation range Change step size Perforation density (number of holes / m) 12-40 7 Perforation phase angle (°) 60-180 30 Internal pressure of the casing (MPa) 60-140 20 Elastic modulus of cement ring (GPa) 2-10 2 Cement ring Poisson's ratio 0.1-0.3 0.1 Formation elastic modulus (GPa) 20-80 20 Poisson's ratio of formations 0.1-0.3 0.1 Maximum horizontal ground stress (MPa) 30-90 20 Minimum horizontal ground stress (MPa) 20-65 15 Vertical ground stress (MPa) 30-90 20

[0086] Based on the parameter variation range and step size indicated in the table above, the effective stress at the center of the double-hole casing under different parameter combinations can be simulated using the double-perforation spring numerical model, generating 5*5*5*5*3*4*3*4*4*4=1,440,000 sets of data, forming a dataset.

[0087] Step 3: Save the casing stress calculation results and corresponding parameters as a sample case data, and continue to perform numerical calculations on the next set of data until all case calculations are completed. Then, establish a dataset for SVM neural network algorithm calculations.

[0088] The flowchart for generating large datasets is as follows: Figure 5 As shown, the specific content is as follows:

[0089] ① Confirm the combined geometric parameters of the dual-perforation shell numerical model in Table 1, including the borehole size, casing outer diameter, and casing wall thickness;

[0090] ② Select the initial values ​​of the main control factors of casing stress in Table 2 (the minimum value in the parameter range), and then use numerical simulation software to establish a numerical model of double-perforated spring. After inputting the selected parameters, calculate the effective stress at the middle position of the double perforation.

[0091] ③ After calculating a set of parameters, keep the geometric parameters unchanged, change the combination of main control factor parameters according to the value range and change step size of the main control factors in Table 2, and continue to calculate the stress results;

[0092] ④ Continue to calculate and combine the stress results of all main control parameter combinations according to the above method to form a dataset;

[0093] ⑤ Change the combined geometric parameters of the numerical model of the double-perforation projectile, complete the casing stress calculation of the main controlling factor parameter combination under all geometric parameter combinations, and thus form a complete dataset.

[0094] Step 4: Divide the dataset into training set, test set and validation set. Divide the data obtained from the numerical calculation of the sample Case according to the ratio of "training set: test set: validation set = 70%: 15%: 15%".

[0095] Step 5: Design the SVM (Support Vector Machine) algorithm model. Based on the input and output parameters, determine the number of neurons in the input layer, output layer, and hidden layer, and determine the activation function.

[0096] A casing perforation stress calculation model was trained using the SVM learning algorithm. The SVM algorithm performs binary classification of data using a supervised learning approach. Its core idea is to find a hyperplane, also known as the decision boundary, that maximizes the margin between the two classes, thereby achieving good classification results. This algorithm model transforms the raw data into a format suitable for the SVM model and performs necessary processing, such as feature scaling and outlier handling. Appropriate SVM parameters, including the kernel function and regularization parameter C, are selected to minimize the sum of squared errors of the network. The SVM model is trained using the training data and the selected parameters. In this stage, the SVM finds an optimal hyperplane that minimizes the distance between the hyperplane and the sample points, thus obtaining a regression function that fits the training set samples as closely as possible.

[0097] like Figure 6 The deep learning model structure includes an input layer and an SVM algorithm, as shown in the figure. The algorithm consists of an input layer, a hidden layer, and an output layer. For each type of perforation density, perforation diameter, perforation depth, and perforation phase angle, each sample vector is output. Then, the SVM algorithm model is used for decision-making and classification.

[0098] SVM finds the hyperplane that maximizes the distance between sample points, also known as the maximum margin hyperplane. Any hyperplane can be described by the following linear equation:

[0099] ω T x+b=0;

[0100] In the formula, ω is the weight vector, T is the training dataset in the feature space, and b is the bias term.

[0101] The formula for the distance from a point (x, y) in two-dimensional space to the line Ax + By + C = 0 is:

[0102]

[0103] In the formula, A, B, and C are the target linear correlation coefficients.

[0104] After extending to n-dimensional space, the point x = (x1, x2, x3, ..., x...). n ) to w T The distance x+b=0 is:

[0105]

[0106] In the formula, ω n For x = x n The weight vector at that point.

[0107] according to Figure 7 We know that the support vector (i.e., the sample point closest to the hyperplane) is d from the hyperplane, and the distances of other points to the hyperplane are greater than d. Therefore, we can conclude that:

[0108]

[0109] Since ||w||d>0, the above expression can be transformed into:

[0110] y(w T x+b)≥1;

[0111] Then we can obtain Figure 7 The upper and lower hyperplanes of the large-margin hyperplane. The distance from each support vector to the hyperplane is:

[0112]

[0113] Therefore, the optimization problem obtained is:

[0114]

[0115] In the formula, yi is the label of the i-th data point, and xi is the corresponding feature vector.

[0116] Step 6: Use the dataset to perform vector regression calculations, tests, and verifications, establish a casing stress calculation system model, and deploy it in the cloud.

[0117] like Figure 8 As shown, the SVM model is defined with input vectors (defining combinations of parameters such as perforation parameters and perforation phase) and desired output (effective stress at the center of the two perforations of the casing). The model calculates the input vectors based on the hidden layer function, solving for the corresponding effective stress at the center of the two perforations of the casing. The output layer displays the calculation results based on different parameter combinations. The output deviation E is obtained by comparing the desired output with the actual calculation results of the output layer, and the output deviation is judged. If the preset conditions are met, which are the user-defined tolerances, the calculation ends; otherwise, the SVM model continues to be trained, and the weights of each function in the hidden layer are continuously adjusted until the accuracy of the SVM model meets the requirements or the maximum number of training iterations is reached, thus completing the SVM model training and obtaining a mature SVM algorithm model for perforated casing stress calculation.

[0118] A mature perforated casing stress SVM algorithm model is deployed in the cloud. Users input actual parameter values ​​in the client and send them to the perforated casing stress model trained in the cloud. The casing stress result is calculated by the cloud perforated casing stress calculation model and then sent to the client.

[0119] Step 7: Determine the condition of the casing based on the casing failure evaluation criteria.

[0120] Because the area between the two perforations is highly sensitive to the stress generated by changes in perforation density, this disclosure selects the effective stress value of the casing at this location as the analysis object, and then analyzes the variation law of casing stress under different perforation parameters. Figure 9 A schematic diagram of the target analysis points.

[0121] The effective stress value of the casing at the center of the two perforation holes is obtained based on the casing perforation stress calculation model. The stress state of the casing is determined according to the Mises yield strength criterion.

[0122] Step 8: Select the optimal perforation parameters based on the effective stress value of the casing.

[0123] After the user inputs the perforation parameters on the client, the data is sent to the cloud for calculation to obtain the effective stress value at the center position of the two holes of the casing after perforation. The cloud-based perforated casing stress model calculates the casing stress result through the perforated casing stress calculation model, determines whether the casing has failed based on the casing strength, and then gives the optimal result.

[0124] like Figure 10 As shown, this disclosure also proposes a system for determining the parameters of the anti-sleeve variable aperture based on an SVM model and numerical simulation, the system comprising:

[0125] Establish a module for building a numerical model of a dual-perforation projectile;

[0126] The dataset module is used to obtain the casing stress through the numerical model of the double-perforation spring and to use the casing stress as the dataset.

[0127] The building block is used to construct a pre-trained casing perforation stress calculation model based on the dataset and the SVM learning algorithm.

[0128] The determination module is used to determine the anti-casing perforation parameters using a pre-trained casing perforation stress calculation model.

[0129] Those skilled in the art should understand that, despite the detailed description of this disclosure with reference to the foregoing embodiments, modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this disclosure.

Claims

1. A method for determining the parameters of the anti-sleeve variable aperture based on SVM model and numerical simulation, characterized in that, The method includes: Establish a numerical model for a double-perforation projectile; The casing stress was obtained through a numerical model of a double-perforated spring, and the casing stress was used as a dataset. Based on the dataset and SVM learning algorithm, a pre-trained casing perforation stress calculation model is constructed. The parameters for preventing casing perforation variation were determined using a pre-trained casing perforation stress calculation model.

2. The method for determining the parameters of the anti-sleeve variable aperture based on SVM model and numerical simulation according to claim 1, characterized in that, The casing stress was obtained using a numerical model of a double-perforated spring, including: Determine the combination of geometric parameters, the range of variation of the main control factor parameters, and the step size of variation of the main control factor parameters in the numerical model of the dual-perforation projectile. Perform the following steps for each set of geometric parameters: Input the initial values ​​of the main control factor parameters into the dual-perforation numerical model to obtain the casing stress corresponding to the current input value; Adjust the values ​​of the main control factor parameters according to the step size of the change, and then input them into the dual-perforation numerical model to obtain the casing stress corresponding to the current input value.

3. The method for determining the parameters of the anti-sleeve variable aperture based on SVM model and numerical simulation according to claim 1, characterized in that, Based on the dataset and SVM learning algorithm, a pre-trained casing perforation stress calculation model is constructed, including: Construct an initial casing perforation stress calculation model; By using a dataset and an SVM learning algorithm, with the goal of minimizing the model loss of the initial casing perforation stress calculation model, the model parameters of the initial casing perforation stress calculation model are fine-tuned to obtain a pre-trained casing perforation stress calculation model.

4. The method for determining the parameters of the anti-sleeve variable aperture based on SVM model and numerical simulation according to claim 3, characterized in that, By using a dataset and an SVM learning algorithm, and aiming to minimize the model loss of the initial casing perforation stress calculation model, the model parameters of the initial casing perforation stress calculation model are fine-tuned to obtain a pre-trained casing perforation stress calculation model, including: Obtain the input vector and the desired output, wherein the input vector is a combination of geometric parameters and the desired output is the sleeve stress; Input the input vector into the initial casing perforation stress calculation model to obtain the actual output value; By comparing the expected output with the actual output, the output deviation is obtained; Determine whether the output deviation meets the preset conditions; If not, determine whether the current training has reached the training count threshold; When the training count threshold is reached, the current training ends, and the pre-trained casing perforation stress calculation model is obtained. If the training count threshold is not reached, the hidden layer unit error is calculated, the error gradient is calculated, and the weights of each function in the hidden layer are adjusted.

5. The method for determining the parameters of the anti-sleeve variable aperture based on SVM model and numerical simulation according to claim 2, characterized in that, The parameters for preventing casing perforation variation are determined using a pre-trained casing perforation stress calculation model, including: By inputting the geometric parameters and the main control parameters into the pre-trained casing perforation stress calculation model, the actual casing stress is obtained. Determine whether the casing strength has failed based on the Mises yield strength criterion and the actual casing stress, and determine the anti-casing perforation parameters.

6. A system for determining the parameters of the anti-sleeve variable aperture based on SVM model and numerical simulation, characterized in that, The system includes: Establish a module for building a numerical model of a dual-perforation projectile; The dataset module is used to obtain the casing stress through the numerical model of the double-perforation spring and to use the casing stress as the dataset. The building block is used to construct a pre-trained casing perforation stress calculation model based on the dataset and the SVM learning algorithm. The determination module is used to determine the anti-casing perforation parameters using a pre-trained casing perforation stress calculation model.

7. The system for determining the parameters of the anti-sleeve variable aperture based on SVM model and numerical simulation according to claim 6, characterized in that, The dataset module, used to obtain casing stress through a dual-perforation spring numerical model, includes: The dataset module is used to determine the combination of geometric parameters, the range of variation of the main control factor parameters, and the step size of variation of the main control factor parameters in the numerical model of the dual-perforation projectile. The following steps are performed for each set of geometric parameters: Input the initial values ​​of the main control factor parameters into the dual-perforation numerical model to obtain the casing stress corresponding to the current input value; Adjust the values ​​of the main control factor parameters according to the step size of the change, and then input them into the dual-perforation numerical model to obtain the casing stress corresponding to the current input value.

8. The system for determining the parameters of the anti-sleeve variable aperture based on SVM model and numerical simulation according to claim 6, characterized in that, The building blocks are used to construct a pre-trained casing perforation stress calculation model based on the dataset and the SVM learning algorithm, including: The building block is used to construct the initial casing perforation stress calculation model; By using a dataset and an SVM learning algorithm, with the goal of minimizing the model loss of the initial casing perforation stress calculation model, the model parameters of the initial casing perforation stress calculation model are fine-tuned to obtain a pre-trained casing perforation stress calculation model.

9. The system for determining the parameters of the anti-sleeve variable aperture based on SVM model and numerical simulation according to claim 6, characterized in that, By using a dataset and an SVM learning algorithm, and aiming to minimize the model loss of the initial casing perforation stress calculation model, the model parameters of the initial casing perforation stress calculation model are fine-tuned to obtain a pre-trained casing perforation stress calculation model, including: Obtain the input vector and the desired output, wherein the input vector is a combination of geometric parameters and the desired output is the sleeve stress; Input the input vector into the initial casing perforation stress calculation model to obtain the actual output value; By comparing the expected output with the actual output, the output deviation is obtained; Determine whether the output deviation meets the preset conditions; If not, determine whether the current training has reached the training count threshold; When the training count threshold is reached, the current training ends, and the pre-trained casing perforation stress calculation model is obtained. If the training count threshold is not reached, the hidden layer unit error is calculated, the error gradient is calculated, and the weights of each function in the hidden layer are adjusted.

10. The system for determining the parameters of the anti-sleeve variable aperture based on SVM model and numerical simulation according to claim 7, characterized in that, The determination module is used to determine the anti-casing perforation parameters using a pre-trained casing perforation stress calculation model, including: The determining module is used to input the combination of geometric parameters and the main control parameter values ​​into the pre-trained casing perforation stress calculation model to obtain the actual casing stress. Determine whether the casing strength has failed based on the Mises yield strength criterion and the actual casing stress, and determine the anti-casing perforation parameters.

Images