Partition material detection method, device and equipment for pipeline ring welding joint and medium

Abstract

The invention discloses a method, device, equipment and medium for detecting partition materials of a pipeline ring welding joint, and relates to the technical field of pipeline structure detection.The method comprises the steps that the elastic-plastic constitutive relation and load-displacement data of materials of all areas in the pipeline ring welding joint are obtained; establishing a small punch test finite element model, and comparing the small punch test finite element model with actually obtained load-displacement data for verification; performing numerical simulation work on the successfully verified small punch test finite element model, and establishing a machine learning database meeting the constitutive relation parameter range of the pipeline steel material; and constructing a deep learning model, training by utilizing the machine learning database, inputting the load-displacement data corresponding to the to-be-detected pipeline ring welding head into the trained deep learning model, and outputting the constitutive relationship of the partition materials of the to-be-detected pipeline ring welding head. Therefore, the material constitutive relation of each area of the pipeline ring welding joint can be accurately and efficiently determined, and the problem that the material tensile constitutive relation of each area of the pipeline ring welding joint cannot be directly tested is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of pipeline structure detection, and in particular to a method, device, equipment and medium for detecting the materials of different zones of a pipeline girth weld joint. Background Art

[0002] Accurate testing and characterization of the constitutive relationship of pipeline girth welds are the basis for determining the constitutive relationships of different material zones of girth welds and conducting safety evaluations. In related technical solutions, there have been many studies on the constitutive relationships of weld materials, including a variety of different testing methods, such as thermal simulation method, hardness conversion method, indentation method, etc. However, these testing methods usually obtain the material constitutive relationship only with individual data points. Due to the uneven distribution of materials in each zone of the pipeline girth weld, it is easy to lead to inaccurate test results. In addition, the sizes of the weld zone and the heat affected zone are extremely small, and it is impossible to directly determine the tensile constitutive relationships of the materials in each zone of the pipeline girth weld joint. Summary of the Invention

[0003] The purpose of the present invention is to provide a method, device, equipment and medium for detecting the materials of different zones of a pipeline girth weld joint, which can accurately and efficiently determine the constitutive relationships of the materials in each zone of the pipeline girth weld joint and solve the problem that the tensile constitutive relationships of the materials in each zone of the pipeline girth weld joint cannot be directly tested.

[0004] To solve the above technical problems, the present invention provides a method for detecting the materials of different zones of a pipeline girth weld joint, the method comprising:

[0005] Obtaining the elastoplastic constitutive relationships and load-displacement data of the materials in each zone of the pipeline girth weld joint;

[0006] Establishing a small punch test finite element model, inputting the obtained elastoplastic constitutive relationships into the small punch test finite element model, and comparing the output results with the obtained load-displacement data to verify the small punch test finite element model;

[0007] Performing numerical simulation on the successfully verified small punch test finite element model, and establishing a machine learning database that meets the parameter range of the constitutive relationships of pipeline steel materials;

[0008] Constructing a deep learning model, and training the deep learning model using the machine learning database;

[0009] Inputting the load-displacement data corresponding to the pipeline girth weld joint to be measured into the trained deep learning model, and outputting the constitutive relationships of the materials in different zones of the pipeline girth weld joint to be measured.

[0010] In the first aspect, in the above method for detecting the materials of different zones of a pipeline girth weld joint provided by the present invention, obtaining the elastoplastic constitutive relationships of the materials in each zone of the pipeline girth weld joint includes:

[0011] Obtain the elastoplastic constitutive relations of the materials in each region of the pipeline girth weld joint obtained from the quasi-static tensile test;

[0012] Among them, the quasi-static tensile test calculates the axial engineering strain of each pipeline girth weld joint specimen through the change in the axial length between virtual extensometers; combines the axial engineering strain with the axial load recorded by the in-situ tensile testing machine to calculate the engineering stress data and engineering strain data of each pipeline girth weld joint specimen; converts the engineering strain data and engineering stress data before the maximum load of each pipeline girth weld joint specimen into true strain and true stress data to obtain the true stress-strain curves of the materials in each region of the pipeline girth weld joint parent sample.

[0013] On the other hand, in the above-mentioned pipeline girth weld joint partition material detection method provided by the present invention, the pipeline girth weld joint specimen is obtained by successively performing grinding and polishing treatment, weld erosion treatment and calibration of each region position on the pipeline girth weld joint parent sample; after the weld erosion treatment, the base metal region material, weld region material and heat-affected region material of the pipeline girth weld joint parent sample are revealed;

[0014] Each of the pipeline girth weld joint specimens corresponds to one of the base metal region material, weld region material and heat-affected region material.

[0015] On the other hand, in the above-mentioned pipeline girth weld joint partition material detection method provided by the present invention, obtaining the load-displacement data of the materials in each region of the pipeline girth weld joint includes:

[0016] Obtain the load-displacement data of the materials in each region of the pipeline girth weld joint obtained from the small punch test;

[0017] Among them, the small punch test uses a small punch testing machine to test the pipeline girth weld joint specimen, monitors the change in the displacement of the punch in the loading direction during the test using a displacement sensor, and records the load-displacement data of the pipeline girth weld joint specimen.

[0018] On the other hand, in the above-mentioned pipeline girth weld joint partition material detection method provided by the present invention, establishing a finite element model of the small punch test, inputting the obtained elastoplastic constitutive relations into the finite element model of the small punch test, and comparing the output results with the obtained load-displacement data to verify the finite element model of the small punch test, including:

[0019] A finite element model for the small punch test is established using finite element software; all components in the small punch test finite element model for simulating the small punch test are consistent with the geometric dimensions of the components in the small punch test; the tangential behavior of the contact friction between the pipe girth weld joint specimen in the small punch test finite element model and the components in the small punch simulation test is defined as a penalty function with a set friction coefficient, and the normal behavior is defined as hard contact. The deformable entities in the small punch test finite element model adopt four-node axisymmetric reduced integration elements;

[0020] The average value of the obtained elastoplastic constitutive relationship is input into the small punch test finite element model, and through numerical simulation, the load-displacement curve during the deformation process of the specimen is output;

[0021] The load-displacement curve output by the simulation is compared with the obtained load-displacement data to verify the rationality of the small punch test finite element model in terms of boundary conditions, mesh division, and model settings.

[0022] On the other hand, in the above-mentioned pipe girth weld joint partition material detection method provided by the present invention, numerical simulation work is carried out on the small punch test finite element model after successful verification, and a machine learning database that meets the constitutive relationship parameter range of pipeline steel materials is established, including:

[0023] Multiple groups of combinations of yield strength and hardening index are generated by means of random value selection;

[0024] According to the multiple groups of combinations of yield strength and hardening index generated, multiple groups of true stress-plastic strain curves of materials that conform to the power law hardening rule are obtained;

[0025] Using the multiple groups of obtained true stress-plastic strain curves, numerical simulation work is carried out on the small punch test finite element model after successful verification to obtain multiple groups of load-displacement curves and their corresponding material constitutive parameters, so as to establish a machine learning database that meets the constitutive relationship parameter range of pipeline steel materials.

[0026] On the other hand, in the above-mentioned pipe girth weld joint partition material detection method provided by the present invention, a deep learning model is constructed, and the machine learning database is used to train the deep learning model, including:

[0027] A deep learning model is constructed by adopting a structure combining a convolutional neural network and a multi-layer fully connected layer;

[0028] The load-displacement data before the maximum load of each group of load-displacement curves is extracted from the machine learning database as the original features, and the difference between adjacent coordinate points of the load data and displacement data in the original features is calculated as the high-order features;

[0029] Concatenate the original features and the high-order features to form an input feature matrix;

[0030] Perform normalization on the two target variables of yield strength and hardening index to construct a corresponding target value matrix;

[0031] Use the input feature matrix and the target value matrix to train the deep learning model.

[0032] To solve the above technical problems, the present invention also provides a device for determining the constitutive relationship of materials in each region of a pipeline girth weld joint. The device includes:

[0033] A data acquisition module for acquiring the elastoplastic constitutive relationship and load-displacement data of the materials in each region of the pipeline girth weld joint;

[0034] A finite element model establishment and verification module for establishing a small punch test finite element model, inputting the acquired elastoplastic constitutive relationship into the small punch test finite element model, and comparing the output results with the acquired load-displacement data to verify the small punch test finite element model;

[0035] A database establishment module for performing numerical simulation on the successfully verified small punch test finite element model to establish a machine learning database that meets the range of constitutive relationship parameters of pipeline steel materials;

[0036] A deep learning module construction and training module for constructing a deep learning model and training the deep learning model using the machine learning database;

[0037] A deep learning module inference module for inputting the load-displacement data corresponding to the pipeline girth weld joint to be measured into the trained deep learning model and outputting the constitutive relationship of the materials in the partitioned regions of the pipeline girth weld joint to be measured.

[0038] To solve the above technical problems, the present invention also provides an electronic device, which includes:

[0039] A memory for storing a computer program;

[0040] A processor for implementing the steps of the above pipeline girth weld joint partitioned material detection method when executing the computer program.

[0041] To solve the above technical problems, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above pipeline girth weld joint partitioned material detection method are implemented.

[0042] As can be seen from the above technical solution, a method for detecting the partition materials of a pipeline girth weld joint provided by the present invention includes: obtaining the elastoplastic constitutive relationship and load-displacement data of the materials in each area of the pipeline girth weld joint; establishing a small punch test finite element model, inputting the obtained elastoplastic constitutive relationship into the small punch test finite element model, and comparing the output result with the obtained load-displacement data to verify the small punch test finite element model; performing numerical simulation on the successfully verified small punch test finite element model to establish a machine learning database that meets the parameter range of the constitutive relationship of pipeline steel materials; constructing a deep learning model and training the deep learning model using the machine learning database; inputting the load-displacement data corresponding to the pipeline girth weld joint to be tested into the trained deep learning model, and outputting the constitutive relationship of the partition materials of the pipeline girth weld joint to be tested.

[0043] The beneficial effects of the present invention are as follows. For the method for detecting the partition materials of a pipeline girth weld joint provided by the present invention, first, the elastoplastic constitutive relationship and load-displacement data of the materials in each area of the pipeline girth weld joint are obtained, providing accurate basic data for subsequent steps. Then, a small punch test finite element model is established and compared with the actually obtained load-displacement data for verification, which can ensure the accuracy and reliability of the finite element model. The verified model can more realistically simulate the small punch test process, providing an effective tool for further numerical simulation and analysis, and reducing the wrong results caused by model errors. After that, numerical simulation is performed on the successfully verified small punch test finite element model to establish a machine learning database that meets the parameter range of the constitutive relationship of pipeline steel materials, providing rich and accurate data support for deep learning and helping the model learn more comprehensive constitutive relationship characteristics of materials. Finally, a deep learning model is constructed and trained using the machine learning database, and the load-displacement data corresponding to the pipeline girth weld joint to be tested is input into the trained deep learning model, which can quickly output the constitutive relationship of the partition materials of the pipeline girth weld joint to be tested. In this way, the constitutive relationship of the materials in each area of the pipeline girth weld joint can be accurately and efficiently determined according to the punch load-displacement data, overcoming the drawback of the traditional empirical formula method that only obtains the constitutive relationship of materials with individual data points, solving the problem that the tensile constitutive relationship of the materials in each area of the pipeline girth weld joint cannot be directly tested, providing another way for the efficient testing and characterization of the constitutive relationship of the materials in each area of the pipeline girth weld joint, and at the same time providing a data basis for the safety assessment and life evaluation of the pipeline girth weld joint and pipeline structure, which has high guiding significance for engineering practice.

[0044] In addition, the present invention also provides a corresponding device, electronic device, and computer-readable storage medium for determining the constitutive relationship of the materials in each area of the pipeline girth weld joint for the method for detecting the partition materials of the pipeline girth weld joint, which have the same or corresponding technical features as the above-mentioned method for detecting the partition materials of the pipeline girth weld joint, and the effects are the same. Brief Description of the Drawings

[0045] To more clearly illustrate the embodiments of the present invention, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0046] Figure 1 It is a flowchart of the pipeline girth weld joint partition material detection method provided by the embodiment of the present invention;

[0047] Figure 2 It is the true stress-strain curve obtained from the small-scale standard tensile test of the base metal, weld seam, and heat-affected zone materials of the X80 pipeline girth weld joint provided by the embodiment of the present invention;

[0048] Figure 3 It is a schematic structural diagram corresponding to the small punch test provided by the embodiment of the present invention;

[0049] Figure 4 It is the load-displacement curve obtained from the small punch test of the base metal, weld seam, and heat-affected zone materials of the X80 pipeline girth weld joint provided by the embodiment of the present invention;

[0050] Figure 5 It is a schematic diagram of the finite element model of the small punch test provided by the embodiment of the present invention;

[0051] Figure 6 It is a comparison chart of the load-displacement curves of the small punch test and the finite element model of the heat-affected zone material provided by the embodiment of the present invention;

[0052] Figure 7 It is a schematic diagram of the distribution of the yield strength and hardening index values of the pipeline steel material provided by the embodiment of the present invention;

[0053] Figure 8 It is a schematic diagram of the change of the loss curve with the number of training rounds during the training process of the deep learning model provided by the embodiment of the present invention;

[0054] Figure 9 It is a comparison chart of the true stress-strain curve of the pipeline steel material output by the deep learning model and the tensile test results provided by the embodiment of the present invention;

[0055] Figure 10 It is a schematic structural diagram of the pipeline girth weld joint partition material detection device provided by the embodiment of the present invention;

[0056] Figure 11 It is a schematic structural diagram of the electronic device provided by the embodiment of the present invention. Detailed Description of the Embodiments

[0057] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present invention.

[0058] To enable those skilled in the art of the present technology to better understand the solution of the present invention, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. Figure 1 It is a flowchart of the method for detecting the partition materials of the pipeline girth weld joint provided by the embodiment of the present invention. As Figure 1 shown, the method includes:

[0059] S101. Obtain the elastoplastic constitutive relationship and load-displacement data of the materials in each area of the pipeline girth weld joint.

[0060] It should be noted that the elastoplastic constitutive relationship is an expression that describes the relationship between stress and strain in the elastic and plastic deformation stages of materials. When the present invention executes step S101, the elastoplastic constitutive relationship and load-displacement data of the materials in each area of the pipeline girth weld joint can be obtained. Here, the elastoplastic constitutive relationship can be understood as elastoplastic mechanical property parameters, which mainly include the yield strength and strain hardening index of the material. Here, the load-displacement data can be understood as the load-displacement data obtained through various tests on the pipeline girth weld joint specimens, such as the Small Punch Test (SPT). The small punch test is a test method for evaluating the mechanical properties of materials.

[0061] S102. Establish a finite element model of the small punch test, input the obtained elastoplastic constitutive relationship into the finite element model of the small punch test, and compare the output result with the obtained load-displacement data to verify the finite element model of the small punch test.

[0062] In implementation, the small punch test finite element model is a numerical model established based on the finite element method for simulating the small punch test process and analyzing related mechanical behaviors. When executing step S102, the elastic-plastic constitutive relationship obtained in step S101 is input into the small punch test finite element model. In the small punch test finite element model, the behavior of the material is simulated and calculated through these constitutive relationships, so that the model can predict the mechanical response of the material under the small punch test conditions according to the input material properties. The load-displacement data calculated by the finite element model is compared with the load-displacement data obtained by the actual small punch test, and the accuracy and reliability of the small punch test finite element model are judged according to the comparison results. If the load-displacement data output by the model and the actual small punch test data are within the set error range, it means that the model verification is successful, the small punch test process can be simulated more accurately, the elastic-plastic constitutive relationship input and other settings of the model are reasonable, the model is reliable, and can be used for further analysis and prediction. If the load-displacement data output by the model and the actual small punch test data are not within the set error range, it means that the model verification has failed.

[0063] S103. Perform numerical simulation on the finite element model of the small punch test after successful verification, and establish a machine learning database that meets the parameter range of the constitutive relationship of pipeline steel materials.

[0064] In implementation, the parameter range of the constitutive relationship of pipeline steel materials refers to the allowed value range of the parameters used to describe the mechanical behavior of pipeline steel materials when subjected to stress. When executing step S103, a large amount of numerical simulation work is carried out with the help of a verified small punch test finite element model to establish a machine learning database that meets the range of mechanical property parameters of common pipeline steel materials. The machine learning database can mainly include: elastic-plastic constitutive relationship (mechanical property parameters) and load-displacement data, etc.

[0065] S104. Build a deep learning model, and use the machine learning database to train the deep learning model.

[0066] It should be noted that the present invention can construct a deep learning model with a complex structure to realize the function of determining the true stress-strain relationship (output) of the material based on the load-displacement data (input).

[0067] S105. Input the load-displacement data corresponding to the girth weld joint of the pipeline to be tested into the trained deep learning model, and output the constitutive relationship of the partition material of the girth weld joint of the pipeline to be tested.

[0068] In the above-mentioned method for detecting the partitioned material of the pipeline girth weld joint provided by the embodiment of the present invention, first, the elastoplastic constitutive relations and load-displacement data of the materials in each region of the pipeline girth weld joint are obtained, providing accurate basic data for subsequent steps. Then, a small punch test finite element model is established and compared with the actually obtained load-displacement data for verification, which can ensure the accuracy and reliability of the finite element model. The verified model can more realistically simulate the small punch test process, providing an effective tool for further numerical simulation and analysis, and reducing the wrong results caused by model errors. After that, numerical simulation work is carried out on the small punch test finite element model after successful verification, and a machine learning database that meets the parameter range of the pipeline steel material constitutive relation is established, providing rich and accurate data support for deep learning and helping the model learn more comprehensive material constitutive relation features. Finally, a deep learning model is constructed and trained using the machine learning database. By inputting the load-displacement data corresponding to the pipeline girth weld joint to be tested into the trained deep learning model, the constitutive relation of the partitioned material of the pipeline girth weld joint to be tested can be quickly output. In this way, the constitutive relations of the materials in each region of the pipeline girth weld joint can be accurately and efficiently determined according to the punch load-displacement data, overcoming the drawback of the traditional empirical formula method that only obtains the material constitutive relation with individual data points, solving the problem that the tensile constitutive relations of the materials in each region of the pipeline girth weld joint cannot be directly tested, providing another way for the efficient test and characterization of the constitutive relations of the materials in each region of the pipeline girth weld joint, providing a data basis for the safety assessment and life evaluation of the pipeline girth weld joint and the pipeline structure, and having high guiding significance for engineering practice.

[0069] Further, in specific implementation, in the above-mentioned method for detecting the partitioned material of the pipeline girth weld joint provided by the embodiment of the present invention, step S101 of obtaining the elastoplastic constitutive relations of the materials in each region of the pipeline girth weld joint may specifically include: obtaining the elastoplastic constitutive relations of the materials in each region of the pipeline girth weld joint obtained through a quasi-static tensile test; wherein, the quasi-static tensile test calculates the axial engineering strain of each pipeline girth weld joint specimen through the change in the axial length between virtual extensometers; combining the axial engineering strain with the axial load recorded by the in-situ tensile testing machine, calculating the engineering stress data and engineering strain data of each pipeline girth weld joint specimen; converting the engineering strain data and engineering stress data before the maximum load of each pipeline girth weld joint specimen into true strain and true stress data to obtain the true stress-strain curve of the materials in each region of the pipeline girth weld joint parent sample.

[0070] In implementation, the above-mentioned pipe girth weld joint specimen is obtained by successively polishing the pipe girth weld joint master sample, etching the weld seam, and calibrating the positions of each area; after the weld seam etching, the base metal area material, weld seam area material, and heat-affected zone material of the pipe girth weld joint master sample are revealed; that is, the materials of each area of the pipe girth weld joint mainly include the base metal area material, weld seam area material, and heat-affected zone material. Each pipe girth weld joint specimen corresponds to one of the base metal area material, weld seam area material, and heat-affected zone material.

[0071] Taking a certain semi-automatic welded X80 pipe girth weld joint as an example, the elastic-plastic mechanical properties of the materials in each area of the pipe girth weld joint and the load-displacement curve of the small punch test can be specifically obtained through small-size standard tensile tests and small punch tests; it mainly includes: grinding and polishing the girth weld joint; etching the weld seam; marking the sampling positions; cutting the specimens; processing the specimens; conducting in-situ tensile tests and small punch tests; obtaining and processing test data, etc. It can be specifically divided into the following steps:

[0072] Step 1: First, grind and polish the large-size master sample, then use a 4% nitric acid alcohol solution to etch the area around the weld seam to reveal the base metal area, weld seam area, and heat-affected zone, calibrate the sampling position of each specimen, so that each specimen only represents a special area (one of the base metal, weld seam, and heat-affected zone) of the welded joint. After calibrating the sampling position of each specimen, use the slow wire electrical discharge machining technology to process the specimen shape, and finally use different types of sandpaper to strictly control the surface roughness and specimen thickness of the specimen to ensure that the specimen thickness and roughness are both within the required range.

[0073] Step 2: Use an in-situ tensile and compression test system to conduct a quasi-static tensile test on the rectangular cross-section specimens of the materials in each area of the X80 girth weld joint. This test system mainly consists of three major modules: a stress loading system, a digital image correlation (Digital Image Correlation, DIC) system, and a control and data acquisition system. Calculate the axial engineering strain of the specimen through the change in the axial length between the virtual extensometers, and combine the axial load recorded by the in-situ tensile testing machine to calculate the engineering stress and engineering strain data of the specimen. Then convert the engineering strain and engineering stress data before the maximum load of the specimen into true strain and true stress data to obtain the true stress-strain curves of the three area materials of this girth weld joint. Figure 2 The true stress-strain curves obtained from the small-size standard tensile tests of the base metal, weld seam, and heat-affected zone materials of the X80 pipe girth weld joint provided by the embodiments of the present invention.

[0074] In the above Step 2, the engineering strain data and engineering stress data before the maximum load of the specimen can be specifically converted into true strain and true stress data by using the first formula and the second formula; the first formula is:

[0075] ; (1)

[0076] The second formula is:

[0077] ; (2)

[0078] Wherein, and are the true strain data and engineering strain data of the material respectively; and are the true stress data and engineering stress data of the material respectively.

[0079] Step 3: The constitutive relation parameters of the material part, such as elastic modulus, yield strength, tensile strength, uniform elongation and fracture elongation, can be directly obtained from the stress-strain curve of the material. The true stress-strain curves of the materials in each region of the pipe girth weld joint specimen can satisfy the power-law hardening rule; the power-law hardening rule is expressed by the third formula and the fourth formula; the third formula is:

[0080] ; (3)

[0081] The fourth formula is:

[0082] ; (4)

[0083] Wherein, is the equivalent plastic strain, is the yield strain, is the strain hardening index, is the yield strength, is the elastic modulus.

[0084] Furthermore, in specific implementation, in the above-mentioned pipe girth weld joint partition material detection method provided by the embodiment of the present invention, step S101 of obtaining the load-displacement data of the materials in each region of the pipe girth weld joint may specifically include: obtaining the load-displacement data of the materials in each region of the pipe girth weld joint obtained by the small punch test; wherein, the small punch test is to test the pipe girth weld joint specimen by using a small punch test machine, and a displacement sensor is used to monitor the change of the displacement of the punch in the loading direction during the test, and record the load-displacement data of the pipe girth weld joint specimen.

[0085] In implementation, the materials in each region of the X80 girth weld joint can be tested by using a small punch test machine. Figure 3 is the structural schematic diagram corresponding to the small punch test provided by the embodiment of the present invention. As Figure 3As shown in the figure, the small punch testing machine mainly includes a punch 1, a rigid ball 2, an upper restraint 3, a receiving hole 5 for placing a specimen 4, and a lower restraint 6. The specimen 4 can be a "thin sheet" with a diameter of 10 mm (or 8 mm) and a thickness of 0.5 mm. During the test, a load parallel to the thickness direction of the specimen 4 can be applied to the center of the specimen 4 through the rigid ball 2 until the specimen fails. During the test, a displacement sensor is used to monitor the change in the displacement of the punch 1 in the loading direction during the test, and the load-displacement data of each specimen 4 is recorded. Figure 4 The load-displacement curve obtained from the small punch test of the base metal, weld seam, and heat affected zone materials of the X80 pipeline girth weld joint provided by the embodiment of the present invention. From Figure 4 the load-displacement curve in it, the corresponding load-displacement data can be obtained.

[0086] Furthermore, in specific implementation, in the above-mentioned pipeline girth weld joint partition material detection method provided by the embodiment of the present invention, in step S102, a finite element model of the small punch test is established, the obtained elastoplastic constitutive relationship is input into the finite element model of the small punch test, and the output result is compared with the obtained load-displacement data to verify the finite element model of the small punch test. Specifically, it may include: establishing a finite element model of the small punch test using finite element software; the geometric dimensions of each component in the finite element model of the small punch test simulation are the same as those of each component in the small punch test; the tangential behavior of the contact friction between the pipeline girth weld joint specimen in the finite element model of the small punch test and each component is defined as a penalty function with a set friction coefficient, and the normal behavior is defined as hard contact. The deformable entity in the finite element model of the small punch test uses a four-node axisymmetric reduced integration element; the average value of the obtained elastoplastic constitutive relationship is input into the finite element model of the small punch test, and through numerical simulation, the load-displacement curve during the specimen deformation process is output; the simulated load-displacement curve is compared with the obtained load-displacement data to verify the rationality of the finite element model of the small punch test in terms of boundary conditions, mesh division, and model settings.

[0087] In implementation, the present invention can establish a finite element model of the small punch test with the help of finite element software, input the elastoplastic mechanical properties of materials in each region, compare the small punch test data and simulation data, and verify the accuracy of the finite element model; it mainly includes: establishing a finite element model; verifying the finite element model and possibly adjusting the finite element model, etc. Specifically, it can be divided into the following steps:

[0088] The first step is to establish a finite element model of the small punch test using the commercial finite element software ABAQUS. Figure 5 The schematic diagram of the finite element model of the small punch test provided by the embodiment of the present invention. As Figure 5As shown in the figure, in the small punch test finite element model, the geometric dimensions of the rigid ball, the upper and lower constraints, and the specimen are all consistent with the actual situation. Both the upper and lower constraints are set to be completely fixed (without degrees of freedom). A reference point is set at the center of the rigid ball, and a displacement along the negative Y-axis is applied to it. The loading method is also consistent with the actual test. The tangential behavior of all contact frictions between the specimen and the two constraints and between the specimen and the rigid ball is defined as a penalty function with a certain friction coefficient (such as 0.3); the normal behavior is defined as "hard" contact. The deformable entities in the finite element model adopt four-node axisymmetric reduced integration elements (CAX4R), and the mesh in the contact area between the rigid ball and the specimen is refined.

[0089] Step 2: Verify the reliability of the established small punch test finite element model by using the in-situ tensile test results and small punch test results of the materials in each region of the X80 pipeline girth weld joint obtained from the test. First, input the average value of the elastic-plastic constitutive relations of the materials in each region obtained from the in-situ tensile test into the established small punch test finite element model. Subsequently, through numerical simulation, obtain the load-displacement curve during the deformation process of the specimen and compare it with the load-displacement curve of the small punch test. Figure 6 It is a comparison diagram of the load-displacement curves of the small punch test and the finite element model for the heat affected zone material provided by the embodiment of the present invention.

[0090] Step 3: The established finite element models can all well reproduce the load-displacement curves during the small punch test. The change trends and coordinate values of the load-displacement curves at each deformation stage of the specimen are in good agreement with the test curves, indicating the rationality of the established SPT finite element model in terms of boundary conditions, mesh division, and model settings.

[0091] Furthermore, in specific implementation, in the above pipeline girth weld joint partition material detection method provided by the embodiment of the present invention, in step S103, numerical simulation work is carried out on the small punch test finite element model after successful verification to establish a machine learning database that meets the constitutive relation parameter range of pipeline steel materials, which can specifically include: generating multiple combinations of yield strength and hardening index by means of random value selection; according to the generated multiple combinations of yield strength and hardening index, obtaining multiple true stress-plastic strain curves of materials that conform to the power law hardening rule; using the obtained multiple true stress-plastic strain curves to carry out numerical simulation work on the small punch test finite element model after successful verification to obtain multiple load-displacement curves and their corresponding material constitutive parameters, so as to establish a machine learning database that meets the constitutive relation parameter range of pipeline steel materials.

[0092] In implementation, the present invention can carry out a large number of numerical simulation works with the help of a verified small punch test finite element model, and establish a machine learning database that meets the mechanical property parameter ranges of common pipeline steel materials; mainly including: determining the mechanical property parameter ranges of common pipeline steel materials; constructing a database of mechanical property parameters (inputs) and small punch load-displacement data (outputs). Specifically, it may include the following steps:

[0093] First, the yield strength range of common pipeline steel materials is 450 - 750 MPa, and the hardening index range is 0.075 - 0.15. To expand the applicable range of the final prediction model, the yield strength value range in the power-law hardening model of Equation (3) is set to 300 - 800 MPa, and the hardening index value range is set to 0.03 - 0.3. A plurality of groups (such as 396 groups) of combinations of yield strength and hardening index are generated by random selection. Figure 7 It is a schematic diagram of the value distribution of the yield strength and hardening index of the pipeline steel material provided by the embodiment of the present invention.

[0094] Then, the obtained multiple groups (such as 396 groups) of combinations of yield strength and hardening index are respectively substituted into Equation (3) to obtain multiple groups (such as 396 groups) of true stress - plastic strain curves of materials that conform to the power-law hardening rule. Finally, with the help of a verified small punch test finite element model, through a large number of simulation works, 396 groups of small punch test load-displacement curves (inputs) and their corresponding material constitutive parameters (yield strength and hardening index, outputs) are obtained to form a training database for the deep learning model.

[0095] Furthermore, in specific implementation, in the above-mentioned pipeline girth weld joint partition material detection method provided by the embodiment of the present invention, in step S104 of constructing a deep learning model and training the deep learning model using the machine learning database, it specifically may include: constructing a deep learning model by adopting a structure combining a convolutional neural network and a multi-layer fully connected layer; extracting the load-displacement data before the maximum load of each group of load-displacement curves from the machine learning database as the original features, calculating the differences between adjacent coordinate points of the load data and displacement data in the original features as high-order features; splicing the original features and high-order features to form an input feature matrix; performing normalization processing on the two target variables of yield strength and hardening index to construct the corresponding target value matrix; training the deep learning model using the input feature matrix and the target value matrix.

[0096] It should be noted that in the present invention, load-displacement data before reaching the maximum load can be selected from the load-displacement curves of each group of small punch bars as the original features. This part of the data contains the basic mechanical response information of the material at different stages during the small punch bar test. To more fully characterize and explore the data features, the differences between adjacent coordinate points of the above-mentioned load data and displacement data are further calculated. For example, for the load data, calculating the difference between adjacent load values can reflect information such as the rate of load change; for the displacement data, calculating the difference between adjacent displacement values can reflect the speed of displacement change and other situations. These differences, as high-order features, can provide richer and deeper information about the changes in the mechanical behavior of the material during the small punch bar test than the original features.

[0097] In implementation, step S104 of the present invention can be divided into three stages, namely the data preprocessing stage, the model construction and training stage, the model evaluation stage, and the model verification stage.

[0098] Data preprocessing stage: To fully characterize and explore the features of the input data (small punch bar load-displacement data), not only the load-displacement data before reaching the maximum load of each group of small punch bar load-displacement curves are extracted as the original features, but also the differences between their adjacent coordinate points are further calculated as high-order features, and the original features and high-order features are concatenated to form a more informative input feature matrix. At the same time, for the two target variables of yield strength and hardening index, normalization processing is performed to eliminate the data dimension difference and enhance the stability of model training, and then the corresponding target value matrix is constructed. The feature data of the divided training set is also normalized, and the corresponding normalizer model is saved for subsequent application to the preprocessing of new data to ensure the consistency of the data processing method.

[0099] Model construction and training stage: To fully capture the local and sequential features in the data (which conform to the characteristics of the load-displacement coordinate points) and construct a complex non-linear mapping relationship to predict the target variables, an architecture combining a convolutional neural network (CNN) and a multi-layer fully connected layer (Dense) can be used to construct a backpropagation (BP) neural network model. At the same time, during the model construction process, an L2 regularization mechanism is introduced to effectively control the model complexity, prevent overfitting, and enhance the generalization ability of the model.

[0100] Model evaluation stage: The same normalization process is performed on the divided test set. The generalization ability of the model is evaluated by calculating the loss value of the model on the test set (using Mean Squared Error as the loss function), and the test loss value is printed out. Through the debugging process of hyperparameters (such as the number of cross-validation folds, regularization coefficient, number of training epochs, batch size, activation function, etc.), the loss value of the test set is continuously reduced until the loss value of the test set is less than 0.01, and the final deep learning model structure and hyperparameter configurations are determined. After a series of attempts, a BP neural network structure that combines CNN layers and multiple fully connected layers is finally determined. Based on the above-established deep learning model structure, the cross-validation model loss curve is output after its operation. Figure 8 It is a schematic diagram showing the variation of the loss curve with the number of training epochs during the training process of the deep learning model provided by the embodiment of the present invention. Figure 8 It shows the variation of the training loss (CV Train Loss) and the validation loss (CV Validation Loss) with the number of training epochs (Epoch). There is no overfitting or underfitting phenomenon during the model training process, and the predetermined training end standard of the model is achieved.

[0101] Model validation stage: After the deep learning model is established, the test data of each region of the X80 pipeline girth weld joint material is used to verify the performance of the model, and the average value of the small punch test load-displacement data is input into the established deep learning model. Figure 9 It is a comparison chart of the true stress-strain curve of the pipeline steel material output by the deep learning model provided by the embodiment of the present invention and the tensile test results. Figure 9 It shows the absolute relative error between the predicted result output by the deep learning model and the test result. When the material enters the strain hardening stage, the output result of the model matches the test result very well, indicating that the deep learning model can basically achieve the expected function.

[0102] In the above embodiment, the detection method for the materials of each region of the pipeline girth weld joint is described in detail. The present invention also provides corresponding embodiments of the device for determining the constitutive relationship of the materials of each region of the pipeline girth weld joint and the electronic device. It should be noted that the present invention describes the embodiments of the device part from two perspectives, one is from the perspective of functional modules, and the other is from the perspective of hardware.

[0103] Figure 10 It is a schematic structural diagram of the device for determining the constitutive relationship of the materials of each region of the pipeline girth weld joint provided by the embodiment of the present invention. In this embodiment, from the perspective of functional modules, as Figure 10 shown, the device includes:

[0104] The data acquisition module 10 is used to acquire the elastoplastic constitutive relations and load-displacement data of the materials in each region of the pipeline girth weld joint;

[0105] The finite element model establishment and verification module 11 is used to establish a small punch test finite element model, input the acquired elastoplastic constitutive relations into the small punch test finite element model, and compare the output results with the acquired load-displacement data to verify the small punch test finite element model;

[0106] The database establishment module 12 is used to perform numerical simulation on the successfully verified small punch test finite element model and establish a machine learning database that meets the constitutive relation parameter range of pipeline steel materials;

[0107] The deep learning module construction and training module 13 is used to construct a deep learning model and train the deep learning model using the machine learning database;

[0108] The deep learning module inference module 14 is used to input the load-displacement data corresponding to the pipeline girth weld joint to be tested into the trained deep learning model and output the constitutive relations of the partition materials of the pipeline girth weld joint to be tested.

[0109] In the above-mentioned device for determining the constitutive relations of the materials in each region of the pipeline girth weld joint provided by the embodiment of the present invention, through the interaction of the above four modules, the constitutive relations of the materials in each region of the pipeline girth weld joint can be accurately and efficiently determined according to the punch load-displacement data, overcoming the drawback that the traditional empirical formula method only obtains the material constitutive relations based on individual data points, solving the problem that the tensile constitutive relations of the materials in each region of the pipeline girth weld joint cannot be directly tested, providing another way for the efficient testing and characterization of the constitutive relations of the materials in each region of the pipeline girth weld joint, and at the same time providing a data basis for carrying out the safety assessment and life assessment of the pipeline girth weld joint and pipeline structure, which has high guiding significance for engineering practice.

[0110] Since the embodiments of the device part correspond to the embodiments of the method part, please refer to the description of the embodiments of the method part for the embodiments of the device part, which will not be elaborated here. And it has the same beneficial effects as the above-mentioned method for detecting the partition materials of the pipeline girth weld joint.

[0111] Figure 11 It is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. Based on the hardware perspective, as Figure 11 shown, the electronic device includes:

[0112] The memory 20 is used to store computer programs;

[0113] The processor 21 is used to implement the steps of the method for detecting the partition materials of the pipeline girth weld joint as mentioned in the above embodiment when executing the computer program.

[0114] Among them, the processor 21 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 21 may be implemented in at least one hardware form of a Digital Signal Processor (DSP), a Field-Programmable Gate Array (FPGA), and a Programmable Logic Array (PLA). The processor 21 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the wake state, also known as the CPU; the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 21 may be integrated with a Graphics Processing Unit (GPU), and the GPU is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 21 may further include an Artificial Intelligence (AI) processor, and the AI processor is used to process computational operations related to machine learning.

[0115] The memory 20 may include one or more computer-readable storage media, and the computer-readable storage media may be non-transitory. The memory 20 may further include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash storage devices. In this embodiment, the memory 20 is at least used to store the following computer program 201. After the computer program is loaded and executed by the processor 21, it can implement the relevant steps of the pipeline girth weld joint partition material detection method disclosed in any of the foregoing embodiments. In addition, the resources stored in the memory 20 may further include an operating system 202 and data 203, etc., and the storage method may be transient storage or permanent storage. Among them, the operating system 202 may include Windows, Unix, Linux, etc. The data 203 may include, but is not limited to, the data involved in the pipeline girth weld joint partition material detection method mentioned above.

[0116] In some embodiments, the electronic device may further include a display screen 22, an input / output interface 23, a communication interface 24, a power supply 25, and a communication bus 26. Those skilled in the art can understand that Figure 11 the structure shown in does not constitute a limitation on the electronic device, and it may include more or fewer components than those shown in the figure. The electronic device provided by the embodiments of the present invention includes a memory and a processor. When the processor executes the program stored in the memory, it can implement the pipeline girth weld joint partition material detection method mentioned above, and the effect is the same.

[0117] Finally, the present invention also provides an embodiment corresponding to a computer-readable storage medium. A computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps recorded in the above method embodiments are implemented.

[0118] It can be understood that if the method in the above embodiments is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and executes all or part of the steps of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes. The computer-readable storage medium provided by the present invention can implement the above-mentioned pipeline girth weld joint partition material detection method, and the effect is the same.

[0119] Finally, the present invention also provides an embodiment corresponding to a computer program product. The computer program product includes computer programs / instructions, and when the computer programs / instructions are executed by a processor, the steps recorded in the above pipeline girth weld joint partition material detection method embodiments are implemented. The computer program product provided by the present invention can implement the above-mentioned pipeline girth weld joint partition material detection method, and the effect is the same.

[0120] It should also be noted that in this specification, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the existence of additional identical elements in the process, method, article or device including the element.

[0121] The above has introduced in detail the pipeline girth weld joint partition material detection method, device, equipment and medium provided by the present invention. The various embodiments in the specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the various embodiments, reference can be made to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and reference can be made to the description in the method part for the relevant parts. It should be noted that for those of ordinary skill in the art in this technical field, without departing from the principle of the present invention, several improvements and modifications can still be made to the present invention, and these improvements and modifications also fall within the protection scope of the present invention.

Claims

1. A method for detecting partition materials of pipeline girth weld joints, characterized in that: The method comprises: Obtain the elastic-plastic constitutive relationship and load-displacement data of the materials in each region of the pipeline girth weld joint; Establishing a small punch test finite element model, inputting the obtained elastic-plastic constitutive relationship into the small punch test finite element model, and comparing the output result with the obtained load-displacement data to verify the small punch test finite element model; Performing numerical simulation on the successfully verified small punch test finite element model to establish a machine learning database that meets the parameter range of the constitutive relationship of pipeline steel materials; Constructing a deep learning model, and training the deep learning model using the machine learning database; The load-displacement data corresponding to the pipeline girth weld joint to be tested is input into the trained deep learning model, and the constitutive relationship of the partition material of the pipeline girth weld joint to be tested is output.

2. The method for detecting partition materials of pipeline girth weld joints according to claim 1 is characterized in that: Obtain the elastic-plastic constitutive relationship of the materials in each region of the pipeline girth weld joint, including: Obtain the elastic-plastic constitutive relationship of the materials in each region of the pipeline girth weld joint obtained through quasi-static tensile test; Among them, the quasi-static tensile test is to calculate the axial engineering strain of each pipe girth weld joint sample by the change in axial length between virtual extensometers; the axial engineering strain is combined with the axial load recorded by the in-situ tensile testing machine to calculate the engineering stress data and engineering strain data of each pipe girth weld joint sample; the engineering strain data and engineering stress data of each pipe girth weld joint sample before the maximum load are converted into true strain and true stress data to obtain the true stress-strain curve of the material in each region of the pipe girth weld joint master sample.

3. The method for detecting partition materials of pipeline girth weld joints according to claim 2 is characterized in that: The pipe girth weld joint sample is obtained by sequentially grinding and polishing the pipe girth weld joint master sample, performing weld erosion treatment, and calibrating the positions of various regions; after the weld erosion treatment, the parent material region material, weld region material, and heat-affected region material of the pipe girth weld joint master sample are revealed; Each of the pipe girth weld joint specimens corresponds to one of the parent material area material, the weld area material and the heat affected area material.

4. The pipeline girth weld joint partition material detection method according to claim 3 is characterized in that: Obtain load-displacement data for materials in various regions of a pipe girth weld, including: Obtain the load-displacement data of the materials in each area of ​​the pipe girth weld joint obtained by the small punch test; Among them, the small punch test is to test the pipe girth weld joint sample using a small punch testing machine, use a displacement sensor to monitor the change of the punch displacement in the loading direction during the test, and record the load-displacement data of the pipe girth weld joint sample.

5. The method for detecting partition materials of pipeline girth weld joints according to claim 1 is characterized in that: Establishing a small punch test finite element model, inputting the obtained elastic-plastic constitutive relationship into the small punch test finite element model, and comparing the output result with the obtained load-displacement data to verify the small punch test finite element model, including: A finite element model of a small punch test is established using finite element software; the geometric dimensions of the various components of the small punch simulation test in the small punch test finite element model are consistent with those of the various components of the small punch test; the tangential behavior of the contact friction between the pipe girth weld joint specimen in the small punch test finite element model and the various components of the small punch simulation test is defined as a penalty function with a set friction coefficient, and the normal behavior is defined as hard contact, and the deformable entity in the small punch test finite element model adopts a four-node axisymmetric simplified integral unit; The obtained average value of the elastic-plastic constitutive relationship is input into the finite element model of the small punch test, and the load-displacement curve during the deformation process of the sample is output through numerical simulation; The load-displacement curve output by the simulation is compared with the acquired load-displacement data to verify the rationality of the small punch test finite element model in terms of boundary conditions, mesh division and model setting.

6. The method for detecting partition materials of pipeline girth weld joints according to claim 1 is characterized in that: The small punch test finite element model after successful verification is numerically simulated to establish a machine learning database that meets the parameter range of the constitutive relationship of pipeline steel materials, including: Multiple groups of yield strength and hardening index combinations are generated by random value selection; According to the generated multiple sets of yield strength and hardening index combinations, multiple sets of true stress-plastic strain curves of materials conforming to the power law hardening rule are obtained; Using the multiple sets of true stress-plastic strain curves obtained, the small punch test finite element model after successful verification was numerically simulated to obtain multiple sets of load-displacement curves and their corresponding material constitutive parameters, so as to establish a machine learning database that meets the parameter range of the constitutive relationship of pipeline steel materials.

7. The method for detecting partition materials of pipeline girth weld joints according to claim 1 is characterized in that: Constructing a deep learning model and training the deep learning model using the machine learning database includes: A deep learning model is constructed by combining a convolutional neural network with a multi-layer fully connected layer. Extracting the load-displacement data before each set of load-displacement curves reaches the maximum load from the machine learning database as original features, and calculating the difference between adjacent coordinate points of the load data and the displacement data in the original features as high-order features; Concatenate the original features with the high-order features to form an input feature matrix; The two target variables, yield strength and hardening index, are normalized and the corresponding target value matrix is ​​constructed; The deep learning model is trained using the input feature matrix and the target value matrix.

8. A device for determining the constitutive relationship of materials in various regions of a pipeline girth weld joint, characterized in that: The device comprises: A data acquisition module is used to obtain the elastic-plastic constitutive relationship and load-displacement data of the materials in each region of the pipeline girth weld joint; A finite element model establishment and verification module, used to establish a small punch test finite element model, input the obtained elastic-plastic constitutive relationship into the small punch test finite element model, and compare the output result with the obtained load-displacement data to verify the small punch test finite element model; A database establishment module is used to perform numerical simulation on the small punch test finite element model after successful verification, and establish a machine learning database that meets the parameter range of the constitutive relationship of pipeline steel materials; A deep learning module construction and training module, used to construct a deep learning model and train the deep learning model using the machine learning database; The deep learning module reasoning module is used to input the load-displacement data corresponding to the pipeline girth weld joint to be tested into the trained deep learning model, and output the constitutive relationship of the partition material of the pipeline girth weld joint to be tested.

9. An electronic device, characterized in that: The device comprises: Memory for storing computer programs; A processor is used to implement the steps of the pipeline girth weld joint partition material detection method as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the pipeline girth weld joint partition material detection method according to any one of claims 1 to 7 are implemented.

Images