A method for monitoring the operating status of aluminum plate stamping equipment

Abstract

This application relates to the field of intelligent control technology, and in particular to a method for monitoring the operating status of an aluminum plate stamping equipment. The method includes: acquiring aluminum plate defect data to identify local defect patterns and determining defect state vectors; learning online and updating the coupling influence matrix describing the cross-influence between parameters and defects; globally calculating an optimal pressure adjustment amount that can coordinately adjust all parameters based on the updated matrix and the current defect state; and applying the optimal pressure adjustment amount to set the parameters for the next cycle and executing the calculation. This invention achieves coordinated optimization of stamping pressure by learning the coupling relationship between various pressure parameters online, effectively avoiding local control conflicts, and enabling rapid and stable global optimal control of overall defects in the aluminum plate, thereby improving the forming quality of the aluminum plate.

Description

Technical Field

[0001] This application relates to the field of intelligent control technology, and in particular to a method for monitoring the operating status of aluminum plate stamping equipment. Background Technology

[0002] Aluminum sheet stamping is a key process in high-end equipment manufacturing fields such as automobile manufacturing and aerospace, mainly used to produce body panels and structural parts. Due to the characteristics of aluminum alloy materials, such as high springback and relatively poor formability, the operating status of the equipment during the stamping process, especially the control precision of process parameters such as blank holder force and hydraulic cushion zone pressure, is extremely high. The operating status of the stamping equipment directly determines the forming quality of the aluminum sheet. Unstable operating status can easily lead to defects such as wrinkling, cracking, or excessive springback on the surface of the aluminum sheet.

[0003] However, in actual stamping production, due to fluctuations in the material properties of aluminum sheets between batches, changes in die temperature, and the nonlinear response of the equipment's hydraulic system, fixed process parameters are often difficult to maintain stable production over a long period. Current stamping equipment control largely relies on open-loop settings based on manual experience, or employs independent PID control for single-point pressure. This traditional approach ignores the complex spatial coupling relationship between pressure in different zones and forming quality during the stamping of large sheet metal parts. For example, adjusting the blank holder force on the left side of the die not only affects the material flow on the left but also influences the forming state of the right side or central region through stress transmission within the sheet metal.

[0004] In existing technologies, the independent controllers of each zone only make isolated adjustments based on feedback from local sensors, lacking a global consideration of the correlation between the overall operating status of the equipment and the final surface quality of the product. This leads to mutual interference between control loops, causing the equipment operating status to oscillate frequently around the optimal operating point, making it difficult to quickly converge to the globally optimal state that can eliminate all surface defects, resulting in a high scrap rate and a long debugging cycle. Summary of the Invention

[0005] To address the technical problem that existing aluminum plate stamping equipment has poor control performance, resulting in the inability to effectively eliminate complex surface defects and converge to the global optimum, this application provides a method for monitoring the operating status of aluminum plate stamping equipment.

[0006] This application provides a method for monitoring the operating status of an aluminum plate stamping equipment, comprising: acquiring three-dimensional surface data of the aluminum plate in the current stamping cycle, identifying multiple local defect patterns, and determining a defect state vector characterizing the current overall operating status; learning and updating a state coupling influence matrix online, wherein the state coupling influence matrix is ​​used to describe the cross-influence of multiple adjustable stamping pressure parameters on the local defect patterns; the update is based on the actual change in defects caused by the actual change in parameters in adjacent stamping cycles, comparing it with the defect change predicted by the state coupling influence matrix to obtain a prediction error, and correcting the state coupling influence matrix according to the prediction error; based on the updated state coupling influence matrix and the current defect state vector, globally calculating an optimal pressure adjustment amount that can coordinately adjust all stamping pressure parameters with the goal of eliminating the current defect; setting the target stamping pressure parameter for the next stamping cycle according to the optimal pressure adjustment amount, thereby achieving adaptive control of the operating status of the aluminum plate stamping equipment through periodic iteration.

[0007] This application can accurately evaluate the direct and cross-effects of each stamping pressure parameter on all defect areas by constructing and learning the state coupling influence matrix online. Based on the control quantity calculation of this global model, the adjustment of each parameter is coordinated and systematic, avoiding mutual interference between independent PID controllers. It can converge to the globally optimal combination of process parameters more quickly and stably, thereby significantly reducing the overall surface deviation of the aluminum plate, rather than just improving the local area, and realizing global optimization control.

[0008] In one embodiment, the method for acquiring the three-dimensional surface data is as follows: acquiring three-dimensional point cloud data of the surface of the stamped aluminum plate through a structured light three-dimensional vision system; aligning and registering the three-dimensional point cloud data with a preset standard CAD model, and calculating the normal distance between each data point in the three-dimensional point cloud data and its corresponding position on the standard CAD model to form the three-dimensional surface data.

[0009] In one embodiment, the identification of multiple local defect patterns is achieved by performing cluster analysis on the three-dimensional surface data.

[0010] In one embodiment, the clustering analysis employs a density-based noise-applied spatial clustering algorithm to aggregate points that are spatially adjacent and have the same deviation direction into defect clusters, with each defect cluster corresponding to a local defect pattern.

[0011] In one embodiment, the average deviation of the local defect mode satisfies the following relationship: ;in, For the first A defect cluster corresponding to a local defect pattern This represents the number of data points contained in the defect cluster. For data points in the defect cluster, For data points The surface deviation value.

[0012] In one embodiment, the update of the state coupling influence matrix satisfies the following relation: ;in, The updated state coupling effect matrix, The state coupling effect matrix before the update. For learning rate, For prediction error, This refers to the actual change in the stamping pressure parameter. Let it be its transpose vector. The square of its Euclidean norm To prevent tiny positive numbers with a denominator of zero.

[0013] By adopting an online incremental learning mechanism based on the least mean square algorithm, the state coupling influence matrix can be continuously corrected and updated after each stamping cycle. This means that the control system can automatically adapt to environmental changes such as material batch fluctuations or mold thermal deformation, and gradually adjust through continuous learning to ensure that the control strategy always matches the most realistic physical forming process, demonstrating strong adaptive capabilities.

[0014] In one embodiment, the prediction error satisfies the following relationship: ;in, and These are the defect state vectors for the current and previous stamping cycles, respectively. and These are the stamping pressure parameter vectors for the current and previous stamping cycles, respectively.

[0015] In one embodiment, the global solution includes: constructing a system of linear equations aimed at eliminating the current defect: ;in, The updated state coupling effect matrix, This represents the defect state vector for the current stamping cycle. The parameter adjustment amount is to be solved; the optimization pressure adjustment amount is obtained by solving the linear equation system.

[0016] In one embodiment, the optimization pressure adjustment amount satisfies the following relationship: ;in, The optimization pressure adjustment amount. for The false rebellion.

[0017] This ensures the universality and reliability of the control algorithm, enabling it to approach the goal of eliminating all defects most effectively with minimal control cost, thus guaranteeing the optimality of control decisions.

[0018] The technical solution of this application has the following beneficial technical effects:

[0019] This application constructs a state coupling influence matrix to accurately evaluate the direct and cross-effects of each stamping pressure parameter on all defect areas. Based on this global model, the coordinated control avoids interference between independent controllers and can converge to the globally optimal parameter combination more quickly and stably, significantly improving the yield of aluminum plate stamping and the stability of equipment operation.

[0020] Furthermore, the state coupling influence matrix is ​​continuously updated online after each stamping cycle. Therefore, the control system can automatically adapt to the slow changes in environmental conditions. When the physical response characteristics of the system change, the influence matrix will gradually self-correct through a continuous learning process, ensuring that the control strategy always matches the most realistic physical process at present, and has strong adaptability and robustness. Attached Figure Description

[0021] Figure 1 This is a flowchart of a method for monitoring the operating status of an aluminum plate stamping equipment according to an embodiment of this application.

[0022] Figure 2 This is a schematic diagram illustrating the defects of an unoptimized initial product.

[0023] Figure 3 This is a schematic diagram of a product modified by collaborative control according to an embodiment of this application. Detailed Implementation

[0024] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0025] Figure 1 This is a flowchart illustrating a method for monitoring the operating status of an aluminum plate stamping equipment according to an embodiment of this application. Figure 1 As shown, the method for monitoring the operating status of aluminum plate stamping equipment includes steps S101 to S104, which are described in detail below.

[0026] S101: Obtain the three-dimensional surface data of the aluminum plate in the current stamping cycle, identify multiple local defect patterns, and determine the defect state vector that characterizes the current overall operating state.

[0027] In one embodiment, a non-contact, high-precision structured light 3D vision system can be deployed at the end of the stamping production line to perform online full-surface scanning of the aluminum sheet produced after each stamping cycle. To improve data quality, the acquired raw point cloud can be preprocessed, for example, by using a statistical outlier removal (SOR) filter to remove isolated noise points caused by ambient light interference or surface reflection.

[0028] In this optional embodiment, the preprocessed point cloud data is aligned and registered with a pre-stored standard CAD design model using an iterative nearest-point algorithm. After registration, the normal distance between each data point in the point cloud and its nearest point on the CAD model is calculated. This distance is the surface deviation value of that point. A positive value may represent springback warping, while a negative value may represent collapse or wrinkling. The deviation values ​​of all points together constitute a three-dimensional defect field covering the entire surface of the aluminum plate. In this scheme, all data constituting the three-dimensional defect field are used as three-dimensional surface data.

[0029] Furthermore, for all data points in the 3D surface data whose absolute deviation values ​​exceed a preset tolerance threshold (for example, the preset tolerance threshold can be set to 0.5mm), the DBSCAN algorithm is applied. The advantage of the DBSCAN algorithm is that it does not require pre-specifying the number of clusters and can identify clusters of arbitrary shapes. By setting an appropriate neighborhood radius and a minimum number of core points, the algorithm can automatically aggregate spatially adjacent points with similar deviation values ​​to form several discrete defect clusters. Each defect cluster physically corresponds to a local defect pattern, such as warping in the upper left corner or concavity in the central area.

[0030] In this optional embodiment, to provide input for subsequent control calculations, it is necessary to determine the severity of each defect mode. This can be achieved by calculating the average deviation value of each identified defect cluster. The average deviation value of local defect modes satisfies the following relationship:

[0031]

[0032] in, For the first A defect cluster corresponding to a local defect pattern This represents the number of data points contained in the defect cluster. For data points in the defect cluster, For data points The surface deviation value.

[0033] Finally, the average deviation values ​​of all n defect modes in the current stamping cycle are combined to form an n-dimensional defect state vector that represents the current overall operating state.

[0034] In this way, by performing three-dimensional scanning and data processing on stamped products, key quality defects can be accurately identified and transformed into mathematical vectors that can be used for closed-loop control, providing accurate input for subsequent adaptive learning and optimization decisions.

[0035] S102, online learning and updating of the state coupling influence matrix, which describes the cross-influence of multiple adjustable stamping pressure parameters on local defect modes; the update is based on the actual change of defects caused by the actual change of parameters in adjacent stamping cycles, compared with the defect change predicted by the state coupling influence matrix, to obtain the prediction error, and the state coupling influence matrix is ​​corrected according to the prediction error.

[0036] In one embodiment, a can be defined The state coupling influence matrix, where, This represents the number of defect clusters currently identified. This refers to the number of all independently adjustable pressure parameters in the stamping equipment, such as the pressure of the eight ejector pins of a hydraulic cushion. Elements in the matrix. The physical meaning is: when the first Pressure parameters When a change of one unit occurs, it is expected to cause the first unit to be affected. The average deviation of each defect cluster How much has changed?

[0037] matrix It fully describes the linear effects of all control inputs on all state outputs, including direct effects and cross-coupling effects. This state coupling effect matrix... An online, data-driven learning approach is used to perform incremental adaptive updates after each stamping cycle. The update algorithm employs the logic of the least mean square adaptive filtering algorithm.

[0038] In one stamping cycle At the end, the system records the actual pressure parameter vector applied during this cycle. and the detected defect state vector Compared to the previous stamping cycle By comparing the data, the actual change in parameters can be obtained. and the actual change in aluminum plate defects .

[0039] Based on the influence matrix known at the end of the previous cycle The system can predict the result of The change in defect that this control action should cause: .

[0040] Prediction error This is the difference between the actual response and the predicted response. Using this error vector Based on the LMS algorithm, the influence matrix After correction, the update of the state coupling effect matrix satisfies the following relationship:

[0041]

[0042] in, This is the learning rate, for example, 0.1, used to control the update step size; It is a very small positive integer; for example, it takes the value of . This is used to prevent the denominator from being zero. This represents the actual change in the stamping pressure parameters. Let it be its transpose vector. It is the square of its Euclidean norm.

[0043] For example, suppose the system has two pressure parameters ( ) and 2 defect modes ( ), the matrix of the previous time step identity matrix The parameters for this period have increased. Predicted defect changes Actual observed changes in defects Then the prediction error Next, calculate and update the numerator: ; Calculate the denominator: Ignore the minimum value ,but The increment is The updated matrix This calculation process demonstrates how the matrix automatically adjusts the coupling weights between parameters based on the error, correcting the initial assumption that there was no correlation or that the correlation was 1.

[0044] This calculation process demonstrates how the matrix automatically adjusts the coupling weights between parameters based on the error, correcting the initial assumption that there was no correlation (0 elements) or that the correlation was 1.

[0045] In this way, through online learning of production data, the system can automatically build and continuously optimize a mathematical model that reflects the complex coupling relationship between various control variables and defect states, laying the foundation for achieving precise global control.

[0046] S103, based on the updated state coupling influence matrix and the current defect state vector, aims to eliminate the current defect and globally calculates the optimal pressure adjustment amount that can coordinately adjust all stamping pressure parameters.

[0047] In one embodiment, in each stamping cycle End, and the influence matrix is ​​now complete. After the update, this latest coupling model can be used to calculate the parameters for the next cycle. Optimal pressure adjustment In this invention, the optimal pressure adjustment amount obtained is used as the optimization pressure adjustment amount.

[0048] In this optional embodiment, the ultimate goal of control is in the next cycle. This ensures that all defect patterns are completely eliminated, i.e., the defect state vector is... It becomes a zero vector. Based on the established linear system model, , to the target This results in a system of linear equations that needs to be solved:

[0049]

[0050] In practice, matrix The system of equations is usually not a square matrix, or even if it is, it may be singular. Therefore, the Moore-Penrose pseudoinverse can be used to solve the least-norm least-squares solution of the system of equations. This ensures that even in complex cases, a unique and physically reasonable solution can be obtained. The optimal pressure adjustment satisfies the following relationship:

[0051]

[0052] in, To find the optimal pressure adjustment amount, for The false rebellion.

[0053] For example, suppose the current defect vector Elimination is required; the updated matrix should be used. Its determinant is approximately 1. To simplify the calculation demonstration, it is approximated as an identity matrix (in actual computer control, a precise pseudo-inverse is used for solving). The pseudo-inverse... It is also approximately an identity matrix, and the calculated adjustment amount This means that to eliminate a positive deviation of 0.5, the first pressure parameter needs to be reduced by 0.5; to eliminate a negative deviation of -0.3, the second pressure parameter needs to be increased by 0.3.

[0054] Furthermore, since there is always an error between the model and the actual physical process, a damping factor is introduced in this embodiment to enhance the stability and smoothness of the system. For example, a value of 0.8 represents the target stamping pressure parameter for the next cycle. Satisfying the relation:

[0055]

[0056] in, The target stamping pressure parameter can be used to coordinately adjust all stamping pressure parameters; These are the stamping pressure parameters for the current stamping cycle. The damping factor, with a value between 0 and 1, is used to scale the optimization pressure adjustment to improve the stability of the control process.

[0057] Thus, through global optimization based on the latest coupling model and combined with a damping mechanism, a set of optimal cooperative control commands can be generated that can effectively approach the control target while taking into account system stability.

[0058] S104 sets the target stamping pressure parameter for the next stamping cycle based on the optimization pressure adjustment amount, thereby achieving adaptive control of the operating status of the aluminum plate stamping equipment through periodic iteration.

[0059] In one embodiment, the calculated target pressure parameter vector Through industrial Ethernet, for example, the programmable logic controller of the stamping equipment is sent to the OPC UA protocol for execution. The equipment adjusts the output of each hydraulic cylinder according to the new set value to stamp the next aluminum plate. The system will automatically repeat the complete process, forming an adaptive closed-loop control system that continuously learns and optimizes.

[0060] like Figure 2 The diagram illustrates the defects of an unoptimized initial product. It shows that the main defect areas are concentrated in the warped areas (represented by red) with higher peaks, and the recessed areas (represented by blue) with deeper depressions. Figure 3 The diagram shown is a product after collaborative control correction according to an embodiment of this application. It can be seen that after collaborative control correction, the defect status of the next product is significantly improved, and the entire surface becomes very flat. Whether it is the initial warping or depression, the deviation is effectively suppressed.

[0061] In this way, by translating optimization decisions into actual equipment control actions and continuously cycling through the entire process, a fully automated process from quality monitoring to state optimization is achieved, significantly reducing reliance on human experience and ensuring long-term stability and high yield in aluminum plate stamping production.

[0062] It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the scope of protection of this application. Therefore, the scope of protection of this patent application shall be determined by the appended claims.

Claims

1. A method for monitoring the operating status of an aluminum plate stamping equipment, characterized in that, include: The three-dimensional surface data of the aluminum plate in the current stamping cycle is obtained, multiple local defect patterns are identified, and a defect state vector representing the current overall operating state is determined. The state coupling influence matrix is ​​updated online. This matrix describes the cross-influence of multiple adjustable stamping pressure parameters on local defect modes. The update is based on the actual defect changes caused by the actual parameter changes within adjacent stamping cycles. These changes are compared with the defect changes predicted by the state coupling influence matrix to obtain the prediction error. The state coupling influence matrix is ​​then corrected based on this prediction error. The update of the state coupling influence matrix satisfies the following relationship: in, The updated state coupling effect matrix, The state coupling effect matrix before the update. For learning rate, For prediction error, This represents the actual change in the stamping pressure parameters. Let it be its transpose vector. The square of its Euclidean norm To prevent the division of tiny positive numbers with a denominator of zero; Based on the updated state coupling influence matrix and the current defect state vector, with the goal of eliminating the current defect, the optimal pressure adjustment amount that can coordinately adjust all stamping pressure parameters is globally calculated. The global solution involves constructing a system of linear equations aimed at eliminating the current defect. ;in, The updated state coupling effect matrix, This represents the defect state vector for the current stamping cycle. The parameter adjustment amount is to be solved; the optimal pressure adjustment amount is obtained by solving the linear equation system. The target stamping pressure parameters for the next stamping cycle are set based on the optimization pressure adjustment amount, thereby achieving adaptive control of the operating status of the aluminum plate stamping equipment through periodic iteration.

2. The method for monitoring the operating status of an aluminum plate stamping equipment according to claim 1, characterized in that, The method for obtaining the three-dimensional surface data is as follows: Three-dimensional point cloud data of the surface of stamped aluminum plate were acquired using a structured light 3D vision system. The three-dimensional point cloud data is aligned and registered with a preset standard CAD model, and the normal distance between each data point in the three-dimensional point cloud data and its corresponding position on the standard CAD model is calculated to form the three-dimensional surface data.

3. The method for monitoring the operating status of an aluminum plate stamping equipment according to claim 1, characterized in that, The identification of multiple local defect patterns is achieved by performing cluster analysis on the three-dimensional surface data.

4. The method for monitoring the operating status of an aluminum plate stamping equipment according to claim 3, characterized in that, The clustering analysis employs a density-based noise-applied spatial clustering algorithm to aggregate points that are spatially adjacent and have the same deviation direction into defect clusters. Each defect cluster corresponds to a local defect pattern.

5. The method for monitoring the operating status of an aluminum plate stamping equipment according to claim 1, characterized in that, The average deviation of the local defect mode satisfies the following relationship: in, For the first A defect cluster corresponding to a local defect pattern This represents the number of data points contained in the defect cluster. For data points in the defect cluster, For data points The surface deviation value.

6. The method for monitoring the operating status of an aluminum plate stamping equipment according to claim 1, characterized in that, The prediction error satisfies the following relationship: in, and These are the defect state vectors for the current and previous stamping cycles, respectively. and These are the stamping pressure parameter vectors for the current and previous stamping cycles, respectively.

7. The method for monitoring the operating status of an aluminum plate stamping equipment according to claim 1, characterized in that, The optimization pressure adjustment amount satisfies the following relationship: in, The optimization pressure adjustment amount. for The false rebellion.

Images