Large heat input high-performance combined welding method suitable for thick-wall high-strength steel

Abstract

The invention provides a large-heat-input high-performance combined welding method suitable for thick-wall high-strength steel, relates to the technical field of high-strength steel welding, and solves various technical problems of thick-wall high-strength steel welding by integrating a machine learning algorithm and an online monitoring technology. The preheating temperature can be accurately predicted through the temperature prediction model constructed through the SVM algorithm, and therefore the heat input amount in the welding process can be more effectively controlled. And the linear regression model further optimizes the relation between the welding parameters and the actual fusion depth, and the accuracy of welding parameter setting is improved. And the welding parameters are adjusted in real time through the PID control model by combining the comprehensive evaluation index generated by the sound wave and appearance detection technology, so that the stability of the welding process and the consistency of the welding quality are ensured. The effects of improving the welding efficiency and the welding seam quality and reducing the material waste and the rework rate are achieved, and good economic and social benefits are achieved in the fields of aerospace, ocean engineering, high-pressure containers and the like with high-quality welding requirements.

Description

Technical Field

[0001] The invention relates to the technical field of high-strength steel welding, and in particular to a large-energy input and high-performance combined welding method suitable for thick-walled high-strength steel. Background Art

[0002] In the traditional thick-walled high-strength steel welding process, controlling the welding quality, especially the accuracy of the penetration depth, is often a challenge. The complexity of the welding environment and the variability of the material properties greatly increase the difficulty of welding. For example, the preheating temperature, equipment power, welding speed and other parameters in the welding process are difficult to adjust, making it difficult for traditional empirical formulas and simple control methods to meet the needs of high-quality welding. Especially in the process of high-energy line welding, the control of heat input is crucial to the formation of the weld, and is easily affected by changes in the preheating temperature, ambient temperature and the thermophysical properties of the material. In addition, the quality evaluation after welding mostly relies on post-testing, and it is impossible to adjust the welding parameters in real time, which leads to instability in the welding process and fluctuations in welding quality.

[0003] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0004] The object of the present invention is to provide a high-energy input and high-performance combined welding method suitable for thick-walled high-strength steel, so as to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A high-energy input and high-performance combined welding method suitable for thick-walled high-strength steel, the specific steps comprising:

[0007] S1: Collect material data and dimension data of steel materials in historical welding information, as well as preheating temperature and ambient temperature during successful welding, actual penetration depth after welding, and welding parameters of welding equipment, build a temperature prediction model based on machine learning algorithm, and divide the material data, dimension data, and preheating temperature into a first training set and a first validation set to train the temperature prediction model;

[0008] S2: After the temperature prediction model is trained, the input heat during welding is calculated based on the preheating temperature generated by the temperature prediction model and the simplified heat balance equation. The input heat is then corrected based on the ambient temperature. A linear regression model for the actual penetration depth is constructed based on the corrected input heat and welding parameters. The collected preheating temperature, welding parameters, and actual penetration depth are divided into a second training set and a second validation set to optimize the linear regression model.

[0009] S3: Collect material data and dimension data of the steel to be welded, substitute them into the temperature prediction model to generate the preheating temperature and calculate the input heat, then substitute the input heat and the theoretical penetration depth preset by the user into the linear regression model to solve, generate the corresponding welding parameters and weld a certain section of the steel to be welded;

[0010] S4: After welding is completed, the weld is subjected to ultrasonic and visual inspections, and a comprehensive evaluation index is generated based on the weld inspection results. The theoretical penetration depth preset by the user is then adjusted based on the comprehensive evaluation index, and the next section of steel to be welded is welded. This step is repeated until all the steel to be welded are welded.

[0011] Preferably, the material data is the content ratio of C, Mn and Si in the steel, the size data includes the thickness, length and width of the steel, and the welding parameters include laser power and welding speed.

[0012] Preferably, the temperature prediction model is constructed using a support vector machine algorithm, and a feature set is constructed based on material data and dimension data, and the feature set is used as the input of the temperature prediction model, and the preheating temperature during successful welding is used as the output. The expression of the temperature prediction model is:

[0013]

[0014] In the formula , , , Respectively represent the content ratio of C, Mn and Si in steel. , Respectively represent the thickness and width of the steel, Indicates the preheat temperature of the output.

[0015] Preferably, the calculation method for correcting the input heat based on the ambient temperature is:

[0016]

[0017] In the formula Indicates the input heat during welding, , , , They represent the density, initial temperature, heat transfer coefficient, and surface area of ​​the steel, respectively. Indicates the ambient temperature.

[0018] Preferably, the linear regression model of the actual penetration depth is expressed as:

[0019]

[0020] In the formula , , Respectively represent the first model coefficient, the second model coefficient, and the third model coefficient, represents the residual constant, , They represent the laser power and welding speed during welding respectively.

[0021] Preferably, the generation logic of the comprehensive evaluation index is:

[0022] Based on the AUT technology, the weld is ultrasonically inspected to obtain the defect length inside the weld, and the internal evaluation index is generated according to the defect type and defect length. The calculation method is:

[0023]

[0024] In the formula represents the internal evaluation index, , represent the defect length and the corresponding weight coefficient respectively, Indicates the total length of the weld, subscript Indicates the number of the defect, ;

[0025] Collect the surface image of the outside of the weld, perform grayscale processing on the surface image, detect the abnormal area in the surface image, determine whether the abnormal area is a crack or a cavity, and calculate the appearance evaluation index according to the area size of the abnormal area. The calculation method is:

[0026]

[0027] In the formula represents the appearance evaluation index, , Respectively represent the area of ​​the crack in the abnormal area and the corresponding weight coefficient, , Respectively represent the area of ​​the hole in the abnormal area and the corresponding weight coefficient, Indicates the total area of ​​the weld, subscript Indicates the number of the crack in the abnormal area, , Indicates the number of the hole in the abnormal area, , Indicates the preset adjustment factor;

[0028] The comprehensive evaluation index is generated based on the internal evaluation index and the appearance evaluation index. The calculation method is:

[0029]

[0030] In the formula , denote the first proportional coefficient and the second proportional coefficient respectively, ,and .

[0031] Preferably, the method for determining whether the abnormal area is a crack or a cavity is:

[0032] Get the gray value of each pixel , subscript Represents the coordinates of the pixel point. , , and use the Prewitt operator to calculate the gradient value of each pixel and the gradient direction ;

[0033] The gradient value of each pixel With the preset segmentation threshold Compare and satisfy The pixel points of are marked as image background, satisfying The pixel points are marked as the target edge, and then a closed figure is formed according to the corresponding gradient direction, and the closed figure is considered to be an abnormal area;

[0034] Calculate the aspect ratio based on the horizontal and vertical lengths of the abnormal area , the calculation method is:

[0035]

[0036] In the formula , Respectively represent the length of the abnormal area in the horizontal and vertical directions;

[0037] Compare the aspect ratio to a preset ratio threshold For comparison, when When When , the abnormal area is considered to be a cavity.

[0038] Preferably, the segmentation threshold The method to determine is:

[0039] Calculate the average gradient value of the image background and the detection target respectively , , the calculation method is:

[0040]

[0041]

[0042] In the formula Represents the number of pixels in the image background. Indicates the number of pixels of the detected target, Represents the gradient value of the pixel marked as the image background, Represents the gradient value of the pixel point marked as the abnormal area;

[0043] Then calculate the inter-class variance of the image background and the detected target , the calculation method is:

[0044]

[0045] Changing the segmentation threshold The size of Gray value range, calculate the average gray value of the corresponding image background and detection target , and between-class variance , will have the largest between-class variance The gray value is used as the segmentation threshold .

[0046] Preferably, the method for adjusting the theoretical penetration depth preset by the user according to the comprehensive evaluation index is:

[0047] The theoretical penetration depth is calibrated as , a control model is constructed based on the PID algorithm, and the expression of the model is:

[0048]

[0049] In the formula Indicates the new theoretical melting depth that needs to be adjusted. represents the preset evaluation threshold, , , Respectively represent the preset proportional gain, integral gain, and differential gain;

[0050] Substitute the comprehensive evaluation index of the weld into the control model, calculate the new theoretical penetration depth and make adjustments.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] The present invention solves the technical problems of insufficient automation and difficulty in setting welding parameters in the thick-walled high-strength steel welding process by integrating machine learning algorithms and online monitoring technologies. First, the temperature prediction model constructed using the SVM algorithm can accurately predict the preheating temperature based on material data and dimensional data, so that the heat input in the welding process can be more effectively controlled. Secondly, the linear regression model further optimizes the relationship between welding parameters and actual penetration depth, and improves the accuracy of welding parameter setting. Finally, the comprehensive evaluation index generated by combining the acoustic wave and appearance detection technology is adjusted in real time through the PID control model to ensure the stability of the welding process and the consistency of the welding quality. It has the effect of improving welding efficiency and weld quality, reducing material waste and rework rate, and has good economic and social benefits for aerospace, marine engineering, high-pressure vessels and other fields that require high-quality welding. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a schematic diagram of the overall method flow of the present invention. DETAILED DESCRIPTION

[0054] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0055] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0056] Example:

[0057] See also Figure 1 , the present invention provides a technical solution:

[0058] A high-energy input and high-performance combined welding method suitable for thick-walled high-strength steel, the specific steps comprising:

[0059] S1: Collect historical welding information, including material data, dimension data of steel, preheating temperature and ambient temperature during successful welding, actual penetration depth after welding, and welding parameters of welding equipment. Historical welding information refers to the various welding parameters and data collected in past welding work. The material data is the content ratio of C, Mn and Si in steel. The dimension data includes the thickness, length and width of steel. The welding parameters include laser power and welding speed. A temperature prediction model is constructed based on a machine learning algorithm, and the material data, dimension data and preheating temperature are divided into a first training set and a first validation set to train the temperature prediction model.

[0060] The temperature prediction model is constructed using the support vector machine algorithm. This is because the support vector machine algorithm is more suitable for processing data with a small number of samples, and a feature set is constructed based on material data and size data, which are used as inputs to the temperature prediction model. The preheating temperature at successful welding is used as output. The expression of the temperature prediction model is:

[0061]

[0062] In the formula , , , Respectively represent the content ratio of C, Mn and Si in steel. , Respectively represent the thickness and width of the steel, Indicates the preheat temperature of the output.

[0063] In this step, the chemical composition of the steel is used and size , The preheating temperature is taken as the input characteristic variable Make predictions, select appropriate kernel functions (such as radial basis functions), and adjust and select optimal hyperparameters (such as penalty parameter C) through cross-validation. Then use the training set to train the model, and find the optimal decision boundary through learning to minimize the prediction error. Error analysis can be measured using statistical indicators such as mean square error (MSE) and R². Finally, use the validation set to evaluate the prediction performance of the model, focusing on the accuracy, stability, and generalization ability of the model.

[0064] By predicting the preheating temperature through the machine learning model, the temperature setting can be personalized according to the actual material composition (C, Mn, Si content) and the size of the steel (thickness, length, width). Compared with traditional empirical formulas or fixed parameter settings, this method can more accurately reflect the needs of the actual welding process and improve the adaptability and accuracy of the welding process. At the same time, through the analysis and learning of historical welding information, the machine learning model can comprehensively consider the impact of various factors on the preheating temperature, thereby optimizing the heat input of the welding process. This strategy can effectively reduce weld quality problems caused by improper heat control during the welding process, such as welding deformation, crack generation, etc., and ensure the consistency and stability of welding quality.

[0065] S2: After the temperature prediction model is trained, the input heat during welding is calculated based on the preheating temperature generated by the temperature prediction model and the simplified heat balance equation. The input heat is then corrected based on the ambient temperature, and a linear regression model for the actual penetration depth is constructed based on the corrected input heat and welding parameters. The collected preheating temperature, welding parameters, and actual penetration depth are divided into the second training set and the second validation set to optimize the linear regression model.

[0066] The calculation method for correcting the input heat based on the ambient temperature is:

[0067]

[0068] In the formula Indicates the input heat during welding, , , , They represent the density, initial temperature, heat transfer coefficient, and surface area of ​​the steel, respectively. Indicates the ambient temperature.

[0069] The first half of the formula represents the theoretical input heat required for the steel to reach the preheating temperature, and the second half represents the input heat wasted by the heat dissipation of the steel. The overall reflection is the input energy required in the preheating stage. By considering the influence of ambient temperature changes, the control of welding heat is more accurate and scientific, the welding process can be better controlled, overheating or underheating can be reduced, and a wider range of working environments can be adapted to ensure the stability and reliability of the welding process, thereby improving welding quality, while reducing energy consumption and material waste, and reducing production costs.

[0070] The linear regression model of the actual penetration depth is expressed as:

[0071]

[0072] In the formula , , Respectively represent the first model coefficient, the second model coefficient, and the third model coefficient, represents the residual constant, , They represent the laser power and welding speed during welding respectively. The formula is used to reflect the actual penetration depth of the steel after the two stages of preheating and welding.

[0073] In this step, the innovation and advancement of the welding process are enhanced by introducing data-based decision support, so that the welding process can be automatically adjusted according to actual conditions (such as different ambient temperatures and material properties), thereby improving the adaptability and flexibility of the welding process.

[0074] S3: Collect material data and dimension data of the steel to be welded, substitute them into the temperature prediction model to generate the preheating temperature and calculate the input heat, then substitute the input heat and the theoretical penetration depth preset by the user into the linear regression model to solve, generate the corresponding welding parameters and weld a section of the steel to be welded.

[0075] In this step, after mastering the specific data of the material, the theoretical melting depth is substituted into the linear regression model. The welding parameters can be directly obtained through the model, which speeds up production preparation and process optimization, enables each weld to achieve the expected penetration depth, ensures product consistency, and reduces possible errors in manually setting parameters, reduces rework and waste, and improves overall production efficiency.

[0076] S4: After welding is completed, the weld is subjected to ultrasonic and visual inspections, and a comprehensive evaluation index is generated based on the weld inspection results. The theoretical penetration depth preset by the user is then adjusted based on the comprehensive evaluation index, and the next section of steel to be welded is welded. This step is repeated until all the steel to be welded are welded.

[0077] The generation logic of the comprehensive evaluation index is:

[0078] Based on the AUT technology, the weld is ultrasonically inspected to obtain the defect length inside the weld, and the internal evaluation index is generated according to the defect type and defect length. The calculation method is:

[0079]

[0080] In the formula represents the internal evaluation index, , represent the defect length and the corresponding weight coefficient respectively, Indicates the total length of the weld, subscript Indicates the number of the defect, .

[0081] For the defect length, there are different severity judgment results in different intervals, and the degree of impact on the overall weld is also different. Therefore, different weight coefficients need to be set to reflect this degree of impact. The judgment interval of the defect length has different standards according to the actual situation. It can be set according to the actual situation, so as to divide defects of different lengths to match the corresponding weight coefficients.

[0082] Collect the surface image of the outside of the weld, perform grayscale processing on the surface image, detect the abnormal area in the surface image, determine whether the abnormal area is a crack or a cavity, and calculate the appearance evaluation index according to the area size of the abnormal area. The calculation method is:

[0083]

[0084] In the formula represents the appearance evaluation index, , Respectively represent the area of ​​the crack in the abnormal area and the corresponding weight coefficient, , Respectively represent the area of ​​the hole in the abnormal area and the corresponding weight coefficient, Indicates the total area of ​​the weld, subscript Indicates the number of the crack in the abnormal area, , Indicates the number of the hole in the abnormal area, , Represents the preset adjustment coefficient, where , They respectively reflect the influence of cracks and voids on the overall quality of the weld, and the specific values ​​can be set by the user.

[0085] The method for determining whether the abnormal area is a crack or a cavity is as follows:

[0086] Get the gray value of each pixel , subscript Represents the coordinates of the pixel point. , , and use the Prewitt operator to calculate the gradient value of each pixel and the gradient direction ;

[0087] The gradient value of each pixel With the preset segmentation threshold Compare and satisfy The pixel points of are marked as image background, satisfying The pixel points are marked as the target edge, and then a closed figure is formed according to the corresponding gradient direction, and the closed figure is considered to be an abnormal area;

[0088] Calculate the aspect ratio based on the horizontal and vertical lengths of the abnormal area , the calculation method is:

[0089]

[0090] Respectively represent the horizontal and vertical lengths of the abnormal area. This formula takes The longer part is used as the length and the shorter part is used as the width, which can ensure that the aspect ratio is always ≥ 1, which is convenient for judgment. For comparison, when When When , the abnormal area is considered to be a cavity.

[0091] This is because cracks are usually elongated, resulting in a larger aspect ratio, while holes are generally round or elliptical, resulting in a smaller aspect ratio. It should be greater than or equal to 1 and close to 1. The specific size is set by the user.

[0092] The segmentation threshold The method for determining is:

[0093] Calculate the average gradient value of the image background and the detection target respectively , , the calculation method is:

[0094]

[0095]

[0096] In the formula Represents the number of pixels in the image background. Indicates the number of pixels of the detected target, Represents the gradient value of the pixel marked as the image background, Represents the gradient value of the pixel point marked as the abnormal area;

[0097] Then calculate the inter-class variance of the image background and the detected target , the calculation method is:

[0098]

[0099] Changing the segmentation threshold The size of Gray value range, calculate the average gray value of the corresponding image background and detection target , and between-class variance , will have the largest between-class variance The gray value is used as the segmentation threshold .

[0100] This method can automatically determine a suitable threshold to segment the image into two categories: image background and abnormal area, so that the inter-class variance is maximized, that is, the difference between the image background and the abnormal area is maximized.

[0101] The comprehensive evaluation index is generated based on the internal evaluation index and the appearance evaluation index. The calculation method is:

[0102]

[0103] In the formula , denote the first proportional coefficient and the second proportional coefficient respectively, ,and .

[0104] The comprehensive evaluation index is used to make a comprehensive judgment based on the quality of the inside and outside of the weld. The longer the internal defect and the larger the abnormal area on the surface, the greater the value of the comprehensive evaluation index and the worse the quality of the weld.

[0105] The method for adjusting the theoretical penetration depth preset by the user according to the comprehensive evaluation index is as follows:

[0106] The theoretical penetration depth is calibrated as , a control model is constructed based on the PID algorithm, and the expression of the model is:

[0107]

[0108] In the formula Indicates the new theoretical melting depth that needs to be adjusted. represents the preset evaluation threshold, , , They represent the preset proportional gain, integral gain, and differential gain respectively. The comprehensive evaluation index of the weld is substituted into the control model, and the new theoretical penetration depth is calculated and adjusted. , , The value can be obtained by the user in advance through simulation software, or by trial welding a batch of steel products.

[0109] Since the comprehensive evaluation index is positively correlated with the defect length and the area of ​​the abnormal region, it needs to be adjusted when it exceeds the preset evaluation threshold. The biggest reason for the appearance of internal defects, cracks, and voids is that the theoretical penetration depth of the steel is not enough, resulting in the solved welding parameters being too small, making the temperature too low during the welding process and the solder not completely melted. When the comprehensive evaluation index decreases (indicating that the weld quality exceeds expectations), the theoretical penetration depth needs to be appropriately reduced to adjust the welding parameters to avoid excessive welding caused by accidental power fluctuations of the equipment; conversely, if the comprehensive evaluation index increases (indicating that the weld quality does not meet expectations), the theoretical penetration depth needs to be increased to improve the welding quality, thereby achieving dynamic optimization and automatic adjustment of the weld quality.

[0110] In summary, the present invention innovatively solves multiple technical problems in the welding process of thick-walled high-strength steel by integrating machine learning algorithms and online monitoring technologies. First, the temperature prediction model constructed using the SVM algorithm can accurately predict the preheating temperature based on material data and dimensional data, so that the heat input in the welding process can be more effectively controlled. Secondly, the linear regression model further optimizes the relationship between welding parameters and actual penetration depth, and improves the accuracy of welding parameter setting. Finally, the comprehensive evaluation index generated by combining the acoustic wave and appearance detection technology is adjusted in real time through the PID control model to ensure the stability of the welding process and the consistency of the welding quality. The implementation of this solution can improve welding efficiency and weld quality, reduce material waste and rework rate, and has obvious economic and social benefits for aerospace, marine engineering, high-pressure vessels and other fields that require high-quality welding.

[0111] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0112] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0113] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0114] 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 high-energy-line high-performance combined welding method suitable for thick-walled high-strength steel, characterized in that: The specific steps include: S1: Collect material data and dimension data of steel materials in historical welding information, as well as preheating temperature and ambient temperature during successful welding, actual penetration depth after welding, and welding parameters of welding equipment, build a temperature prediction model based on machine learning algorithm, and divide the material data, dimension data, and preheating temperature into a first training set and a first validation set to train the temperature prediction model; S2: After the temperature prediction model is trained, the input heat during welding is calculated based on the preheating temperature generated by the temperature prediction model and the simplified heat balance equation. The input heat is then corrected based on the ambient temperature. A linear regression model for the actual penetration depth is constructed based on the corrected input heat and welding parameters. The collected preheating temperature, welding parameters, and actual penetration depth are divided into a second training set and a second validation set to optimize the linear regression model. S3: Collect material data and dimension data of the steel to be welded, substitute them into the temperature prediction model to generate the preheating temperature and calculate the input heat, then substitute the input heat and the theoretical penetration depth preset by the user into the linear regression model to solve, generate the corresponding welding parameters and weld a certain section of the steel to be welded; S4: After welding is completed, the weld is subjected to ultrasonic and visual inspections, and a comprehensive evaluation index is generated based on the weld inspection results. The theoretical penetration depth preset by the user is then adjusted based on the comprehensive evaluation index, and the next section of steel to be welded is welded. This step is repeated until all the steel to be welded are welded.

2. A high-energy input and high-performance combined welding method suitable for thick-walled high-strength steel according to claim 1, characterized in that: The material data is the content ratio of C, Mn and Si in the steel, the size data includes the thickness, length and width of the steel, and the welding parameters include laser power and welding speed.

3. A high-energy input and high-performance combined welding method suitable for thick-walled high-strength steel according to claim 2, characterized in that: The temperature prediction model is constructed using a support vector machine algorithm, and a feature set is constructed based on material data and dimension data, which are used as inputs to the temperature prediction model, and the preheating temperature at successful welding is used as output. The expression of the temperature prediction model is: In the formula , , , Respectively represent the content ratio of C, Mn and Si in steel. , Respectively represent the thickness and width of the steel, Indicates the preheat temperature of the output.

4. The high-energy input and high-performance combined welding method for thick-walled high-strength steel according to claim 3 is characterized in that: The calculation method for correcting the input heat based on the ambient temperature is: In the formula Indicates the input heat during welding, , , , They represent the density, initial temperature, heat transfer coefficient, and surface area of ​​the steel, respectively. Indicates the ambient temperature.

5. The high-energy input and high-performance combined welding method for thick-walled high-strength steel according to claim 2, characterized in that: The linear regression model of the actual penetration depth is expressed as: In the formula , , Respectively represent the first model coefficient, the second model coefficient, and the third model coefficient, represents the residual constant, , They represent the laser power and welding speed during welding respectively.

6. The high-energy input and high-performance combined welding method for thick-walled high-strength steel according to claim 1, characterized in that: The generation logic of the comprehensive evaluation index is: Based on the AUT technology, the weld is ultrasonically inspected to obtain the defect length inside the weld, and the internal evaluation index is generated according to the defect type and defect length. The calculation method is: In the formula represents the internal evaluation index, , represent the defect length and the corresponding weight coefficient respectively, Indicates the total length of the weld, subscript Indicates the number of the defect, ; Collect the surface image of the outside of the weld, perform grayscale processing on the surface image, detect the abnormal area in the surface image, determine whether the abnormal area is a crack or a cavity, and calculate the appearance evaluation index according to the area size of the abnormal area. The calculation method is: In the formula represents the appearance evaluation index, , Respectively represent the area of ​​the crack in the abnormal area and the corresponding weight coefficient, , Respectively represent the area of ​​the hole in the abnormal area and the corresponding weight coefficient, Indicates the total area of ​​the weld, subscript Indicates the number of the crack in the abnormal area, , Indicates the number of the hole in the abnormal area, , Indicates the preset adjustment factor; The comprehensive evaluation index is generated based on the internal evaluation index and the appearance evaluation index. The calculation method is: In the formula , denote the first proportional coefficient and the second proportional coefficient respectively, ,and .

7. A high-energy input and high-performance combined welding method suitable for thick-walled high-strength steel according to claim 6, characterized in that: The method for determining whether the abnormal area is a crack or a cavity is as follows: Get the gray value of each pixel , subscript Represents the coordinates of the pixel point. , , and use the Prewitt operator to calculate the gradient value of each pixel and the gradient direction ; The gradient value of each pixel With the preset segmentation threshold Compare and satisfy The pixel points of are marked as image background, satisfying The pixel points are marked as the target edge, and then a closed figure is formed according to the corresponding gradient direction, and the closed figure is considered to be an abnormal area; Calculate the aspect ratio based on the horizontal and vertical lengths of the abnormal area , the calculation method is: In the formula , Respectively represent the length of the abnormal area in the horizontal and vertical directions; Compare the aspect ratio to a preset ratio threshold For comparison, when When When , the abnormal area is considered to be a cavity.

8. The high-energy input and high-performance combined welding method for thick-walled high-strength steel according to claim 7, characterized in that: The segmentation threshold The method for determining is: Calculate the average gradient value of the image background and the detection target respectively , , the calculation method is: In the formula Represents the number of pixels in the image background. Indicates the number of pixels of the detected target, Represents the gradient value of the pixel marked as the image background, Represents the gradient value of the pixel point marked as the abnormal area; Then calculate the inter-class variance of the image background and the detected target , the calculation method is: Changing the segmentation threshold The size of Gray value range, calculate the average gray value of the corresponding image background and detection target , and between-class variance , will have the largest between-class variance The gray value is used as the segmentation threshold .

9. The high-energy input and high-performance combined welding method for thick-walled high-strength steel according to claim 6, characterized in that: The method for adjusting the theoretical penetration depth preset by the user according to the comprehensive evaluation index is as follows: The theoretical penetration depth is calibrated as , a control model is constructed based on the PID algorithm, and the expression of the model is: In the formula Indicates the new theoretical melting depth that needs to be adjusted. represents the preset evaluation threshold, , , Respectively represent the preset proportional gain, integral gain, and differential gain; Substitute the comprehensive evaluation index of the weld into the control model, calculate the new theoretical penetration depth and make adjustments.

Images