Submerged-arc welding weld metal oxygen content prediction method and device, terminal and storage medium

The prediction of metal oxygen content of submerged arc welds through BP neural network solves the problem of blindness and time-consuming prediction in the existing technology, and achieves more accurate and efficient welding process management, supporting flux automation design and real-time monitoring.

CN118969137APending Publication Date: 2024-11-15NORTHEASTERN UNIV CHINA
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202411226286.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-03
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

In the existing submerged arc welding technology, the prediction of metal oxygen content of welds is blind and time-consuming, and the influencing factors considered are incomplete, resulting in low flux design and optimization efficiency.

Method used

The BP neural network training model is used to calculate the welding heat input through flux composition and welding parameters, predict the metal oxygen content of the weld, and optimize the model with the determination coefficient to ensure prediction accuracy.

Benefits of technology

It realizes accurate prediction of the metal oxygen content of welds, improves the efficiency and adaptability of the welding process, supports dynamic flux automation design and real-time monitoring, and meets the changing needs of the welding process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118969137A_ABST
    Figure CN118969137A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of industrial welding, and discloses a submerged-arc welding weld metal oxygen content prediction method and device, a terminal and a storage medium, and the method comprises the steps that set welding flux components and welding parameters are obtained; welding heat obtained through calculation of the welding flux components and the welding parameters is input and input into a trained submerged-arc welding seam metal oxygen content prediction model, and a predicted value of the welding seam metal oxygen content obtained after welding is conducted through the welding flux components and the welding parameters is obtained; the submerged-arc welding weld metal oxygen content prediction model is a prediction model obtained by inputting training data into a BP neural network for training. The submerged-arc welding seam metal oxygen content is predicted based on the BP neural network, and parameters are learned and adjusted through the BP neural network, so that the submerged-arc welding seam metal oxygen content prediction model has better adaptability and generalization ability when facing different types of data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of industrial welding technology, and in particular to a method, device, terminal and storage medium for predicting the metal oxygen content of a submerged arc welding weld. Background Art

[0002] In submerged arc welding, oxygen is one of the key elements that affect the microstructure and mechanical properties of submerged arc welding weld metal. At present, researchers generally believe that in submerged arc welding, the oxygen content in the weld metal should be controlled between 200 and 500 ppm. Too high or too low oxygen content will reduce the low-temperature impact toughness of the weld metal. The oxides in the flux will decompose and release oxygen under the action of arc plasma, resulting in an increase in the oxygen content in the droplet and the molten pool. Welding heat input is also an important parameter affecting the oxygen content of submerged arc welding weld metal. Changes in heat input will affect the solute diffusion rate of the metal in the weld and the kinetic behavior of the molten pool, thereby changing the oxygen content of the submerged arc welding weld metal. At present, the research and development of flux is mainly based on a large amount of experimental data to establish the empirical relationship between flux composition and weld mechanical properties, and the flux is designed and optimized accordingly. However, this method has the disadvantages of blindness and time-consuming, and the influencing factors considered are not comprehensive. Summary of the invention

[0003] In view of this, an embodiment of the present application provides a method for predicting the metal oxygen content of a submerged arc welding weld, which can effectively solve the problems of blind prediction and time-consuming prediction.

[0004] In a first aspect, an embodiment of the present application provides a method for predicting the metal oxygen content of a submerged arc welding weld, comprising:

[0005] Get the set flux composition and welding parameters;

[0006] Inputting the welding heat input calculated by the flux composition and the welding parameters into a trained prediction model for oxygen content in weld metal of submerged arc welding to obtain a predicted value of oxygen content in weld metal after welding using the flux composition and the welding parameters;

[0007] The submerged arc welding weld metal oxygen content prediction model is a prediction model obtained by inputting training data into a BP neural network for training.

[0008] In some embodiments, the prediction model obtained by inputting training data into the BP neural network training includes:

[0009] Get training data;

[0010] Through the training data and the BP neural network, a prediction model for the metal oxygen content of the submerged arc welding weld is obtained by training;

[0011] The submerged arc welding weld metal oxygen content prediction model is evaluated by the determination coefficient. If the determination coefficient is not within the set qualified range, the submerged arc welding weld metal oxygen content prediction model is optimized until the determination coefficient obtained after optimization is within the qualified range, thereby obtaining the submerged arc welding weld metal oxygen content prediction model.

[0012] In some embodiments, the training data includes flux composition, welding heat input, and weld metal oxygen content measurements;

[0013] The obtaining of training data comprises:

[0014] Determine the flux composition and welding parameters during welding;

[0015] Obtaining a welding sample after welding using the flux and the welding parameters, measuring the oxygen content of the weld, then performing wire cutting on the weld of the welding sample to obtain weld metal, sampling the center of the weld metal to measure the oxygen content, obtaining the oxygen content measurement value, and calculating the welding heat input according to the welding parameters;

[0016] The flux composition is adjusted, and the step of obtaining welding samples after welding using the flux composition and the welding parameters is repeatedly performed to obtain multiple sets of training data consisting of flux composition, welding heat input and weld metal oxygen content measurement values.

[0017] In some embodiments, the BP neural network includes an input layer, a hidden layer and an output layer; the submerged arc welding weld metal oxygen content prediction model is obtained by training the training data and the BP neural network, including:

[0018] Using the welding heat input calculated from the flux composition and the welding parameters as input data of the input layer, and the measured value of the weld metal oxygen content as input data of the output layer;

[0019] The input data is transferred to the output layer through the hidden layer by forward propagation and the gradient is calculated by back propagation to optimize the parameters of the BP neural network and obtain a prediction model for the metal oxygen content of the submerged arc welding weld.

[0020] In some embodiments, the step of transferring the input data through the hidden layer to the output layer by forward propagation and calculating the gradient by back propagation to optimize the parameters of the BP neural network includes:

[0021] During each training, the output layer error and hidden layer error are calculated;

[0022] According to the output layer error and the hidden layer error, a first weight from the input layer to the hidden layer is corrected, and a second weight from the hidden layer to the output layer is corrected;

[0023] When the number of training times of the parameters of the BP neural network reaches a preset value or the output layer error is lower than a preset minimum error requirement, the optimization operation of the first weight and the second weight is completed.

[0024] In some embodiments, the calculation expression of the determination coefficient is:

[0025]

[0026] In the formula, R 2 is the determination coefficient, SS res is the residual sum of squares, SS tot is the total sum of squares;

[0027] The submerged arc welding weld metal oxygen content prediction model is evaluated by the determination coefficient, and if the determination coefficient is not within a set qualified range, the submerged arc welding weld metal oxygen content prediction model is optimized, including:

[0028] Determine whether the determination coefficient is within the qualified range. If it is not within the qualified range, the fitting effect of the submerged arc welding weld metal oxygen content prediction model is unqualified;

[0029] If the determination coefficient is not within the qualified range, the training data is re-input into the submerged arc welding weld metal oxygen content prediction model for training to optimize the submerged arc welding weld metal oxygen content prediction model.

[0030] In a second aspect, the present application provides a device for predicting the metal oxygen content of a submerged arc welding weld, comprising:

[0031] A data acquisition module, used to obtain set flux composition and welding parameters;

[0032] A prediction module, for inputting the welding heat input calculated by the flux composition and the welding parameters into a trained prediction model for the oxygen content of weld metal in submerged arc welding to obtain a predicted value of the oxygen content of weld metal after welding using the flux composition and the welding parameters;

[0033] The submerged arc welding weld metal oxygen content prediction model is a prediction model obtained by inputting training data into a BP neural network for training.

[0034] In a third aspect, the present application provides a terminal device, which includes a processor and a memory, wherein the memory stores a computer program, and the processor is used to execute the computer program to implement the method for predicting the metal oxygen content of a submerged arc welding weld.

[0035] In a fourth aspect, the present application provides a computer-readable storage medium storing a computer program, which, when executed on a processor, implements the method for predicting the metal oxygen content of a submerged arc welding weld.

[0036] The embodiments of the present application have the following beneficial effects:

[0037] This application comprehensively considers multiple influencing factors such as flux composition and welding heat input obtained under submerged arc welding conditions, and the complexity of the welding process is more fully reflected, making the prediction more comprehensive and accurate. By learning and adjusting parameters through the BP neural network, the prediction model of metal oxygen content in submerged arc welding welds has better adaptability and generalization ability when facing different types of data. Using the BP neural network for auxiliary design, the automatic design of dynamic flux can be realized. The prediction model of metal oxygen content in submerged arc welding welds using the BP neural network is used to monitor the metal oxygen content in submerged arc welding welds in real time. And adjust the flux in time according to the prediction results to meet the changes and needs in the welding process. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0039] Figure 1 A flow chart of a method for predicting metal oxygen content in a submerged arc welding seam according to an embodiment of the present application is shown;

[0040] Figure 2 A relationship diagram between a predicted value and a true value of metal oxygen content in a submerged arc welding seam according to an embodiment of the present application is shown;

[0041] Figure 3 A flow chart of a prediction model training for metal oxygen content in a submerged arc welding seam according to an embodiment of the present application is shown;

[0042] Figure 4 A schematic diagram of the structure of a neural network model according to an embodiment of the present application is shown;

[0043] Figure 5 A schematic diagram of the structure of a submerged arc welding weld metal oxygen content prediction device implemented in the present application is shown. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments.

[0045] The components of the embodiments of the present application generally described and shown in the drawings herein may be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0046] Hereinafter, the terms "including", "having" and their cognates that can be used in various embodiments of the present application are intended only to indicate specific features, numbers, steps, operations, elements, components or a combination of the foregoing items, and should not be understood as first excluding the existence of one or more other features, numbers, steps, operations, elements, components or a combination of the foregoing items or increasing the possibility of one or more features, numbers, steps, operations, elements, components or a combination of the foregoing items. In addition, the terms "first", "second", "third" and the like are only used to distinguish descriptions and cannot be understood as indicating or implying relative importance.

[0047] Unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meanings as those generally understood by those skilled in the art to which the various embodiments of the present application belong. The terms (such as those defined in generally used dictionaries) will be interpreted as having the same meanings as the contextual meanings in the relevant technical field and will not be interpreted as having idealized meanings or overly formal meanings unless clearly defined in the various embodiments of the present application.

[0048] The technical solution of the present application is applied to the submerged arc welding scene to predict the value of the oxygen content of the weld metal after welding with a certain flux. The present application obtains the set flux composition and welding parameters; the welding heat input calculated by the flux composition and the welding parameters is input into the trained submerged arc welding weld metal oxygen content prediction model to obtain the predicted value of the weld metal oxygen content after welding with the flux composition and the welding parameters. In this way, the oxygen content of the weld metal after welding can be directly predicted by inputting the flux composition and the welding heat input, which saves the steps of actual welding operation and measurement operation and increases the experimental efficiency.

[0049] Example 1

[0050] The following is an explanation of the method for predicting the metal oxygen content in the submerged arc welding seam in conjunction with some specific embodiments.

[0051] Figure 1 A flow chart of a method for predicting the oxygen content of a submerged arc welding weld metal according to an embodiment of the present application is shown. Exemplarily, the method for predicting the oxygen content of a submerged arc welding weld metal comprises the following steps:

[0052] Step S100, obtaining set flux composition and welding parameters.

[0053] The present application is used to predict the metal oxygen content of a submerged arc welding weld, wherein the present embodiment uses both the flux composition and the welding heat input calculated by the welding parameters as input values ​​for prediction.

[0054] Among them, welding parameters include voltage, current and welding speed. These parameters can be set at the beginning of welding. Generally speaking, after setting, the entire welding process will not change. Therefore, the welding heat input during the welding process can be calculated based on the welding parameters.

[0055] The calculation expression of welding heat input is:

[0056]

[0057] Where I is the welding current, in A; U is the arc voltage, in V; υ is the welding speed, in cm / s; Q is the welding heat input, in kJ / cm.

[0058] The flux composition refers to the flux components and the ratio of each component, generally including two or more compounds. For example, the mass percentage range of the flux components is as follows: CaF2 (0-100%), SiO2 (0-60%), MnO (0-60%), CaO (0-80%), TiO2 (0-60%), Al2O3 (0-60%), Na2O (0-5%), MgO (0-40%), FeO (0-60%).

[0059] More specifically, a flux component may be SiO2, MnO, CaO, and the ratio thereof may be 4:1:5 or 2:1:2.

[0060] In this embodiment, the oxygen content of the weld metal is predicted, so the above-mentioned flux composition and welding parameters can actually be set according to the requirements. It can also be the actual existing flux composition and welding parameters, which are set when predicting before actual welding.

[0061] Among them, this embodiment takes into account multiple influencing factors such as flux composition and welding heat input, among which the flux composition affects the melting behavior and interaction between metals during welding, while the welding heat input affects the heat input and molten pool formation during welding, ensuring that the input data includes most of the influencing factors and is more comprehensive.

[0062] Step S200, inputting the welding heat input calculated by the flux composition and the welding parameters into a trained submerged arc welding weld metal oxygen content prediction model to obtain a predicted value of the weld metal oxygen content after welding using the flux composition and the welding parameters.

[0063] The embodiment uses a trained prediction model for oxygen content of submerged arc welding weld metal for prediction. The model is a prediction model obtained by inputting training data into a BP neural network for training. In this embodiment, the model is obtained through BP neural network training, and the oxygen content of the weld metal is predicted by the input flux composition and welding heat input.

[0064] Among them, this embodiment also provides the following SiO2-MnO-CaO flux to verify the accuracy of the BP neural network model:

[0065] Composition of SiO2-MnO-CaO flux (mass%)

[0066]

[0067] The oxygen content values ​​of the above-mentioned flux components under certain welding heat input are shown in the following table:

[0068] Welding heat input and oxygen content of SiO2-MnO-CaO flux

[0069]

[0070]

[0071] The above oxygen content can be measured by using a LECO element analyzer (ONH836, LECO Co.) after actual welding experiments to obtain the actual oxygen content.

[0072] Then, the flux composition and welding heat input are fed into the trained model as input data to obtain the corresponding predicted oxygen content.

[0073] like Figure 2 As shown, it is the relationship between the true value and the predicted value of the above example data, wherein the data points in the figure are the coordinates formed by the true value and the predicted value belonging to the same flux. It can be seen that these points are very consistent with the straight line y=x.

[0074] It means that the predicted value and the true value are basically the same, and the prediction result of the model in this embodiment is close to the actual data. It can be seen that the prediction method of the submerged arc welding weld metal oxygen content in this embodiment can accurately predict the submerged arc welding weld metal oxygen content in the range of 40mass%SiO2, 10-60mass%MnO, and 0-50mass%CaO of SiO2-MnO-CaO series weld metal.

[0075] In addition, this embodiment also provides a training method for the above-mentioned submerged arc welding weld metal oxygen content prediction model, such as Figure 2 As shown, the method includes:

[0076] Step S300, obtaining training data.

[0077] The training data of this embodiment is the data of flux composition, welding heat input and weld metal oxygen content obtained under submerged arc welding conditions.

[0078] Select CaF2, SiO2, CaO chemically pure reagents, and prepare the test samples by changing the CaF2 / CaO ratio and welding heat input, respectively, and record them as WM7-WM11. The specific composition ratios are as follows:

[0079] Composition of CaF2-SiO2-CaO flux (mass%)

[0080]

[0081] In addition to the above six types of flux, more types of flux can be prepared. The data format is similar to the above table and is not listed here. Set appropriate welding parameters and perform specific welding operations through experimental operations.

[0082] As mentioned above, the welding heat input can be calculated through the welding parameters.

[0083] In this way, the flux composition and welding heat input are obtained.

[0084] As for the oxygen content, after welding using the flux and the welding parameters, a corresponding welding sample can be obtained. Subsequently, a wire cutting is performed on the weld of the welding sample to obtain the weld metal, a sample is taken at the center of the weld metal to measure the oxygen content, and the welding heat input is calculated based on the welding parameters. In this embodiment, a LECO element analyzer can be used to measure the total oxygen content in the LECO molecular weight table, so that the oxygen content of the welding sample can be measured.

[0085] The specific results are shown in the following table:

[0086] Welding heat input and oxygen content of CaF2-SiO2-CaO flux

[0087]

[0088]

[0089] It is understandable that a corresponding oxygen content value can be measured for each flux component and a welding heat input, so by repeating the above welding and measuring operations, training data consisting of multiple sets of data can be obtained. For example, the 6 sets of data obtained in this embodiment are training data.

[0090] It is understandable that in order to train the model, the training data required is huge, so thousands of data can be prepared as training data for training.

[0091] Step S400: obtaining a prediction model for the metal oxygen content of a submerged arc welding weld by training the training data and the BP neural network.

[0092] In this embodiment, the BP neural network is used for training. Figure 3 As shown in FIG. , it is a structural diagram of the BP network model, which includes an input layer, a hidden layer and an output layer.

[0093] The flux composition and welding heat input are used as input data of the input layer, and the oxygen content measurement value is used as input data of the output layer.

[0094] During training, the input data is passed to the output layer through forward propagation and the gradient is calculated using back propagation, thereby effectively optimizing the network parameters and improving the model's predictive performance and generalization ability.

[0095] Specifically, suppose the data input to the input layer is X = [x1, x2, x3…x i ], hidden layer S = [s1, s2, s3…s j ], output layer O = [o1].

[0096] The mapping relationship between the input layer and the output layer can be calculated according to the following calculation expression.

[0097]

[0098]

[0099] Where S i is the value of the hidden layer, O j is the output of the output layer, ω i,j is the weight from the input layer to the hidden layer; ψ j,1 is the connection weight from the hidden layer to the output layer; h is the activation function from the input layer to the hidden layer; g is the activation function from the hidden layer to the output layer.

[0100] Set the activation function of the neural network to the Sigmoid function. The specific expression is as follows:

[0101]

[0102] Where exp is the exponential function with the natural constant e as the base, and x is the input data of the input layer.

[0103] The back propagation of BP neural network first needs to calculate the output layer error and hidden layer error. The output layer error calculation equation is as follows:

[0104] δ = E-O1;

[0105] Among them, O1 is the real output of the network, E is the expected output of the network, and δ is the error between the real output and the expected output. If δ is greater than the given error requirement, δ will be back-propagated along the BP neural network.

[0106] The hidden layer error is calculated using the following formula:

[0107] δ sj =ψ j,1 *δ;

[0108] In the formula, δ sj is the hidden layer error, which is used to adjust the weights ω between layers. i,j , ψ j,1 After correction, the corrected weight ω′ i,j ,ψ′ j,1 Will be used as the new weight in the next training process, the new weight ω′ i,j ,ψ′ j,1 It will continue to be iteratively optimized until the entire BP neural network training exceeds the maximum cycle limit or the error δ is lower than the minimum error requirement.

[0109] Among them, the correction calculation formula of the above two weights is:

[0110]

[0111]

[0112] After the training is completed, a prediction model for the oxygen content of weld metal in submerged arc welding can be obtained. At this point, the model can theoretically predict the oxygen content of weld metal in submerged arc welding.

[0113] Step S500, evaluating the submerged arc welding weld metal oxygen content prediction model through the determination coefficient. If the determination coefficient is not within the set qualified range, optimizing the submerged arc welding weld metal oxygen content prediction model until the determination coefficient is within the qualified range, thereby obtaining the submerged arc welding weld metal oxygen content prediction model.

[0114] After training, the model can also be evaluated to determine whether the model's degree of fit meets the requirements. To this end, this embodiment uses the determination coefficient R 2 To evaluate the trained model, the formula for calculating the coefficient of determination is:

[0115]

[0116] Where SSres is the residual sum of squares (the sum of squares of the differences between the predicted values ​​and the true values), and SStot is the total sum of squares (the sum of squares of the differences between the true values ​​and their means).

[0117] Specifically, when obtaining training data, this part of the data can be divided into a training set and a test set. When training the model, use the data in the training set for training, and when evaluating, use the data in the test set for testing.

[0118] During the test, the flux composition and welding heat input parameters in the test set are input into the prediction model of the metal oxygen content of the submerged arc welding weld to obtain the predicted value of the metal oxygen content of the submerged arc welding weld output by the model. Correspondingly, the measured value of the metal oxygen content of the weld in the test set is the true value. In this way, the above-mentioned residual sum of squares and total sum of squares can be calculated, and then the above-mentioned determination coefficient can be calculated.

[0119] R 2 It is used to measure the degree to which the model explains the variability of the data. Its value range is between 0 and 1. The closer it is to 1, the better the model fits the data. Therefore, a qualified range can be set, such as setting the interval from 0.8 to 1 as the qualified range. If higher accuracy is sought, 0.9 to 1 can be set as the qualified range, or 0.95 to 1 can be set as the qualified range. If the determination coefficient falls within this range, it means that the model is qualified, otherwise it is necessary to continue iterative optimization.

[0120] The prediction method of the metal oxygen content of the submerged arc welding weld of this embodiment comprehensively considers the multiple influencing factors of the flux composition and welding heat input obtained under the submerged arc welding conditions, and the complexity of the welding process is more fully reflected, so that the prediction is more comprehensive and accurate. By learning and adjusting parameters through the BP neural network, the prediction model of the metal oxygen content of the submerged arc welding weld has better adaptability and generalization ability when facing different types of data. In addition, the BP neural network can capture the nonlinear relationship that may exist between the flux composition, welding heat input and the metal oxygen content of the weld under the submerged arc welding conditions, so that this relationship can be better simulated and the prediction accuracy can be improved.

[0121] Moreover, the use of BP neural network for auxiliary design can realize the automatic design of dynamic flux. The prediction model of metal oxygen content in submerged arc welding seam using BP neural network can monitor the metal oxygen content in submerged arc welding seam in real time. And adjust the flux in time according to the prediction results to meet the changes and needs in the welding process. It reduces the time used for the experiment and improves the experimental efficiency.

[0122] Example 2

[0123] Figure 4 A schematic diagram of the structure of a submerged arc welding weld metal oxygen content prediction device according to an embodiment of the present application is shown. Exemplarily, the submerged arc welding weld metal oxygen content prediction device 100 comprises:

[0124] The data acquisition module 110 is used to acquire the set flux composition and welding parameters;

[0125] The prediction module 120 is used to input the welding heat input calculated by the flux composition and the welding parameters into a trained submerged arc welding weld metal oxygen content prediction model to obtain a predicted value of the weld metal oxygen content after welding using the flux composition and the welding parameters; the submerged arc welding weld metal oxygen content prediction model is a prediction model obtained by inputting training data into a BP neural network for training.

[0126] The present application provides a terminal device, which includes a processor and a memory, wherein the memory stores a computer program, and the processor is used to execute the computer program to implement the method for predicting the metal oxygen content of a submerged arc welding weld.

[0127] The present application provides a computer-readable storage medium storing a computer program. When the computer program is executed on a processor, the method for predicting the metal oxygen content of a submerged arc welding weld is implemented.

[0128] It can be understood that the device of this embodiment corresponds to the method of the above embodiment, and the options in the above embodiment are also applicable to this embodiment, so they will not be described repeatedly here.

[0129] The present application also provides a terminal device. Exemplarily, the terminal device includes a processor and a memory, wherein the memory stores a computer program, and the processor runs the computer program, thereby enabling the terminal device to execute the functions of each module in the above method or the above apparatus.

[0130] Among them, the processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including a central processing unit (CPU), a graphics processing unit (GPU) and a network processor (NP), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or at least one of other programmable logic devices, discrete gates or transistor logic devices, and discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc., which can implement or execute the disclosed methods, steps and logic block diagrams in the embodiments of the present application.

[0131] The memory may be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), etc. The memory is used to store a computer program, and the processor may execute the computer program accordingly after receiving an execution instruction.

[0132] The present application also provides a computer-readable storage medium for storing the computer program used in the above-mentioned terminal device. For example, the computer-readable storage medium may include, but is not limited to, various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0133] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and structure diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in an alternative implementation, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the structure diagram and / or the flow diagram, and the combination of boxes in the structure diagram and / or the flow diagram, can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0134] In addition, the functional modules or units in the various embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

[0135] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a smart phone, a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application.

[0136] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for predicting the metal oxygen content of a submerged arc welding weld, characterized in that: include: Get the set flux composition and welding parameters; Inputting the welding heat input calculated by the flux composition and the welding parameters into a trained prediction model for oxygen content in weld metal of submerged arc welding to obtain a predicted value of oxygen content in weld metal after welding using the flux composition and the welding parameters; The submerged arc welding weld metal oxygen content prediction model is a prediction model obtained by inputting training data into a BP neural network for training.

2. The method for predicting the metal oxygen content of submerged arc welding weld according to claim 1, characterized in that: The prediction model obtained by inputting the training data into the BP neural network training includes: Get training data; Through the training data and the BP neural network, a prediction model for the metal oxygen content of the submerged arc welding weld is obtained by training; The submerged arc welding weld metal oxygen content prediction model is evaluated by the determination coefficient. If the determination coefficient is not within the set qualified range, the submerged arc welding weld metal oxygen content prediction model is optimized until the determination coefficient obtained after optimization is within the qualified range, thereby obtaining the submerged arc welding weld metal oxygen content prediction model.

3. The method for predicting the metal oxygen content of submerged arc welding weld according to claim 2, characterized in that: The training data includes flux composition, welding heat input, and weld metal oxygen content measurements; The obtaining of training data comprises: Determine the flux composition and welding parameters during welding; Obtaining a welding sample after welding using the flux and the welding parameters, measuring the oxygen content of the weld, then performing wire cutting on the weld of the welding sample to obtain weld metal, sampling the center of the weld metal to measure the oxygen content, obtaining the oxygen content measurement value, and calculating the welding heat input according to the welding parameters; The flux composition is adjusted, and the step of obtaining welding samples after welding using the flux composition and the welding parameters is repeatedly performed to obtain multiple sets of training data consisting of flux composition, welding heat input and weld metal oxygen content measurement values.

4. The method for predicting the metal oxygen content of submerged arc welding weld according to claim 3, characterized in that: The BP neural network includes an input layer, a hidden layer and an output layer; the prediction model of metal oxygen content in a submerged arc welding weld is obtained by training the training data and the BP neural network, including: Using the welding heat input calculated from the flux composition and the welding parameters as input data of the input layer, and the measured value of the weld metal oxygen content as input data of the output layer; The input data is transferred to the output layer through the hidden layer by forward propagation and the gradient is calculated by back propagation to optimize the parameters of the BP neural network and obtain a prediction model for the metal oxygen content of the submerged arc welding weld.

5. The method for predicting the metal oxygen content of submerged arc welding weld according to claim 4, characterized in that: The step of transferring the input data through the hidden layer to the output layer by forward propagation and calculating the gradient by back propagation to optimize the parameters of the BP neural network includes: During each training, the output layer error and hidden layer error are calculated; According to the output layer error and the hidden layer error, a first weight from the input layer to the hidden layer is corrected, and a second weight from the hidden layer to the output layer is corrected; When the number of training times of the parameters of the BP neural network reaches a preset value or the output layer error is lower than a preset minimum error requirement, the optimization operation of the first weight and the second weight is completed.

6. The method for predicting the metal oxygen content of submerged arc welding weld according to claim 2, characterized in that: The calculation expression of the determination coefficient is: In the formula, R 2 is the determination coefficient, SS res is the residual sum of squares, SS tot is the total sum of squares; The submerged arc welding weld metal oxygen content prediction model is evaluated by the determination coefficient, and if the determination coefficient is not within a set qualified range, the submerged arc welding weld metal oxygen content prediction model is optimized, including: Determine whether the determination coefficient is within the qualified range. If it is not within the qualified range, the fitting effect of the submerged arc welding weld metal oxygen content prediction model is unqualified; If the determination coefficient is not within the qualified range, the training data is re-input into the submerged arc welding weld metal oxygen content prediction model for training to optimize the submerged arc welding weld metal oxygen content prediction model.

7. The method for predicting the metal oxygen content of submerged arc welding weld according to claim 1, characterized in that: The welding heat input calculated by the flux composition and the welding parameters is input into a trained prediction model for the oxygen content of weld metal in submerged arc welding to obtain a predicted value of the oxygen content of weld metal after welding using the flux composition and the welding parameters, including: Normalizing the flux composition and the welding heat input respectively; The normalized flux composition and the welding heat input are input into the submerged arc welding weld metal oxygen content prediction model to obtain a predicted value of the weld metal oxygen content after welding using the flux composition and the welding parameters.

8. A device for predicting oxygen content of weld metal in submerged arc welding, characterized in that: include: A data acquisition module, used to obtain set flux composition and welding parameters; A prediction module, for inputting the welding heat input calculated by the flux composition and the welding parameters into a trained prediction model for the oxygen content of weld metal in submerged arc welding to obtain a predicted value of the oxygen content of weld metal after welding using the flux composition and the welding parameters; The submerged arc welding weld metal oxygen content prediction model is a prediction model obtained by inputting training data into a BP neural network for training.

9. A terminal device, characterized in that: The terminal device includes a processor and a memory, the memory stores a computer program, and the processor is used to execute the computer program to implement the method for predicting the metal oxygen content of a submerged arc welding weld according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer program is stored therein, and when the computer program is executed on a processor, the method for predicting the metal oxygen content of a submerged arc welding seam according to any one of claims 1 to 7 is implemented.

Citation Information

Patent Citations

  • Welding line forming quality monitoring method

    CN108500498A

  • Prediction method for submerged-arc welding weld metal components and application of prediction method

    CN117464237A

  • Soldering flux viscosity prediction method and device based on data-driven neural network and terminal

    CN118553323A