Arc-shaped plate welding procedure optimization analysis method based on artificial intelligence

Abstract

The invention belongs to the technical field of welding process optimization, and particularly discloses an arc-shaped plate welding process optimization analysis method based on artificial intelligence, which comprises the following steps: performing position and geometric contour recognition on a to-be-welded area on the surface of an arc-shaped plate workpiece, and determining a first welding area based on position information; then stress distribution detection is conducted on the surface of the arc-shaped plate workpiece after welding of the first welding area is completed, then the welding sequence of the follow-up to-be-welded areas is determined according to the stress distribution detection till welding of all the to-be-welded areas is completed, and deformation and stress concentration generated in the welding process of the workpiece can be reduced to the maximum extent; meanwhile, welding temperature is collected in real time and temperature gradient analysis is carried out during welding of all the to-be-welded areas, and then welding parameters are dynamically adjusted according to the temperature gradient so as to maintain the stable temperature gradient; and the thermal stress and the shrinkage effect of the workpiece can be reduced to a certain extent, so that the risk of welding deformation is reduced.

Description

Technical Field

[0001] The invention belongs to the technical field of welding process optimization, and specifically discloses an optimization analysis method for arc plate welding process based on artificial intelligence. Background Art

[0002] As a key manufacturing process, welding has a wide range of application needs in modern society. Since the welding process involves highly complex physical phenomena, significant uncertainty, variability, and multivariable coupling characteristics, it is difficult to ensure consistent and high-quality welding results every time if the welding process is not optimized and adjusted in a targeted manner. Therefore, in actual operation, it is usually necessary to optimize and adjust the welding process to achieve high-quality welding.

[0003] There are many optimization schemes for welding processes in the prior art. For example, the Chinese invention patent with publication number CN111461398A proposes a welding process parameter optimization method based on a neural network model. The neural network is trained through training samples to establish a nonlinear mapping relationship between actual welding process parameters and penetration data. The prediction ability of the trained neural network model is used to continuously adjust the actual process parameter combination to make it close to the optimal process parameter combination corresponding to the target penetration data, thereby simplifying the welding parameter adjustment process and improving the welding quality.

[0004] Another example is the Chinese invention patent with publication number CN118863159A, which proposes a method and system for dynamic optimization of welding processes combined with tracking and monitoring. The method obtains a historical welding record set, parses the record set based on the target welding scenario, determines a calibrated welding process set, configures the forward acquisition step, initializes and activates the weld acquisition component for continuous acquisition, and obtains weld contour data. A displacement weld model containing multiple weld sections is generated through three-dimensional reconstruction, and a weld feature parameter set is extracted. This is used as an index constraint to traverse the calibrated welding process set for matching and obtain a welding process sequence. Based on the welding process sequence, continuous welding control of the target weld is performed, and the tracking and monitoring component is activated to continuously monitor the welding results, generate welding feedback information, and perform welding process responses accordingly to improve the dynamic adjustment capability of the welding process.

[0005] Although the above technical solutions have made significant progress in optimizing welding process parameters, they mainly focus on the optimization of process parameters such as welding speed and current, and ignore the importance of welding sequence. Specifically, in actual welding operations, the workpiece usually contains multiple welding areas. If the welding sequence of these areas is not reasonably planned and the welding sequence is selected arbitrarily, the heat input and cooling rate of different welding areas may be inconsistent, which will cause local stress concentration and increase the risk of crack formation. In addition, the thermal stress and shrinkage effect generated during the welding process will cause the workpiece to deform, and additional workpiece correction is required after the welding is completed, which increases production costs and time. This feature is particularly evident in the welding operation of curved plates. Curved plates have complex geometric shapes, and the stress distribution at different positions during the welding process is quite different, which is easy to cause stress concentration. Curved plates are not only widely used in traditional manufacturing, but also widely used in emerging technology fields such as aerospace, shipbuilding, and energy facilities. Therefore, the optimization of welding sequence in the optimization of curved plate welding process is particularly necessary. Summary of the invention

[0006] To this end, one purpose of an embodiment of the present application is to provide an artificial intelligence-based arc plate welding process optimization analysis method, which solves the problems mentioned in the background technology by optimizing the welding sequence for multiple welding areas of the arc plate workpiece.

[0007] The objective of the present invention can be achieved by the following technical scheme: An artificial intelligence-based arc plate welding process optimization analysis method, comprising the following steps: (1) marking the area to be welded on the surface of the arc plate workpiece, and performing high-resolution image acquisition on the marked workpiece surface, then locating the position of each area to be welded from the acquired image, and capturing the geometric contour of each area to be welded.

[0008] (2) Determine the first welding area based on the location of each area to be welded.

[0009] (3) During the welding process in the first welding area, the welding temperature of the arc plate workpiece is collected, and the temperature gradient analysis is performed based on the collected welding temperature, thereby dynamically adjusting the welding parameters.

[0010] (4) After the welding is completed in the first welding area, the stress distribution of the arc plate workpiece is detected, and the subsequent welding areas are determined in sequence according to the stress distribution detection results. Similarly, the temperature gradient analysis and dynamic adjustment of welding parameters are performed on each subsequent welding area until the welding of all welding areas is completed.

[0011] (5) After the welding of all the areas to be welded of the curved plate is completed, images of the curved plate surface are collected, and stress defect detection and arc plate welding sequence optimization quality evaluation are performed from the collected curved plate surface images.

[0012] Combining all the above technical solutions, the positive effects of the present invention are as follows: 1. The present invention accurately identifies the position and geometric contour of the area to be welded on the surface of the arc plate workpiece, and determines the first welding area based on this information, and then performs stress distribution detection on the surface of the arc plate workpiece after the welding of the first welding area is completed, and then further optimizes and determines the welding sequence of subsequent areas to be welded according to the stress distribution detection results until the welding of all areas to be welded is completed. This method can minimize the deformation and stress concentration of the workpiece during the welding process, thereby significantly reducing the incidence of additional correction required for the workpiece, and ultimately achieving the goal of reducing welding costs and time.

[0013] 2. The present invention collects the welding temperature in real time and performs temperature gradient analysis when welding each area to be welded, and then dynamically adjusts the welding parameters according to the temperature gradient to maintain a stable temperature gradient. This can reduce the thermal stress and shrinkage effect of the workpiece to a certain extent, thereby reducing the risk of welding deformation. In addition, closed-loop control can be achieved by feeding back the temperature gradient analysis to the welding parameter adjustment, ensuring that the parameters in the welding process are always in the optimal state, which is conducive to maintaining the stability of the welding quality.

[0014] 3. After completing the welding of all the areas to be welded on the arc plate, the present invention collects images of the surface of the arc plate, and performs stress defect detection from the collected surface images of the arc plate, thereby performing quality evaluation of the arc plate welding sequence optimization, realizing the feasibility evaluation of the arc plate welding sequence after optimization, and can intuitively provide specific optimization basis for welding management personnel, so that welding management personnel can perform process optimization and management decisions more scientifically. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0016] Figure 1 It is a diagram of the steps for implementing the method of the present invention.

[0017] Figure 2 It is a plan view schematically showing the corresponding center distance and boundary distance of the area to be welded on the surface of the arc plate in the present invention.

[0018] Figure 3 This is a flow chart for dynamically adjusting welding parameters of the area to be welded in the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] See also Figure 1 As shown, the present invention proposes an optimization and analysis method for arc plate welding process based on artificial intelligence, comprising the following steps: (1) marking the area to be welded on the surface of the arc plate workpiece, and performing high-resolution image acquisition on the marked workpiece surface, then locating the position of each area to be welded from the acquired image, and capturing the geometric contour of each area to be welded.

[0021] It is important to know that marking the area to be welded on the surface of the curved plate workpiece according to the welding design drawing before welding can ensure that the welder can accurately find and perform the welding operation, avoiding welding deviations caused by inaccurate positioning.

[0022] When marking the area to be welded, a laser line marker can be used to project the outline of the weld on the surface of the workpiece. Compared with the traditional manual marking method, the laser line marker reduces human errors and can accurately control the size and spacing of the weld to ensure that the position of each welding area meets the design requirements. In addition, the laser line marker can complete the marking of the weld outline in a short time, significantly improving the marking efficiency and shortening the preparation time.

[0023] Applied to the above scheme, the geometric contours of each area to be welded are captured by the following process: focusing the surface image of the curved plate workpiece on each area to be welded, and extracting the contour lines of each area to be welded by contour detection.

[0024] It should be added that after the surface image of the curved plate workpiece is acquired, image preprocessing is required, such as denoising and contrast enhancement. Denoising is to remove noise in the image and improve image quality, and contrast enhancement is to enhance the image contrast through methods such as histogram equalization to make the mark more obvious.

[0025] The geometric contour of the area to be welded is obtained according to the contour lines of each area to be welded.

[0026] (2) Determine the first welding area according to the position of each area to be welded. The specific process is as follows: extract the boundary contour of the arc plate from the surface image of the arc plate workpiece, and mark the center point and boundary turning point according to the boundary contour.

[0027] The center point mentioned above is the geometric center point (center of mass) for calculating the boundary contour of the workpiece. The boundary turning point is to identify the key turning points on the boundary contour (such as corners, bulges, etc.), which are usually locations with large structural changes.

[0028] For each area to be welded, calculate the distance between its geometric center and the center point of the workpiece and record it as the center distance.

[0029] Calculate the distance between the geometric center of each area to be welded and all boundary turning points, and select the minimum value as the boundary distance. Figure 2 shown.

[0030] The center distance and boundary distance corresponding to each area to be welded are compared. If the center distance of a certain area to be welded is smaller than the boundary distance, the position bias of the area to be welded is the center bias, otherwise the position bias of the area to be welded is the boundary bias.

[0031] It should be understood that the center-biased area to be welded is usually located near the center of gravity of the workpiece, which has less impact on the overall structure during welding and helps maintain the stability of the workpiece.

[0032] The area to be welded with a boundary bias is usually located at the edge of the workpiece or at a location with large structural changes, which may have a significant impact on the local stress distribution during welding.

[0033] From the position deviations of each area to be welded, it is identified whether there is an area to be welded with a center deviation. If so, the number of the areas to be welded with a center deviation is counted.

[0034] If there is only one, the area to be welded is taken as the first welding area. If there is more than one, the area to be welded with the shortest distance from the center of each center to the area to be welded is selected as the first welding area.

[0035] It should be emphasized that by prioritizing the area near the center of gravity as the first welding area, the workpiece can maintain a high stability throughout the welding process, thereby improving the welding accuracy. In addition, starting welding near the center of gravity can make the stress generated during the welding process more evenly transferred to the entire workpiece, avoiding structural problems caused by excessive local stress.

[0036] If there is no center-biased area to be welded, the distance between the geometric center of each area to be welded and all boundary turning points is extracted, and the center proximity is calculated by combining it with the boundary distance. The specific calculation formula is: , where represents the center proximity, , They represent the maximum distance between the geometric center of the area to be welded and all boundary turning points, and the boundary distance, respectively. It represents the sum of the boundary distances of all the areas to be welded, and then the area to be welded with the maximum center proximity is taken as the first welding area.

[0037] It should be noted that when there is no center-biased welding area, it means that all welding areas are boundary-biased welding areas. In this case, simply selecting the welding area with a longer boundary distance as the first welding area is a feasible method, because the farther the boundary distance, the closer the area is to the geometric center of the arc plate workpiece surface. However, in order to more accurately determine the first welding area, the center proximity of the welding area is evaluated by comprehensively analyzing the boundary distance of the welding area and the distance difference between the geometric center and all boundary turning points, where the boundary distance is defined as the shortest distance between the geometric center of the welding area and all boundary turning points, reflecting the position of the area relative to the edge of the workpiece, and the distance difference is defined as the distance difference between the geometric center of the welding area and all boundary turning points, reflecting the uniformity of the distance between the area and each boundary turning point. The larger the distance difference, the more uneven the distance between the welding area and each boundary turning point, indicating that the area is more biased toward a certain boundary turning point, and the symmetry between its geometric center and the overall structure of the workpiece is poor. By comprehensively analyzing the boundary distance and distance difference of each area to be welded, the center proximity of each area to be welded can be calculated. The area to be welded with a higher center proximity means that its geometric center is closer to the overall geometric center of the workpiece, and the distance between each boundary turning point is more uniform. Selecting the area to be welded with a high center proximity as the first welding area helps to reduce the stress concentration phenomenon generated during the welding process and ensure the overall stability of the workpiece.

[0038] (3) During the welding process in the first welding area, the welding temperature of the arc plate workpiece is collected, and the temperature gradient analysis is performed based on the collected welding temperature, thereby dynamically adjusting the welding parameters.

[0039] In the manner in which the above scheme can be implemented, welding temperature collection is implemented as follows: during the welding process in the first welding area, a thermal image of the curved plate workpiece is collected using an infrared thermal imager, and the area is marked from the thermal image according to the geometric contour of the first welding area.

[0040] The welding temperature of the first welding area is determined by comparing the chromaticity value of the marked area with the corresponding relationship between chromaticity and temperature in the thermal image.

[0041] The correspondence between chromaticity and temperature in the above-mentioned medium-temperature image can be obtained by establishing a calibration curve or a lookup table between chromaticity value and temperature, which is usually provided by the infrared thermal imager manufacturer, or obtained through laboratory calibration.

[0042] It should be understood that infrared thermal imagers can measure temperature without contacting the workpiece, avoiding interference with the welding process and ensuring that the welding quality is not affected. In addition, infrared thermal imagers can provide high-resolution temperature distribution maps to ensure the accuracy of welding temperature.

[0043] In a further implementation of the above scheme, the temperature gradient analysis refers to the following process: subtracting the welding temperature of the first welding area at the current acquisition time from the welding temperature at the previous acquisition time to obtain the temperature gradient of the first welding area at different acquisition times.

[0044] In a further embodiment of the above scheme, see Figure 3 As shown, the dynamic adjustment of welding parameters refers to the following process: when welding the first welding area using welding equipment according to preset welding parameters, the temperature gradient at each acquisition moment is divided by the interval between adjacent acquisition moments to obtain the temperature increase amplitude at each acquisition moment.

[0045] The welding parameters mentioned above refer to welding speed and welding current. The preset welding parameters depend on the material type, material thickness, weld type, etc. of the arc plate workpiece, and can be determined based on historical welding experience before welding.

[0046] The temperature increase amplitude of the first welding area at each acquisition moment is compared with the set increase amplitude threshold. For example, the increase amplitude threshold is 40°C / min. If the temperature increase amplitude at a certain acquisition moment is higher than the increase amplitude threshold, the first welding area is focused on from the thermal image of the arc plate workpiece at the current acquisition moment to identify whether there is oxidation in the weld of the first welding area. If oxidation exists, the welding current is reduced at the current acquisition moment, otherwise the welding speed is increased at the current acquisition moment. If there is no temperature increase amplitude at a certain acquisition moment higher than the increase amplitude threshold, welding continues to be performed using the preset welding parameters until the welding is completed.

[0047] It should be pointed out that in the welding process, the welding temperature often increases too quickly. For example, if the welding current, voltage or welding speed are set unreasonably, the temperature in the local area may rise rapidly. For example, when welding thick plates, the heat capacity of the workpiece is large and the heat dissipation is slow, which makes it difficult for the temperature in the welding area to dissipate, resulting in a phenomenon of excessive temperature increase. For example, some materials have high thermal conductivity, and heat is easy to diffuse during welding. However, if it is not properly controlled, local overheating may still occur. In this case, the welding parameters need to be adjusted to alleviate the excessive increase in welding temperature.

[0048] It should be explained that when the temperature increase reaches the increase threshold at a certain moment, the focus is on the first weld area to identify whether the weld in the first weld area is oxidized. This is because metals easily react with oxygen in the air at high temperatures to form an oxide layer. Oxidation not only affects the quality and appearance of the weld, but also reduces the mechanical properties of the material. When oxidation is detected, it means that the temperature in the welding area is too high and an oxidation reaction has occurred. At this time, continuing to use the current welding parameters will cause the oxide layer to further expand, affecting the quality of the weld and material properties. Reducing the welding current is one of the most direct and effective means to quickly reduce the heat input in the welding area and prevent further oxidation. In addition, lower currents can also reduce the risk of welding stress and deformation.

[0049] If no oxidation is detected, it means that the temperature of the current welding area is still within the controllable range and has not yet reached the critical temperature for oxidation. In this case, the heat input can be optimized by increasing the welding speed. Increasing the welding speed means that the welding head stays at each position for a shorter time, and the heat input per unit area is also reduced. This can shorten the time the welding area is exposed to high temperature and avoid local overheating. In addition, moving the welding head quickly can make the heat more evenly distributed over a larger area, preventing the local area from being exposed to high temperature for a long time and reducing the risk of oxidation. At the same time, increasing the welding speed can increase production efficiency and shorten welding time.

[0050] The above-mentioned identification of oxidation phenomenon can be achieved by real-time monitoring of the temperature distribution of the welding area through an infrared thermal imager, and high-temperature areas (usually areas with higher oxidation risks) can be identified.

[0051] In the above operation example, assume the initial settings for welding a curved plate: set the welding current to 150A, the welding speed to 5mm / s, and the preset temperature increase threshold to 40°C / min.

[0052] If the temperature increase at the current acquisition moment exceeds 40°C / min, extract the first weld area from the thermal image to analyze whether there is oxidation. If there is oxidation (assuming the temperature exceeds 450°C), reduce the welding current to 135A (a 10% decrease) to reduce heat input and prevent further oxidation. If there is no oxidation, increase the welding speed to 6mm / s (a 20% increase) to shorten the heat input time per unit area and avoid local overheating. Continue to monitor the temperature changes during the welding process, evaluate the effect of the adjustment, and further fine-tune the welding parameters based on the feedback results to ensure welding quality.

[0053] (4) After the welding is completed in the first welding area, the stress distribution of the arc plate workpiece is detected, and the subsequent welding areas are determined in sequence according to the stress distribution detection results. Similarly, the temperature gradient analysis and dynamic adjustment of welding parameters are performed on each subsequent welding area until the operation of all welding areas is completed.

[0054] Preferably, the stress distribution detection refers to the following process: after completing welding in the first welding area, use multiple ultrasonic probes to emit ultrasonic pulses to the curved plate workpiece to ensure that the entire surface of the curved plate workpiece can be covered, and receive the reflected ultrasonic signal, and at the same time record the ultrasonic signal data to perform stress calculation to obtain the stress value of the curved plate workpiece at different ultrasonic probe emission positions.

[0055] It should be pointed out that the use of ultrasound for distribution detection of curved plate workpieces is due to the fact that the propagation speed and attenuation of ultrasound in the material are affected by the internal stress state of the material. Specifically: in the absence of stress, the propagation speed and path of ultrasound are fixed. When stress exists, the propagation speed of ultrasound changes (usually slows down) and may be offset or deformed. Therefore, by analyzing the changes in ultrasonic signals, the stress state inside the material can be inferred.

[0056] Ultrasonic stress distribution testing is a non-destructive testing method that does not cause any physical damage to the workpiece, ensuring the integrity of the workpiece and the safety of subsequent use.

[0057] Further preferably, the subsequent welding area is determined according to the stress distribution detection result as follows: a stress distribution diagram is drawn according to the stress values ​​of the arc-shaped plate workpiece at different ultrasonic probe emission positions.

[0058] The boundary lines of each stress area are extracted from the stress distribution diagram, and the stress values ​​of each stress area are marked.

[0059] The stress region where the remaining area to be welded is located is determined by comparing the geometric contour of the remaining area to be welded with the boundary line of the stress region, and the stress value of the stress region where the remaining area to be welded is located is obtained.

[0060] In the specific implementation of the above scheme, the stress value of the stress region where the remaining area to be welded is located is obtained as follows: the number of stress regions involved in the remaining area to be welded is counted. If a remaining area to be welded is only located in a specific stress region, the stress value of the stress region is directly used as the stress value of the area to be welded. If a remaining area to be welded spans multiple stress regions, the stress values ​​of all stress regions involved in the area to be welded are compared, and the maximum value of these stress values ​​is selected as the stress value of the stress region where the area to be welded is located. This is because in the welding process, higher stress values ​​often correspond to higher risks (such as crack formation, deformation, etc.), and selecting the maximum stress value can obtain the stress condition of the area to be welded under the most unfavorable conditions.

[0061] The symmetric area of ​​the first welding area on the surface of the curved plate workpiece is determined based on the position of the first welding area on the surface of the curved plate workpiece.

[0062] The exemplary method for determining the symmetric region of the first welding area on the surface of the curved plate workpiece is as follows: If the workpiece has symmetry, the corresponding symmetry axis can be defined. For example, assuming that the workpiece is a circular plate, the symmetry axis can be defined by the center of the circle, and the symmetric region of the first welding area can be determined by geometric transformation (such as reflection transformation). For example, if the symmetry axis is a straight line perpendicular to a certain direction, the coordinates of the symmetric region can be obtained by reflection transformation.

[0063] The geometric center of the remaining area to be welded is compared with the geometric center of the symmetrical area corresponding to the first welding area to obtain the symmetrical distance between the remaining area to be welded and the first welding area.

[0064] Substitute the stress value of the stress zone where the remaining area to be welded is located into the evaluation formula combined with the symmetrical distance between the remaining area to be welded and the first welding area. Get the selection value of the remaining area to be welded , where Indicates the stress value of the stress zone where the remaining area to be welded is located. Indicates the symmetrical distance between the remaining area to be welded and the first welding area. It represents the sum of stress values ​​of all remaining stress areas to be welded. It represents the sum of the symmetrical distances between all remaining areas to be welded and the first welding area.

[0065] The remaining welding area corresponding to the maximum selection value is selected from the selection values ​​of the remaining areas to be welded as the subsequent welding area.

[0066] After the welding of the first welding area is completed, the present invention determines the subsequent welding area based on the stress distribution detection of the arc plate workpiece, and comprehensively determines the stress value of the stress area where the remaining area to be welded is located and the distance between the remaining area to be welded and the symmetrical area of ​​the first welding area, and preferentially selects the remaining strip welding area with the smallest stress value in the stress area and the shortest distance between it and the symmetrical area of ​​the first welding area as the subsequent welding area. The reason for doing this is that the area to be welded with the smallest stress value is preferentially selected for welding, which can reduce local stress concentration and avoid problems such as cracks and deformation caused by high stress, and can provide a relatively stable starting point for subsequent welding, and reduce quality problems caused by improper initial welding. Prioritizing welding with the area to be welded that is symmetrical to the first welding area can maintain symmetry during the welding process, help to evenly distribute the heat and stress generated during the welding process, and reduce the deformation risk of the overall structure. Therefore, comprehensive evaluation based on the two factors of stress value and distance can more accurately select the subsequent welding area, ensure that each welding operation is carried out under optimal conditions, and improve welding accuracy.

[0067] (5) After the welding of all the areas to be welded on the curved plate is completed, images of the curved plate surface are collected, and stress defect detection is performed from the collected curved plate surface images to evaluate the quality of the arc plate welding sequence optimization.

[0068] In the manner in which the above scheme can be implemented, the stress defect detection process is as follows: the surface image of the arc-shaped plate workpiece after welding and the surface image of the arc-shaped plate workpiece before welding are respectively subjected to arc-shaped plate shape contour extraction.

[0069] The shape contour of the arc plate after welding is overlapped and aligned with the shape contour of the arc plate before welding, and the shape deformation degree is calculated.

[0070] Specifically, when aligning the arc-shaped plate shape contours before and after welding, geometric transformations (such as translation, rotation, and scaling) can be used to make the two contours overlap as much as possible. After the overlap and alignment, the area difference of the area enclosed by the contours before and after welding can be calculated and divided by the area of ​​the area enclosed by the arc-shaped plate contours before welding to obtain the shape deformation degree.

[0071] Crack identification is performed on the surface images of the curved plate workpiece after welding and before welding, respectively, and the crack generation degree is calculated.

[0072] Specifically, crack recognition can use grayscale co-occurrence matrix, wavelet transform and other methods to extract image features and identify potential crack areas. After crack recognition, the number of cracks can be counted and the crack generation area can be marked. Then, the cracks identified on the surface of the curved plate workpiece after welding can be compared with the cracks identified on the surface of the curved plate workpiece before welding. The number of newly generated cracks is summarized and divided by the number of cracks identified on the surface of the curved plate workpiece after welding to obtain the crack generation degree.

[0073] In a further implementation of the above scheme, the quality of arc plate welding sequence optimization is evaluated as follows: the arc plate shape deformation and crack newness before welding sequence optimization are compared with the arc plate shape deformation and crack newness after welding sequence optimization, and the evaluation formula is used to evaluate the quality of arc plate welding sequence optimization. Get the optimal quality of arc plate welding sequence , where , Respectively represent the deformation of the arc plate shape before and after the welding sequence optimization, , They respectively represent the crack initiation degree of the arc plate before and after the optimization of welding sequence.

[0074] It should be pointed out that by determining the welding method of the subsequent welding area according to the stress distribution on the workpiece surface after each welding, compared with the arbitrary welding sequence setting, this stress analysis-based welding sequence optimization method can effectively reduce the stress concentration generated during the welding process, and further reduce the incidence of workpiece deformation and cracks. Therefore, when conducting welding sequence optimization quality analysis, the shape deformation and crack initiation of the arc plate after the welding sequence of the arc plate workpiece of the same specification and model is optimized. Compare with the shape deformation and crack initiation of the arc plate before the welding sequence optimization, the quality of welding sequence optimization can be quantified.

[0075] The present invention collects images of the surface of the arc plate after completing the welding of all areas to be welded of the arc plate, and performs stress defect detection from the collected surface images of the arc plate, thereby performing quality evaluation of the arc plate welding sequence optimization, realizing the feasibility evaluation of the arc plate welding sequence after optimization, and can intuitively provide specific optimization basis for welding management personnel, so that welding management personnel can perform process optimization and management decisions more scientifically.

[0076] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the protection scope of the present invention.

Claims

1. An optimization analysis method for arc plate welding process based on artificial intelligence, characterized in that: The following steps are involved: (1) Marking the areas to be welded on the surface of the curved plate workpiece, and performing high-resolution image acquisition on the marked workpiece surface. Then, the position of each area to be welded is located from the acquired image, and the geometric contour of each area to be welded is captured; (2) Determine the first welding area based on the location of each area to be welded; (3) During the welding process of the first welding area, the welding temperature of the curved plate workpiece is collected, and the temperature gradient analysis is performed based on the collected welding temperature, thereby dynamically adjusting the welding parameters; (4) After the welding is completed in the first welding area, the stress distribution of the arc plate workpiece is detected, and the subsequent welding areas are determined in sequence according to the stress distribution detection results. The temperature gradient analysis and dynamic adjustment of welding parameters are performed on each subsequent welding area in the same way until the welding of all welding areas is completed; (5) After the welding of all the areas to be welded of the curved plate is completed, images of the curved plate surface are collected, and stress defect detection and arc plate welding sequence optimization quality evaluation are performed from the collected curved plate surface images.

2. The method for optimizing and analyzing arc plate welding process based on artificial intelligence according to claim 1, characterized in that: The process of capturing the geometric contours of each area to be welded is as follows: The surface image of the arc-shaped plate workpiece is focused on each area to be welded, and the contour line of each area to be welded is extracted by contour detection; The geometric contour of the area to be welded is obtained according to the contour lines of each area to be welded.

3. The method for optimizing and analyzing arc plate welding process based on artificial intelligence according to claim 1, characterized in that: The process of determining the first welding area is as follows: Extracting the boundary contour of the curved plate from the surface image of the curved plate workpiece, and marking the center point and boundary turning point according to the boundary contour; For each area to be welded, calculate the distance between its geometric center and the center point of the workpiece, which is recorded as the center distance; Calculate the distance between the geometric center of each area to be welded and all boundary turning points, and select the minimum value as the boundary distance; Compare the center distance and boundary distance corresponding to each area to be welded. If the center distance of a certain area to be welded is smaller than the boundary distance, the position deviation of the area to be welded is the center deviation, otherwise the position deviation of the area to be welded is the boundary deviation. Identify whether there is a center-biased area to be welded from the position deviations of each area to be welded, and if so, count the number of the center-biased areas to be welded; If there is only one, the area to be welded is taken as the first welding area. If there is more than one, the area to be welded with the shortest distance from the center of each center to the area to be welded is selected as the first welding area. If there is no center-biased area to be welded, the distance between the geometric center of each area to be welded and all boundary turning points is extracted, and the center proximity is calculated by combining it with the boundary distance. The specific calculation formula is: , where represents the center proximity, , They represent the maximum distance between the geometric center of the area to be welded and all boundary turning points, and the boundary distance, respectively. It represents the sum of the boundary distances of all the areas to be welded, and then the area to be welded with the maximum center proximity is taken as the first welding area.

4. The method for optimizing and analyzing arc plate welding process based on artificial intelligence according to claim 2, characterized in that: The welding temperature collection is implemented as follows: During the welding process of the first welding area, an infrared thermal imager is used to collect thermal images of the curved plate workpiece, and the area is marked from the thermal image according to the geometric contour of the first welding area; The welding temperature of the first welding area is determined by comparing the chromaticity value of the marked area with the corresponding relationship between chromaticity and temperature in the thermal image.

5. The method for optimizing and analyzing arc plate welding process based on artificial intelligence according to claim 4, characterized in that: The temperature gradient analysis is described in the following process: The temperature gradient of the first welding area at different acquisition times is obtained by subtracting the welding temperature of the first welding area at the current acquisition time from the welding temperature at the previous acquisition time.

6. The method for optimizing and analyzing arc plate welding process based on artificial intelligence according to claim 5, characterized in that: The dynamic adjustment of welding parameters refers to the following process: When welding the first welding area using welding equipment according to preset welding parameters, the temperature gradient at each acquisition moment is divided by the interval between adjacent acquisition moments to obtain the temperature increase amplitude at each acquisition moment; The temperature increase amplitude of the first welding area at each acquisition moment is compared with the set increase amplitude threshold. If the temperature increase amplitude at a certain acquisition moment is higher than the increase amplitude threshold, the first welding area is focused on from the thermal image of the arc plate workpiece at the current acquisition moment to identify whether the weld in the first welding area is oxidized. If oxidation exists, the welding current is reduced at the current acquisition moment, otherwise the welding speed is increased at the current acquisition moment. If the temperature increase amplitude at a certain acquisition moment is not higher than the increase amplitude threshold, welding continues to be performed using the preset welding parameters until the welding is completed.

7. The method for optimizing and analyzing arc plate welding process based on artificial intelligence according to claim 1, characterized in that: The stress distribution detection refers to the following process: After welding is completed in the first welding area, multiple ultrasonic probes are used to transmit ultrasonic pulses to the curved plate workpiece, and the reflected ultrasonic signals are received. At the same time, the ultrasonic signal data is recorded to perform stress calculation to obtain the stress values ​​of the curved plate workpiece at different ultrasonic probe emission positions.

8. The method for optimizing and analyzing arc plate welding process based on artificial intelligence according to claim 7, characterized in that: The following process is used to determine the subsequent welding area according to the stress distribution detection result: A stress distribution diagram is drawn based on the stress values ​​of the curved plate workpiece at different ultrasonic probe emission positions; Extract the boundary lines of each stress area from the stress distribution diagram and mark the stress value of each stress area; Determine the stress region where the remaining area to be welded is located by comparing the geometric contour of the remaining area to be welded with the boundary line of the stress region, and obtain the stress value of the stress region where the remaining area to be welded is located; Determine a symmetrical area of ​​the first welding area on the surface of the curved plate workpiece based on the position of the first welding area on the surface of the curved plate workpiece; Compare the geometric center of the remaining area to be welded with the geometric center of the symmetrical area corresponding to the first welding area to obtain the symmetrical distance between the remaining area to be welded and the first welding area; Substitute the stress value of the stress zone where the remaining area to be welded is located into the evaluation formula combined with the symmetrical distance between the remaining area to be welded and the first welding area. Get the selection value of the remaining area to be welded , where Indicates the stress value of the stress zone where the remaining area to be welded is located. Indicates the symmetrical distance between the remaining area to be welded and the first welding area. It represents the sum of stress values ​​of all remaining stress areas to be welded. It represents the sum of the symmetrical distances between all remaining areas to be welded and the first welding area; The remaining welding area corresponding to the maximum selection value is selected from the selection values ​​of the remaining areas to be welded as the subsequent welding area.

9. The method for optimizing and analyzing arc plate welding process based on artificial intelligence according to claim 8, characterized in that: The process of obtaining the stress value of the stress area where the remaining area to be welded is located is as follows: The number of stress areas involved in the remaining area to be welded is counted. If a remaining area to be welded is only located in a specific stress area, the stress value of the stress area is directly used as the stress value of the area to be welded. If a remaining area to be welded spans multiple stress areas, the stress values ​​of all stress areas involved in the area to be welded are compared, and the maximum value of these stress values ​​is selected as the stress value of the stress area where the area to be welded is located.

10. The method for optimizing and analyzing arc plate welding process based on artificial intelligence according to claim 1, characterized in that: The stress defect detection and arc plate welding sequence optimization quality evaluation refer to the following process: The curved plate shape contour is extracted from the curved plate workpiece surface image after welding and the curved plate workpiece surface image before welding respectively; The shape contour of the arc plate after welding is overlapped and aligned with the shape contour of the arc plate before welding, thereby calculating the shape deformation degree; Crack identification is performed on the surface image of the arc plate workpiece after welding and the surface image of the arc plate workpiece before welding, thereby calculating the crack generation degree; The deformation and crack initiation of the arc plate before and after the welding sequence optimization were compared. Get the optimal quality of arc plate welding sequence , where , Respectively represent the deformation of the arc plate shape before and after the welding sequence optimization, , They respectively represent the crack initiation degree of the arc plate before and after the optimization of welding sequence.

Images