A laser inspection system for welding quality of steel box girders in municipal ramp bridges

Abstract

This invention relates to the field of signal detection technology, specifically a laser inspection system for the welding quality of steel box girders in municipal ramp bridges. The system includes a signal analysis module, a defect identification module, a morphological analysis module, a risk assessment module, and a decision support module. In this invention, the accuracy of welding defect identification is improved by analyzing the light intensity distribution in the welding area and the light intensity gradient changes in adjacent scanning segments. The accuracy of identification is optimized by combining the proportion of the defect area. The calculation of numerical indicators of welding quality characteristics makes the assessment more detailed and quantitative, reducing reliance on manual experience and improving detection efficiency and accuracy. By analyzing changes in welding morphology and fluctuations in defect edges, abnormal trends can be identified in a timely manner, and risk points overlapping with the welding stress monitoring area can be screened, optimizing the accuracy of risk assessment, improving predictive ability, and preventing the deterioration of welding quality problems. The optimization of the real-time feedback cycle makes welding quality monitoring more dynamic, adaptable, and accurate.

Description

Technical Field

[0001] This invention relates to the field of signal detection technology, and in particular to a laser detection system for the welding quality of steel box girders in municipal ramp bridges. Background Technology

[0002] Signal detection technology is a comprehensive technical field that studies the identification, analysis, and judgment of the state of a measured object using the characteristics of physical quantity signals. It mainly involves core aspects such as signal acquisition, signal conversion, signal transmission, signal analysis, and judgment of detection results. This field is widely used in industrial manufacturing, structural health monitoring, transportation engineering, and automated inspection. It often uses sensors, laser measuring instruments, ultrasonic probes, or electromagnetic wave devices to acquire physical signals of the target surface or internal structure, and combines these with changes in signal characteristics to assess the integrity and stability of the measured object. In the field of engineering structural inspection, signal detection technology plays an important role in welding quality inspection, and can be used to identify weld defects, welding deformation, and material discontinuities. Specifically, the traditional laser inspection system for the welding quality of steel box girders in municipal ramp bridges refers to the process of manually or semi-automatically inspecting the welds of steel box girders using ultrasonic flaw detectors, magnetic particle detectors, or X-ray detectors, relying on probe scanning or test block comparison to achieve defect identification. This type of inspection method requires operators to manually control the position of the inspection equipment and inspect the weld surface segment by segment. The presence or absence of welding defects is determined by changes in reflected waveforms or magnetic traces. The inspection process is greatly affected by human experience, resulting in limited inspection efficiency and difficulty in achieving continuous and high-precision inspection.

[0003] Current technologies rely on manual or semi-automated methods for weld inspection. Operators need to manually control the equipment to scan the weld area segment by segment. This process not only requires a high level of operator experience but is also prone to omissions and discontinuous inspections due to the limited inspection range, affecting the comprehensiveness and accuracy of the inspection results. Traditional inspection methods are limited by equipment and technical capabilities, resulting in low inspection efficiency and the inability to provide real-time feedback on defect expansion trends. This leads to insufficient timeliness and accuracy in defect diagnosis, especially in the inspection of complex weld areas or large-scale structures, where traditional methods struggle to achieve efficient and accurate defect identification, posing significant operational risks. Summary of the Invention

[0004] To address the technical problems existing in the prior art, this invention provides a laser inspection system for the welding quality of steel box girders used in municipal ramp bridges. The technical solution is as follows:

[0005] On the one hand, a laser inspection system for the welding quality of steel box girders of municipal ramp bridges is provided, the system comprising:

[0006] The signal analysis module analyzes the light intensity distribution of the welding area within a continuous scanning area based on the reflected signal of the weld surface collected by the laser detection equipment, compares the changes in light intensity gradient between adjacent scanning segments, and judges the boundary consistency of the defect area to obtain a welding feature distribution dataset.

[0007] Based on the welding feature distribution dataset, the defect identification module filters the welding key point data at the corresponding time, compares the corresponding positional relationship of the key points in space, combines the ratio of the length to the width of the defect area, and obtains the numerical index of welding quality characteristics based on the number and distribution density of key points.

[0008] Based on the numerical index of the welding quality characteristics, the morphology analysis module optimizes the selection of target areas, analyzes the changes in welding morphology and the fluctuation of defect edges in five consecutive scans, filters out abnormal deviations in morphology fitting, and obtains the results of welding morphology change deviation analysis.

[0009] Based on the analysis results of the welding morphology change deviation, the risk assessment module filters out time periods in the predicted trajectory that exceed the current welding stress range, compares the difference between the welding stress change and the defect area expansion trend, and obtains a list of welding risk assessment and stress matching parameters.

[0010] As a further aspect of the present invention, the welding feature distribution dataset includes regional light intensity distribution characteristics, gradient change characteristics, and boundary continuity index; the numerical index of welding quality characteristics includes symmetry error value, defect morphology ratio, and key point density; the welding morphology change deviation analysis results include edge displacement amplitude, morphology fluctuation frequency, and fitting anomaly point distribution; and the welding risk assessment and stress matching parameter list includes welding stress overlap coefficient, risk period range, and expansion trend deviation value.

[0011] As a further aspect of the present invention, the signal parsing module includes:

[0012] The light intensity analysis submodule analyzes the light intensity distribution characteristics of the welding area within three consecutive scans based on the weld surface reflection signal collected by the laser detection equipment, compares the spatial distribution characteristics of the light intensity curve within each scan, determines whether the sampling segment is continuous and representative, and generates a periodic sampling light intensity sequence.

[0013] The gradient analysis submodule calculates the rate of change between the end and the beginning of the light intensity in the light intensity sequence between adjacent scans based on the periodically sampled light intensity sequence, compares the difference in the magnitude of the rate of change, identifies the data combination of defect features, and obtains the light intensity gradient interval.

[0014] The boundary consistency determination submodule calls the light intensity gradient interval to determine whether the continuity of the defect region boundary between adjacent scans is consistent, and filters out the segments that meet the conditions of boundary consistency and defect amplitude, thereby obtaining the welding feature distribution dataset.

[0015] As a further aspect of the present invention, the defect identification module includes:

[0016] The data filtering submodule analyzes the welding key point data at the corresponding time based on the welding feature distribution dataset, filters the monitoring results of welding nodes within the same scan, determines whether the data segments have completeness and continuity, and arranges them in spatial order to obtain the welding key point distribution sequence.

[0017] The position comparison submodule calls the welding key point distribution sequence, compares the spatial positional relationship of each group of welding key points, analyzes the corresponding distance between the different nodes, determines whether the distance falls within the reference range, filters the nodes that meet the requirements, and obtains the welding key point distribution characteristics.

[0018] The defect area quantification submodule analyzes the ratio of the length to the width of the node defect area based on the distribution characteristics of the welding key points, calculates the morphological feature value of the defect area, determines the number of nodes corresponding to the defect area, integrates the welding network topology, and obtains numerical indicators of welding quality characteristics.

[0019] As a further aspect of the present invention, the morphological analysis module includes:

[0020] The target screening submodule filters target areas with associated characteristics based on the numerical index of welding quality features, organizes welding morphology and defect edge data within five consecutive scans of the target area, completes data collection according to the scanning order, and obtains the scanned target area dataset.

[0021] The morphological analysis submodule analyzes the correspondence between the changes in welding morphology and the fluctuations of defect edges in each scan based on the scan target area dataset, compares the synchronous changes of defect edges in each scan, and classifies the fluctuation states in different time periods to obtain a morphological correspondence distribution dataset.

[0022] The anomaly filtering submodule judges the scan data in the morphology correspondence distribution dataset, filters out scan anomalies whose fluctuation amplitude deviates in the morphology fitting analysis, and identifies their position and change status in the monitoring sequence to obtain the welding morphology change deviation analysis results.

[0023] As a further aspect of the present invention, the target area refers to receiving numerical indicators representing the quality of welding as input, and analyzing and screening the entire welding monitoring range based on the indicators to identify and delineate local areas with high correlation to key quality characteristics as the target area to be analyzed.

[0024] The scan anomaly points refer to the scan data points in the morphological correspondence distribution dataset. The fluctuation range of each data point is compared with the preset or normal fluctuation range obtained through fitting analysis. Data points whose fluctuation range deviates from the normal range are filtered out and defined as scan anomaly points.

[0025] As a further aspect of the present invention, the risk assessment module includes:

[0026] Based on the welding morphology change deviation analysis results, the interval judgment submodule determines the coverage of the offset characteristics in the current interval of the welding stress monitoring unit, filters overlapping welding stress change segments, integrates the corresponding time sequence according to the segment order, and obtains the interval coverage sequence.

[0027] The trajectory filtering submodule filters time periods in which the predicted trajectory continuously exceeds the current welding stress range based on the interval coverage sequence, analyzes the changes in welding stress and the expansion trend of the defect area within the segment, summarizes the characteristics of the difference segment, and obtains the trajectory offset segment.

[0028] The risk parameter generation submodule compares the difference between the welding stress change and the defect area expansion trend within the trajectory offset segment, determines the correspondence between the welding stress change and the defect expansion trend within the key segment, identifies the corresponding index, and integrates the segment fluctuation data to obtain a welding risk assessment and stress matching parameter list.

[0029] As a further aspect of the present invention, the system also includes a decision support module:

[0030] Based on the welding risk assessment and stress matching parameter list, the auxiliary decision-making module adjusts the diagnostic data output, analyzes the change range of the target data within the current scan, selects the target data with the optimal change range, executes the output, and collects the expansion range of the defect area within the output cycle to obtain a welding quality diagnostic feedback list.

[0031] As a further aspect of the present invention, the welding quality diagnostic feedback list includes diagnostic data adjustment range, regional expansion intensity, and output stability index.

[0032] As a further aspect of the present invention, the auxiliary decision-making module includes:

[0033] The diagnostic optimization submodule, based on the welding risk assessment and stress matching parameter list, determines the relationship between the current diagnostic data and risk characteristics, optimizes the diagnostic data adjustment configuration, and adjusts the available diagnostic data according to the scan data to obtain the range of diagnostic data change.

[0034] The target comparison submodule compares the change characteristics of the corresponding target data within the scan based on the change range of the diagnostic data, analyzes the response performance of each target data, and selects the target data with the best change range performance to obtain the target data response difference dataset.

[0035] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0036] By analyzing the light intensity distribution in the welding area and comparing the light intensity gradient changes in adjacent scanning segments, the accuracy of welding defect identification is effectively improved. Through the screening and analysis of welding feature datasets, welding symmetry can be more accurately determined. Combining the length and width ratio of the defect area optimizes the accuracy of defect identification. Simultaneously, the calculation of numerical indicators based on welding quality characteristics makes welding quality assessment more detailed and quantifiable, forming stable and reliable welding quality diagnostic feedback, reducing reliance on human experience, and improving detection efficiency and accuracy. Analysis of welding morphology changes and defect edge fluctuations allows for timely identification of abnormal trends in welding defects, screening out risk points overlapping with the welding stress monitoring area, and further optimizing the accuracy of risk assessment. Combining the differences between stress changes and defect expansion trends forms precise welding risk assessment parameters, improving the ability to predict potential risks and effectively preventing the deterioration of welding quality problems. Furthermore, the optimization of diagnostic data output and real-time feedback cycles enables welding quality monitoring to react quickly based on real-time data, improving the dynamic adaptability and accuracy of the entire welding quality management system. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is a schematic diagram of a laser inspection system for the welding quality of steel box girders of municipal ramp bridges provided in an embodiment of the present invention;

[0039] Figure 2 This is a schematic diagram of the system framework of the present invention;

[0040] Figure 3 This is a flowchart of the signal analysis module in this invention;

[0041] Figure 4 This is a flowchart of the defect identification module in this invention;

[0042] Figure 5 This is a flowchart of the morphological analysis module in this invention;

[0043] Figure 6 This is a flowchart of the risk assessment module in this invention;

[0044] Figure 7 This is a flowchart of the decision support module in this invention. Detailed Implementation

[0045] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0046] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0047] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0048] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0049] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0050] This invention provides a laser inspection system for the welding quality of steel box girders in municipal ramp bridges, such as... Figure 1-2 The diagram shown illustrates a laser inspection system for the welding quality of steel box girders in a municipal ramp bridge. This system includes:

[0051] The signal analysis module analyzes the light intensity distribution of the welding area within a continuous scanning area based on the reflected signal of the weld surface collected by the laser detection equipment, compares the changes in light intensity gradient between adjacent scanning segments, and judges the boundary consistency of the defect area to obtain a welding feature distribution dataset.

[0052] The defect identification module is based on the welding feature distribution dataset, filters the welding key point data at the corresponding time, compares the corresponding positional relationship of the key points in space, determines whether it meets the welding symmetry standard, and obtains the numerical index of welding quality characteristics based on the ratio of the length and width of the defect area and the number and distribution density of key points.

[0053] The morphology analysis module optimizes the selection of target areas based on numerical indicators of welding quality characteristics, analyzes the changes in welding morphology and the fluctuation of defect edges in five consecutive scans, compares the change amplitude of defect edges in each scan, filters out abnormal deviations in morphology fitting, and obtains the results of welding morphology change deviation analysis.

[0054] Based on the analysis results of welding morphology change deviation, the risk assessment module determines the overlap with the current interval of the welding stress monitoring unit, filters the time periods in the predicted trajectory that exceed the current welding stress interval, compares the difference between welding stress change and defect area expansion trend, and obtains a list of welding risk assessment and stress matching parameters.

[0055] The auxiliary decision-making module adjusts the diagnostic data output based on the welding risk assessment and stress matching parameter list, analyzes the change range of the target data within the current scan, selects the target data with the optimal change range, executes the output, and collects the expansion range of the defect area within the output cycle to obtain a welding quality diagnostic feedback list.

[0056] The welding feature distribution dataset includes regional light intensity distribution characteristics, gradient change characteristics, and boundary continuity indicators. The numerical indicators of welding quality characteristics include symmetry error values, defect morphology proportions, and key point density. The welding morphology change deviation analysis results include edge displacement amplitude, morphology fluctuation frequency, and fitting outlier distribution. The welding risk assessment and stress matching parameter list includes welding stress overlap coefficient, risk period range, and expansion trend deviation value. The welding quality diagnosis feedback list includes diagnostic data adjustment range, regional expansion intensity, and output stability indicators.

[0057] Specifically, such as Figure 2 , 3 As shown, the signal analysis module includes:

[0058] The light intensity analysis submodule analyzes the light intensity distribution characteristics of the welding area within three consecutive scans based on the weld surface reflection signal collected by the laser detection equipment, compares the spatial distribution characteristics of the light intensity curve within each scan, determines whether the sampling segment is continuous and representative, and generates a periodic sampling light intensity sequence.

[0059] Obtain the data of the laser detection device, collect the reflected signals on the surface of the device weld, analyze the light intensity distribution characteristics in the welding area within three consecutive scans (Scan 1, 2, 3), each scan contains 1000 sampling points. Divide the light intensity sequence of Scan 1 into 10 sub-segments, calculate the average light intensity of each sub-segment, and obtain the average value sequence M1 = {125.3, 128.1, 130.5, 129.8, 127.6, 126.9, 131.2, 130.8, 129.5, 128.3}. Calculate the standard deviation SD1 = 2.05 of this sequence. Set the standard deviation threshold SD_th to 3.5. This threshold is the 95% quantile of the statistical distribution of the standard deviation of the light intensity sequence under 100 sets of standard processes. Since SD1 < SD_th, it is determined that Scan 1 has spatial representativeness. Similarly, analyze Scan 2 and 3, and obtain SD2 = 2.15 and SD3 = 1.99, both of which are less than the threshold, so it is determined that all three scans have internal representativeness. Then calculate the average value of the absolute value of the difference in the average light intensity of the corresponding sub-segments between adjacent scans as the difference degree D between scans. The difference degree D12 between Scan 1 and 2 is 1.85, and the difference degree D23 between Scan 2 and 3 is 1.91. Set the difference degree threshold D_th to 5.0. This threshold is determined based on the upper bound of the historical normal fluctuation range. Since both D12 and D23 are less than D_th, it is determined that the sampling segment has continuity and representativeness. Concatenate the light intensity sequences of the three scans to generate a periodic sampling light intensity sequence.

[0060] Based on the periodic sampling light intensity sequence, the gradient analysis sub-module calculates the change rate between the end and start light intensities in the light intensity sequence between adjacent scans, compares the amplitude differences in the rate changes, identifies the data combinations of defect features, and obtains the light intensity gradient interval.

[0061] Call the periodic sampling light intensity sequence, extract the last 5 light intensity values {128.9, 129.1, 128.5, 128.8, 129.0} of Scan 1 and the first 5 light intensity values {95.2, 94.8, 95.5, 96.1, 95.9} of Scan 2 from the sequence. Calculate the average light intensity at the end of Scan 1, I_end1 = 128.86, and the average light intensity at the start of Scan 2, I_start2 = 95.5. Calculate the light intensity change rate V12 = I_start2 - I_end1 = -33.36. Process Scan 2 and Scan 3 in the same way, and obtain I_end2 = 130.2 and I_start3 = 128.82, with a change rate V23 = -1.38. Compare the amplitude differences in the rate changes, ΔV = |V12| - |V23| = 31.98. Set the rate change amplitude difference threshold ΔV_th to 20.0. This value is determined by analyzing the differences in the light intensity change rates between scans of defective and non-defective samples. Since ΔV > ΔV_th, identify the data combination in the transition area between Scan 1 and Scan 2 as a defect feature, record the interval where the light intensity drops suddenly from 128.86 to 95.5, and obtain the light intensity gradient interval.

[0062] The boundary consistency determination submodule calls the light intensity gradient interval to determine whether the continuity of the defect region boundary is consistent between adjacent scans, and filters out the segments that meet the conditions of boundary consistency and defect amplitude, thereby obtaining the welding feature distribution dataset.

[0063] Based on the light intensity gradient interval, in the light intensity sequence of scan 1, the boundary of the defect region is marked as spatial coordinates [55.2, 58.3] mm by the light intensity decrease slope. Similarly, in scan 2, the defect boundary is marked as [55.5, 58.9] mm. Scan 3 has no defect features and does not form a defect region. The difference between the starting boundary Δ_start = 0.3 mm and the difference between the ending boundary Δ_end = 0.6 mm of the defect region in scans 1 and 2 are calculated. The boundary offset threshold Δ_th is set to 0.8 mm. This value is determined based on the single scan advance distance determined by the welding speed and scanning frequency, and combined with the upper limit of the 98% confidence interval of normal fluctuation statistical analysis. Since Δ_start and Δ_end are both less than Δ_th, it is determined that the defect region boundaries of scans 1 and 2 are continuous. Two segments of scans 1 and 2 that meet the boundary consistency and have defect amplitude are selected. The defect-related data within the segments are integrated to obtain the welding feature distribution dataset.

[0064] Specifically, such as Figure 2 , 4 As shown, the defect identification module includes:

[0065] The data filtering submodule analyzes the welding key point data at the corresponding time based on the welding feature distribution dataset, filters the monitoring results of welding nodes within the same scan, determines whether the data segments have completeness and continuity, and arranges them in spatial order to obtain the welding key point distribution sequence.

[0066] Based on the welding feature distribution dataset, which originates from arc voltage sensors and welding current sensors that work synchronously with the laser detection equipment, the dataset indicates defects in scans 1 and 2. The corresponding arc voltage of 19.5V and welding current of 150A are retrieved at this moment. The monitoring results of welding nodes within the same scan are filtered out. The nodes are preset monitoring points along the weld direction. Five monitoring nodes are set within a 5mm length covered by scan 1. The completeness of the data of the five nodes is checked by comparing the timestamp of each node data record with the start and end timestamp of scan 1. It is confirmed that all node data are within the time range and there are no missing values. The data segment is judged to have completeness and continuity. Then, the node data is arranged in spatial order. The spatial coordinates of nodes 1 to 5 are (10.1, 2.3), (11.2, 2.4), (12.0, 2.2), (13.1, 2.5), and (14.0, 2.4), respectively, in mm. The monitoring data of the weld pool width, weld depth, etc. of the nodes are organized in this spatial order to obtain the welding key point distribution sequence.

[0067] The position comparison submodule calls the welding key point distribution sequence, compares the spatial positional relationship of each group of welding key points, analyzes the corresponding distance between the different nodes, determines whether the distance falls within the reference range, filters the nodes that meet the requirements, and obtains the distribution characteristics of welding key points.

[0068] The welding key point distribution sequence is called up, and the spatial positional relationship of the welding key points is compared. For example, the spatial positions of node 2 (11.2, 2.4) and node 3 (12.0, 2.2) are compared. The corresponding distance between the different nodes is analyzed, and the Euclidean distance D23 between the two nodes is calculated to be 0.82 mm. It is then determined whether this distance falls within the reference range [0.8, 1.2] mm. This range is set based on the standard node spacing determined by the welding parameters. Its upper and lower limits are obtained by statistical analysis of the node spacing continuously monitored in normal welding experiments, taking the mean ± 3 times the standard deviation. Since D23 = 0.82 mm falls within the reference range, it indicates that the positional relationship between node 2 and node 3 is normal. The same distance calculation and judgment are performed on all adjacent node pairs in the sequence. All nodes whose distances meet the reference range requirements are screened, and the nodes are integrated to obtain the welding key point distribution characteristics.

[0069] The defect area quantification submodule analyzes the ratio of the length to the width of the node defect area based on the distribution characteristics of welding key points, calculates the morphological feature value of the defect area, determines the number of nodes corresponding to the defect area, integrates the welding network topology, and obtains numerical indicators of welding quality characteristics.

[0070] The morphological characteristic value of the defect area is calculated using the following formula:

[0071] ;

[0072] in, Represents the morphological characteristics of the defect area. Represents the length of the defective region. N represents the width of the defect area, N represents the number of nodes corresponding to the defect area, and A represents the area of ​​the welded area.

[0073] Based on the distribution characteristics of key welding points, defect areas are identified by recognizing consecutive sets of nodes with abnormal parameters within these distribution characteristics. For example, if the penetration depth monitoring values ​​of nodes 2, 3, and 4 are continuously below 80% of the normal threshold, then these three nodes are identified as defect areas, and the length of this defect area is determined. It is determined by calculating the straight-line distance between the first and last two nodes (node ​​2 and node 4). = ≈1.90mm, width of the defect area The estimate is made by calculating the average width of the molten pool at these three nodes. If the measured widths are 1.8mm, 1.7mm, and 1.8mm respectively, then... =(1.8+1.7+1.8) / 3≈1.77mm, determine the number of nodes corresponding to the defect area. In this example, N=3. Calculate the morphological characteristic value of the defect area. Formulas are used to quantify the severity and morphological irregularity of defects. Calculations are performed, in which, This represents the length of the defective area, which is 1.90 mm in this example. The width of the defect area is 1.77mm in this example. N represents the number of nodes corresponding to the defect area, which is 3 in this example. A represents the area of ​​the welding area, which is the nominal welding area covered by a single scan. It is set according to the welding parameters. For example, if the laser spot diameter is 2mm and the scanning line length is 5mm, then A = 2mm × 5mm = 10mm².

[0074] The innovation of the formula lies in incorporating the geometric dimensions of the defect ( This is combined with the number of discrete monitoring points (N) within the defect area and the macroscopic scale (A) of the welding process, through... The study assessed the elongation of the defects, a shape that is more prone to stress concentration than circular defects. This method correlates the number of microscopic nodes with the macroscopic welding area, so that the eigenvalue G not only reflects the absolute size of the defect, but also its relative significance in the entire welding area. It comprehensively considers the geometric shape of the defect, the discrete sampling density, and the process scale, thereby more comprehensively assessing the potential risk of the defect.

[0075] Substitute the numerical values ​​into the formula to calculate: ;

[0076] A baseline value G_th = 5.0 was set. This baseline value was determined through finite element simulation analysis of a large number of samples. It was found that when the G value exceeds 5.0, the stress concentration factor will exceed the allowable range of the material. The calculation result G≈5.877>5.0 indicates that the morphological risk of this defect area is high. Subsequently, the welding network topology relationship, including information such as G value, node position, arc voltage, and welding current, was integrated to obtain a numerical index of welding quality characteristics.

[0077] Specifically, such as Figure 2 , 5 As shown, the morphological analysis module includes:

[0078] The target screening submodule filters target areas with associated characteristics based on numerical indicators of welding quality features, organizes welding morphology and defect edge data within five consecutive scans of the target area, completes data collection according to the scanning order, and obtains the scanned target area dataset.

[0079] The target area refers to the area that receives numerical indicators representing the quality of welding as input, and analyzes and filters them in the entire welding monitoring range based on the indicators. The local areas with high correlation to key quality characteristics are identified and delineated as the target areas to be analyzed.

[0080] Based on the numerical indicators of welding quality characteristics, a value of G=5.877 was received, exceeding the risk benchmark of 5.0, triggering a screening procedure. G=5.877 and its associated defect area location (determined by the spatial coordinates of nodes 2, 3, and 4) were used as input for analysis and screening across the entire welding monitoring range. The local area containing the high G value, i.e., the rectangular area surrounding nodes 2, 3, and 4, was identified and designated as the target area. Then, the welding morphology and defect edge data of this target area from five consecutive scans (scans 0 to 4) before and after the defect occurred were compiled. The welding morphology data includes the width, length, and depth of the molten pool, while the defect edge data refers to the set of defect contour coordinate points extracted through image processing. Data was collected in the scanning order, and the data from scans 0 to 4 were sequentially stored in the data structure. For example, the dataset for scan 1 includes its molten pool dimensions {width: 1.8mm, length: 4.2mm, depth: 0.8mm} and the set of defect edge coordinate points, thus obtaining the target area dataset.

[0081] The morphology analysis submodule analyzes the correspondence between the changes in welding morphology and the fluctuations of defect edges in each scan based on the scan target area dataset, compares the synchronous changes of defect edges in each scan, and classifies the fluctuation states in different time periods to obtain a morphology correspondence distribution dataset.

[0082] Based on the scanned target area dataset, the welding morphology data {width: 1.8mm, depth: 0.8mm} of scan 1 is extracted. The changes relative to the previous normal scan (scan 0, {width: 3.0mm, depth: 1.5mm}) are calculated: width change -1.2mm, depth change -0.7mm. Simultaneously, the defect edge data of scan 1 is analyzed, and the average deviation of its edge contour from the theoretical weld centerline is calculated to be 0.6mm. A correspondence between morphology changes (width -1.2mm, depth -0.7mm) and edge fluctuations (deviation 0.6mm) is established. This process is repeated for scans 2, 3, and 4 to obtain a... The sequence of morphological changes and edge fluctuations from five scans was analyzed, and then the synchronous changes of the defect edge in each scan were compared. Defects were detected in both scan 1 and scan 2. The average deviation distance of the defect edge from scan 1 to scan 2 was calculated to be 0.1 mm. At the same time, the molten pool width changed from 1.8 mm to 1.7 mm, a change of -0.1 mm. It was found that the expansion of the defect edge and the contraction of the molten pool width were synchronous. The fluctuation state of the difference period was classified as follows: the transition from normal to defect was marked as "mutation type", the period of defect persistence was marked as "stable expansion type", and the period after the defect disappeared was marked as "recovery type", thus obtaining the morphological correspondence distribution dataset.

[0083] The anomaly point screening submodule judges the scan data in the morphology correspondence distribution dataset, filters the scan anomalies whose fluctuation amplitude deviates in the morphology fitting analysis, and marks their position and change status in the monitoring sequence to obtain the welding morphology change deviation analysis results.

[0084] Scanning outliers refers to examining scanned data points in a morphological correspondence distribution dataset, comparing the fluctuation range of each data point with a preset or normal fluctuation range obtained through fitting analysis, filtering out data points whose fluctuation range deviates from the normal range, and defining them as scanning outliers.

[0085] The scan data in the morphological correspondence distribution dataset are judged, and the weld pool width of the "recovery" period (scans 3 and 4) is linearly fitted to obtain the expected recovery trend line. For example, the width recovery formula is W(t) = 2.2 + 0.3t (t is the scan number). Based on this, the width of scan 5 is predicted to be 3.7 mm. Then, the fluctuation range of each data point is compared with the normal fluctuation range [3.33, 4.07] mm obtained by fitting analysis (defined as the fitted value fluctuating by 10%). If the actual monitored weld pool width of scan 5 is 3.1 mm, and its fluctuation range deviates from the normal range, the data point of scan 5 is screened out and defined as a scan anomaly point. Its position in the monitoring sequence is marked as "scan number 5", and its change status is recorded as "insufficient width recovery, actual value 3.1 mm is lower than the lower limit of the predicted range 3.33 mm". The result of the welding morphological change deviation analysis is obtained.

[0086] Specifically, such as Figure 2 , 6 As shown, the risk assessment module includes:

[0087] Based on the analysis results of welding morphology change deviation, the interval judgment submodule determines the coverage of the offset characteristics in the current interval of the welding stress monitoring unit, filters overlapping welding stress change segments, integrates the corresponding time sequence according to the segment order, and obtains the interval coverage sequence.

[0088] Based on the analysis results of welding morphology change deviation, the results indicate that the width recovery is insufficient at the position of scan number 5. The stress range [150MPa, 180MPa] at the time of scan 5 of the welding stress monitoring unit is retrieved to determine the coverage of the offset characteristic in the current range. The offset characteristic of "insufficient width recovery" is associated with an additional stress of 25MPa. The predicted stress value is the median of the current range of 165MPa plus the additional stress of 25MPa, which equals 190MPa. This predicted value exceeds the upper limit of the current stress range of 180MPa, indicating that the offset characteristic has caused the stress to exceed the limit. Welding stress change segments that overlap due to this offset characteristic are screened, that is, the area where the actual stress reaches 190MPa overlaps with the next higher stress level range [180MPa, 210MPa]. The time period starting from scan 5 is recorded, and the corresponding time series data starting from scan 5 are integrated according to the segment order to obtain the range coverage sequence.

[0089] The trajectory filtering submodule filters time periods in which the predicted trajectory continuously exceeds the current welding stress range based on the interval coverage sequence, analyzes the changes in welding stress and the expansion trend of the defect area within the segment, summarizes the characteristics of the difference segment, and obtains the trajectory offset segment.

[0090] Based on the interval coverage sequence, this sequence indicates that the welding stress starting from scan 5 is at risk of exceeding the existing range. Tracking the stress prediction trajectory of subsequent scans (6, 7, 8), the predicted stress values ​​are 192MPa, 195MPa, and 198MPa, respectively. The stress values ​​continuously exceed the upper limit of the current welding stress range [150MPa, 180MPa]. The time period from scan 5 to scan 8 is selected, and the stress change (from 190MPa to 198MPa) and the defect area expansion trend in this segment are analyzed. By retrieving the synchronous laser detection signal, it is found that the density of microcracks on the weld surface increased by 15% in this segment. The change characteristics of the difference segment are summarized as "the stress continues to rise, accompanied by an increase in the density of microcracks". The data set from scan 5 to scan 8 is used to obtain the trajectory offset segment.

[0091] The risk parameter generation submodule compares the difference between the welding stress change and the defect area expansion trend within the trajectory offset segment, determines the correspondence between the welding stress change and the defect expansion trend within the key segment, identifies the corresponding index, and integrates the segment fluctuation data to obtain a list of welding risk assessment and stress matching parameters.

[0092] By comparing the differences between welding stress changes and defect expansion trends within the trajectory offset segments, in the trajectory offset segments from scan 5 to 8, the welding stress change rate was 2.67 MPa / scan, and the microcrack density growth rate was 5% / scan. Comparing these two rates, the correspondence between welding stress changes and defect expansion trends in the critical segments was determined. By querying a preset "stress-damage" correlation matrix (established based on material fatigue experimental data), the cells corresponding to the input stress change rate of 2.67 MPa / scan and damage expansion rate of 5% / scan were found, resulting in a correlation index CI=0.85. This index ranges from 0 to 1, with a value closer to 1 indicating that stress growth is the driving factor for defect expansion. After identifying this index, the fluctuation data within the segments, such as the maximum stress value of 198 MPa, the minimum stress value of 190 MPa, and the final value of microcrack density, were integrated to obtain a list of welding risk assessment and stress matching parameters.

[0093] Specifically, such as Figure 2 , 7 As shown, the decision support module includes:

[0094] The diagnostic optimization submodule, based on the welding risk assessment and stress matching parameter list, determines the relationship between the current diagnostic data and risk characteristics, optimizes the diagnostic data adjustment configuration, and adjusts the available diagnostic data according to the scan data to obtain the magnitude of the diagnostic data change.

[0095] Based on the welding risk assessment and stress matching parameter list, which includes parameters {correspondence index: 0.85, maximum stress: 198MPa, microcrack density: 15%}, the relationship between the current diagnostic data and risk characteristics was determined. High correspondence index (0.85) and high stress (198MPa) were identified as risk characteristics, indicating that welding parameters need to be adjusted to reduce welding heat input, thereby reducing residual stress. The diagnostic data adjustment configuration was optimized. Based on the rule base recommendation, the welding current was reduced and the welding speed was increased. The available diagnostic data was adjusted according to the scan data. It was decided to reduce the welding current by 2% to 147A and increase the welding speed by 2% to 5.1mm / s, and the change range of the diagnostic data was obtained.

[0096] The target comparison submodule compares the change characteristics of the corresponding target data within the scan based on the change range of the diagnostic data, analyzes the response performance of each target data, and selects the target data with the best change range performance to obtain the target data response difference dataset.

[0097] Based on the variation range of the diagnostic data {current change: -3A, velocity change: +0.1mm / s}, the variation characteristics of the corresponding target data in subsequent scans 9 were compared. The target data included weld width, reinforcement height, stress, and microcrack density. The response performance of each target data was analyzed. In the adjusted scan 9, the weld width was monitored to decrease from 3.1mm to 3.0mm, still within the acceptable range [2.8mm, 3.5mm]. The stress monitoring value decreased from 198MPa to 185MPa, and the microcrack density growth rate decreased to 1% / scan. The target data with the best variation range was selected. The adjustment effect was quantitatively evaluated using an evaluation function J=w1×Δσ+w2×Δρ+w3×ΔW containing weights w1, w2, and w3. The calculated comprehensive score of the current adjustment was positive and better than the simulated score of other alternative schemes. Therefore, this group {current change: -3A, velocity change: +0.1mm / s} was selected as the optimal target data, and the target data response difference dataset was obtained.

[0098] The output response acquisition submodule performs optimal target data output based on the target data response difference dataset, acquires the changes in the defect area within the output cycle, analyzes the defect expansion range, and obtains a welding quality diagnosis feedback list.

[0099] Based on the target data response difference dataset, the optimal adjustment strategy was identified as {current change: -3A, velocity change: +0.1mm / s}. This optimal target data output was executed, and the welding parameters were adjusted and applied to subsequent welding starting from scan 10. Within one cycle after the output (10 consecutive scans), the changes in the defect area were continuously collected. Using laser detection equipment and stress monitoring unit, the weld surface from scan 10 to scan 19 was analyzed. No new microcracks were found to be generated, and the original microcrack areas did not expand. The stress value was stable within the range of 180MPa±5MPa. The defect expansion range was analyzed, and it was confirmed that the defect expansion had been suppressed, with the expansion range being zero. The diagnosis, adjustment, and verification process, as well as the final quality status (stress stability, no defect expansion), were recorded to obtain the welding quality diagnosis feedback list.

[0100] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A laser inspection system for the welding quality of steel box girders in municipal ramp bridges, characterized in that, The system includes: The signal analysis module analyzes the light intensity distribution of the welding area within a continuous scanning area based on the reflected signal of the weld surface collected by the laser detection equipment, compares the changes in light intensity gradient between adjacent scanning segments, and judges the boundary consistency of the defect area to obtain a welding feature distribution dataset. Based on the welding feature distribution dataset, the defect identification module filters the welding key point data at the corresponding time, compares the corresponding positional relationship of the key points in space, combines the ratio of the length to the width of the defect area, and obtains the numerical index of welding quality characteristics based on the number and distribution density of key points. Based on the numerical index of the welding quality characteristics, the morphology analysis module optimizes the selection of target areas, analyzes the changes in welding morphology and the fluctuation of defect edges in five consecutive scans, filters out abnormal deviations in morphology fitting, and obtains the results of welding morphology change deviation analysis. Based on the analysis results of the welding morphology change deviation, the risk assessment module filters out time periods in the predicted trajectory that exceed the current welding stress range, compares the difference between the welding stress change and the defect area expansion trend, and obtains a list of welding risk assessment and stress matching parameters.

2. The laser inspection system for welding quality of steel box girders for municipal ramp bridges according to claim 1, characterized in that: The welding feature distribution dataset includes regional light intensity distribution characteristics, gradient change characteristics, and boundary continuity indicators. The numerical indicators of welding quality characteristics include symmetry error value, defect morphology ratio, and key point density. The welding morphology change deviation analysis results include edge displacement amplitude, morphology fluctuation frequency, and fitting anomaly point distribution. The welding risk assessment and stress matching parameter list includes welding stress overlap coefficient, risk period range, and expansion trend deviation value.

3. The laser inspection system for welding quality of steel box girders for municipal ramp bridges according to claim 1, characterized in that: The signal analysis module includes: The light intensity analysis submodule analyzes the light intensity distribution characteristics of the welding area within three consecutive scans based on the weld surface reflection signal collected by the laser detection equipment, compares the spatial distribution characteristics of the light intensity curve within each scan, determines whether the sampling segment is continuous and representative, and generates a periodic sampling light intensity sequence. The gradient analysis submodule calculates the rate of change between the end and the beginning of the light intensity in the light intensity sequence between adjacent scans based on the periodically sampled light intensity sequence, compares the difference in the magnitude of the rate of change, identifies the data combination of defect features, and obtains the light intensity gradient interval. The boundary consistency determination submodule calls the light intensity gradient interval to determine whether the continuity of the defect region boundary between adjacent scans is consistent, and filters out the segments that meet the conditions of boundary consistency and defect amplitude, thereby obtaining the welding feature distribution dataset.

4. The laser inspection system for welding quality of steel box girders for municipal ramp bridges according to claim 3, characterized in that: The defect identification module includes: The data filtering submodule analyzes the welding key point data at the corresponding time based on the welding feature distribution dataset, filters the monitoring results of welding nodes within the same scan, determines whether the data segments have completeness and continuity, and arranges them in spatial order to obtain the welding key point distribution sequence. The position comparison submodule calls the welding key point distribution sequence, compares the spatial positional relationship of each group of welding key points, analyzes the corresponding distance between the different nodes, determines whether the distance falls within the reference range, filters the nodes that meet the requirements, and obtains the welding key point distribution characteristics. The defect area quantification submodule analyzes the ratio of the length to the width of the node defect area based on the distribution characteristics of the welding key points, calculates the morphological feature value of the defect area, determines the number of nodes corresponding to the defect area, integrates the welding network topology, and obtains numerical indicators of welding quality characteristics.

5. The laser inspection system for welding quality of steel box girders for municipal ramp bridges according to claim 4, characterized in that: The morphological analysis module includes: The target screening submodule filters target areas with associated characteristics based on the numerical index of welding quality features, organizes welding morphology and defect edge data within five consecutive scans of the target area, completes data collection according to the scanning order, and obtains the scanned target area dataset. The morphological analysis submodule analyzes the correspondence between the changes in welding morphology and the fluctuations of defect edges in each scan based on the scan target area dataset, compares the synchronous changes of defect edges in each scan, and classifies the fluctuation states in different time periods to obtain a morphological correspondence distribution dataset. The anomaly filtering submodule judges the scan data in the morphology correspondence distribution dataset, filters out scan anomalies whose fluctuation amplitude deviates in the morphology fitting analysis, and identifies their position and change status in the monitoring sequence to obtain the welding morphology change deviation analysis results.

6. The laser inspection system for welding quality of steel box girders for municipal ramp bridges according to claim 5, characterized in that: The target area refers to receiving numerical indicators representing the quality of welding as input, and analyzing and screening them in the entire welding monitoring range based on the indicators to identify and delineate local areas with high correlation to key quality characteristics as the target area to be analyzed. The scan anomaly points refer to the scan data points in the morphological correspondence distribution dataset. The fluctuation range of each data point is compared with the preset or normal fluctuation range obtained through fitting analysis. Data points whose fluctuation range deviates from the normal range are filtered out and defined as scan anomaly points.

7. The laser inspection system for welding quality of steel box girders for municipal ramp bridges according to claim 5, characterized in that: The risk assessment module includes: Based on the welding morphology change deviation analysis results, the interval judgment submodule determines the coverage of the offset characteristics in the current interval of the welding stress monitoring unit, filters overlapping welding stress change segments, integrates the corresponding time sequence according to the segment order, and obtains the interval coverage sequence. The trajectory filtering submodule filters time periods in which the predicted trajectory continuously exceeds the current welding stress range based on the interval coverage sequence, analyzes the changes in welding stress and the expansion trend of the defect area within the segment, summarizes the characteristics of the difference segment, and obtains the trajectory offset segment. The risk parameter generation submodule compares the difference between the welding stress change and the defect area expansion trend within the trajectory offset segment, determines the correspondence between the welding stress change and the defect expansion trend within the key segment, identifies the corresponding index, and integrates the segment fluctuation data to obtain a welding risk assessment and stress matching parameter list.

8. The laser inspection system for welding quality of steel box girders for municipal ramp bridges according to claim 1, characterized in that: The system also includes a decision support module: Based on the welding risk assessment and stress matching parameter list, the auxiliary decision-making module adjusts the diagnostic data output, analyzes the change range of the target data within the current scan, selects the target data with the optimal change range, executes the output, and collects the expansion range of the defect area within the output cycle to obtain a welding quality diagnostic feedback list.

9. The laser inspection system for welding quality of steel box girders for municipal ramp bridges according to claim 8, characterized in that: The welding quality diagnostic feedback list includes the diagnostic data adjustment range, the intensity of regional expansion, and the output stability index.

10. The laser inspection system for welding quality of steel box girders for municipal ramp bridges according to claim 8, characterized in that: The decision support module includes: The diagnostic optimization submodule, based on the welding risk assessment and stress matching parameter list, determines the relationship between the current diagnostic data and risk characteristics, optimizes the diagnostic data adjustment configuration, and adjusts the available diagnostic data according to the scan data to obtain the range of diagnostic data change. The target comparison submodule compares the change characteristics of the corresponding target data within the scan based on the change range of the diagnostic data, analyzes the response performance of each target data, and filters the target data with the best change range performance to obtain the target data response difference dataset. The output response acquisition submodule performs optimal target data output based on the target data response difference dataset, acquires the changes in the defect area within the output cycle, analyzes the defect expansion range, and obtains a welding quality diagnosis feedback list.

Images