An intelligent control system based on a steel column installation, calibration and welding integrated machine

Abstract

This invention relates to the field of welding technology, specifically an intelligent control system based on a steel column installation and calibration welding integrated machine. The system includes a position sensing module, a region division module, a path coupling module, and an execution feedback module. By acquiring the spatial posture of the steel column, the system uses the current steel column butt joint angle, weld position, and column elevation as input data to generate a welding parameter matrix. Based on the welding parameter matrix, the weld position is divided into multiple working areas, and the welding displacement corresponding to each welding parameter in the matrix is ​​recorded. A displacement time series is formed based on the welding displacements, and the deformation is predicted based on the displacement time series, determining the path trajectory line and deviation evolution curve corresponding to the deformation deviation. Based on the deformation deviation in the deviation evolution curve, the system analyzes the changing trend of the deviation evolution curve and executes a welding early warning feedback strategy for each weld position. This achieves accurate welding early warning and timely welding response.

Description

Technical Field

[0001] This invention relates to the field of welding technology, specifically to an intelligent control system based on an integrated welding machine for steel column installation and calibration. Background Technology

[0002] In the field of steel structure construction, the installation, calibration, and welding of steel columns is a core process for ensuring building safety. During the alignment and welding of steel columns, it is necessary to detect any deformation that occurs after welding to ensure the structural quality and safety of the building after installation. However, current technologies for steel column welding rely heavily on manual experience, and deformation monitoring often involves delayed processing, leading to excessive thermal deformation. This results in the accuracy of the steel column welding installation and the quality of the welds failing to meet specifications, affecting the installation of subsequent steel components.

[0003] For example, Chinese Patent Publication No. CN118513745A discloses a welding method, welding system, and storage medium. The welding method includes: acquiring weld information of a workpiece; controlling a welding device to perform arc positioning detection on the workpiece and acquiring the detection result after the arc positioning detection; controlling an image acquisition device to perform weld scanning operation on the workpiece based on the weld information and detection result, and acquiring the coordinate information of the weld of the workpiece; and controlling the welding part of the welding device to perform welding operation on the workpiece based on the weld information, detection result, and coordinate information.

[0004] For example, Chinese Patent Publication No. CN117900712A discloses a welding method, apparatus, electronic device, and storage medium, which relates to the field of welding technology. The key points of its technical solution are: when a welding defect is detected in the workpiece and a re-welding area is determined to correspond to multiple first welding parameters, a second welding parameter for the subsequent welding station is obtained; based on the second welding parameter and the first welding parameter, the re-welding area is allocated to a sub-area to the corresponding subsequent welding station, so that when the subsequent welding station welds the sub-area, the adjustment range from the second welding parameter to the first welding parameter is minimized.

[0005] Existing technologies describe how welding can be adjusted based on the relationship between rotation speed and preset angle, and how welding requirements can be processed using time information of the minimum difference in the welding area. However, these technologies neglect the heat accumulation effect during long weld seam welding and the amount of deformation that occurs during welding. These factors are then used to predict the timeliness and accuracy of welding early warning response by using the deviation between the current welding state and the amount of deformation during welding. Summary of the Invention

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: an intelligent control system based on a steel column installation calibration welding integrated machine, including: a position sensing module, used to collect the spatial posture of the steel column, and use the current steel column docking angle, weld position and column elevation as input data to generate a welding parameter matrix.

[0007] The region division module is used to divide the weld position into multiple working regions based on the welding parameter matrix, identify the welding length and welding swing width in each working region, and record the welding displacement corresponding to each welding parameter in the welding parameter matrix based on the welding length and welding swing width.

[0008] The path coupling module is used to form a displacement time series based on the welding displacement, predict the deformation based on the displacement time series, determine the path trajectory line corresponding to the deformation deviation, and determine whether there is column deformation at each weld position based on the deviation evolution curve on the path trajectory line.

[0009] The execution feedback module is used to analyze the trend of the deviation evolution curve in the form of time-deformation integral based on the deformation deviation of the deviation evolution curve, and to implement welding early warning feedback strategy for each weld position based on the analysis results of the trend.

[0010] The beneficial effects of this invention are as follows: First, this invention collects the spatial posture of the steel column and uses parameters such as the relative angle of the steel column and the position of the weld to detect data such as current and voltage during the welding of the steel column, thereby describing the parameter matrix existing under the current welding condition; at the same time, it introduces a misalignment recognition method to determine whether the current angle position is accurate, thereby setting relevant parameters during the steel column welding process.

[0011] Second, this invention divides multiple working areas based on the geometry of the weld seam, identifies deviation points during welding based on the welding length and welding swing width within each working area, and outputs the displacement of these deviation points as welding displacement to describe the cumulative effect of thermal deformation between steel columns during long weld seam welding. This describes the deformation situation of the current welding process. Subsequently, the deformation deviation is calculated based on the generated displacement, common risk moments in the displacement time series are identified, and closed deformation areas and strip-shaped abnormal areas with welding deformation are viewed using displacement contour maps and other methods, thus completing the identification and monitoring of welding deformation in each working area.

[0012] Third, this invention processes the deviation evolution curve related to the current deformation deviation using a tree structure, describes the hierarchical features related to the time sequence during the welding process using a collaborative state tree, and matches the time nodes under the current time sequence with pre-similarity to obtain the deviation evolution situation corresponding to the current scenario. It also examines the relevant welding early warning feedback strategies to complete the setting of welding early warning feedback strategies for the corresponding time nodes under the time sequence evolution and frequency statistics, thereby improving the accuracy and timeliness of welding early warnings. Attached Figure Description

[0013] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0014] Figure 1 This is a system framework diagram of an intelligent control system based on a steel column installation, calibration, and welding integrated machine.

[0015] Figure 2 This is a flowchart illustrating the area division module of an intelligent control system based on a steel column installation, calibration, and welding integrated machine.

[0016] Figure 3 This is a flowchart illustrating the path coupling module of an intelligent control system based on a steel column installation, calibration, and welding integrated machine.

[0017] Figure 4 This is a flowchart illustrating the execution feedback module of an intelligent control system based on a steel column installation, calibration, and welding integrated machine. Detailed Implementation

[0018] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. Where specific techniques or conditions are not specified in the embodiments, they shall be performed in accordance with the techniques or conditions described in the literature in the art or in accordance with the product manual.

[0019] See Figure 1 An intelligent control system based on a steel column installation, calibration, and welding integrated machine includes: a position sensing module, a region division module, a path coupling module, and an execution feedback module; wherein, the output end of the position sensing module is connected to the region division module, the output end of the region division module is connected to the path coupling module, and the output end of the path coupling module is connected to the execution feedback module.

[0020] The position awareness module is used to collect the spatial posture of the steel column. It takes the current steel column docking angle, weld position and column elevation as input data to generate a welding parameter matrix.

[0021] The region division module is used to divide the weld position into multiple working regions based on the welding parameter matrix, identify the welding length and welding swing width in each working region, and record the welding displacement corresponding to each welding parameter in the welding parameter matrix based on the welding length and welding swing width.

[0022] The path coupling module is used to form a displacement time series based on the welding displacement, predict the deformation based on the displacement time series, determine the path trajectory line corresponding to the deformation deviation, and determine whether there is column deformation at each weld position based on the deviation evolution curve on the path trajectory line.

[0023] The execution feedback module is used to analyze the trend of the deviation evolution curve in the form of time-deformation integral based on the deformation deviation of the deviation evolution curve, and to implement welding early warning feedback strategy for each weld position based on the analysis results of the trend.

[0024] The aforementioned welding parameter matrix includes the coordinates, welding speed, voltage, and current at the steel column butt joint and the weld location.

[0025] The aforementioned spatial posture of the steel column represents the position coordinates, direction angle, butt joint angle, weld position, and column elevation of the steel column during welding. Extracting this positional information helps determine whether the steel column is properly aligned during welding.

[0026] The implementation of the position sensing module includes: controlling the current steel column to align its position, identifying the misalignment of the weld position based on the distribution of the weld position on the butt joint angle of the steel column during welding, and using the coordinates, welding speed, voltage and current after misalignment identification as the output welding parameter matrix.

[0027] The steel columns are marked with their current docking angle and position coordinates to determine if they are aligned. Welding data is extracted based on the aligned position to generate a welding parameter matrix. During alignment, a fixed auxiliary device is used to align the coordinates of the two steel columns using a dual-arm robotic arm according to the current docking angle and direction. The shape of the weld seam at the corresponding docking angle is then identified. For example, if the two steel columns can overlap directly after calibration, welding is performed along the column circumference. If the steel columns have an angle after alignment, such as a triangular shape on a plane, welding is required layer by layer from the inside out to complete the current steel column welding installation. The parameters obtained from the welding parameter matrix indicate the location of the weld layer at each welding stage. It is then determined whether this position coordinate can cover the space corresponding to the steel column docking angle after welding, and the data sensed at this position serves as the basis for subsequent judgments of the current welding status.

[0028] At this point, the misalignment detection is used to compare the actual scanned weld position with the theoretical weld position, and then find out if there is a deviation in the current weld. The deviation and the current coordinates are both input into the welding parameter matrix.

[0029] At this point, image recognition and welding torch / weld seam recognition methods are used to view the current location of the weld seam and its relevant coordinates. A laser vision system or arc sensor can be used to capture the three-dimensional coordinates of the weld seam in real time. By comparing historical weld seam data, a local coordinate system for the steel column is established, and the vector offset between the current weld seam and the previous location is calculated.

[0030] At this point, the welding volume is dynamically correlated with the rotation angle of the steel column. When the steel column rotates, the weld bevel angle changes, leading to a change in the distribution of welding heat input. For example, angular deformation is positively correlated with the transverse shrinkage of the weld. The system monitors welding current fluctuations; if the current change exceeds a 5% threshold during submerged arc welding, it combines this with changes in the number of weld layers and shrinkage (e.g., the second layer shrinkage is 20% of the first layer) to calculate the geometric deformation caused by rotation. The rotation implemented here is identified according to different steel column welding scenarios. When welding steel columns, welding is generally achieved by rotating the welding torch or the steel column, which changes the relative weld bevel angle, making it easy to identify shrinkage when identifying the weld position. The rotation angle represents the adjustment of the rotation angle of the welding robot or positioner by calculating the relative position of the weld and the welding torch in real time, ensuring that the welding torch is always aligned with the center of the weld.

[0031] When performing misalignment identification, the implementation method also includes: determining the relative positional relationship between the previous weld position and the current weld position; based on the relative positional relationship, checking the weld bevel angle at the current weld position; and associating the current weld bevel angle with the angle between the steel column and the butt joint angle to identify the rotation angle of the steel column during welding.

[0032] The deviation of the current weld position is detected, and the rotation angle is adjusted based on the detected weld offset.

[0033] At this point, the adjustment method involves querying the database based on the deviation identified by the weld position to find the rotation angle that needs to be adjusted, thereby completing the welding on the relative curved surface.

[0034] At this point, a three-dimensional coordinate system can be established with the steel column joint point as the origin. The coordinates of any point on the weld and the coordinates of the welding torch can be set in the coordinate system, and the weld vector and welding torch vector can be set respectively. Then, the current included angle can be identified based on these two vectors, and the rotation angle of the welding torch during welding can be obtained according to the process requirements during welding.

[0035] The method for adjusting the rotation angle can be based on the deviation between the weld vector and the welding torch vector in the current three-dimensional coordinate system. Data can be queried from the database to determine whether the rotation angle needs to be adjusted at the current position. If adjustment is needed, the rotation angle queried from the database is used as the adjustment angle; otherwise, no adjustment is made.

[0036] In one embodiment of the present invention, the welding length represents the length value when performing one welding operation, and the welding swing width refers to the maximum distance that the welding torch or welding rod swings laterally from one side to the other during welding.

[0037] At this point, an image acquisition device will be used to view the welding process after the welding torch is aligned with the steel column in the current welding scenario, in order to identify whether there are changes in the welding torch trajectory and whether the welding electrode position swings during the process.

[0038] As for the welding displacement, it will represent the movement trajectory of the welding torch during welding, and the movement trajectory will be displayed in the form of a time series.

[0039] The region division module primarily identifies the weld position based on the welding parameter matrix, recognizing the length and swing width of each weld. This facilitates viewing the trajectory after misalignment identification of the current weld position, illustrating potential deviations and locational offsets during steel column welding. Based on the welding parameter matrix, existing deviations are marked on multiple working areas divided according to weld position. Path displacement is then identified using the deviations within each working area's weld to check their correlation with the current weld length and swing width, and to verify if the welding method is normal.

[0040] like Figure 2 As shown, the implementation of the region division module includes: viewing the region division rules based on the current cross-sectional shape of the steel column, and dividing the weld into multiple working areas according to the region division rules.

[0041] Based on the welding length and welding swing width in each working area, weld deviation points at the weld location are identified. At this time, based on the preset welding length value and welding swing width, the position coordinates of the current weld and the expected weld are identified. It is checked whether there are positions with excessive swing width and increased range. The displacement value of each weld deviation is identified as the output welding displacement. The output welding displacement will include the displacement of the weld with deviation, so as to facilitate checking whether the current welding torch is completed in the expected form. By identifying the form of the corresponding trajectory, the deformation that may exist during welding is identified to obtain the part that needs to be adjusted under the current welding.

[0042] Preferably, the area division rule is expressed as dividing the weld into equal-length areas based on the weld length, curvature, and bevel angle; for example, different length values ​​are used for equal-interval division in the case of cylindrical surfaces with curvature radii and in the case of straight welds; for example, cylindrical surfaces are divided in the form of 100mm or other sizes, while square columns are divided in the form of 500mm or other sizes. The size of the divided area is not limited here. It should be noted that the working area must contain at least one fully welded weld.

[0043] Simultaneously, when identifying weld deviation points at the weld location, the implementation method includes: using the two ends of each working area as standard points, identifying the Euclidean distance of the standard points before and after welding, and describing the weld deviation points using the Euclidean distance of the standard points before and after welding. If there is an Euclidean distance greater than a preset threshold, the corresponding standard point is regarded as a weld deviation point, and the data corresponding to the weld deviation point is used as the output welding displacement. At this time, it describes whether there are deformation or expansion phenomena such as standard points divided in each working area before and after welding. If so, the weld corresponding to this standard point is output to describe the corresponding welding torch trajectory and process it to identify the welding path and position of the current welding torch in the scenario of deformation. The preset threshold will be set according to the average width value of the welding swing width under multiple welding. This welding swing width is the maximum distance of the welding torch's lateral swing, which directly affects the coverage area and welding quality of the weld. Setting the preset threshold by the average width value under multiple welding allows the threshold to dynamically adapt to the swing width requirements under different welding scenarios. For example, in curved welds, the swing width may be smaller due to the greater welding difficulty, so the preset threshold will be adjusted accordingly; while in straight welds, the swing width may be larger, so the preset threshold will also increase. At the same time, this preset threshold can reduce misjudgments and interference when identifying the current weld. When the value exceeds this preset threshold, it indicates that the area has generated a significant amount of deformation, which needs to trigger the corresponding adjustment mechanism. Then, the welding trajectory of the deformed area is identified, and the deformation trend may be more obvious at which coordinate in time sequence. The welding torch trajectory is then adjusted accordingly to complete the welding process between the steel columns.

[0044] In one embodiment of the present invention, when the welding displacement is composed into a displacement time series, the displacement time series represents the movement trajectory information of the welding torch during the welding process. Then, it is necessary to identify the displacement generated by the welding torch and its space, and check whether there is obvious deformation at multiple coordinates under the current displacement time series, so as to identify the current welding situation.

[0045] like Figure 3 As shown, the path coupling module is implemented by recording the moments in the displacement time series where deviations exist, and using these moments as the target moments.

[0046] Based on the target time, feature recognition is performed at the welding speed and welding position in the welding parameter matrix to examine the differences between the target time and adjacent time periods, and to obtain the trend feature value at each target time.

[0047] The characteristic value of the change trend at the current target time is compared with the characteristic value of the change trend at other target times to identify the common risk time under multiple welding batches. The data corresponding to the deformation amount is segmented using the obtained common risk time. If there is no common risk time in the current displacement time series, the maximum deformation amount corresponding to the current displacement time series is recorded.

[0048] The aforementioned trend characteristics are described based on the fluctuations in welding speed and positional deviations between the target time and adjacent time periods, with the positional deviation and speed fluctuations used as the trend characteristics.

[0049] The aforementioned common risk moments refer to the time points in multiple independent welding batches that are identified by comparing the trend characteristics of the target moment when displacement deviation occurs. These time points have similar welding behavior patterns or deformation risks. Then, multiple work areas are identified with data from historical batches to find out whether there are unstable risks at certain moments in the current welding process. This is determined by whether the position deviation and velocity fluctuation at the time of displacement deviation exceed the mean plus the standard deviation in historical data. If they do, the corresponding moment is judged as a common risk moment.

[0050] When segmenting the data corresponding to the deformation amount using the common risk moment, the implementation method includes: obtaining the range value corresponding to the deformation amount, and setting the displacement contour map according to the deformation direction of the deformation amount at the common risk moment.

[0051] By using the difference in duration of displacement contour maps at different common risk moments, deformation deviations are identified, and deformation deviations are used to monitor the location of each weld.

[0052] When identifying deformation deviations, the deformation at the common risk moment will be displayed, along with the length of the current duration. The longer the duration, the more significant the cumulative deformation effect in that area, which may lead to greater structural instability.

[0053] At the same time, it is necessary to determine whether the deviation of the generated deformation exceeds the normal deformation. When it exceeds the normal deformation, the relevant positions need to be marked to explain the deformation state in the current scene.

[0054] When identifying deformation deviations, a high-speed infrared thermal imager and a laser displacement sensor array are needed to identify the displacement of the weld at various locations after welding. Then, the identified displacements are processed, such as closed areas with displacement ≥ 1.5 mm and strip areas with displacement gradient ≥ 0.3 mm / mm. The identified displacements are marked on the identified displacement contour map, especially closed areas with certain displacements and strip areas with obvious gradient changes. Then, it is determined whether the displacement generated during the displacement duration at the time of common risk exceeds the design allowable displacement. This part is combined with the marked displacement form to determine the displacement deviation generated at each weld position. Finally, the path coupled by the displacement at each weld position in the current scenario is defined to illustrate the dynamic change form of the column welding process.

[0055] The method for identifying deformation deviations also includes: based on the acquired displacement contour map, in the displacement field corresponding to the common risk moment, sequentially extracting the closed deformation region corresponding to the displacement amount under the current duration difference and the strip-shaped anomaly region corresponding to the displacement gradient. At this point, the actual displacement generated during steel column welding is used to describe the currently identified deformation amount. Then, the region with specific deformation is divided by displacement gradient and specific displacement amount. This is mainly based on whether the displacement amount has a closed region, and using the significant displacement amount that can be generated during steel column welding as the marker. The average displacement amount that appears at the common risk moment in historical data is used as the significant displacement amount, or points on the steel column can be selected as the markers. The maximum allowable displacement of 80% is used as the standard for judging the closed deformation region corresponding to the displacement. The closed region that is greater than the one considered to be a significant displacement is identified to indicate that there is a significant displacement closure in the current scene. At the same time, the closed region can be extracted from the displacement contour map by the contour detection method in the image recognition algorithm to identify whether there is a corresponding displacement gradient. In this case, the displacement gradient identification needs to select 80% of the maximum allowable displacement gradient at each point on the steel column as the judgment standard. If there are two or more consecutive displacement gradients that reach and exceed 80% of the maximum displacement gradient, they are identified as strip-shaped abnormal regions.

[0056] Meanwhile, the extracted closed deformation regions and strip-shaped anomaly regions will be labeled with the duration value of the current common risk moment. The duration difference mainly indicates that there are multiple value situations within the time period corresponding to different common risk moments, and the closed deformation regions and strip-shaped anomaly regions appear under different duration length values.

[0057] The closed deformation region and the strip-shaped anomaly region are compared to identify whether there is an overlap. If there is, the displacement deviation of the overlap between the closed deformation region and the strip-shaped anomaly region is used as the deformation deviation, based on the direction of the overlap. If there is no overlap, the maximum displacement deviation of the closed deformation region and the gradient deviation of the strip-shaped anomaly region are used as the deformation deviation.

[0058] The output deformation is described in the form of displacement and is divided into several levels, such as normal, warning, and risk, to describe the deformation situation during the welding of the steel column. If the deformation deviation is less than 20% of the theoretical deformation, it is considered normal. If the deformation deviation is between 20% and 50% of the theoretical deformation, it is considered to require a warning. If the deformation deviation is greater than 50% of the theoretical deformation, it is considered to be risky. Corresponding measures are then taken to complete the integrated processing of the positioning welding of the steel column.

[0059] The obtained deformation deviations are then combined according to the values ​​obtained at multiple common risk moments. That is, when the data corresponding to the deformation is segmented using common risk moments, the implementation method also includes: using the deformation deviations at each common risk moment for path coupling to establish a multi-moment deformation index table. This multi-moment deformation index table represents forming an index table for the region corresponding to the deformation deviation at each common risk moment, which facilitates querying the region index and position at different times.

[0060] Multiple regions corresponding to the deformation index table at multiple time points are coupled to form a path trajectory line corresponding to the deformation deviation. Deformation is statistically analyzed using the path trajectory line to form multiple path nodes corresponding to the path trajectory line. The deviation evolution curve is generated by the change value of the deformation deviation on the path node with the displacement time series.

[0061] When coupling multiple regions corresponding to the multi-time deformation index table, closed deformation regions are coupled in the form of displacement fields. Using spatiotemporal displacement, the positions corresponding to the deformation deviations at different risk deformation times are connected to form a trajectory line for displaying the region's displacement. For example, using the center point of the closed deformation region as the connection point, multiple regions are connected into a path trajectory line, resulting in a path trajectory line about the closed deformation region. As for strip-shaped anomaly regions, the corresponding strip-shaped anomaly regions in the current index are coupled in a tiered direction, using their gradient direction and value as individual gradient vectors. Then, each gradient vector is used to connect the strip-shaped anomaly regions at the corresponding common risk times, resulting in a key... The path trajectory line for the strip-shaped anomaly area also needs to be fitted to the path trajectory lines of the closed deformation area and the strip-shaped anomaly area in the form of a fitted path trajectory line, forming a relatively complete path trajectory line. The position of the path trajectory line at different common risk moments when it is fitted is regarded as the path node. The deformation deviation is accumulated. Then, according to the change value of the accumulated deformation deviation, a deviation evolution curve is generated with time as the horizontal axis and deformation deviation as the vertical axis. At this time, the deviation evolution curve can be viewed to directly understand whether there is corresponding column deformation at each weld position. The corresponding curve and other data are output to the database for subsequent staff to view the current value status.

[0062] Preferably, the aforementioned path nodes are obtained by fitting the center point of the closed deformation region and the gradient vector of the strip-shaped anomaly region with respect to the coordinate position of the point during the path trajectory fitting process. Then, the deformation deviation of the corresponding path node is checked. For the path fitting method, the least squares method is used to quantize the path trajectory of the closed deformation region and the strip-shaped anomaly region, and the path nodes on the path trajectory are connected in the form of the shortest path to obtain the current output path trajectory.

[0063] In one embodiment of the present invention, the execution feedback module needs to identify the deformation deviation in the deviation evolution curve as a time series at adjacent positions and the deformation direction, and associate the deviation evolution curve with the deformation deviation in adjacent time periods in the form of time integration to see if there is a scenario where the deformation deviation is large, resulting in the current steel column welding instability.

[0064] For example, at this time, the amplitude difference and slope value of the change in deformation deviation in the corresponding time period are used to construct the relevant time integral value. Then, the cumulative integral value is viewed in the time period corresponding to the multi-level welding during welding to set the welding quality judgment decision matrix, thereby judging the changing trend of the current deviation evolution curve. Then, the changing trend of multiple positions on the path trajectory line is used to explain the current welding status, to explain whether there are faults and abnormalities during the current welding implementation. Finally, the existing faults and abnormalities are output to express the relative situation of the current welding process.

[0065] like Figure 4 As shown, the implementation of the feedback module includes: matching the deviation evolution curve with time nodes, calculating the maximum amplitude difference of the deviation evolution curve at adjacent time points, and performing time integration with the slope value of each time node to determine the cooperative state tree corresponding to each adjacent time point.

[0066] Each node in the collaborative state tree is used as the processing source for the welding early warning feedback strategy. The welding early warning feedback strategy is sent sequentially to the child nodes and leaf nodes of the collaborative state tree to obtain the strategy execution path.

[0067] The welding early warning feedback strategy executed under the strategy execution path is used to perform interactive analysis on each strategy execution path, and the trigger value of each strategy execution path after interactive analysis is used to set the welding early warning feedback strategy output to each weld position.

[0068] The aforementioned welding early warning feedback strategy refers to the strategy executed after the maximum amplitude difference, deformation deviation, and slope value on the deviation evolution curve reach a certain value. Such strategies include reducing the welding speed, adjusting the welding sequence of steel columns, stopping steel column welding, and readjusting the steel column alignment coordinates. These strategies adjust the current steel column welding. To obtain the welding early warning feedback strategy, the values ​​of the deviation evolution curve at adjacent time nodes can be used to query the database and obtain the welding early warning feedback strategy corresponding to the current value.

[0069] The above-mentioned implementation of the cooperative state tree is as follows: the initial welding state corresponding to the current deviation evolution curve is taken as the root node of the cooperative state tree, and the time node after matching is taken as the child node of the cooperative state tree. The values ​​of the child nodes and leaf nodes in the cooperative state tree are represented as the cumulative amount of the maximum amplitude difference and slope value of the corresponding time node under time integration. The adjacent time nodes corresponding to the child nodes are taken as leaf nodes, and the child nodes and leaf nodes are connected in time order to obtain the cooperative state tree.

[0070] Preferably, the initial value of the aforementioned initial welding state when deformation deviation occurs. Regarding the matching of time nodes, the deviation evolution curve between the time node and the preset historical data is calculated using Euclidean distance. For example, the cumulative amount of the maximum amplitude difference and slope value corresponding to the time node under time integration, and the deformation deviation corresponding to the time node are used as the state vector of each time node. After normalizing these three values, the cosine similarity value between the time node and the state vector of the preset deviation evolution curve is used as the matching method for the time node. When there are time nodes with a cosine similarity value greater than 0.7, these time nodes are used as child nodes of the collaborative state tree, and the adjacent time nodes of these time nodes are used as leaf nodes. The child nodes are sorted by the cosine similarity value, so that the leaf nodes are connected to the child nodes in chronological order, ultimately forming a picture of the evolution of the state deviation curve at different time nodes. Then, the welding early warning feedback strategy viewed by each child node and leaf node is used, and the path from the root node to the leaf node is used as the strategy execution path to illustrate how to adjust the current steel column welding under different conditions.

[0071] Preferably, the above-mentioned implementation method for interactive analysis of each strategy execution path is to determine whether the triggering conditions of the welding early warning feedback strategy are consistent on the same strategy execution path, and whether there are scenarios where the triggering conditions of multiple strategy execution paths are consistent. In this case, the triggering condition on the strategy execution path with the highest frequency of occurrence is taken as the triggering value of each strategy execution path, and the welding early warning feedback strategy of the corresponding strategy execution path is output.

[0072] At this point, the interactive analysis mainly analyzes all nodes on the same strategy execution path, checking whether their triggering conditions meet the requirements of logical consistency and parameter coupling. Logical consistency means that there are no mutually exclusive conditions in the welding early warning feedback strategy. For example, if there is a conflict between deformation > 1.5 mm and deformation < 1.0 mm, the corresponding welding early warning feedback strategy will not be adopted. Parameter coupling means that the condition values ​​in the welding early warning feedback strategy meet the value range of the normal process.

[0073] Next, each strategy execution path is compared to see if there are cases of complete overlap, partial overlap, or mutual exclusion. Then, based on the frequency of triggering conditions on the strategy execution path and the weld location, all strategy execution paths related to that weld location are examined, along with the frequency of occurrence of the corresponding triggering conditions. The frequency of occurrence represents the ratio of the occurrence of that triggering condition to the total occurrence of all triggering conditions. Then, by selecting the maximum occurrence frequency to filter out occasional anomalies and by adaptively adjusting the trigger thresholds for different processing strategies in the current scenario, the sensitivity of the corresponding warnings is improved. Finally, the output welding warning feedback strategy can demonstrate how to handle the steel column welding in the current scenario, so as to quickly complete the alignment welding of the steel column and improve the welding quality of the steel column.

[0074] 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, which are still covered within the protection scope of the present invention.

Claims

1. An intelligent control system based on a steel column installation, calibration, and welding integrated machine, characterized in that, include: The position awareness module is used to collect the spatial posture of the steel column, and uses the current steel column docking angle, weld position and column elevation as input data to generate a welding parameter matrix; The region division module is used to divide the weld position into multiple working regions according to the welding parameter matrix, identify the welding length and welding swing width in each working region, and record the welding displacement corresponding to each welding parameter in the welding parameter matrix based on the welding length and welding swing width. The path coupling module is used to form a displacement time series based on the welding displacement, predict the deformation based on the displacement time series, determine the path trajectory line corresponding to the deformation deviation, and determine whether there is column deformation at each weld position based on the deviation evolution curve on the path trajectory line. The execution feedback module is used to analyze the trend of the deviation evolution curve in the form of time-deformation integral based on the deformation deviation of the deviation evolution curve, and to implement welding early warning feedback strategy for each weld position based on the analysis results of the trend. The implementation methods of the region division module include: Based on the current cross-sectional shape of the steel column, examine the area division rules and divide the weld into multiple working areas according to the area division rules; Based on the welding length and welding swing width in each working area, the weld deviation points at the weld position are identified, and the displacement value of each weld deviation is identified as the output welding displacement. When identifying weld deviation points at weld locations, the methods also include: Using the two ends of each working area as standard points, the Euclidean distance between the standard points before and after welding is identified, and the weld deviation points are described by the Euclidean distance between the standard points before and after welding. If there is an Euclidean distance greater than a preset threshold, the corresponding standard point is regarded as a weld deviation point, and the data corresponding to the weld deviation point is used as the output welding displacement.

2. The intelligent control system based on the integrated steel column installation, calibration, and welding machine according to claim 1, characterized in that, The location awareness module can be implemented in the following ways: The current steel column is aligned in position. The distribution of weld position on the butt joint angle of the steel column during welding is used to identify the misalignment of the weld position. The coordinates, welding speed, voltage and current after misalignment identification are used as the output welding parameter matrix. Marking is performed using the current steel column's butt joint angle and position coordinates to determine if the current steel column is aligned in terms of position and angle. Then, the welding data is extracted based on the aligned position to generate a welding parameter matrix.

3. The intelligent control system based on the integrated steel column installation, calibration, and welding machine according to claim 2, characterized in that, When performing misalignment detection, the implementation methods also include: Based on the previous weld position, check the weld bevel angle at the current weld position, and associate it with the angle between the current weld bevel angle and the steel column butt angle to identify the rotation angle of the steel column during welding. The deviation of the current weld position is detected, and the rotation angle is adjusted based on the detected weld offset.

4. The intelligent control system based on the integrated steel column installation, calibration, and welding machine according to claim 1, characterized in that, The implementation methods of the path coupling module include: Record the moments in the displacement time series where deviations occur, and use these moments as the target moments; Based on the target time, feature recognition is performed at the welding speed and welding position in the welding parameter matrix, and the differences between the target time and adjacent time periods are examined to obtain the trend feature value at each target time. The characteristic value of the change trend at the current target time is compared with the characteristic value of the change trend at other target times to identify the common risk time under multiple welding batches. The data corresponding to the deformation amount is segmented using the obtained common risk time. If there is no common risk time in the current displacement time series, the maximum deformation amount corresponding to the current displacement time series is recorded.

5. The intelligent control system based on the integrated steel column installation, calibration, and welding machine according to claim 4, characterized in that, When segmenting the data corresponding to the deformation amount using common risk moments, the implementation methods include: Obtain the range value corresponding to the deformation amount, and set the displacement contour map according to the deformation direction of the deformation amount at the common risk moment; By using the difference in duration of displacement contour maps at different common risk moments, deformation deviations are identified, and deformation deviations are used to monitor the location of each weld. Path coupling is performed based on the deformation deviation at each common risk moment to establish a multi-time deformation index table; Multiple regions corresponding to the deformation index table at multiple time points are coupled to form a path trajectory line corresponding to the deformation deviation. Deformation is statistically analyzed using the path trajectory line to form multiple path nodes corresponding to the path trajectory line. The deviation evolution curve is generated by the change value of the deformation deviation on the path node with the displacement time series.

6. The intelligent control system based on the integrated steel column installation, calibration, and welding machine according to claim 5, characterized in that, Other methods for identifying deformation deviations include: Based on the obtained displacement contour map, in the displacement field corresponding to the common risk moment, the closed deformation region corresponding to the displacement amount under the current duration difference and the strip-shaped anomaly region corresponding to the displacement gradient are extracted sequentially. The closed deformation region and the strip-shaped anomaly region are compared to identify whether there is an overlap. If there is, the displacement deviation of the overlap between the closed deformation region and the strip-shaped anomaly region is used as the deformation deviation, based on the direction of the overlap. If there is no overlap, the maximum displacement deviation of the closed deformation region and the gradient deviation of the strip-shaped anomaly region are used as the deformation deviation.

7. The intelligent control system based on the integrated steel column installation, calibration, and welding machine according to claim 1, characterized in that, The implementation methods for the execution feedback module include: The deviation evolution curve is matched with time nodes, the maximum amplitude difference of the deviation evolution curve at adjacent time nodes is calculated, and the slope value of each time node is integrated over time to determine the cooperative state tree corresponding to each adjacent time point. Each node in the collaborative state tree is used as the processing source for the welding early warning feedback strategy. The welding early warning feedback strategy is sent sequentially to the child nodes and leaf nodes of the collaborative state tree to obtain the strategy execution path. The welding early warning feedback strategy executed under the strategy execution path is used to perform interactive analysis on each strategy execution path, and the trigger value of each strategy execution path after interactive analysis is used to set the welding early warning feedback strategy output to each weld position.

8. The intelligent control system based on the integrated steel column installation, calibration, and welding machine according to claim 7, characterized in that, The implementation of a collaborative state tree is represented as follows: The initial welding state corresponding to the current deviation evolution curve is taken as the root node of the cooperative state tree. The time node after matching is taken as the child node of the cooperative state tree. The values ​​of the child nodes and leaf nodes in the cooperative state tree are represented as the cumulative amount of the maximum amplitude difference and slope value of the corresponding time node under time integration. The adjacent time nodes of the child nodes are taken as leaf nodes. The child nodes and leaf nodes are connected in time order to obtain the cooperative state tree.

Images