Intelligent control method and system for mold production process

Abstract

The application relates to the technical field of intelligent control, and discloses an intelligent control method and system for a mold production process. The method comprises the following steps: collecting material characteristic distribution data of a plate material, and constructing an initial material characteristic map; simulating the deformation behavior of the plate material under a forming load, and establishing a deformation behavior model; if it is detected that the stress exceeds a preset threshold value in the deformation behavior model, adjusting simulation parameters, obtaining a preliminary mold pose correction value, integrating a real-time coupling algorithm to calculate a dynamic pose parameter, generating an adaptive optimization path of the mold, and evaluating feasibility; processing the adaptive optimization path through a neural network to generate a refined mold pose sequence, and driving an actuator to adjust the position and posture of the mold in real time, monitoring the plate material forming process to determine a quality stability index; if the index shows that defects are reduced, recording the current real-time coupling configuration, and obtaining a control parameter configuration set. The application improves the control accuracy of the mold in the plate material forming process.

Description

Technical Field

[0001] This application relates to the field of intelligent control technology, and in particular to an intelligent control method and system for mold production processes. Background Technology

[0002] In traditional mold manufacturing processes, especially in sheet metal forming, mold adjustments typically rely on fixed design parameters or empirical rules. However, the material properties of sheet metal, such as thickness, hardness, and tensile strength, can vary significantly across different regions. These differences can lead to localized stress concentrations during the forming process, thereby affecting the quality of the finished product. Traditional methods fail to effectively and dynamically correlate the actual properties of the sheet metal with mold adjustments, potentially resulting in forming defects such as cracks, wrinkles, or dimensional inaccuracies.

[0003] With the development of industrial automation and intelligent technologies, utilizing sensors to collect sheet metal characteristic data and combining it with finite element analysis for simulation has become a trend in optimizing mold production processes. However, current technologies still face some challenges. First, the various characteristics of sheet metal, such as hardness, thickness, and temperature, have complex influences on mold adjustment, a point often overlooked by existing technologies, leading to inaccurate path optimization. Second, traditional path optimization methods fail to fully consider the physical limitations of machines, such as actuator response time and mold precision, thus limiting the practical feasibility of optimization effects. To overcome these shortcomings, in recent years, intelligent mold control methods based on real-time feedback and intelligent algorithms such as neural networks and morphological processing have gradually become a research hotspot. By accurately simulating the behavior of sheet metal during the forming process and combining it with dynamic adjustment of control parameters such as mold position and pressure, the limitations of traditional methods can be effectively addressed, significantly improving the quality and efficiency of mold production.

[0004] Therefore, how to achieve more efficient and precise mold adjustment through intelligent control has become an important direction for the technological development in this field. Summary of the Invention

[0005] To address the aforementioned technical problems, this application provides an intelligent control method and system for the mold production process, which improves the control accuracy in the sheet metal forming mold production process.

[0006] In a first aspect, this application provides an intelligent control method for a mold manufacturing process, the method comprising:

[0007] Step S1: Collect material property distribution data of the sheet metal through a sensor array to construct an initial material property map;

[0008] Step S2: Based on the initial material property map, simulate the deformation behavior of the sheet metal under forming load and establish a deformation behavior model;

[0009] Step S3: If stress exceeds a preset threshold is detected in the deformation behavior model, the simulation parameters are adjusted to obtain preliminary mold pose correction values;

[0010] Step S4: After obtaining the preliminary mold pose correction value, integrate the real-time coupling algorithm to calculate the dynamic pose parameters, generate the mold's adaptive optimization path, and evaluate the feasibility of the adaptive optimization path;

[0011] Step S5: The adaptive optimization path is processed by a neural network to generate a refining die pose sequence, and the actuator is driven to adjust the die position and posture in real time to monitor the sheet metal forming process and determine quality stability indicators.

[0012] Step S6: If the quality stability index shows a reduction in defects, record the current real-time coupling configuration and obtain the control parameter configuration set.

[0013] Secondly, this application provides an intelligent control system for a mold production process, the system comprising:

[0014] The data acquisition module is used to collect material property distribution data of the sheet metal through a sensor array and construct an initial material property map.

[0015] The deformation simulation module is used to simulate the deformation behavior of the sheet metal under forming load based on the initial material property map and to establish a deformation behavior model.

[0016] The pose adjustment module is used to adjust the simulation parameters and obtain the initial mold pose correction value if the stress detected in the deformation behavior model exceeds the preset threshold.

[0017] The path generation module is used to obtain the preliminary mold pose correction value, integrate the real-time coupling algorithm to calculate the dynamic pose parameters, generate the adaptive optimization path of the mold, and evaluate the feasibility of the adaptive optimization path.

[0018] The monitoring module is used to process the adaptive optimization path through a neural network, generate a refining die pose sequence, and drive the actuator to adjust the die position and posture in real time, thereby monitoring the sheet metal forming process and determining quality stability indicators.

[0019] The control module is used to record the current real-time coupling configuration and obtain a set of control parameter configurations if the quality stability index shows a reduction in defects.

[0020] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0021] This application provides an intelligent control method and system for mold manufacturing processes. First, a sensor array collects material property data of the sheet metal and constructs an initial material property map, enabling the mold to accurately reflect the physical differences in the sheet metal during forming, avoiding defects caused by uneven thickness and hardness variations. Second, in the deformation behavior simulation stage, finite element analysis is used to accurately predict the stress distribution and potential defect locations of the sheet metal, helping to identify stress concentration areas and thus optimizing the mold adjustment path, reducing problems such as cracks or wrinkling caused by excessive stress.

[0022] After obtaining the initial mold pose correction values, the real-time coupling algorithm calculates dynamic pose parameters based on hardness variation data and sheet metal physical properties, enabling real-time adjustment of the mold path to ensure the mold can adapt to changes in the sheet metal during actual production. Through neural network processing to optimize the path, the mold pose sequence is further refined, resulting in more precise mold adjustments and reduced forming defects. Finally, quality stability indicators are calculated using real-time monitoring feedback data. If defects decrease, the effectiveness of the adjustment is verified, and the optimized mold pose configuration is recorded as a reference template, forming a continuously improving intelligent production closed loop to ensure stable and efficient mold adjustments in subsequent production cycles. This method effectively improves the consistency of forming quality, reduces production defects, optimizes the mold path, and enhances production efficiency and quality stability. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram of one embodiment of the intelligent control method for the mold production process in this application.

[0025] Figure 2 This is a schematic diagram of the thickness distribution in the initial material property map of this application embodiment;

[0026] Figure 3 This is a schematic diagram of the hardness distribution in the initial material property map of the embodiments of this application;

[0027] Figure 4 This is a flowchart illustrating the feasibility assessment of the adaptive optimization path in the embodiments of this application;

[0028] Figure 5 This is a schematic diagram of an embodiment of an intelligent control system for the mold production process in this application. Detailed Implementation

[0029] This application provides an intelligent control method and system for a mold production process. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0030] In mold manufacturing processes, such as the stamping of high-end components like automotive body panels and aerospace skins, the final quality and production efficiency highly depend on the precise control of sheet metal deformation by the mold. Traditional mold control uses preset, fixed motion trajectories and blank holder forces, which cannot cope with the uneven distribution of thickness and hardness caused by batch variations in sheet metal. This material inhomogeneity can lead to inconsistent results when forming different sheet metals with the same set of mold parameters. Sometimes the mold produces qualified parts, while at other times it may crack due to excessive local stress or wrinkle due to uneven material flow. This forces frequent production line shutdowns for adjustments, severely hindering the realization of intelligent and stable production. Therefore, this application provides an intelligent control method for the mold manufacturing process. By collecting sheet metal characteristic data in real time to construct a material map, and combining finite element simulation and intelligent algorithms, it dynamically predicts and optimizes the mold pose adjustment path, effectively improving the quality consistency of complex sheet metal forming, reducing the defect rate, and constructing a continuously improving intelligent production closed loop through self-learning optimization of production data.

[0031] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the intelligent control method for the mold production process in this application includes:

[0032] Step S1: Collect material property distribution data of the sheet material through a sensor array to construct an initial material property map.

[0033] The step S1 of constructing the initial material property map includes: collecting thickness unevenness data and hardness variation data of each region of the sheet as material property distribution data; generating a numerical matrix of material property distribution based on the collected thickness unevenness data and hardness variation data; constructing an initial material property map based on the numerical matrix, which includes the thickness value and hardness value of each region of the sheet; identifying abnormal regions of material property distribution through the initial material property map, and adjusting the accuracy of the initial material property map based on the abnormal regions.

[0034] Specifically, sheet metal refers to sheet-shaped raw materials of metal or other materials used in the mold production process. It is commonly used to manufacture key components in the automotive, aerospace, and home appliance industries. Sheet metal typically possesses physical properties such as uniform or non-uniform thickness, hardness, and tensile strength. During manufacturing, these physical properties are crucial to the mold forming process, influencing mold design and forming quality. To accurately simulate the forming behavior of sheet metal and optimize the mold path, this invention first collects material property distribution data of the sheet metal using a sensor array to construct an initial material property map. Specifically, the sensor array consists of multiple ultrasonic thickness sensors and micro-indentation hardness sensors arranged in a grid above the sheet metal. This allows for real-time collection of material property distribution data in different areas of the sheet metal, including thickness unevenness data and hardness variation data. The ultrasonic thickness sensors measure the thickness differences in various areas of the sheet metal with an accuracy of ±0.02 mm, reflecting the thickness distribution of the sheet metal in different regions. The micro-indentation hardness sensors collect hardness variation data, reflecting the hardness differences of the sheet metal and affecting its deformation capacity during the forming process. The collected discrete data are fused using a spatial interpolation algorithm (such as Kriging) to generate a numerical matrix, for example, 20×20. Each element of this matrix corresponds to the thickness and hardness values ​​of a region of the sheet metal. Based on this matrix, a visualized initial material property map is constructed, such as... Figure 2 The image shows a schematic diagram of the thickness distribution in the initial material property map. This grayscale map clearly shows the thickness distribution within a 300mm × 200mm area on the sheet surface. The thickness values ​​vary between 1.0mm and 2.0mm, presented intuitively through different grayscale gradients. The image clearly identifies two typical thickness anomaly areas: one is a thickness anomaly area (thicker), located near coordinates (180, 120), characterized by a locally concentrated area of ​​darker grayscale, with an average thickness of approximately 1.75mm; the other is a thickness anomaly area (thinner), located near coordinates (80, 50), characterized by a locally concentrated area of ​​lighter grayscale, with an average thickness of approximately 1.2mm. Figure 3The image shows a schematic diagram of the hardness distribution in the initial material property map. This grayscale map clearly displays the hardness distribution within a 300mm × 200mm area on the sheet surface, with hardness values ​​varying between HV100 and HV240, presented intuitively through different grayscale gradients. Two typical hardness anomaly areas are clearly identified in the image: one is a hardness anomaly area (higher), located near coordinates (40, 140), characterized by a locally concentrated area of ​​darker grayscale, with an average hardness of approximately HV220; the other is a hardness anomaly area (lower), located near coordinates (200, 30), characterized by a locally concentrated area of ​​lighter grayscale, with an average hardness of approximately HV140. Through these measures, the final generated initial material property map can provide accurate physical property data of the sheet material, ensuring the accuracy of subsequent deformation behavior simulation and providing a reliable basis for optimizing the mold adjustment path.

[0035] Subsequently, the system automatically identifies abnormal areas in the map. First, a threshold filtering method is applied to identify abnormal areas in the initial material property map. These abnormal areas typically exhibit significant deviations in sheet metal thickness or hardness. Based on set thresholds, such as a thickness deviation exceeding 10% or a hardness deviation exceeding 15%, these abnormal areas are marked. This allows for early identification of areas that may affect forming quality, providing a basis for subsequent accuracy adjustments. For these abnormal areas, the accuracy of the initial material property map is adjusted according to specific circumstances. Specifically, within abnormal areas, the resolution of the local map is refined by increasing the sampling point density. For example, the sampling point density is increased from 10 points / cm² to 20 points / cm² to re-collect data and ensure clearer details in the abnormal areas. Furthermore, a local mesh refinement method is used to further improve the resolution of this area, achieving a mesh size of 0.5mm. This accuracy adjustment ensures that the details of the abnormal areas are fully captured, thus providing higher-precision input data for subsequent deformation simulations. Through these measures, the final generated initial material property map provides accurate physical property data of the sheet metal, ensuring the accuracy of subsequent deformation behavior simulations and providing a reliable basis for optimizing the mold adjustment path.

[0036] Step S2: Based on the initial material property map, simulate the deformation behavior of the sheet metal under forming load and establish a deformation behavior model.

[0037] In step S2, the deformation behavior of the sheet metal under forming load is simulated, and the deformation behavior model is established by: extracting material parameters of each region of the sheet metal from the initial material property map as input for finite element analysis; simulating the deformation of the sheet metal under forming load using finite element analysis to obtain stress distribution data of each region; predicting the location of potential forming defects based on the stress distribution data; and constructing a deformation behavior model based on the predicted location of potential forming defects and stress distribution data.

[0038] Specifically, the deformation behavior of sheet metal during the forming process is not only affected by its initial material properties, but also closely related to external loads, mold shape, and pressure. In order to accurately predict potential defects such as cracks and wrinkles during the forming process, it is necessary to simulate the deformation behavior of sheet metal under forming loads. This application uses finite element analysis to combine the material properties of sheet metal with forming loads, calculates the stress distribution in each region, and identifies possible stress concentration points. This allows for more accurate prediction of the location of potential defects and provides a reliable basis for optimizing the mold path, thereby minimizing forming defects and improving product quality.

[0039] Specifically, the simulation process first extracts material parameters for each region of the sheet metal from an initial material property map. These parameters include sheet thickness, hardness, tensile strength, etc. This data is used as input for finite element analysis (FEM) to simulate the deformation behavior of the sheet metal under forming loads. Specifically, FEM calculates the stress distribution in each region of the sheet metal under different forming loads. This process accurately reflects the stress concentration and strain distribution in different regions of the sheet metal, helping to assess the sheet metal's load-bearing capacity and deformation trend. Using the stress distribution data obtained from FEM, stress thresholds are set to identify high-risk areas where stress exceeds the threshold. By analyzing these stress concentration areas, potential forming defects such as cracks and wrinkles can be predicted. Specifically, cracks typically occur in areas of stress concentration, especially in weak areas of the sheet metal, such as thinner or less hard sections, while wrinkles usually appear in compression areas. Based on the coordinates of the high-risk areas, combined with the material's physical properties, such as yield strength and fracture toughness, the exact locations of cracks and wrinkles are further determined. For example, if uneven sheet metal thickness leads to excessive local stress in a certain area, cracks may appear in that area; while wrinkles may appear in pressure areas. Identifying these potential defect locations provides a basis for mold adjustment, optimizing the mold path to reduce stress concentration and defect occurrence, thereby improving forming quality. Based on the predicted potential forming defect locations and stress distribution data, a deformation behavior model is constructed. This model comprehensively considers the material properties, stress distribution, and deformation behavior of the sheet metal. It integrates stress distribution data obtained through finite element analysis to ultimately form a model describing the overall deformation of the sheet metal during forming. In addition, the model extracts various physical property data of the sheet metal, including thickness, hardness, and tensile strength, and combines them with stress distribution data to simulate the deformation process of the sheet metal under forming loads. Through detailed analysis of stress concentration in each area, the model can accurately predict potential forming defect areas, such as cracks and wrinkles, especially in areas with excessive local stress. It reflects the overall deformation trend of the sheet metal, helping to identify areas with a high risk of defects, thus providing a basis for subsequent mold path optimization.

[0040] Step S3: If stress exceeds the preset threshold is detected in the deformation behavior model, the simulation parameters are adjusted to obtain the preliminary mold pose correction value.

[0041] The step S3, obtaining the preliminary mold pose correction value, includes: analyzing the stress values ​​of each region using a deformation behavior model and comparing them with a preset threshold; if the comparison results show that the stress values ​​exceed the preset threshold, identifying regions with uneven thickness and modifying the geometric parameters of the finite element analysis based on these regions; resimulating the deformation behavior using the modified geometric parameters to obtain an optimized stress distribution; calculating the preliminary mold pose correction value using the optimized stress distribution, where the preliminary mold pose correction value represents the initial adjustment amount of the mold position; determining whether the preliminary mold pose correction value meets the requirements of the geometric matching scheme; if not, iteratively adjusting the simulation parameters until the preliminary mold pose correction value meets the requirements.

[0042] Specifically, in order to further determine the potential problem areas of the sheet metal, it is necessary to identify areas where the sheet metal has uneven thickness or large changes in hardness during the stress process, because these areas are prone to local stress concentration, which can lead to defects. Therefore, this application uses a deformation behavior model to analyze the stress values ​​of each area and compare them with preset thresholds to identify areas where the stress exceeds the standard. Based on these areas, the simulation parameters are adjusted to optimize the mold pose. In this way, potential problem areas can be identified in a timely manner, and preliminary mold pose correction values ​​can be calculated, thereby providing a precise adjustment basis for the optimization of the mold path and the subsequent forming process.

[0043] Specifically, the process begins by scanning each region in the deformation behavior model point by point to obtain the stress value for each region. These stress values ​​are then compared with preset stress thresholds. If the stress value in a region exceeds the threshold, that region is identified as a potential problem area. This process effectively identifies areas in the sheet metal prone to defects, providing foundational data for subsequent adjustments. Once regions with excessive stress are identified, areas with uneven thickness are further identified. By extracting the coordinates of the stress points exceeding the threshold from the deformation behavior model and mapping these coordinates to the initial material property map, the thickness data for the corresponding region is queried. If the thickness of this region deviates from the overall average thickness by more than a preset threshold, such as 0.2 mm, it is marked as an area with uneven thickness. Such areas often lead to localized stress concentration, amplifying the imbalance in stress distribution. Therefore, identifying these areas is crucial for subsequent mold adjustments and forming process optimization.

[0044] After identifying regions with uneven thickness, their geometric parameters are adjusted. In finite element analysis, geometric parameters typically include mesh size, density, and node distribution. Boundary data for these uneven regions is identified, and based on this data, the finite element mesh model for that region is adjusted. For example, increasing the mesh density to make the mesh finer allows for a more accurate simulation of the deformation behavior in these regions. This adjustment helps improve simulation accuracy and ensures optimized stress distribution. After modifying the geometric parameters, the finite element analysis is rerun, and optimized stress distribution data is output based on the updated geometric parameters and material properties. Based on the optimized stress distribution data, preliminary mold pose correction values ​​are calculated. These values ​​represent the initial adjustment required for the mold during the forming process. Specifically, displacement vectors are extracted from high-stress areas, which are regions where stress values ​​exceed a preset threshold. A reverse calculation method is used to convert these displacement vectors into mold position offsets. For example, if the stress in a high-stress area is concentrated at the intersection of the X and Y axes, the mold adjustment amounts on the X and Y axes are calculated based on the stress distribution at that location, such as a 0.5mm X-axis offset and a 0.3mm Y-axis offset. This calculation provides a precise reference for mold adjustment. Subsequently, the calculated preliminary mold pose correction values ​​are further verified to ensure they meet the set geometric matching requirements. Specifically, it is determined whether the adjusted displacement is within a preset tolerance range, such as whether the offset does not exceed 1mm. If the requirements are met, the process proceeds to the subsequent optimization stage.

[0045] If the initial mold pose correction value does not meet the geometric matching requirements, the simulation parameters are iteratively adjusted until the correction value meets the requirements. During this process, simulation parameters such as mesh density and load step size may be adjusted to further optimize the simulation results. After each adjustment, the simulation is rerun to calculate the new stress distribution and recalculate the mold pose correction value. This process continues until the initial mold pose correction value meets the requirements of the geometric matching scheme. By analyzing the stress distribution in the deformation behavior model, high-stress areas are identified, leading to problems such as uneven thickness. A refined finite element analysis method is used to adjust the simulation parameters, optimize the stress distribution, and calculate the initial mold pose correction value. This series of steps effectively identifies potential defects, optimizes the mold adjustment path, and ensures that defects are minimized during the forming process, improving production efficiency and product quality.

[0046] Step S4: After obtaining the initial mold pose correction value, integrate the real-time coupling algorithm to calculate the dynamic pose parameters, generate the mold's adaptive optimization path, and evaluate the feasibility of the adaptive optimization path.

[0047] In step S4, generating the adaptive optimization path for the mold includes: fusing the initial mold pose correction value with the hardness change data obtained from the initial material property map, and inputting it together with the sheet thickness change and tensile strength into the real-time coupling algorithm. The real-time coupling algorithm processes the input data, calculates the dynamic pose parameters of the mold during the forming process, arranges the dynamic pose parameters at multiple time points, and generates the adaptive optimization path for the mold.

[0048] Specifically, to ensure precise mold adjustment and improve forming quality during the forming process, dynamic pose parameters must be calculated in real-time using a coupled algorithm, based on the initial mold pose correction values. This algorithm allows for real-time adjustment of the mold position and orientation according to actual changes in the sheet metal, such as uneven thickness and hardness variations, to adapt to different forming conditions. The generated adaptive optimized path not only helps reduce defects caused by stress concentration but also maximizes the optimization of the mold adjustment path, ensuring the smooth progress of the entire forming process.

[0049] Specifically, after obtaining the initial mold pose correction value, it is fused with the hardness change data collected from the initial material property map to form an input dataset. This dataset also includes sheet thickness variation and tensile strength data. This information serves as input to the real-time coupling algorithm. The algorithm calculates dynamic pose parameters based on the input data. These parameters reflect the real-time adjustments of the mold during sheet forming, ensuring the mold can accurately adapt to the physical changes of the sheet and optimize the forming path. Specifically, the real-time coupling algorithm uses a Kalman filter to process the input data. The Kalman filter first calculates the mold's state prediction value through a prediction step, i.e., the initial position and orientation of the mold at the current moment. This initial state prediction value is derived from the mold's position and orientation at the previous time point, as well as the applied load or mechanical environment. For example, based on the mold's displacement and rotation angle at the previous moment, the filter uses physical models, such as the contact state between the mold and the sheet, and material properties, to predict the possible position and orientation of the mold at the current moment. In this way, the Kalman filter can provide a preliminary prediction value for the mold position at each time step, serving as the basis for subsequent updates and optimizations. The next step involves updating the Kalman filter. Specifically, this involves combining the actual hardness change data with the preliminary mold pose correction value to refine the previous state prediction, resulting in more accurate dynamic pose parameters. The Kalman filter uses a weighted average method to update the previously predicted mold pose by combining the hardness change data. The hardness change data reflects the physical changes in the sheet metal during the forming process and may affect the mold's adjustment path. Through Kalman gain, the model's prediction accuracy is adjusted based on the reliability of the hardness data. The updated mold state is defined as dynamic pose parameters, including new translational amounts (displacement and rotation angles in the X, Y, and Z directions), such as rotational adjustments expressed using Euler angles or quaternions. These new dynamic pose parameters more accurately reflect the mold's real-time adjustments to the sheet metal during forming, providing accurate mold displacement and rotation information. This provides a precise adjustment reference for subsequent mold path optimization, ensuring the mold adapts to the physical changes in the sheet metal. The dynamic pose parameters at each time point reflect the mold's adjustment requirements for adapting to the sheet metal's physical properties at different stages. By arranging the dynamic pose parameters at multiple time points, an optimized adaptation path for the mold is generated. This optimized path considers factors such as changes in sheet metal hardness, thickness unevenness, and tensile strength during the forming process, ensuring that the mold adjustment can cope with these changes and thus avoid defects such as stress concentration. Next, the feasibility of the optimized adaptation path is determined by evaluating these dynamic pose parameters, because not all optimized paths can be successfully executed in the actual forming process. Some paths may be infeasible due to mold physical limitations, equipment response speed, or differences in sheet metal characteristics. For example, some paths may cause excessive mold displacement, exceeding the equipment's adjustment range, or generate excessive stress in specific areas, leading to defects.Therefore, by evaluating dynamic pose parameters, we can ensure that the optimized path can be effectively executed in actual operation and avoid potential problems caused by infeasible paths.

[0050] The feasibility assessment of the adaptive optimization path in step S4 includes: simulating the execution of the adaptive optimization path in a simulation environment and obtaining stress distribution data from the simulation results; assessing the feasibility of the adaptive optimization path based on the stress distribution data; if there are stress values ​​in the stress distribution data that exceed a preset threshold, it is determined that there is a path deviation, and the coupling coefficient in the real-time coupling algorithm is modified, and the adaptive optimization path is regenerated using the modified real-time coupling algorithm. The simulation and assessment steps are repeated until a feasible adaptive optimization path is obtained.

[0051] Specifically, assessing the feasibility of the adaptive optimization path includes the following steps: such as Figure 4 As shown in the flowchart for assessing the feasibility of the adaptive optimization path, the adaptive optimization path is first simulated in a simulation environment, and stress distribution data from the simulation results are obtained. This process is initially completed using finite element analysis or other numerical simulation methods. Based on the path generated from the initially calculated dynamic pose parameters, the movement of the mold during the actual forming process is simulated, and the stress distribution of the sheet metal is evaluated. The simulation environment can accurately reflect the stress situation of the sheet metal under the mold adjustment path, especially how mold adjustment affects the stress distribution of the sheet metal under the influence of complex factors such as hardness changes and thickness unevenness. Based on the stress distribution data in the simulation results, it is analyzed whether there are stress values ​​exceeding a preset threshold. If the stress value exceeds the preset threshold, it is determined that the path has a deviation, that is, the mold path has failed to effectively avoid problems such as stress concentration or uneven deformation, which may lead to forming defects such as cracks and wrinkles. Therefore, it is necessary to further adjust the optimization path based on these data.

[0052] After path deviation is detected, the coupling coefficient in the real-time coupling algorithm is modified. Specifically, the coupling coefficient, a crucial parameter in the Kalman filter, determines the weight between hardness variation data and mold pose correction values ​​during mold adjustment. By increasing the process noise covariance or adjusting the ratio of measurement noise covariance, the algorithm prioritizes hardness variation data, thereby improving path optimization accuracy and ensuring more precise and detailed path adjustment. The modified real-time coupling algorithm regenerates the adaptive optimized path and performs simulation again to calculate new stress distribution data. This process is iterative, adjusting the coupling coefficient, updating the path, and using new simulation data for feedback until a feasible adaptive optimized path is obtained. After each simulation and evaluation, the feasibility of the path is assessed to confirm whether it meets all stress and deformation requirements until the path fully conforms to the preset standards, ensuring that no new defects are introduced after path optimization. Through this iterative and optimization process, the final adaptive optimized path not only solves the stress concentration problem but also maximizes the accuracy of the mold adjustment path, ensuring high quality and efficiency in the forming process.

[0053] Step S5: Through neural network processing, an adaptive optimization path is generated to produce a refining mold pose sequence, and the actuator is driven to adjust the mold position and posture in real time to monitor the sheet metal forming process and determine quality stability indicators.

[0054] The step S5, generating the refining mold pose sequence, includes: acquiring historical forming defect data and inputting it, along with hardness change data and adaptive optimization path, into a neural network model; the neural network model learns the real-time coupling relationship between hardness change data and dynamic pose parameters, obtains the refining mold pose sequence based on the real-time coupling relationship, and outputs it, wherein the refining mold pose sequence includes the temporal adjustment value of the mold position; it is determined whether the refining mold pose sequence reduces the number of potential defects, and if so, the parameters of the neural network model are optimized, and the final refining mold pose sequence is output through the optimized neural network.

[0055] Specifically, in order to further optimize the mold adjustment path, i.e. the adaptive optimization path, this application uses neural network processing to generate a refining mold pose sequence, and uses this sequence to drive the actuator to adjust the mold pose in real time. Through this process, the mold adjustment path can be controlled more precisely, ensuring that defects are minimized and forming quality is improved during the sheet metal forming process.

[0056] Specifically, the process begins by acquiring historical forming defect data, which, along with hardness variation data and an adaptive optimization path, is input into the neural network model. The historical forming defect data includes the types, locations, and severity of defects that occurred during past forming processes, helping the neural network understand which factors, such as hardness variations and thickness unevenness, are prone to causing defects. Hardness variation data reflects the hardness distribution in different areas of the sheet metal. The adaptive optimization path is a mold adjustment path obtained through a real-time coupling algorithm, reflecting the mold's adjustment status at different time points. The neural network calculates the weights of the relationship between the input data and the dynamic pose of the mold through feedforward propagation. This step allows the neural network to understand how different sheet metal characteristics, such as hardness and thickness unevenness, affect the mold adjustment path. After training, the network can capture how hardness variations affect the translation and rotation of the mold, obtaining the real-time coupling relationship between hardness variations and dynamic pose parameters. Based on the learned coupling relationship, the network predicts the dynamic pose adjustment of the mold according to hardness variations and sheet metal characteristics. For example, in areas with higher hardness, the mold may need to adjust a smaller displacement, while in areas with thinner thickness, the mold may need a larger adjustment. In this way, the pose sequence generated by the neural network can accurately adapt to changes in the sheet metal. By optimizing the weights of the neural network through backpropagation, the neural network adjusts the calculated dynamic pose parameters to generate a refined mold pose sequence. This sequence contains the adjustment values ​​of the mold at every moment in the entire forming process, such as displacement and rotation angle, ensuring that the mold can accurately respond to the physical properties of the sheet metal at every stage of the forming process.

[0057] After generating the refining die pose sequence, it is further determined whether the sequence reduces the number of potential defects. This determination is made by comparing the number of simulated defects before and after optimization. If the refining die pose sequence can effectively reduce the number of potential defects during the forming process, such as cracks and wrinkles, the sequence is considered effective. At this point, the parameters of the neural network model are optimized by further adjusting the learning rate and weight biases to improve the accuracy and stability of path optimization. The optimization process uses the gradient descent algorithm to adjust the learning rate and weights of the network. Specifically, the amount of defect reduction is calculated as a reward signal. By optimizing the parameters of the neural network, its responsiveness to die adjustment paths is enhanced. During the optimization process, the network gradually updates the weights to reduce the number of defects in the simulation results, ensuring that the refining die pose sequence can effectively cope with the physical changes of different sheet materials.

[0058] The optimized neural network outputs the final refining die pose sequence, which is then used as input data for the actuator. The actuator adjusts the die's position and orientation based on this optimized sequence, ensuring the die accurately responds to changes in the sheet metal throughout the forming process, avoiding excessive stress or uneven deformation. Furthermore, it monitors quality stability indicators in real time, such as the number of defects and dimensional accuracy, using feedback data to ensure efficient and stable production. Ultimately, through this series of optimizations and adjustments, the system can adjust the die path in real time according to changes in the sheet metal during each forming cycle, reducing defects, improving production efficiency, and ensuring consistent final product quality.

[0059] In step S5, monitoring the sheet metal forming process to determine quality stability indicators includes: generating control signals for the actuator based on the refining die pose sequence; adjusting the actual position and posture of the die based on the control signals; monitoring feedback data of the sheet metal forming process, wherein the feedback data includes real-time indicators of forming defects; and calculating quality stability indicators based on the number of defects in the feedback data.

[0060] Specifically, after the actuator adjusts the die position and orientation in real time, the sheet metal forming process is monitored to determine quality stability indicators, ensuring quality control and minimizing defects during the forming process. First, control signals for the actuator are generated based on the refining die pose sequence. This sequence is obtained through neural network processing and adaptive optimization, encompassing a series of die position and orientation adjustments. The die position and orientation at each time point are converted into corresponding control signals, typically transmitted to the actuator in voltage or current form to ensure interface compatibility. The control signals adjust the die position by mapping the pose data to the actuator's drive commands, achieving precise die path control. Next, the actuator adjusts the actual position and orientation of the die according to the control signals. After receiving these control signals, the actuator drives a hydraulic or electric mechanism to move the die, fine-tuning the die position and angle to achieve precise forming of the sheet metal. During this process, the die position adjustment ensures optimized matching between the sheet metal geometry and the die shape, reducing defects caused by uneven forming.

[0061] While the mold is being adjusted, a sensor array collects feedback data in real time. This feedback data includes real-time indicators of forming defects, such as crack length, deformation deviation, and stress value. The defect indicators in the feedback data help to understand the impact of mold adjustment on forming quality, thereby enabling timely monitoring and handling of defects. Based on the feedback data collected from the sensors, the quality stability index is calculated by statistically analyzing the number of defects. The quality stability index mainly reflects the change in the number of defects during the forming process, especially the reduction of defects such as cracks and wrinkles. Specifically, defects in each monitoring area are counted, and by comparing the change in the number of defects before and after adjustment, it can be determined whether the mold path adjustment has effectively reduced defects. If the number of defects is significantly reduced, it indicates that the mold adjustment and path optimization have effectively improved the forming quality and verified the effectiveness of the adjustment. The current real-time coupling configuration, namely the refining mold pose sequence and the quality stability index, is recorded as a reference template and updated in the control parameter configuration set to provide data support for subsequent production.

[0062] Step S6: If the quality stability index shows a reduction in defects, record the current real-time coupling configuration and obtain the control parameter configuration set.

[0063] In step S6, obtaining the control parameter configuration set includes: analyzing and confirming the degree of defect reduction based on the quality stability index; if the defect reduction is confirmed, extracting the current real-time coupling configuration as a reference template; constructing the control parameter configuration set according to the reference template; applying the control parameter configuration set in the continuous production cycle; judging the stability of the control parameter configuration set; if the stability meets the requirements, storing the control parameter configuration set for mold control during the sheet metal forming process.

[0064] Specifically, the degree of defect reduction is confirmed based on the quality stability index, and the current real-time coupling configuration is recorded to obtain a set of control parameter configurations for subsequent production cycles and mold control. By analyzing the quality stability index, it is confirmed whether defects have been effectively reduced, and the current real-time coupling configuration is extracted based on this result to ensure that the subsequent mold adjustment path can stably improve the forming quality.

[0065] Specifically, based on the analysis of quality stability indicators, the degree of defect reduction during the forming process is confirmed. This process includes collecting the number of defects in the feedback data of the forming process and comparing it with the number of defects identified in the initial material property map. The percentage of defect reduction is calculated by dividing the difference in the number of defects by the initial number of defects. For example, in the forming process of a sheet metal, the initial material property map shows that the stress in a certain thickness uneven area exceeds the threshold, resulting in 10 potential defects. After mold adjustment, the number of defects is reduced to 5 through analysis of feedback data, a reduction percentage of 50%. This indicates that the adjustment is effective and can significantly improve the stability of forming quality. Once the defect reduction is confirmed, the current real-time coupling configuration is extracted as a reference template. The real-time coupling configuration refers to the mold adjustment path and related parameters obtained through the real-time coupling algorithm, such as the refining mold pose sequence and quality stability indicators. This configuration reflects how the mold adapts to the physical changes of the sheet metal at different stages, especially in response to defects such as stress concentration, cracks, and wrinkles. The real-time coupling configuration is generated based on input data such as mold pose correction values, hardness changes, and thickness unevenness. Based on this reference template, the system constructs a set of control parameter configurations. The final verified control parameter configuration set is applied to continuous production cycles, and the model's stability is judged by the fluctuation range of statistical quality stability indicators. Specifically, feedback data is collected in multiple production cycles, and the variance of the number of defects is calculated. If the variance is lower than a preset threshold, such as 5%, it indicates that the control parameter configuration set is stable and can continuously and effectively reduce defects and optimize the forming process. The optimized control parameter configuration set is saved to a database and associated with the initial material property map and deformation behavior model to form a closed-loop optimization process. Finally, the closed-loop optimization process will be stored and used in subsequent production cycles to ensure the stability and efficiency of the overall optimization process. Real-time coupling configurations obtained through real-time coupling algorithms record the mold adjustment path and parameters. Combined with the analysis of quality stability indicators, a control parameter configuration set is constructed and used in subsequent production cycles to ensure that the mold can adapt to the physical changes of the sheet metal, reduce defects, and guarantee the continuous stability of forming quality.

[0066] The above describes an intelligent control method for a mold production process according to embodiments of this application. The following describes an intelligent control system for a mold production process according to embodiments of this application. Please refer to [link / reference]. Figure 5 One embodiment of an intelligent control system for mold production process in this application includes:

[0067] The data acquisition module is used to collect material property distribution data of the sheet metal through a sensor array and construct an initial material property map.

[0068] The deformation simulation module is used to simulate the deformation behavior of sheet metal under forming load based on the initial material property map and to establish a deformation behavior model.

[0069] The pose adjustment module is used to adjust the simulation parameters and obtain preliminary mold pose correction values ​​if the stress detected in the deformation behavior model exceeds a preset threshold.

[0070] The path generation module is used to obtain the initial mold pose correction value, integrate the real-time coupling algorithm to calculate the dynamic pose parameters, generate the adaptive optimization path of the mold, and evaluate the feasibility of the adaptive optimization path.

[0071] The monitoring module is used to generate a refining die pose sequence by using neural network processing to adapt and optimize the path, and drive the actuator to adjust the die position and posture in real time, thereby monitoring the sheet metal forming process and determining quality stability indicators.

[0072] The control module is used to record the current real-time coupling configuration and obtain a set of control parameter configurations if the quality stability index shows a reduction in defects.

[0073] In summary, the intelligent control method and system for mold production provided in this application firstly collects material property data of the sheet metal through a sensor array and constructs an initial material property map, enabling the mold to accurately reflect the physical differences of the sheet metal during the forming process and avoiding defects caused by uneven thickness and hardness differences. Secondly, in the deformation behavior simulation stage, finite element analysis is used to accurately predict the stress distribution and potential defect locations of the sheet metal, helping to identify stress concentration areas, thereby optimizing the mold adjustment path and reducing problems such as cracks or wrinkles caused by excessive stress.

[0074] After obtaining the initial mold pose correction values, the real-time coupling algorithm calculates dynamic pose parameters based on hardness variation data and sheet metal physical properties, enabling real-time adjustment of the mold path to ensure the mold can adapt to changes in the sheet metal during actual production. Through neural network processing to optimize the path, the mold pose sequence is further refined, resulting in more precise mold adjustments and reduced forming defects. Finally, quality stability indicators are calculated using real-time monitoring feedback data. If defects decrease, the effectiveness of the adjustment is verified, and the optimized mold pose configuration is recorded as a reference template, forming a continuously improving intelligent production closed loop to ensure stable and efficient mold adjustments in subsequent production cycles. This method effectively improves the consistency of forming quality, reduces production defects, optimizes the mold path, and enhances production efficiency and quality stability.

[0075] To ensure that those skilled in the art can clearly understand the specific working process of the systems and units described above, they can refer to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

[0076] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0077] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. An intelligent control method for mold production process, characterized in that, The method includes: Step S1: Collect material property distribution data of the sheet metal through a sensor array to construct an initial material property map; Step S2: Based on the initial material property map, simulate the deformation behavior of the sheet metal under forming load and establish a deformation behavior model; Step S3: If stress exceeds a preset threshold is detected in the deformation behavior model, the simulation parameters are adjusted to obtain preliminary mold pose correction values; Step S4: After obtaining the preliminary mold pose correction value, the real-time coupling algorithm is integrated to calculate the dynamic pose parameters, generate the mold's adaptive optimization path, and evaluate the feasibility of the adaptive optimization path. The generation of the mold's adaptive optimization path in step S4 includes: fusing the preliminary mold pose correction value with the hardness change data obtained from the initial material property map, and inputting it along with the sheet thickness change and tensile strength into the real-time coupling algorithm. The real-time coupling algorithm processes the input data, calculates the dynamic pose parameters of the mold during the forming process, arranges the dynamic pose parameters at multiple time points, and generates the mold's adaptive optimization path. Step S5: The adaptive optimization path is processed by a neural network to generate a refining die pose sequence, and the actuator is driven to adjust the die position and posture in real time to monitor the sheet metal forming process and determine quality stability indicators. Step S6: If the quality stability index shows a reduction in defects, record the current real-time coupling configuration and obtain the control parameter configuration set.

2. The method according to claim 1, characterized in that, The construction of the initial material property map in step S1 includes: Data on thickness variation and hardness variation in various regions of the sheet are collected as material property distribution data; a numerical matrix of material property distribution is generated based on the collected thickness variation data and hardness variation data; an initial material property map is constructed based on the numerical matrix, including thickness and hardness values ​​in various regions of the sheet. Abnormal regions in the distribution of material properties are identified by an initial material property map, and the accuracy of the initial material property map is adjusted based on these abnormal regions.

3. The method according to claim 1, characterized in that, In step S2, the deformation behavior of the sheet metal under forming load is simulated, and a deformation behavior model is established, including: Material parameters for each region of the sheet metal are extracted from the initial material property map and used as input for finite element analysis. Finite element analysis is then used to simulate the deformation of the sheet metal under forming load to obtain stress distribution data for each region. Based on the stress distribution data, the location of potential forming defects is predicted. Based on the predicted location of potential forming defects and the stress distribution data, a deformation behavior model is constructed.

4. The method according to claim 1, characterized in that, The preliminary mold pose correction value obtained in step S3 includes: The stress values ​​of each region are compared with the preset threshold by analyzing the deformation behavior model. If the comparison results show that the stress value exceeds the preset threshold, the uneven thickness region is identified, and the geometric parameters of the finite element analysis are modified based on the uneven thickness region. The deformation behavior is re-simulated using the modified geometric parameters to obtain an optimized stress distribution. The preliminary mold pose correction value is calculated using the optimized stress distribution, where the preliminary mold pose correction value represents the initial adjustment amount of the mold position. It is then determined whether the preliminary mold pose correction value meets the requirements of the geometric matching scheme. If it does not meet the requirements, the simulation parameters are iteratively adjusted until the preliminary mold pose correction value meets the requirements.

5. The method according to claim 1, characterized in that, The feasibility assessment of the adaptive optimization path in step S4 includes: The adaptive optimization path is simulated and executed in a simulation environment, and stress distribution data in the simulation results is obtained. The feasibility of the adaptive optimization path is evaluated based on the stress distribution data. If there is a stress value in the stress distribution data that exceeds a preset threshold, it is determined that there is a path deviation. The coupling coefficient in the real-time coupling algorithm is modified, and the adaptive optimization path is regenerated using the modified real-time coupling algorithm. The simulation and evaluation steps are repeated until a feasible adaptive optimization path is obtained.

6. The method according to claim 1, characterized in that, The step S5 of generating the refining mold pose sequence includes: Historical forming defect data is acquired and input into a neural network model along with hardness change data and an adaptive optimization path. The neural network model learns the real-time coupling relationship between hardness change data and dynamic pose parameters, obtains and outputs a refining die pose sequence based on the real-time coupling relationship, wherein the refining die pose sequence includes the temporal adjustment value of the die position. It is determined whether the refining die pose sequence reduces the number of potential defects. If so, the parameters of the neural network model are optimized, and the final refining die pose sequence is output through the optimized neural network.

7. The method according to claim 6, characterized in that, The step S5, which involves monitoring the sheet metal forming process and determining quality stability indicators, includes: The actuator generates control signals based on the refining die pose sequence, and the actuator adjusts the actual position and posture of the die according to the control signals; it monitors feedback data of the sheet metal forming process, wherein the feedback data includes real-time indicators of forming defects; and it calculates quality stability indicators based on the number of defects in the feedback data.

8. The method according to claim 1, characterized in that, The control parameter configuration set obtained in step S6 includes: The degree of defect reduction is confirmed based on the analysis of the quality stability index. If the defect reduction is confirmed, the current real-time coupling configuration is extracted as a reference template. A control parameter configuration set is constructed based on the reference template. The control parameter configuration set is applied in the continuous production cycle, and the stability of the control parameter configuration set is judged. If the stability meets the requirements, the control parameter configuration set is stored for mold control in the sheet metal forming process.

9. An intelligent control system for a mold production process, used to implement the intelligent control method for a mold production process as described in any one of claims 1-8, characterized in that, The system includes: The data acquisition module is used to collect material property distribution data of the sheet metal through a sensor array and construct an initial material property map. The deformation simulation module is used to simulate the deformation behavior of the sheet metal under forming load based on the initial material property map and to establish a deformation behavior model. The pose adjustment module is used to adjust the simulation parameters and obtain the initial mold pose correction value if the stress detected in the deformation behavior model exceeds the preset threshold. The path generation module is used to obtain the preliminary mold pose correction value, integrate the real-time coupling algorithm to calculate the dynamic pose parameters, generate the mold's adaptive optimization path, and evaluate the feasibility of the adaptive optimization path. The generation of the mold's adaptive optimization path includes: fusing the preliminary mold pose correction value with the hardness change data obtained from the initial material property map, and inputting it together with the sheet thickness change and tensile strength into the real-time coupling algorithm. The real-time coupling algorithm processes the input data, calculates the dynamic pose parameters of the mold during the forming process, arranges the dynamic pose parameters at multiple time points, and generates the mold's adaptive optimization path. The monitoring module is used to process the adaptive optimization path through a neural network, generate a refining die pose sequence, and drive the actuator to adjust the die position and posture in real time, thereby monitoring the sheet metal forming process and determining quality stability indicators. The control module is used to record the current real-time coupling configuration and obtain a set of control parameter configurations if the quality stability index shows a reduction in defects.

Images