Welded joint yield strength prediction method, system, storage medium and electronic device

Abstract

The present application relates to the technical field of welding, and particularly relates to a welding joint yield strength prediction method, a system, a storage medium and an electronic device. In the method, a prediction value of yield strength of a welding joint to be predicted is predicted according to a normalized value of a chemical element proportion of a base material of the welding joint to be predicted, a normalized value of a chemical element of a welding material of the welding joint to be predicted, a normalized value of a welding process parameter of the welding joint to be predicted, a trained mass transfer model, a mathematical relationship model and a final weld joint yield strength prediction model. The prediction accuracy of the yield strength of the welding joint can be improved, and the method has a great application prospect in the prediction of automatic welding weld strength and the evaluation of the overall quality of the weld.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of welding, in particular to a welding joint yield strength prediction method, system, storage medium and electronic device. BACKGROUND

[0002] Long-distance pipelines have high pressure, large caliber and long line distance, and the geological conditions of the areas they pass through are diverse and complex. The failure of girth welds is caused by the interaction of factors such as defects, load, strength matching, weld geometry and residual stress. The difficulty of welding technology increases for high-grade pipelines, which increases the possibility of girth weld failure. The cause analysis of girth weld failure accidents has formed some conclusions that are helpful to production operation. By combining the problems found in the safety hazard investigation, the metallurgy, plate making, pipe making, design, construction, inspection and acceptable risk and reliability of each link are comprehensively sorted out, compared and analyzed, and evaluated, and the reliability evaluation factors of high-grade pipelines are studied, which provides certain technical support for the reliability protection of in-service and future new high-grade pipelines. However, some problems of the cause of girth weld failure have not been fully solved, and the support for the integrity management of in-service pipelines and the risk control of new pipelines is insufficient.

[0003] As a weak link in the pipeline system, the performance of girth weld directly affects the overall reliability of the pipeline system. Girth weld performance prediction technology can predict girth weld performance based on welding process parameters and other information, which is beneficial to ensuring the quality of new pipelines, reducing the failure probability of girth welds, and providing direct technical support for the integrity management of girth welds of in-service pipelines. However, the performance of girth weld is influenced by many factors such as the chemical composition of the base material and welding material, welding process parameters (including welding method, welding current, voltage, welding speed, etc.), welding structure (pipe diameter, wall thickness, etc.), welding environment (temperature, humidity), etc. Although some factors affecting the performance of girth welds and the correlation between factors have been identified in the early stage, a systematic quantitative correlation model between girth weld composition, process and weld structure and performance has not been established, and it is not possible to accurately predict the performance of the weld based on the welding process, base material and welding material. SUMMARY

[0004] The technical problem to be solved by the present application is to provide a welding joint yield strength prediction method, system, storage medium and electronic device to overcome the shortcomings of the prior art.

[0005] The technical scheme of a welding joint yield strength prediction method of the present application is as follows:

[0006] An initial data set is obtained, each sample in the initial data set including a normalized value of the chemical element proportion of the base material, a normalized value of the chemical element proportion of the welding material, a normalized value of the welding process parameter, and a normalized value of the weld yield strength.

[0007] The mass transfer model is trained by taking the normalized values of the chemical element proportions of the base material, the normalized values of the chemical element proportions of the welding material and the normalized values of the welding process parameters of the samples in the initial data set as input data, to obtain a trained mass transfer model, and the normalized values of the chemical element proportions of the welds are output data of the mass transfer model, and the normalized values of the chemical element proportions of multiple welds are output by using the trained mass transfer model;

[0008] A mathematical relationship model between the welding process parameters and the weld cooling speed is established.

[0009] The normalized values of the chemical element proportions of the welds and the normalized values of the weld cooling speeds are taken as input data of a preset deep learning model, and the normalized values of the weld cooling speeds of the welds are taken as output data of the preset deep learning model, the preset deep learning model is trained, and an initial weld yield strength prediction model is obtained.

[0010] The initial weld yield strength prediction model is trained by using a real data set, and a final weld yield strength prediction model is obtained.

[0011] According to the normalized values of the chemical element proportions of the base material of the to-be-predicted welded joint, the normalized values of the chemical elements of the welding material of the predicted welded joint, the normalized values of the welding process parameters of the predicted welded joint, the trained mass transfer model, the mathematical relationship model and the final weld yield strength prediction model, a predicted value of the yield strength of the to-be-predicted welded joint is obtained.

[0012] The technical scheme of the welded joint yield strength prediction system of the present application is as follows:

[0013] The system comprises an acquisition module, a training module, an establishment module and a prediction module.

[0014] The acquisition module is used to acquire an initial data set, and each sample in the initial data set comprises normalized values of the chemical element proportions of the base material, normalized values of the chemical element proportions of the welding material, normalized values of the welding process parameters and normalized values of the weld yield strength.

[0015] The training module is used to train a mass transfer model by taking the normalized values of the chemical element proportions of the base material, the normalized values of the chemical element proportions of the welding material and the normalized values of the welding process parameters of the samples in the initial data set as input data, to obtain a trained mass transfer model, and the normalized values of the chemical element proportions of the welds are output data of the mass transfer model, and the normalized values of the chemical element proportions of multiple welds are output by using the trained mass transfer model.

[0016] The establishment module is used to establish a mathematical relationship model between the welding process parameters and the weld cooling speed.

[0017] The training module is also used to: use the normalized values ​​of the chemical element ratio of the weld and the normalized values ​​of the weld cooling rate as input data of the preset deep learning model, and use the normalized values ​​of the weld cooling rate as output data of the preset deep learning model to train the preset deep learning model and obtain an initial weld yield strength prediction model.

[0018] The training module is also used to: train the initial weld yield strength prediction model using a real dataset to obtain the final weld yield strength prediction model;

[0019] The prediction module is used to: predict the yield strength of the weld joint to be predicted based on the normalized value of the chemical element ratio of the base material of the weld joint to be predicted, the normalized value of the chemical element of the welding material of the weld joint to be predicted, the normalized value of the welding process parameters of the weld joint to be predicted, the trained mass transfer model, the mathematical relationship model and the final weld yield strength prediction model.

[0020] The present invention provides a storage medium storing instructions, wherein when a computer reads the instructions, the computer executes the method for predicting the yield strength of a welded joint as described above.

[0021] An electronic device according to the present invention includes a processor and the above-described storage medium, wherein the processor executes instructions in the storage medium.

[0022] The beneficial effects of this invention are as follows:

[0023] It can improve the accuracy of predicting the yield strength of welded joints and has great application prospects in predicting the strength of automatic welds and evaluating the overall quality of welds. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating a method for predicting the yield strength of a welded joint according to an embodiment of the present invention.

[0025] Figure 2 This is a geometric diagram of a single V-groove welded joint.

[0026] Figure 3 This is a schematic diagram of a welded joint yield strength prediction system according to an embodiment of the present invention. Detailed Implementation

[0027] like Figure 1 As shown, an embodiment of the present invention provides a method for predicting the yield strength of a welded joint, comprising the following steps:

[0028] S1. Obtain the initial dataset. Each sample in the initial dataset includes the normalized value of the chemical element ratio of the base material, the normalized value of the chemical element ratio of the welding material, the normalized value of the welding process parameters, and the normalized value of the weld yield strength.

[0029] S2. Using the normalized values ​​of the chemical element ratios of the base material, the welding material, and the welding process parameters in the initial dataset as input data, the mass transfer model is trained to obtain a trained mass transfer model. The output data of the mass transfer model is the normalized value of the chemical element ratio of the weld. The trained mass transfer model is then used to output the normalized values ​​of the chemical element ratios of multiple welds.

[0030] S3. Establish a mathematical relationship model between welding process parameters and weld cooling rate;

[0031] S4. Using the normalized values ​​of the chemical element ratio of the weld and the normalized value of the weld cooling rate as input data of the preset deep learning model, and using the normalized value of the weld cooling rate as output data of the preset deep learning model, the preset deep learning model is trained to obtain the initial weld yield strength prediction model.

[0032] S5. Train the initial weld yield strength prediction model using real datasets to obtain the final weld yield strength prediction model.

[0033] S6. Based on the normalized values ​​of the chemical element ratio of the base material of the weld joint to be predicted, the normalized values ​​of the chemical elements of the welding material of the weld joint to be predicted, the normalized values ​​of the welding process parameters of the weld joint to be predicted, the trained mass transfer model, the mathematical relationship model, and the final weld yield strength prediction model, the predicted value of the yield strength of the weld joint to be predicted is obtained.

[0034] The default deep learning model is the RBF neural network.

[0035] Among them, the weld joint to be predicted is the circumferential weld on the pipeline.

[0036] In another embodiment, it includes:

[0037] S11. Obtain the range values ​​of chemical elements and welding process parameters of the base material and welding material to be tested, perform data preprocessing, and obtain initial data collection.

[0038] S12. Train the mass transfer model of the weld seam based on the above initial dataset to obtain the initial fusion ratio of each element in the weld seam to be tested.

[0039] S13. Based on the above initial dataset, establish a mathematical relationship model between the initial data and the weld cooling rate (T8 / 5) to obtain the initial cooling rate (T8 / 5) of the weld to be tested.

[0040] S14. The RBF neural network model is trained based on the initial data of the fusion ratio and cooling rate (T8 / 5) of each element of the weld to be tested, so as to obtain the initial weld strength prediction model.

[0041] S15. Based on the prediction results of the initial strength prediction model of the weld to be tested, N sets of data on the fusion ratio, cooling rate (T8 / 5) and corresponding strength data of each element of the weld to be tested are obtained in a targeted manner to obtain the real dataset.

[0042] S16. Train and test the RBF neural network model using real datasets to obtain the final strength prediction model for the weld to be tested.

[0043] S17. The yield strength of the material to be tested is predicted by using the chemical elements of the base material and welding material, welding process parameters, and the final weld strength prediction model.

[0044] Step S11 includes:

[0045] S110. Using elemental spectroscopy, obtain the chemical elements and welding process parameters of the P groups of base materials and welding materials to be tested. Each group of base materials, welding materials, and welding process parameters is used as a set of raw data, resulting in P groups of raw data as the original dataset. Normalize each data point in the original dataset to form the initial dataset. The normalization formula for each data point in the original dataset is: X = (X... old -X min ) / (X max -X min ).

[0046] Where X is the normalized data, X old This is the data before normalization, X min It is data X old The minimum value of the dimension, X max It is data X old The maximum value in the dimension;

[0047] S111. Randomly divide the initial dataset into 70% initial training set and 30% initial training set.

[0048] In step S12, the mass transfer model of the weld is obtained through a mathematical model for calculating the fusion ratio. This mathematical model requires initial datasets of the base material, welding material, welding process parameters, and bevel form mentioned above. Step S12 includes:

[0049] S120. Fusion ratio refers to the proportion of the base metal that is melted in the weld joint after fusion welding to the total weld metal. It has a direct impact on the composition of the weld, and the calculation formula is expressed by equation (1):

[0050]

[0051] In the formula, θ is the fusion ratio; m X80M The base metal region within the weld section; m fill This refers to the cross-sectional area occupied by the filler metal within the weld section.

[0052] The fusion ratio can be calculated from the geometry of the determined bevel shape, using the geometry of a single V-groove welded joint. Figure 2 For example, the calculation process of its melt ratio is as follows: Let curves AOB and A01B be parabolas, with equations as follows:

[0053] AOB:y=PX 2 (2)

[0054] AO1B:y=aX 2 +bX+c (3)

[0055] Since both curves pass through the rectangular coordinate point B(W / 2, T), and AO1B passes through O1(0, T+h), substituting these values ​​yields SBDE and SOO1B.

[0056] Therefore, the fusion ratio θ = S BDE / S OO1B

[0057] S121: The chemical composition of the weld metal can be derived from the chemical composition of the base metal, welding materials, and fusion ratio. The calculation formula is as follows:

[0058] The weld metal consists of filler metal and locally melted base metal. The chemical composition and fusion ratio of the base metal and filler metal directly affect the composition of the weld. Assuming no alloy loss during welding, the mass fraction of alloying elements in the weld can be expressed by equation (4).

[0059] C0=θC b +(1-θ)C e (4)

[0060] In the formula, C0 is the mass fraction of a certain element in the weld; C b C represents the mass fraction of a certain element in the parent material. e This represents the mass fraction of a certain element in the weld metal.

[0061] From equation (5), we can see that the mass fraction of elements in the weld metal can be calculated from the fusion ratio of the weld, base metal and weld bead.

[0062] Ce=(C0-θC b ) / (1-θ)(5)

[0063] Step S13 includes:

[0064] S130, Establish initial data and weld cooling rate (T) 8 / 5 The mathematical relationship model of ).

[0065] Post-weld cooling rate (T) 8 / 5 T is another major factor affecting the microstructure and properties of welded joints. 8 / 5 For the calculation, this paper selects the following model:

[0066] For three-dimensional heat conduction:

[0067]

[0068] For two-dimensional heat conduction:

[0069]

[0070] Where n is the effective line energy coefficient, E is the welding line energy (J / mm), λ is the thermal conductivity (0.42W / (cm.℃)), To is the preheating temperature (℃), Qc is the specific heat capacity (J / cm2), d is the steel pipe thickness (mm), F2 is the correction coefficient for two-dimensional heat conduction, and F3 is the correction coefficient for three-dimensional heat conduction.

[0071] S131, Cooling rate of weld (T) 8 / 5 A mathematical relationship model was established to determine the welding process parameters, base material, and weld cooling rate under the basic physical and chemical properties of the weld to be tested.

[0072] Step S14 includes:

[0073] S140. Based on the initial fusion ratio and cooling rate (T) of each element in the weld to be tested. 8 / 5 The data results were used to train the RBF neural network model to obtain an initial prediction model for the strength of the weld to be tested.

[0074] In the above embodiments, although the steps are numbered S1, S2, etc., they are only specific embodiments given in this application. Those skilled in the art can adjust the execution order of S1, S2, etc. according to the actual situation, which is also within the protection scope of this invention. It can be understood that in some embodiments, some or all of the above embodiments may be included.

[0075] like Figure 3 As shown, a welded joint yield strength prediction system 200 according to an embodiment of the present invention includes an acquisition module 210, a training module 220, an establishment module 230 and a prediction module 240;

[0076] The acquisition module 210 is used to: acquire an initial dataset, wherein each sample in the initial dataset includes a normalized value of the chemical element ratio of the base material, a normalized value of the chemical element ratio of the welding material, a normalized value of the welding process parameters, and a normalized value of the weld yield strength.

[0077] The training module 220 is used to: train the mass transfer model using the normalized values ​​of the chemical element ratios of the base material, the normalized values ​​of the chemical element ratios of the welding material, and the normalized values ​​of the welding process parameters in the initial dataset as input data, to obtain a trained mass transfer model. The output data of the mass transfer model is the normalized value of the chemical element ratio of the weld, and the trained mass transfer model is used to output the normalized values ​​of the chemical element ratios of multiple welds.

[0078] Module 230 is used to: establish a mathematical relationship model between welding process parameters and weld cooling rate;

[0079] The training module 220 is also used to: use the normalized values ​​of the chemical element ratio of the weld and the normalized values ​​of the weld cooling rate as input data of the preset deep learning model, and use the normalized values ​​of the weld cooling rate as output data of the preset deep learning model to train the preset deep learning model and obtain the initial weld yield strength prediction model.

[0080] Training module 220 is also used to: train the initial weld yield strength prediction model using a real dataset to obtain the final weld yield strength prediction model;

[0081] The prediction module 240 is used to: predict the yield strength of the weld joint to be predicted based on the normalized value of the chemical element ratio of the base material of the weld joint to be predicted, the normalized value of the chemical element of the welding material of the weld joint to be predicted, the normalized value of the welding process parameters of the weld joint to be predicted, the trained mass transfer model, the mathematical relationship model and the final weld yield strength prediction model.

[0082] Optionally, in the above technical solution, the preset deep learning model is an RBF neural network.

[0083] Optionally, in the above technical solution, the weld joint to be predicted is the circumferential weld on the pipeline.

[0084] The parameters and steps of each unit module in the welded joint yield strength prediction system 200 of the present invention described above for realizing their respective functions can be referred to the parameters and steps in the embodiments of the welded joint yield strength prediction method described above, and will not be repeated here.

[0085] An embodiment of the present invention provides a storage medium storing instructions, which, when read by a computer, cause the computer to execute any of the above-described methods for predicting the yield strength of a welded joint.

[0086] An electronic device according to an embodiment of the present invention includes a processor and the aforementioned storage medium. The processor executes instructions in the storage medium. The electronic device may be a computer, a mobile phone, or the like.

[0087] Those skilled in the art will know that this invention can be implemented as a system, method, or computer program product.

[0088] Therefore, this disclosure can be implemented in the following forms: it can be entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software, generally referred to herein as a "circuit," "module," or "system." Furthermore, in some embodiments, the invention can also be implemented as a computer program product in one or more computer-readable media, the computer-readable medium containing computer-readable program code.

[0089] Any combination of one or more computer-readable media may be used. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device.

[0090] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for predicting the yield strength of a welded joint, characterized in that, include: Obtain an initial dataset, wherein each sample in the initial dataset includes a normalized value of the chemical element ratio of the base material, a normalized value of the chemical element ratio of the welding material, a normalized value of the welding process parameters, and a normalized value of the weld yield strength. Using the normalized values ​​of the chemical element ratios of the base material, the welding material, and the welding process parameters in the initial dataset as input data, the mass transfer model is trained to obtain a trained mass transfer model. The output data of the mass transfer model is the normalized value of the chemical element ratio of the weld, and the trained mass transfer model is used to output the normalized values ​​of the chemical element ratios of multiple welds. Establish a mathematical model relating welding process parameters to weld cooling rate; The normalized values ​​of the chemical element ratio of the weld and the normalized value of the weld cooling rate are used as input data for a preset deep learning model, and the normalized value of the weld cooling rate is used as output data for the preset deep learning model. The preset deep learning model is trained to obtain an initial weld yield strength prediction model. The initial weld yield strength prediction model was trained using a real dataset to obtain the final weld yield strength prediction model. The predicted yield strength of the weld joint is obtained by using the normalized values ​​of the chemical element ratio of the base material of the weld joint to be predicted, the normalized values ​​of the chemical elements of the welding material of the weld joint to be predicted, the normalized values ​​of the welding process parameters of the weld joint to be predicted, the trained mass transfer model, the mathematical relationship model, and the final weld yield strength prediction model.

2. The method for predicting the yield strength of a welded joint according to claim 1, characterized in that, The preset deep learning model is an RBF neural network.

3. A method for predicting the yield strength of a welded joint according to claim 1 or 2, characterized in that, The weld joint to be predicted is the circumferential weld on the pipeline.

4. A system for predicting the yield strength of welded joints, characterized in that, It includes an acquisition module, a training module, a building module, and a prediction module; The acquisition module is used to: acquire an initial dataset, wherein each sample in the initial dataset includes a normalized value of the chemical element ratio of the base material, a normalized value of the chemical element ratio of the welding material, a normalized value of the welding process parameters, and a normalized value of the weld yield strength. The training module is used to: train the mass transfer model using the normalized values ​​of the chemical element ratios of the base material, the normalized values ​​of the chemical element ratios of the welding material, and the normalized values ​​of the welding process parameters in the initial dataset as input data, to obtain a trained mass transfer model. The output data of the mass transfer model is the normalized value of the chemical element ratio of the weld, and the trained mass transfer model is used to output the normalized values ​​of the chemical element ratios of multiple welds. The establishment module is used to: establish a mathematical relationship model between welding process parameters and weld cooling rate; The training module is also used to: use the normalized values ​​of the chemical element ratio of the weld and the normalized values ​​of the weld cooling rate as input data of the preset deep learning model, and use the normalized values ​​of the weld cooling rate as output data of the preset deep learning model to train the preset deep learning model and obtain an initial weld yield strength prediction model. The training module is also used to: train the initial weld yield strength prediction model using a real dataset to obtain the final weld yield strength prediction model; The prediction module is used to: predict the yield strength of the weld joint to be predicted based on the normalized value of the chemical element ratio of the base material of the weld joint to be predicted, the normalized value of the chemical element of the welding material of the weld joint to be predicted, the normalized value of the welding process parameters of the weld joint to be predicted, the trained mass transfer model, the mathematical relationship model and the final weld yield strength prediction model.

5. The welded joint yield strength prediction system according to claim 4, characterized in that, The preset deep learning model is an RBF neural network.

6. A welded joint yield strength prediction system according to claim 4 or 5, characterized in that, The weld joint to be predicted is the circumferential weld on the pipeline.

7. A storage medium, characterized in that, The storage medium stores instructions that, when read by a computer, cause the computer to execute a method for predicting the yield strength of a welded joint as described in any one of claims 1 to 3.

8. An electronic device, characterized in that, It includes a processor and the storage medium of claim 7, wherein the processor executes instructions in the storage medium.

Images