Image processing-based quality monitoring method and device

Abstract

The application discloses a quality monitoring method and device based on image processing, and realizes rapid acquisition of workpiece data on a pipeline, is low in equipment cost, high in efficiency, applicable to batch data acquisition and processing, applicable to deformation caused by free deformation or constraint state, analyzes by respectively calculating the influence of temperature and deformation and highlighting stress concentration points, and avoids that stress analysis results cannot reflect actual conditions due to inconsistent change trends when temperature and deformation are simultaneously considered in a common model, so that calculation amount is reduced, and the accuracy of analysis and early warning is improved. The application can be widely applied to the field of engineering quality monitoring as a quality monitoring method and system device based on image processing.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of engineering quality monitoring, and in particular to a quality monitoring method and device based on image processing. BACKGROUND

[0002] In engineering quality monitoring, the residual stress of the construction workpiece needs to be concerned, and fracture failure analysis is usually performed. Residual stress is the action and influence that remains in the component after the action and influence of various process factors and the like during the manufacturing process of the component, and the action and influence that remains in the component after the action and influence of various process factors and the like during the manufacturing process of the component.

[0003] Taking a welded workpiece as an example, the internal stress remaining in the welded structure due to the constraint of deformation during the welding heat process, among which the thermal stress generated by the constraint of the cooling shrinkage after the weld metal is melted and solidified is the most significant, is the main part of the residual stress. The internal stress generated when the internal metallographic structure changes during the welding cooling process is a secondary part of the residual stress. In the prior art, the residual stress is usually eliminated by post-weld heat treatment, overload treatment, vibration treatment and the like, or the generation of the residual stress is reduced by welding process, temperature control and the like during the welding process. In the prior art, the residual stress is usually calculated by finite element analysis, but the existing calculation model only establishes a model for temperature, deformation and the like, and it is difficult to meet the requirements when the quality monitoring precision requirement is high. For example, CN114528733A is a method for controlling residual stress of a multi-point distributed heat source for welding of a steel bridge deck, which is used to adjust the residual stress at different positions of the steel bridge deck weld to reduce the generation of residual stress during the welding process. However, this method is used for continuous control during the welding process, and it is difficult to apply to rapid quality monitoring of batch workpieces, and the final state of the workpiece after welding is not monitored and pressure tested. SUMMARY

[0004] In order to solve the above technical problems, the purpose of the present application is to provide a quality monitoring data acquisition method and system based on image processing, which has less calculation amount and improves the analysis and early warning accuracy.

[0005] One of the technical solutions adopted by the present application is a quality monitoring method based on image processing, comprising the following steps:

[0006] S1, acquiring first three-dimensional scanning data of a to-be-tested object at a first time and second three-dimensional scanning data of the to-be-tested object at a second time, and establishing a first three-dimensional finite element analysis model based on the second three-dimensional scanning data;

[0007] S2, acquire infrared image data of the to-be-tested object at at least two time points within a first time to a second time, obtain infrared image data at multiple time points by interpolation based on the infrared image data at the at least two time points, and input the infrared image data at the multiple time points into a three-dimensional finite element analysis model to obtain a first stress field feature;

[0008] S3, input the first three-dimensional scanning data into the first three-dimensional finite element analysis model to obtain a second stress field feature;

[0009] S4, obtain third stress field feature data according to the first stress field feature and the second stress field feature;

[0010] S5, monitor according to the third stress field feature, and issue a warning when stress greater than a set stress threshold exists.

[0011] Another technical solution adopted by the application is: a quality monitoring device based on image processing, comprising:

[0012] An image data acquisition device is configured to acquire first three-dimensional scanning data of a to-be-tested object at a first time and second three-dimensional scanning data of the to-be-tested object at a second time.

[0013] An infrared data acquisition device is configured to acquire infrared image data of the to-be-tested object at at least two time points within a first time to a second time.

[0014] A data generation device is configured to

[0015] A first three-dimensional finite element analysis model is established based on the second three-dimensional scanning data.

[0016] Infrared image data at multiple time points is obtained by interpolation based on the infrared image data at the at least two time points, and the infrared image data at the multiple time points is input into a three-dimensional finite element analysis model to obtain a first stress field feature.

[0017] The first three-dimensional scanning data is input into the first three-dimensional finite element analysis model to obtain a second stress field feature.

[0018] Third stress field feature data is obtained according to the first stress field feature and the second stress field feature.

[0019] A monitoring and warning device is configured to monitor according to the third stress field feature, and issue a warning when stress greater than a set stress threshold exists.

[0020] The beneficial effects of the present application are: through the method and system, the rapid acquisition of workpiece data on the pipeline is realized, the device cost is low, the efficiency is high, it is suitable for batch data acquisition and processing, whether it is free deformation or deformation caused by constraint state, it is applicable, the influence of temperature and deformation is calculated respectively, and the stress concentration point is highlighted for analysis, and the stress analysis result caused by inconsistent change trend when considering temperature and deformation at the same time in the conventional model cannot reflect the actual situation, the calculation amount is reduced, and the accuracy of analysis and early warning is improved. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 The step flow chart of the first specific embodiment of the present application is shown in the figure.

[0022] Figure 2 The step flow chart of the second specific embodiment of the present application is shown in the figure.

[0023] Figure 3 The residual stress schematic diagram of the welding workpiece of the specific embodiment of the present application is shown in the figure.

[0024] Figure 4 The mesh schematic diagram before segmentation and merging of the specific embodiment of the present application is shown in the figure.

[0025] Figure 5 The mesh schematic diagram after segmentation and merging of the specific embodiment of the present application is shown in the figure.

[0026] Figure 6 The schematic diagram of one segmentation mode of the merged mesh of the specific embodiment of the present application is shown in the figure.

[0027] Figure 7 The schematic diagram of another segmentation mode of the merged mesh of the specific embodiment of the present application is shown in the figure.

[0028] Figure 8 The schematic diagram of the first proportional segmentation mode of the mesh unit of the specific embodiment of the present application is shown in the figure.

[0029] Figure 9 The schematic diagram of the second proportional segmentation mode of the mesh unit of the specific embodiment of the present application is shown in the figure.

[0030] Figure 10 The schematic diagram of the third proportional segmentation mode of the mesh unit of the specific embodiment of the present application is shown in the figure.

[0031] Figure 11 The temperature fitting curve of the specific embodiment of the present application is shown in the figure.

[0032] Figure 12 The temperature fitting curve of the adjacent mesh node in the specific embodiment of the present application is shown in the figure.

[0033] Figure 13 The architecture schematic diagram of the quality monitoring device of the present application is shown in the figure. DETAILED DESCRIPTION

[0034] In the prior art, the residual stress at different positions of the steel bridge deck plate weld is adjusted to reduce the generation of residual stress during welding. However, the existing method continuously regulates the welding process, and only one workpiece can be operated in one welding process, which is difficult to apply to batch workpiece rapid data acquisition and monitoring. In addition, the internal metallographic structure changes or cracks during the cooling process after welding, and the final stress distribution is different from the model prediction, so the quality of the welded workpiece must be monitored and analyzed.

[0035] In the prior art, stress data is usually obtained through finite element analysis, and the effects of elastic deformation and temperature distribution characteristics are considered during the analysis process. However, in practice, it is found that the analysis result of the workpiece welding residual stress is not sensitive to the monitoring of workpiece defects, especially when there are internal empty welds in the weld and the surface is complete. Only professional equipment such as X-ray flaw detector can detect weld defects, so there is a great safety hazard in quality monitoring using finite element analysis. The reason is that when there are empty welds or cracks in the weld, the temperature affects the change trend of stress at the empty weld or crack, and the deformation affects the change trend of stress at the empty weld or crack, that is, the effect of stress on temperature conduction at the empty weld or crack is inconsistent with the effect of stress on deformation stress conduction. Therefore, the stress data obtained by finite element analysis cannot reflect the actual situation. The empty weld or crack may be caused by incomplete welding or may be caused by shrinkage during cooling after welding.

[0036] To make the purpose, technical scheme and advantages of the present application clearer, the technical scheme of the present application will be described clearly and completely through embodiments with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0037] Reference Figure 1 The first specific embodiment of the present application provides a quality monitoring method based on image processing, comprising the following steps:

[0038] S1, obtaining first three-dimensional scanning data of a to-be-tested object at a first time and second three-dimensional scanning data of the to-be-tested object at a second time, and establishing a first three-dimensional finite element analysis model based on the second three-dimensional scanning data;

[0039] The first-time object-to-be-tested first three-dimensional scanning data specifically refers to three-dimensional scanning data of the object-to-be-tested (i.e., the welding workpiece) when the welding is completed; the second-time object-to-be-tested second three-dimensional scanning data specifically refers to three-dimensional scanning data of the object-to-be-tested after cooling and setting, and the basic model for calculation is established based on the three-dimensional scanning data after cooling and setting. The three-dimensional scanning data of the welding workpiece after cooling and setting may be different from the standard workpiece, i.e., the surface of some parts of the welding workpiece may be a plane with a certain degree of curvature due to cooling deformation, and a typical modeling method is to generate a uniform triangular mesh in advance and then project it onto the surface of the welding workpiece to generate a triangular mesh model.

[0040] S2, obtain infrared image data of the object-to-be-tested at at least two time points from the first time to the second time, obtain infrared image data at multiple time points based on a difference value of the infrared image data at the at least two time points, and input the infrared image data at the multiple time points into the three-dimensional finite element analysis model to obtain first stress field characteristics; the first stress field characteristics reflect the influence of stress distribution caused by temperature change of the welding workpiece during the cooling process.

[0041] In the specific implementation, the infrared image data of the object-to-be-tested is obtained at at least two time points of the first time and the second time; the operation of obtaining the infrared image data at some time points between the first time and the second time can also be added, but the data processing amount or the cost of the acquisition equipment will also be increased accordingly, therefore, the infrared image data at multiple time points between the first time and the second time can also be generated through data interpolation, and the specific interpolation method is not limited, linear interpolation or nonlinear interpolation based on a temperature conduction model can be used, the former has a fast data generation speed, and the latter has more accurate data, and a suitable interpolation method can be selected according to actual requirements.

[0042] S3, input the first three-dimensional scanning data into the first three-dimensional finite element analysis model to obtain second stress field characteristics; the second stress field characteristics reflect the influence of stress distribution caused by deformation of the welding workpiece before and after cooling. The first three-dimensional scanning data and the second three-dimensional scanning data are corresponding data of the welding workpiece before and after cooling, respectively, therefore, inputting the first three-dimensional scanning data into the first three-dimensional finite element analysis model established based on the second three-dimensional scanning data can reflect the deformation of the welding workpiece, and the position with greater deformation causes greater stress influence; especially, a crack may be generated inside the position with greater deformation, and therefore, stress conduction at this position is often discontinuous, and therefore, a conventional finite element analysis model often has difficulty in calculating real data.

[0043] In a specific embodiment of this application, since the influence of stress on temperature conduction and on deformation stress conduction at the point of open weld or crack is inconsistent, when the welded workpiece is externally intact but may have internal cracks, the influence of temperature conduction and on deformation stress conduction are calculated separately by finite element analysis using the three-dimensional data of the workpiece surface obtained by scanning, thereby capturing stress singularities and discovering possible hidden internal cracks.

[0044] S4. Obtain the third stress field characteristic data based on the characteristics of the first and second stress fields;

[0045] When calculating the residual stress of a welded workpiece, it is often necessary to consider whether it is free deformation or deformation under constraint. If welding is under constraint, bending or torsional deformation may not occur. In this case, the stress is not released and is concentrated internally. In the finite element model, this is reflected in the stress amplification when the first and second stress field characteristics are superimposed, indicating the possible presence of stress concentration points nearby. If it is free deformation, the stress acts on the workpiece, causing external deformation and thus releasing some of the stress. In the finite element model, this is reflected in the stress cancellation when the first and second stress field characteristics are superimposed.

[0046] S5. Monitor based on the characteristics of the third stress field, and issue an early warning when the stress exceeds the set stress threshold.

[0047] In the finite element model, this manifests as a partial stress increase exceeding a set stress threshold when the first and second stress field characteristics are superimposed, necessitating a warning. At this point, the user can use specialized equipment for further testing based on the warning information.

[0048] As a further preferred embodiment, step S2 specifically includes:

[0049] S21. Acquire infrared image data of the object under test at least two time points from the first time point to the second time point, and obtain infrared image data of multiple time points by interpolation based on the infrared image data of the at least two time points.

[0050] S22. Based on the grid nodes, calculate the temperature fitting curve of each node in the time dimension;

[0051] Mesh nodes are the grid nodes established in the finite element model. Each mesh node corresponds to a point in the infrared image data, and each point has corresponding temperature data. Since the temperature data corresponding to each mesh node changes continuously over time, the data can be interpolated based on the time-temperature curve to simulate the temperature change over time at each mesh node, which can then be used for subsequent calculations. Figure 11As shown, each point in the infrared image has corresponding temperature data. The temperatures corresponding to the first time t1 and the second time t2 for each point are Ti1 and Ti2, respectively. Since the temperature difference between the welded workpiece and the ambient temperature is large when the temperature is high, the heat dissipation efficiency is higher. Temperature changes are usually not linear. Therefore, nonlinear interpolation methods and nonlinear fitting methods are commonly used to fit the temperature curve, such as the most basic least squares method.

[0052] Theoretically, since temperature changes may cause deformation, the points corresponding to the mesh nodes may also experience displacement. Ignoring the displacement of mesh nodes would introduce calculation errors. However, in practice, firstly, the high thermal conductivity of welded metal workpieces in engineering construction means that the influence of welds on temperature conduction can be ignored. Secondly, the displacement in actual welding is usually not large, and the effect of deformation is calculated independently in subsequent step S3. Therefore, the error in the actual displacement of mesh nodes can be ignored. Furthermore, the above-mentioned defects are compensated for by adjusting the mesh node density established in the finite element model.

[0053] S23. Calculate the integral of the temperature change rate difference between each two adjacent nodes per unit time based on the three-dimensional finite element analysis model, and calculate the directional characteristics of the first stress field.

[0054] like Figure 12 As shown, suppose there are two adjacent grid nodes A and B, and during the first time interval [t1,t2] between the first and second time intervals... m The temperature change rate at node A is greater. Since both are cooling processes, the contraction at point A is greater than that at point B, and the stress direction is from B to A; after a certain period of time [t m The temperature change rate at node B is greater, and the contraction at point B is greater than at point A, with the stress direction pointing from A to B. Therefore, from a macroscopic perspective, the final direction of the accumulated stress over each time period is determined by the integral of the difference in the temperature change rates between nodes A and B.

[0055] S24. Calculate the integral of the difference in temperature change rate per unit time between each two adjacent nodes based on the three-dimensional finite element analysis model, and calculate the stress change characteristics of the first stress field.

[0056] At the same time, due to the previous period of [t1,t] m The stress direction is from B to A, and after a certain period of time [t] mThe stress direction of t1 and t2 is A pointing to B, which can cause stress offset, but actually observed from the micro perspective, since the workpiece also has interlayer stress in three-dimensional space, the model used can only reflect the surface condition according to the three-dimensional data of the workpiece surface. If simulated by solid elements, at least three layers of grid in the thickness direction are needed to capture the stress gradient in the thickness direction, and the number of grids will be very large, so simplifying the single-layer grid of the workpiece surface will greatly reduce the number of grids and improve the solving efficiency, but the stress generated by interlayer slip needs to be considered. The change of stress direction in different time periods may bring stress accumulation rather than stress offset at the micro level. Therefore, the three-dimensional finite element analysis model (the first three-dimensional finite element analysis model, the second three-dimensional finite element analysis model, and the third three-dimensional finite element analysis model) of the embodiments of the application are all single-layer grid models based on three-dimensional data modeling of the workpiece surface, so the integral of the difference degree of the temperature change rate per unit time between each adjacent two nodes needs to be calculated in step S24. The difference degree of the temperature change rate per unit time between each adjacent two nodes is positively correlated with the absolute value of the difference value of the temperature change rate per unit time between each adjacent two nodes, so as to obtain the stress accumulation at the micro level under extreme conditions.

[0057] Further as a preferred embodiment, the step S22 specifically comprises:

[0058] S221, based on the grid nodes, calculate the temperature fitting curve of each node in the time dimension; here the grid nodes are the initial grid nodes, i.e. the grid nodes established in the first three-dimensional finite element analysis model;

[0059] S222, calculate the temperature change rate per unit time between each adjacent two nodes, and calculate the absolute value of the difference value of the temperature change rate per unit time, when the average value M of the absolute value of the difference value of the temperature change rate per unit time between adjacent two nodes is in the corresponding change rate threshold interval, increase the nodes in the corresponding grid of the adjacent two nodes to divide or merge the grid, and obtain the second three-dimensional finite element analysis model;

[0060] In the actual implementation process, the change rate threshold interval can be divided into three intervals: [0, M1], (M1, M2), [M2, +∞), and the corresponding division ratios are 3, 4, and 6, respectively, i.e. the number of grid elements after division of each grid element, as shown in Figures 8-10 The threshold interval can also be set according to the corresponding division ratio, and the value size corresponding to the threshold interval is positively correlated with the division ratio.

[0061] The absolute value of the difference in temperature change rate per unit time reflects the stress impact caused by temperature changes during cooling. If the average value is within a relatively small threshold range, it indicates that the temperature difference between two adjacent nodes is small, and the stress change can be ignored. Mesh merging can be used to reduce mesh density and computational load. If the average value is within a relatively large threshold range, it indicates that the temperature difference between two adjacent nodes is large, potentially leading to stress concentration. Mesh splitting to increase mesh density is necessary for analysis and to capture stress singularities. For nodes with a moderate threshold range, no change in mesh density is required. Figures 3-5 As shown, Figure 3 The image illustrates the residual stress generated when two components are welded together via a butt weld. However, since the two components are now welded as a single unit, the contraction displacement along the X and Y axes around the weld in this state is difficult to represent using 3D scanning data. Temperature changes at the mesh nodes, however, can reflect this. Figure X Stress in the Y-axis direction, therefore for example Figure 4 The grid cells are divided or merged, as shown in the following figure. Figure 5 As shown.

[0062] S223. Calculate the temperature fitting curve of the added node in the time dimension; for the newly added grid node, it is necessary to calculate and generate its corresponding temperature fitting curve.

[0063] As a further preferred embodiment, the three-dimensional finite element analysis model in steps S23 and S24 is a second three-dimensional finite element analysis model; the first stress field direction characteristics and the first stress field stress change characteristics are calculated using the second three-dimensional finite element analysis model after updating the mesh nodes. The first stress field characteristics include the first stress field direction characteristics and the first stress field stress change characteristics.

[0064] As a further preferred embodiment, step S222 specifically includes:

[0065] S2221. Calculate the rate of temperature change per unit time between any two adjacent nodes, and calculate the absolute value of the difference in the rate of temperature change per unit time.

[0066] S2222: When the mean absolute value of the difference in temperature change rate per unit time between two adjacent nodes is greater than the change rate threshold, add nodes to the grid corresponding to the two adjacent nodes to divide the grid.

[0067] S2223. When the mean absolute value of the difference in temperature change rate per unit time between two adjacent nodes is less than the change rate threshold, the meshes corresponding to the two adjacent nodes are merged to obtain the second three-dimensional finite element analysis model.

[0068] Since the combined grid cells may form polygonal irregular grid cells, which are not conducive to analysis and calculation, it is necessary to segment them to adapt to subsequent numerical calculation, and therefore, as a further preferred embodiment, the step S2223 is specifically:

[0069] S2223, when the average of the absolute values of the rate of change of temperature per unit time between the two adjacent nodes is less than the rate of change threshold, the grid corresponding to the two adjacent nodes is combined, and the combined grid cell is segmented into a plurality of sub-grid cells, the number of the sub-grid cells is less than that of the grid cells before the combination, and a second three-dimensional finite element analysis model is obtained.

[0070] As a further preferred embodiment, the plurality of sub-grid cells in the step S2223 are triangular sub-grid cells and / or quadrilateral sub-grid cells, and the combined grid cell is preferentially segmented into quadrilateral sub-grid cells.

[0071] As shown in Figure 6 , when a plurality of grid cells are combined into a grid cell as shown by the solid line, they need to be divided into triangular sub-grid cells and / or quadrilateral sub-grid cells, and the division mode is as shown by the dashed line in Figure 6 . In addition, the division mode can also be as shown by the dashed line in Figure 7 . Since the change values of the adjacent grid cells to be combined are all less than the preset threshold, the difference in calculation results caused by the division modes of Figure 6 , Figure 7 or other different division modes can be ignored.

[0072] As a further preferred embodiment, the step S3 is specifically:

[0073] S31, adjusting the grid nodes in the three-dimensional finite element analysis model based on the first three-dimensional scanning data and the second three-dimensional scanning data to obtain a third three-dimensional finite element analysis model, and inputting the first three-dimensional scanning data into the third three-dimensional finite element analysis model to obtain a second stress field distribution feature.

[0074] In finite element analysis software, such as ANSYS program, error estimation and adaptive mesh division functions are usually provided to reduce the error caused by mesh division, such as using a predefined macro ADAPT.MAC to realize adaptive mesh division to reduce error, for example, to avoid using too large grid density in the concentrated load area and near the boundary constraint condition to cause stress concentration in the calculation and analysis results. However, the function of adaptive mesh division still needs the user to manually select the adaptive improvement area in the initial grid, and the ADAPT process usually increases the grid in the area with too dense grid near the concentrated load area and the boundary constraint condition, thereby excluding some stress singular points, the numerical value of the stress singular points is usually larger than that of other areas, for example, the concentrated load area.

[0075] In a specific embodiment of this application, since the influence of stress on temperature conduction and the influence on deformation stress conduction are inconsistent at the point of open weld or crack, it is necessary to increase the mesh density in the area with large deformation in order to capture stress singularities. When the welded workpiece is externally intact but may have internal cracks, the influence of temperature conduction and the influence on deformation stress conduction are calculated separately by finite element analysis using the three-dimensional data of the workpiece surface obtained by scanning, thereby discovering possible hidden internal cracks.

[0076] like Figure 3 As shown, this illustrates the residual stress generated when two components are welded together with a butt weld. The shaded area represents the middle butt weld. The cooling and shrinkage of the weld causes the two components to tend to bend towards each other. Figure 3 As shown by the dashed line; and the closer to the center, the greater the tendency to bend. Figure 3 The arrow lengths are used to represent the corresponding components. However, since the two parts are now welded together, the shrinkage displacement along the XY axes around the weld in the diagram is difficult to represent using 3D scanning data. Instead, the shrinkage displacement along the Z-axis of the 3D Cartesian coordinate system can be represented by the deformation region distribution characteristics (i.e., the difference between the first and second 3D scanning data along the Z-axis). The shrinkage displacement along the XY axes is only visible at the edges, but the stress at the edges is often small and can be ignored. When the welded workpiece has a complex shape, the deformation region distribution characteristics are the set of differences between the first and second 3D scanning data along the normal direction of the corresponding scanning point. The principle is the same as the calculation of the difference along the Z-axis of the 3D Cartesian coordinate system, and will not be elaborated here.

[0077] Since the mesh nodes in the second and third three-dimensional finite element analysis models are different, as a preferred embodiment, step S4 specifically includes:

[0078] A merged mesh node is generated based on the mesh nodes of the second and third 3D finite element analysis models; the merged mesh node includes all mesh nodes of the second and third 3D finite element analysis models.

[0079] Based on the correspondence between the first stress field characteristics and the mesh nodes of the second three-dimensional finite element analysis model, the first stress field characteristics corresponding to the merged mesh nodes are updated; that is, interpolation is performed to supplement the first stress field characteristic data corresponding to mesh nodes that are not in the second three-dimensional finite element analysis model among all merged mesh nodes. Since this data is interpolated and not calculated by increasing the mesh density, no additional stress singularity data is introduced.

[0080] Similarly, according to the correspondence between the second stress field characteristics and the grid nodes of the third three-dimensional finite element analysis model, the second stress field characteristics corresponding to the grid nodes are updated and merged;

[0081] According to the superposition of the first stress field characteristics and the second stress field characteristics, third stress field characteristic data is obtained. When the stress corresponding to the grid nodes in the third stress field characteristic data exceeds the set stress threshold, the data may indicate that the welding workpiece quality has a problem, and a warning information needs to be issued.

[0082] Referring to Figure 2 The second specific embodiment of the present application provides a quality monitoring method based on image processing, comprising the following steps:

[0083] S1, obtaining first three-dimensional scanning data of a to-be-tested object at a first time and second three-dimensional scanning data of the to-be-tested object at a second time, and establishing a first three-dimensional finite element analysis model based on the second three-dimensional scanning data;

[0084] The first three-dimensional scanning data of the to-be-tested object at the first time specifically refers to the three-dimensional scanning data of the to-be-tested object (i.e., the welding workpiece) when the welding is completed; the second three-dimensional scanning data of the to-be-tested object at the second time specifically refers to the three-dimensional scanning data of the to-be-tested object after cooling and setting, and the basic model for calculation is established based on the three-dimensional scanning data after cooling and setting. The three-dimensional scanning data of the welding workpiece after cooling and setting may be different from the standard workpiece, that is, the surface of some parts of the welding workpiece may be a certain curved plane due to cooling deformation, and a typical modeling method is to generate a uniform triangular mesh in advance and then project it onto the surface of the welding workpiece to generate a triangular mesh model.

[0085] S2, obtaining infrared image data of the to-be-tested object at at least two time points from the first time to the second time, interpolating the infrared image data at the at least two time points to obtain infrared image data at multiple time points, and inputting the infrared image data at the multiple time points into the three-dimensional finite element analysis model to obtain first stress field characteristics; the first stress field characteristics reflect the influence of stress distribution caused by temperature change of the welding workpiece during the cooling process.

[0086] In the specific embodiment, the infrared image data of the to-be-tested object is obtained at at least the first time and the second time; the operation of obtaining the infrared image data at some time points between the first time and the second time can also be added, but the data processing amount or the cost of the acquisition equipment will also be increased accordingly, so the infrared image data at multiple time points between the first time and the second time can also be generated through data interpolation, and the specific interpolation method is not limited, linear interpolation or nonlinear interpolation based on a temperature conduction model can be used, the former generates data quickly, and the latter is more accurate, and a suitable interpolation method can be selected according to actual needs.

[0087] S3, adjusting the grid nodes in the three-dimensional finite element analysis model based on the first three-dimensional scanning data and the second three-dimensional scanning data to obtain a third three-dimensional finite element analysis model, the first three-dimensional scanning data being input into the third three-dimensional finite element analysis model to obtain a second stress field feature; the second stress field feature reflecting the influence of the deformation of the welded workpiece before and after cooling on the stress distribution.

[0088] The first three-dimensional scanning data and the second three-dimensional scanning data are respective data of the welded workpiece before and after cooling, so that the deformation of the welded workpiece can be reflected through the two data, and the position with greater deformation causes greater stress influence; especially, cracks may be generated inside the position with greater deformation, and stress conduction at this position is often discontinuous, so that a conventional finite element analysis model often fails to calculate real data.

[0089] In a finite element analysis software, such as an ANSYS program, error estimation and adaptive mesh division functions are usually provided to reduce the error caused by mesh division, that is, a predefined macro ADAPT.MAC is used to realize adaptive mesh division to reduce the error, for example, to avoid using too large grid density in the concentrated load area and near the boundary constraint condition to cause the calculation and analysis result to show that there is stress concentration. However, the function of adaptive mesh division also requires the user to manually select the adaptive improvement area in the initial mesh, and the ADAPT process usually increases the grid in the concentrated load area and near the boundary constraint condition, thereby excluding some stress singular points, the numerical value of the stress singular point is usually larger than that of other regions, for example, the concentrated load area.

[0090] In the specific embodiments of the present application, because the influence of the empty welding or crack on the temperature conduction is inconsistent with the influence on the deformation stress conduction, it is necessary to increase the grid density in the area with greater deformation to capture the stress singular point, and in the case that the welded workpiece is complete outside but may have cracks inside, the three-dimensional data of the workpiece surface obtained by scanning is used to calculate the influence of temperature conduction and the influence on the deformation stress conduction through finite element analysis, so as to find the possible hidden internal cracks.

[0091] S4, obtaining third stress field feature data according to the first stress field feature and the second stress field feature;

[0092] In the calculation of residual stress of the welded workpiece, whether the deformation is in a free state or in a constrained state is often considered. If the welding is in a constrained state, no bending deformation or twisting deformation may be generated, the stress is not released and concentrated in the interior, and in the finite element model, when the first stress field feature and the second stress field feature are superimposed, part of the stress is enhanced, so the data at this position may reflect the existence of a stress concentration point nearby; if the deformation is in a free state, the stress acts on the workpiece to generate external deformation, thereby releasing part of the stress, and in the finite element model, when the first stress field feature and the second stress field feature are superimposed, part of the stress is offset.

[0093] S5, monitoring according to the third stress field feature, and issuing a warning when the stress is greater than a set stress threshold.

[0094] Further preferably, the step S2 is specifically:

[0095] S21, acquiring infrared image data of the to-be-tested object at at least two time points within a first time to a second time, and obtaining infrared image data at a plurality of time points by interpolation based on the infrared image data at the at least two time points;

[0096] S22, calculating a temperature fitting curve of each node in the time dimension based on the grid nodes;

[0097] The grid nodes are grid nodes established in the finite element model, each grid node corresponds to a point in the infrared image data, and each point has corresponding temperature data; since the corresponding temperature data of each grid node in the time dimension is constantly changing, the data can be interpolated based on the time-temperature curve, so as to obtain the simulated temperature change for subsequent calculation. Figure 11 As shown in the figure, each point of the infrared image has corresponding temperature data, and the temperatures of each point at the first time t1 and the second time t2 are T i1 and T i2 respectively. Since the temperature difference between the welded workpiece and the environment is large when the temperature is high, the heat dissipation efficiency is higher, and the temperature change is usually not linear, a nonlinear interpolation method and a nonlinear fitting method are usually used to fit the temperature curve, for example, the most basic least squares method.

[0098] In theory, since the temperature change may also cause deformation, the points corresponding to the grid nodes may also produce displacement, and the calculation error may be caused by not considering the displacement of the grid nodes. However, in practice, on the one hand, the thermal conductivity of the welded metal workpiece in engineering construction is high, and even if there is a weld, the influence on temperature conduction can be ignored, on the other hand, the displacement in actual welding is usually not large, and the influence of the displacement is calculated independently in the subsequent step S3, so the error of the actual displacement of the grid nodes can be ignored; at the same time, the above defects are also compensated by adjusting the density of the grid nodes established in the finite element model.

[0099] S23. Calculate the integral of the temperature change rate difference between each two adjacent nodes per unit time based on the three-dimensional finite element analysis model, and calculate the directional characteristics of the first stress field.

[0100] like Figure 12 As shown, suppose there are two adjacent grid nodes A and B. During the first time interval [t1, t2] from the first time interval t1 to the second time interval t2, ... m The temperature change rate at node A is greater. Since both are cooling processes, the contraction at point A is greater than that at point B, and the stress direction is from B to A; after a certain period of time [t m The temperature change rate at node B is greater, and the contraction at point B is greater than at point A, with the stress direction pointing from A to B. Therefore, from a macroscopic perspective, the final direction of the accumulated stress over each time period is determined by the integral of the difference in the temperature change rates between nodes A and B.

[0101] S24. Calculate the integral of the difference in temperature change rate per unit time between each two adjacent nodes based on the three-dimensional finite element analysis model, and calculate the stress change characteristics of the first stress field.

[0102] Meanwhile, due to the previous period of [t1,t] m The stress direction is from B to A, and after a certain period of time [t] m The stress direction [t2] is from A to B, which may cause stress cancellation. However, from a microscopic perspective, since there is interlayer stress in the workpiece in three-dimensional space, the model used can only reflect the surface condition based on the three-dimensional data of the workpiece surface. If solid elements are used for simulation, at least three layers of mesh in the thickness direction are needed to capture the stress gradient in the thickness direction, and the number of meshes will be very large. Therefore, simplifying it to a single layer of mesh on the workpiece surface will greatly reduce the number of meshes and improve the solution efficiency. However, the stress generated by interlayer slip needs to be considered. The change in stress direction over different time periods may lead to stress accumulation rather than stress cancellation at the microscopic level. Therefore, the three-dimensional finite element analysis models (first three-dimensional finite element analysis model, second three-dimensional finite element analysis model, and third three-dimensional finite element analysis model) of the specific embodiments of this application are all single-layer mesh-based models based on the three-dimensional data of the workpiece surface. Therefore, in step S24, it is necessary to calculate the integral of the difference in the unit time temperature change rate between each two adjacent nodes. The difference in the unit time temperature change rate between each two adjacent nodes is positively correlated with the absolute value of the difference in the unit time temperature change rate between each two adjacent nodes, thereby obtaining the stress accumulation brought about at the microscopic level under extreme conditions.

[0103] As a further preferred embodiment, step S22 specifically includes:

[0104] S221. Based on the grid nodes, calculate the temperature fitting curve of each node in the time dimension;

[0105] The grid nodes here are initial grid nodes, that is, the grid nodes established in the first three-dimensional finite element analysis model;

[0106] S222, calculate the temperature change rate per unit time between each adjacent two nodes, and calculate the absolute value of the temperature change rate per unit time difference, when the average value of the absolute value of the temperature change rate per unit time difference between the adjacent two nodes is in the corresponding change rate threshold interval, increase the node in the corresponding grid between the adjacent two nodes to split or merge the grid, and obtain the second three-dimensional finite element analysis model.

[0107] In actual implementation, the change rate threshold interval can be divided into three intervals: [0, M1], (M1, M2), [M2, +∞), and the corresponding split ratios are 3, 4, and 6, respectively, that is, the number of grid elements after each grid element is split, as shown in Figures 8-10 The threshold interval can also be set according to the corresponding split ratio, and the value size corresponding to the threshold interval is positively correlated with the split ratio.

[0108] The absolute value of the temperature change rate per unit time difference reflects the stress influence caused by temperature change during the cooling process. If the above average value is in the corresponding smaller change rate threshold interval, it indicates that the temperature change difference between the adjacent two nodes is small, and the stress change can be ignored. The grid can be merged to reduce the grid density for analysis and calculation, and the calculation amount is reduced. If the above average value is in the corresponding larger change rate threshold interval, it indicates that the temperature change difference between the adjacent two nodes is large, and stress concentration may occur, so the grid needs to be split to increase the grid density for analysis and calculation, which is beneficial to capture stress singular points. For nodes with moderate change rate threshold interval, the grid density does not need to be changed. As shown in Figures 3-5 Figure 3 The residual stress generated when two components are welded by butt weld is shown, but at this time the two components have been welded into one, so the shrinkage displacement in the XY axis direction of the weld periphery position in the figure is difficult to reflect through the three-dimensional scanning data, but the temperature change of the grid node can reflect the stress in the Y axis direction, so the grid element such as Figure X is split or merged, and the effect diagram is as shown in Figure 4 Figure 5

[0109] S223, calculate the temperature fitting curve of the added node in the time dimension; for the newly added grid node, the corresponding temperature fitting curve needs to be added to generate.

[0110] Further as a preferred implementation manner, the three-dimensional finite element analysis model in steps S23 and S24 is the second three-dimensional finite element analysis model;

[0111] ​​​Calculate a first stress field direction feature and a first stress field stress variation feature based on the second three-dimensional finite element analysis model after updating the grid nodes.

[0112] Further as a preferred implementation, the step S222 is specifically:

[0113] S2221, calculate the temperature change rate per unit time between each adjacent two nodes, and calculate the absolute value of the temperature change rate per unit time difference;

[0114] S2222, when the average value of the absolute value of the temperature change rate per unit time difference between adjacent two nodes is greater than the change rate threshold, increase the nodes in the grid corresponding to the adjacent two nodes to segment the grid;

[0115] S2223, when the average value of the absolute value of the temperature change rate per unit time difference between adjacent two nodes is less than the change rate threshold, merge the grid corresponding to the adjacent two nodes to obtain a second three-dimensional finite element analysis model.

[0116] Since the polygon irregular grid unit may be formed after merging the grid units, which is not conducive to analysis and calculation, it needs to be segmented to adapt to subsequent numerical calculation, therefore, further as a preferred implementation, the step S2223 is specifically:

[0117] S2223, when the average value of the absolute value of the temperature change rate per unit time difference between adjacent two nodes is less than the change rate threshold, merge the grid corresponding to the adjacent two nodes, and segment the merged grid unit into a plurality of sub-grid units, the number of the sub-grid units is less than that of the grid units before merging, to obtain a second three-dimensional finite element analysis model.

[0118] Further as a preferred implementation, the plurality of sub-grid units in the step S2223 are triangular sub-grid and / or quadrilateral sub-grid, and when the merged grid unit is segmented into a plurality of sub-grid units, the quadrilateral sub-grid is preferentially selected.

[0119] As shown in FIG. 1, Figure 6 wherein when a plurality of grid units are merged into a grid unit as shown by the solid line, the grid unit needs to be divided into triangular sub-grid and / or quadrilateral sub-grid, and the division manner is as shown by the dashed line in FIG. 2. Figure 6 In addition, the division manner can also be as shown by the dashed line in FIG. 3. Figure 7 Since the change values of the adjacent grid units to be merged are all less than the preset threshold, the calculation result difference caused by Figure 6 , Figure 7 or other different division manners can be ignored.

[0120] Further as a preferred implementation, the step S3 is specifically:

[0121] S31, obtaining a deformation region distribution feature based on the first three-dimensional scanning data and the second three-dimensional scanning data;

[0122] The second stress field feature reflects the influence of the deformation of the welded workpiece before and after cooling on the stress distribution, so the deformation region distribution feature obtained based on the first three-dimensional scanning data and the second three-dimensional scanning data is used as the basis for subsequent grid adjustment.

[0123] As shown in Figure 3 , which shows the residual stress generated when two components are welded by butt weld, the shaded part is the middle butt weld, and the cooling shrinkage of the weld makes the two components have a tendency to bend towards each other, as shown by the dashed line in Figure 3 ; and the closer to the center, the greater the bending tendency, Figure 3 correspondingly represented by the length of the arrow, but at this time the two components have been welded together, so the shrinkage displacement of the XY axis direction at the periphery of the weld in the figure at this time is difficult to reflect through the three-dimensional scanning data, but the shrinkage displacement on the Z axis of the three-dimensional rectangular coordinate system at this time can be reflected through the deformation region distribution feature (i.e. the difference between the first three-dimensional scanning data and the second three-dimensional scanning data on the Z axis), the shrinkage displacement in the XY axis direction can only be reflected at the edge position, but the stress at the edge position is often small and can be ignored. When the welded workpiece is of a complex shape, the deformation region distribution feature is the set of differences between the first three-dimensional scanning data and the second three-dimensional scanning data in the normal direction of the corresponding scanning point, and the principle is the same as the difference calculation on the Z axis of the three-dimensional rectangular coordinate system described above.

[0124] S32, adjusting the grid nodes in the corresponding first three-dimensional finite element analysis model based on the deformation region distribution feature to obtain a third three-dimensional finite element analysis model;

[0125] S33, inputting the first three-dimensional scanning data into the third three-dimensional finite element analysis model to obtain the second stress field feature.

[0126] Further as a preferred implementation, the step S33 is specifically:

[0127] S331, calculating the change value of the corresponding grid element in the deformation region distribution feature calculation model;

[0128] The change value of the grid cell reflects the change of the corresponding grid cell before and after the welding workpiece is cooled. For example, after the welding workpiece is cooled, the welding part shrinks, and the grid position corresponding to the shrinkage part will have a greater displacement relative to the position before cooling. Specifically, the change value of the grid cell can use the average of the position difference of each node on the grid before and after cooling or the position difference of the center of the grid (such as the center of gravity of the triangular grid) before and after cooling, which can reflect the change of the corresponding grid cell position before and after the welding workpiece is cooled.

[0129] S332, when the change value is greater than the preset threshold value, the corresponding grid cell is segmented;

[0130] When the change value is greater than the preset threshold value, it indicates that there may be a large stress at this position and in the corresponding workpiece, and the grid density needs to be increased so that the stress singular point can be found in the obtained data. For example, as shown in Figure 4 、 5 The middle part corresponds to the shadow part of the weld position in Figure 3 .

[0131] Further, as a preferred embodiment, the step S33 specifically further comprises:

[0132] S333, when the change value is less than the preset threshold value, the adjacent grid cells of the corresponding grid cell are traversed, and if there is an adjacent grid cell whose change value is less than the preset threshold value, the corresponding grid cell and the adjacent grid cell are merged.

[0133] When the change values of the adjacent grid cells are all less than the preset threshold value, the grid cells can be merged to reduce the calculation amount in the finite element analysis and improve the calculation speed. The effect diagram of merging the corresponding grid cells is shown in Figure 5 , which corresponds to the positions on both sides of the shadow part of the weld in Figure 3 . It should be noted that steps S2 and S3 both use Figure 4 and Figure 5 for principle explanation, but since the temperature change and the deformation change have different influences on the stress, the actual results of the segmentation and merging operations based on the same Figure 4 in steps S2 and S3 will not produce the same Figure 5 result as shown in

[0134] Further, as a preferred embodiment, the step S33 specifically further comprises:

[0135] S334, the merged grid cell is segmented into a plurality of sub-grid cells, and the number of the sub-grid cells is less than that of the grid cells before merging.

[0136] After the merging of the grid cells in step S333, polygonal irregular grid cells can be formed, which are not conducive to analysis and calculation, and thus need to be segmented for subsequent numerical calculation. Therefore, as a further preferred embodiment, step S334 is specifically:

[0137] The plurality of sub-grid cells are triangular sub-grid cells and / or quadrilateral sub-grid cells. When the merged grid cells are segmented into a plurality of sub-grid cells, the quadrilateral sub-grid cells are preferentially selected.

[0138] As shown in Figure 6 , when a plurality of grid cells are merged into a grid cell as shown by the solid line, the grid cell needs to be divided into triangular sub-grid cells and / or quadrilateral sub-grid cells, and the division manner is as shown by the dashed line in Figure 6 . In addition, the division manner can also be as shown by the dashed line in Figure 7 . Since the change values of the adjacent grid cells to be merged are all less than the preset threshold value, the calculation results caused by the different division manners Figure 6 , Figure 7 or other different division manners can be negligible.

[0139] As a further preferred embodiment, step S332 is specifically:

[0140] S3321, when the change value is greater than the preset threshold value T, determining the threshold interval corresponding to the change value;

[0141] S3322, segmenting the grid cells according to the segmentation ratio corresponding to the threshold interval;

[0142] In the actual implementation process, the numerical interval [T, +∞) greater than the preset threshold value can be divided into three intervals [T, T1), [T1, T2), [T2, +∞), and the segmentation ratios are 2, 4, and 6, respectively, as shown in the figure. The corresponding segmentation ratio can also be set according to the threshold values T1 and T2. The numerical size corresponding to the threshold interval is positively correlated with the segmentation ratio.

[0143] Since the grid nodes in the second three-dimensional finite element analysis model and the third three-dimensional finite element analysis model are different, as a further preferred embodiment, step S4 is specifically:

[0144] Generating a merged grid node according to the grid nodes of the second three-dimensional finite element analysis model and the grid nodes of the third three-dimensional finite element analysis model; the merged grid node includes all the grid nodes of the second three-dimensional finite element analysis model and the grid nodes of the third three-dimensional finite element analysis model;

[0145] According to the correspondence between the first stress field characteristics and the grid nodes of the second three-dimensional finite element analysis model, the first stress field characteristics corresponding to the merged grid nodes are updated; that is, the first stress field characteristic data corresponding to the grid nodes in the second three-dimensional finite element analysis model in all the merged grid nodes are interpolated and supplemented. Since the data is interpolated and supplemented, rather than calculated by increasing the grid density, stress singular point data is not introduced additionally.

[0146] Similarly, according to the correspondence between the second stress field characteristics and the grid nodes of the third three-dimensional finite element analysis model, the second stress field characteristics corresponding to the merged grid nodes are updated;

[0147] The third stress field characteristic data is obtained by superimposing the first stress field characteristics and the second stress field characteristics. When the stress corresponding to the grid nodes in the third stress field characteristic data exceeds the set stress threshold, the data may indicate that there is a problem with the welding workpiece quality, and a warning information needs to be issued.

[0148] With reference to Figure 13 The third specific embodiment of the present application provides a quality monitoring device based on image processing, which corresponds to the quality monitoring method based on image processing provided in the first and second specific embodiments. The device comprises:

[0149] Image data acquisition devices 11 and 12 are used to obtain the first three-dimensional scanning data of the object 60 at the first time and the second three-dimensional scanning data of the object 60 at the second time;

[0150] Infrared data acquisition devices 21 and 22 are used to obtain the infrared image data of the object 60 at at least two time points within the first time to the second time;

[0151] A data generation device 30 is used to

[0152] A first three-dimensional finite element analysis model is established based on the second three-dimensional scanning data;

[0153] The infrared image data at a plurality of time points is obtained by interpolation based on the infrared image data at at least two time points, and is input into the three-dimensional finite element analysis model to obtain the first stress field characteristics;

[0154] The first three-dimensional scanning data is input into the first three-dimensional finite element analysis model to obtain the second stress field characteristics;

[0155] The third stress field characteristic data is obtained according to the first stress field characteristics and the second stress field characteristics

[0156] A monitoring and warning device 40 is used to monitor according to the third stress field characteristics, and issues a warning when the stress is greater than the set stress threshold.

[0157] AsFigure 13 As shown, further as a preferred embodiment, the quality monitoring device further comprises a workpiece conveying device 50, the image data acquisition device comprises at least a first image data acquisition device 11 and a second image data acquisition device 12, the infrared data acquisition device comprises at least a first infrared data acquisition device 21 and a second infrared data acquisition device 22, the first image data acquisition device 11 and the first infrared data acquisition device 21 are arranged at a first conveying point of the workpiece conveying device 50, and the second image data acquisition device 12 and the second infrared data acquisition device 22 are arranged at a second conveying point of the workpiece conveying device 50.

[0158] The first image data acquisition device and the first infrared data acquisition device are arranged at the first conveying point of the workpiece conveying device to acquire the first three-dimensional scanning data at a first time and the first infrared image data at the first time of the welded workpiece after the welding is completed; the welded workpiece is conveyed on the conveying device for a period of time from the first conveying point to the second conveying point, and cooling is performed within the period of time, the second image data acquisition device and the second infrared data acquisition device are arranged at the second conveying point of the workpiece conveying device to acquire the second three-dimensional scanning data at a second time and the second infrared image data at the second time of the welded workpiece after the welding is completed and the cooling is performed for a period of time.

[0159] The quality monitoring device in the specific embodiments of the present application does not need to stop the equipment when collecting data, and can continuously collect data of batch welded workpieces, and has high processing efficiency.

[0160] Further as a preferred embodiment, when the data generation device inputs the first three-dimensional scanning data into the first three-dimensional finite element analysis model to obtain the second stress field feature, specifically:

[0161] The grid nodes in the three-dimensional finite element analysis model are adjusted based on the first three-dimensional scanning data and the second three-dimensional scanning data to obtain a third three-dimensional finite element analysis model, and the first three-dimensional scanning data is input into the third three-dimensional finite element analysis model to obtain the second stress field feature.

[0162] The above is a specific description of the preferred embodiments of the present application, and it should be noted that the above is only a preferred embodiment, and those skilled in the art will understand that the present application is not limited to the above specific embodiments, and various obvious changes, re-adjustments and substitutions can be made without departing from the protection scope of the present application. Therefore, although the above embodiments are described in detail, the present application is not limited to the above embodiments, and more other equivalent embodiments can be included without departing from the concept of the present application, and the scope of the present application is determined by the scope of the appended claims.

Claims

1. A quality monitoring method based on image processing, characterized in that, The method comprises the following steps: S1, obtaining first three-dimensional scanning data of a to-be-tested object at a first time and second three-dimensional scanning data of the to-be-tested object at a second time, and establishing a first three-dimensional finite element analysis model based on the second three-dimensional scanning data; the first time refers to the time when the to-be-tested object is completed welding, and the second time refers to the time when the to-be-tested object is cooled and shaped; S21, obtaining infrared image data of the to-be-tested object at at least two time points from the first time to the second time, and obtaining infrared image data at multiple time points by interpolation based on the infrared image data at the at least two time points; S221, calculating a temperature fitting curve of each node in the time dimension based on the grid nodes; S222, calculating a unit time temperature change rate between each adjacent two nodes, and calculating an absolute value of a unit time temperature change rate difference value, when a mean value of the absolute value of the unit time temperature change rate difference value between the adjacent two nodes is in a corresponding change rate threshold interval, adding a node in a grid corresponding to the adjacent two nodes to perform a splitting or merging operation on the grid, and obtaining a second three-dimensional finite element analysis model; S223, calculating a temperature fitting curve of the added node in the time dimension; S23, calculating an integral of a unit time temperature change rate difference value between each adjacent two nodes based on the second three-dimensional finite element analysis model, and calculating a first stress field direction feature; S24, calculating an integral of a unit time temperature change rate difference degree between each adjacent two nodes based on the second three-dimensional finite element analysis model, and calculating a first stress field stress change feature; S3, adjusting grid nodes in a three-dimensional finite element analysis model based on the first three-dimensional scanning data and the second three-dimensional scanning data to obtain a third three-dimensional finite element analysis model, and inputting the first three-dimensional scanning data into the third three-dimensional finite element analysis model to obtain a second stress field feature; S4, generating a merged grid node based on the grid nodes of the second three-dimensional finite element analysis model and the grid nodes of the third three-dimensional finite element analysis model; The merged grid node comprises all the grid nodes of the second three-dimensional finite element analysis model and the grid nodes of the third three-dimensional finite element analysis model; According to a corresponding relationship between the first stress field feature and the grid nodes of the second three-dimensional finite element analysis model, updating the first stress field feature corresponding to the merged grid node; According to a corresponding relationship between the second stress field feature and the grid nodes of the third three-dimensional finite element analysis model, updating the second stress field feature corresponding to the merged grid node; Superimposing the first stress field feature and the second stress field feature to obtain third stress field feature data; S5, monitoring according to the third stress field feature, and issuing a warning when a stress greater than a set stress threshold exists.

2. The method of claim 1, wherein: The step S222 specifically comprises: S2221, calculating a unit time temperature change rate between each adjacent two nodes, and calculating an absolute value of a unit time temperature change rate difference value; S2222, when a mean value of the absolute value of the unit time temperature change rate difference value between the adjacent two nodes is greater than a change rate threshold, adding a node in a grid corresponding to the adjacent two nodes to perform a splitting operation on the grid. S2223, when the average of the absolute values of the difference values of the temperature change rates per unit time between the two adjacent nodes is less than the change rate threshold, merging the grids corresponding to the two adjacent nodes to obtain a second three-dimensional finite element analysis model.

3. The method of claim 2, wherein: The step S2223 is specifically: S2223, when the average of the absolute values of the difference values of the temperature change rates per unit time between the two adjacent nodes is less than the change rate threshold, merging the grids corresponding to the two adjacent nodes, and dividing the merged grid elements into a plurality of sub-grid elements, the number of the sub-grid elements being less than that of the grid elements before merging, to obtain a second three-dimensional finite element analysis model.

4. The method of claim 3, wherein: The plurality of sub-grid elements in the step S2223 are triangular sub-grids and / or quadrilateral sub-grids, and the quadrilateral sub-grids are preferentially selected when the merged grid elements are divided into a plurality of sub-grid elements.

5. An image processing-based quality monitoring apparatus, characterized by comprising: Comprise: An image data acquisition device for acquiring first three-dimensional scanning data of a to-be-tested object at a first time and second three-dimensional scanning data of the to-be-tested object at a second time; An infrared data acquisition device for acquiring infrared image data of the to-be-tested object at at least two time points within the first time to the second time; A data generation device for establishing a first three-dimensional finite element analysis model based on the second three-dimensional scanning data; acquiring infrared image data of the to-be-tested object at at least two time points within the first time to the second time, and interpolating the infrared image data at the at least two time points to obtain infrared image data at a plurality of time points; calculating a temperature fitting curve of each node in the time dimension based on the grid nodes; calculating the temperature change rate per unit time between each two adjacent nodes, and calculating the absolute values of the difference values of the temperature change rates per unit time, when the average of the absolute values of the difference values of the temperature change rates per unit time between the two adjacent nodes is in a corresponding change rate threshold interval, adding nodes in the grid corresponding to the two adjacent nodes to segment or merge the grid, to obtain a second three-dimensional finite element analysis model; calculating the temperature fitting curve of the added node in the time dimension; calculating the integral of the difference values of the temperature change rates per unit time between each two adjacent nodes based on the second three-dimensional finite element analysis model, and calculating a first stress field direction feature; calculating the integral of the difference values of the temperature change rates per unit time between each two adjacent nodes based on the second three-dimensional finite element analysis model, and calculating a first stress field stress change feature; adjusting the grid nodes in the three-dimensional finite element analysis model based on the first three-dimensional scanning data and the second three-dimensional scanning data to obtain a third three-dimensional finite element analysis model, and inputting the first three-dimensional scanning data into the third three-dimensional finite element analysis model to obtain a second stress field feature; generating a merged grid node based on the grid nodes of the second three-dimensional finite element analysis model and the grid nodes of the third three-dimensional finite element analysis model; the merged grid node comprises all the grid nodes of the second three-dimensional finite element analysis model and the grid nodes of the third three-dimensional finite element analysis model; updating the first stress field feature corresponding to the merged grid node according to the corresponding relationship between the first stress field feature and the grid nodes of the second three-dimensional finite element analysis model; According to the corresponding relationship between the second stress field characteristics and the grid nodes of the third three-dimensional finite element analysis model, the second stress field characteristics corresponding to the grid nodes are updated and merged; According to the superposition of the first stress field characteristics and the second stress field characteristics, third stress field characteristic data are obtained; The monitoring and early warning device is used for monitoring according to the third stress field characteristics, and issuing a warning when the stress is greater than a set stress threshold.

6. The quality monitoring device based on image processing according to claim 5, characterized in that The image data acquisition device includes at least a first image data acquisition device and a second image data acquisition device, the infrared data acquisition device includes at least a first infrared data acquisition device and a second infrared data acquisition device, the first image data acquisition device and the first infrared data acquisition device are arranged at a first conveying point of the conveying device, and the second image data acquisition device and the second infrared data acquisition device are arranged at a second conveying point of the conveying device.

Images