Stamping die fault decision-making system based on dynamic evolution partition

Abstract

The invention discloses a stamping die fault decision-making system based on dynamic evolution partitioning, and relates to the field of metal processing, and the system comprises a stamping control parameter obtaining module which is used for obtaining stamping control parameters; the multi-source monitoring data acquisition module is used for acquiring multi-source monitoring data; the weak point coordinate positioning module is used for positioning U weak point coordinates; the dynamic evolution partition setting module is used for setting U dynamic evolution partitions; the mold loss analysis channel setting module is used for setting a mold loss analysis channel; and the fault diagnosis decision repair module is used for carrying out stamping die fault decision based on the dynamic evolution partition. The technical problems that in the prior art, due to the fact that a diagnosis method depends on a single data source, a diagnosis result has large errors, and the optimal repairing effect is difficult to achieve are solved, the fault type and position of a mold are accurately recognized through a weak point coordinate positioning module and a dynamic evolution partition setting module, and the fault diagnosis accuracy is improved. The technical effect of providing powerful support for fault diagnosis is achieved.

Description

Technical Field

[0001] This application relates to the field of metal processing technology, specifically to a stamping die fault decision system based on dynamic evolution partitioning. Background Technology

[0002] Stamping dies play a crucial role in metal processing, enabling the efficient and accurate production of large quantities of parts. However, during the stamping process, dies are susceptible to failure and wear due to various factors such as material hardness, stamping speed, and lubrication conditions. Existing stamping die fault diagnosis and repair technologies often rely on a single data source, such as manual inspection or simple sensor data, which fails to comprehensively and accurately reflect the actual condition of the die. Due to a lack of effective analytical methods and tools, traditional diagnostic methods often exhibit significant errors in identifying the type and location of die faults. Furthermore, the lack of systematic analysis and evaluation means that traditional repair decisions are often based on experience or intuition, making it difficult to achieve optimal repair results.

[0003] In summary, existing diagnostic methods rely on a single data source, leading to significant errors in diagnostic results and making it difficult to achieve optimal repair outcomes. Summary of the Invention

[0004] Therefore, it is necessary to provide a stamping die fault decision system based on dynamic evolution partitioning to address the above-mentioned technical problems. This system can solve the technical problem that the existing diagnostic methods rely on a single data source, resulting in large errors in the diagnostic results and making it difficult to achieve the best repair effect. Through the weak point coordinate positioning module and the dynamic evolution partitioning setting module, the fault type and location of the die can be accurately identified, thus providing strong technical support for fault diagnosis.

[0005] Based on this, a stamping die fault decision system based on dynamic evolution partitioning is provided. The system includes: a stamping control parameter acquisition module for connecting to a stamping press to acquire stamping control parameters, including pressure, stamping path, and stamping speed; a multi-source monitoring data acquisition module for real-time monitoring of the stamping die using multi-source monitoring equipment to acquire multi-source monitoring data, including temperature data, pressure data, and vibration amplitude; a weak point coordinate location module for acquiring stamping die failure instances, establishing a working twin model of the stamping die, and locating U weak point coordinates; and a dynamic evolution partitioning setting module for setting the system based on the U weak point coordinates. U stamping loss micro-regions are set up, and associated with the temperature changes, pressure distributions, and vibration frequencies corresponding to the multi-source monitoring data. U dynamic evolution partitions are also set up. The die loss analysis channel setting module is used to set up an upper die loss analysis channel based on the U dynamic evolution partitions, with the protruding top of the upper die punch as the center; and to set up a lower die loss analysis channel based on the U dynamic evolution partitions, with the concave bottom of the lower die cavity as the center. The fault diagnosis decision and repair module is used to connect the upper die loss analysis channel and the lower die loss analysis channel, and, in conjunction with the stamping control parameters, to make stamping die fault decisions based on the dynamic evolution partitions.

[0006] Preferably, the system connects to a stamping die database, uses the model parameters of the stamping die as constraints, and obtains stamping die failure instances. Based on the failure duration corresponding to the stamping die failure instances, it divides them into stamping die failure instances that have exceeded their service life and stamping die failure instances that have not exceeded their service life. Using the stamping die failure instances that have exceeded their service life as positive markers and the stamping die failure instances that have not exceeded their service life as negative markers, the system performs iterative training of feature labels by referring to the working twin model of the stamping die.

[0007] Preferably, based on the failure instance of the stamping die that has not exceeded its time limit, the failure time is located; based on the failure time, abnormal features are extracted and key parameter sequences are collected; first-order difference calculation is performed through the key parameter sequences to analyze weak degradation signals, and the negative markers include weak degradation signals.

[0008] Preferably, first-order difference calculation is performed using the key parameter sequence to output first-order difference results, which include the average, standard deviation, maximum, and minimum values ​​corresponding to continuous differences; periodic characteristics are analyzed based on the first-order difference results to identify anomalies; based on the anomalies, in-depth analysis is performed using the key parameter sequence to determine early failure indicators, and the weak degradation signal is used to trigger the mining of associated features of the early failure indicators.

[0009] Preferably, the equipment operation mode of the punch press is obtained, including servo drive mode, frequency conversion drive mode, pressure closed-loop control mode, and multi-level pressure control mode; based on the equipment operation mode, a joint analysis is performed on the failure examples of the punching die that have not exceeded their expiration period, and a risk warning is issued.

[0010] Preferably, based on the failure instances of the stamping dies that have not exceeded their expiration period and the proportion of such failure instances, a stability analysis is performed on the equipment operation mode to obtain a stability index; based on the equipment operation mode, the pressure switching smoothness is calculated; feature matching is performed based on the early failure symptoms, and combined with the stability index and pressure switching smoothness, a joint analysis result is obtained.

[0011] Preferably, a stamping die repair record is obtained, which includes the number of repairs, repair period, repair location, repair method, repair material, and repair reason; after the repair is completed, the degree of restoration of each stamping die repair task in the stamping die repair record is evaluated to obtain a restoration degree score; based on the restoration degree score, the stamping die repair record is marked.

[0012] Preferably, the mloptimizer database is invoked, and each stamping die repair task in the stamping die repair record is combined to optimize the repair decision. The repair decision is optimized with the goal of maximizing the degree of restoration score.

[0013] The aforementioned stamping die fault decision system based on dynamic evolution partitioning solves the technical problem in existing technologies where diagnostic methods rely on a single data source, leading to significant errors in diagnostic results and difficulty in achieving optimal repair effects. Through the weak point coordinate positioning module and the dynamic evolution partitioning setting module, it accurately identifies the fault type and location of the die, thus providing strong technical support for fault diagnosis.

[0014] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0015] Figure 1 This is a structural block diagram of a stamping die fault decision system based on dynamic evolution partitioning in one embodiment; Figure 2 This is a flowchart illustrating the analysis of weak degradation signals in a stamping die fault decision system based on dynamic evolution partitioning in one embodiment.

[0016] Figure labeling: 11 Stamping control parameter acquisition module, 12 Multi-source monitoring data acquisition module, 13 Weak point coordinate positioning module, 14 Dynamic evolution partition setting module, 15 Die loss analysis channel setting module, 16 Fault diagnosis decision repair module. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0018] like Figure 1 As shown, this application provides a stamping die fault decision system based on dynamic evolution partitioning, the system comprising: The stamping control parameter acquisition module 11 is used to connect to the punch press and acquire stamping control parameters, including pressure, stamping path, and stamping speed.

[0019] Stamping dies are process equipment used in cold stamping to process materials into parts. Also known as cold stamping dies or cold stamping molds, they are used at room temperature to apply pressure to materials via a die mounted on a press, causing separation or plastic deformation to obtain the desired parts. The die directly acts on the material to produce the expected deformation or separation, thus obtaining the required shape and size. Internal fault diagnosis refers to the process of identifying, analyzing, and determining the causes of potential faults within equipment or systems. In the context of stamping dies, internal fault diagnosis specifically refers to diagnosing faults occurring within the die's own components or system. Decision repair, also known as corrective decision-making, refers to modifying, supplementing, or adjusting existing decision-making plans after a fault occurs, based on the results of fault diagnosis, to eliminate the fault or prevent the recurrence of similar faults. Internal fault diagnosis is the process of identifying and analyzing the causes of faults within equipment or systems, while decision repair is the process of modifying, supplementing, or adjusting existing decision-making plans based on the results of fault diagnosis to eliminate the fault or prevent the recurrence of similar faults. In the maintenance and management of stamping dies, internal fault diagnosis and decision-making repair are key to ensuring normal operation and extending the service life of the dies.

[0020] The stamping control parameter acquisition module 11 is specifically designed to connect to the stamping press and acquire key stamping control parameters. During stamping, the maximum force applied by the press to the workpiece ensures proper forming within the die. The magnitude of the stamping force directly affects the forming quality of the workpiece and the lifespan of the die. Insufficient stamping force may result in incomplete workpiece forming or inaccurate dimensions, while excessive stamping force may damage the die or workpiece. By adjusting the press's pressure control system, the magnitude of the stamping force can be precisely controlled. In actual operation, the appropriate stamping force needs to be determined based on factors such as the workpiece's material, thickness, and shape. The stamping path refers to the movement trajectory of the punch relative to the workpiece and die during the stamping process. This includes the punch's starting position, travel path, and ending position. Optimizing the stamping path is crucial for improving production efficiency and reducing production costs. A reasonable stamping path can reduce press running time while minimizing die wear and the risk of workpiece deformation. Stamping path control is typically achieved through a CNC system. The punch's movement trajectory can be programmed according to the workpiece's shape and size. Stamping speed is the speed at which the punch moves relative to the workpiece and die during the stamping process. Stamping speed has a significant impact on the surface quality and dimensional accuracy of workpieces. Excessive stamping speed can lead to scratches and tears on the workpiece surface, while insufficient stamping speed can negatively impact production efficiency. Stamping speed can be controlled by adjusting the motor speed of the press or the flow rate of the hydraulic system. In practice, the appropriate stamping speed needs to be determined based on factors such as the material, thickness, and shape of the workpiece. The stamping control parameter acquisition module 11, connected to the press, can acquire and monitor key parameters during the stamping process in real time, such as pressure, stamping path, and stamping speed. Precise control of these parameters is crucial for ensuring workpiece quality, improving production efficiency, and reducing production costs.

[0021] The multi-source monitoring data acquisition module 12 is used to monitor the stamping die in real time based on the multi-source monitoring equipment and acquire multi-source monitoring data, including temperature data, pressure data, and vibration amplitude.

[0022] The multi-source monitoring data acquisition module 12 is specifically designed for real-time monitoring of stamping dies based on multi-source monitoring equipment and for acquiring multi-source monitoring data. Multi-source monitoring equipment refers to monitoring devices capable of simultaneously acquiring data from different sources and of different types. In stamping die monitoring, these devices may include temperature sensors, pressure sensors, vibration sensors, etc. Multi-source monitoring data includes temperature data, pressure data, and vibration amplitude. Temperature data reflects the temperature changes of the stamping die during operation. Excessively high temperatures may cause softening, deformation, or even damage to the die material, while excessively low temperatures may affect the die's forming performance and lifespan. Temperature sensors monitor the temperature of key parts of the die in real time and transmit the data to the multi-source monitoring data acquisition module 12 for processing and analysis. Temperature data is one of the important indicators for evaluating the die's working status and predicting potential failures. By monitoring and analyzing temperature data in real time, overheating or undercooling problems of the die can be detected in a timely manner, and corresponding measures can be taken for adjustment and optimization. Pressure data reflects the magnitude and distribution of pressure on the die during the stamping process. Appropriate pressure distribution is crucial to ensuring workpiece forming quality and die lifespan. Pressure sensors monitor pressure changes in the die during the stamping process in real time, and transmit the data to the multi-source monitoring data acquisition module 12 for processing and analysis. Pressure data is a crucial basis for evaluating the stability of the stamping process and optimizing stamping parameters. By monitoring and analyzing pressure data in real time, problems such as uneven or excessive pressure distribution can be detected promptly, and stamping parameters can be adjusted to improve forming quality and extend die life. Vibration amplitude reflects the vibration of the die during the stamping process. Excessive vibration may lead to die damage and poor workpiece forming. Vibration sensors monitor the vibration of the die during the stamping process in real time, and transmit the data to the multi-source monitoring data acquisition module 12 for processing and analysis. Vibration amplitude is one of the important indicators for evaluating the working status of the die and predicting potential failures. By monitoring and analyzing vibration amplitude data in real time, abnormal vibration of the die can be detected promptly, and corresponding measures can be taken for adjustment and optimization. The multi-source monitoring data acquisition module 12, by connecting to multi-source monitoring equipment, can acquire key parameters such as temperature, pressure, and vibration amplitude of the stamping die in real time during operation. Real-time monitoring and analysis of these parameters are of great significance for evaluating the working status of the die, predicting potential failures, and optimizing the stamping process.

[0023] The weak point coordinate localization module 13 is used to obtain stamping die failure instances, establish a working twin model of the stamping die, and locate the coordinates of U weak points.

[0024] The Weak Point Coordinate Location Module 13 is designed to address stamping die failure issues. Its main function is to acquire failure instances of stamping dies and, by establishing a working twin model of the die, further locate the coordinates of U weak points within the die. These weak point coordinates represent areas in the die with potential performance degradation or a high risk of failure. The module collects and analyzes failure cases of stamping dies in actual production, including detailed information such as the damaged location, damage type, and cause of damage. These failure instances provide crucial data support for subsequent weak point coordinate location. Based on the acquired failure instances, the module uses finite element analysis to establish a working twin model of the stamping die. The working twin model is a mapping of the actual die in a virtual environment, capable of simulating the stress, strain, temperature, and other state changes of the die under real working conditions. Based on the working twin model, the module identifies areas in the die prone to failure by simulating and analyzing the stress and deformation of the die during the stamping process. These areas are the weak point coordinates. The module can locate U specific weak point coordinates and provide detailed information for each coordinate, including location, size, and type. The introduction of the weak point coordinate positioning module 13 makes the design and manufacturing process of stamping dies more precise and reliable. By identifying and locating the weak point coordinates in the die in advance, targeted measures can be taken to optimize the die design, improve die quality, and extend die life, thereby reducing production costs and increasing production efficiency. In addition, it provides important reference for die fault diagnosis and repair, which helps to realize intelligent management and maintenance of the die.

[0025] The dynamic evolution partition setting module 14 is used to set U stamping loss micro-regions based on the U weak point coordinates, and associate and map the temperature change, pressure distribution and vibration frequency corresponding to the multi-source monitoring data to set U dynamic evolution partitions.

[0026] The dynamic evolution partitioning module 14 is used to set up stamping loss micro-regions based on the weak point coordinates of the stamping die and to associate and map them with multi-source monitoring data. It can dynamically manage and optimize the weak areas of the die based on real-time acquisition of multi-source monitoring data, including temperature, pressure, and vibration, ensuring the stability of the stamping process and the reliability of the die. Based on the U weak point coordinates provided by the weak point coordinate positioning module 13, the dynamic evolution partitioning module 14 sets these coordinate areas as stamping loss micro-regions. These micro-regions are areas in the die with a high risk of potential performance degradation or failure and require key attention and management. The module associates and maps real-time temperature, pressure, and vibration monitoring data provided by the multi-source monitoring data acquisition module 12 with the stamping loss micro-regions. This means that changes in the state of each micro-region can be reflected in real time through multi-source monitoring data. Common changes in the state of micro-regions include temperature changes, uneven pressure distribution, and abnormal vibration. Based on the associated and mapped multi-source monitoring data, the module sets up U dynamic evolution partitions for each stamping loss micro-region. These partitions can be dynamically adjusted according to the actual working state of the die to cope with different stamping requirements and changes in die state. For example, when the temperature of a certain micro-region exceeds a preset threshold, the module can automatically adjust the cooling system of that region to reduce the temperature; when the pressure distribution in a certain micro-region is uneven, the module can adjust the stamping parameters or die structure to improve the pressure distribution. The dynamic evolution zoning setting module 14 achieves dynamic management and optimization of weak areas of the stamping die by associating and mapping multi-source monitoring data with weak point coordinates. This module can reflect the working status of the die in real time and make dynamic adjustments as needed, ensuring the stability of the stamping process and the reliability of the die. This intelligent management method is of great significance for improving production efficiency and reducing production costs.

[0027] The mold loss analysis channel setting module 15 is used to set an upper mold loss analysis channel based on the U dynamic evolution partitions, with the protruding top of the upper mold punch as the center; and to set a lower mold loss analysis channel based on the U dynamic evolution partitions, with the concave bottom of the lower mold cavity as the center.

[0028] The die loss analysis channel setting module 15 is mainly used to further set loss analysis channels for the upper and lower dies based on the pre-defined U dynamic evolution partitions. These channels are designed to perform more accurate die loss analysis based on a specific center point. Specifically, the protruding top of the upper die punch or the recessed bottom of the lower die cavity is used as the center point; further, the protruding top of the upper die punch is used as the center. Based on the U dynamic evolution partitions, the partitions related to the upper die are selected and associated. A loss analysis channel is set around the center point. This channel may have a specific shape and size to facilitate the collection and analysis of various data from the upper die during the stamping process. The recessed bottom of the lower die cavity is used as the center. Similarly, based on the U dynamic evolution partitions, the partitions related to the lower die are selected and associated. A loss analysis channel for the lower die is designed around the center point. This channel is similar to the upper die channel, but the specific design may be adjusted according to the characteristics and needs of the lower die. In summary, the mold loss analysis channel setting module 15 is an important component of the mold management system. Through precise setting and analysis channels, it provides strong support for mold loss assessment and optimized maintenance.

[0029] The fault diagnosis and repair module 16 is used to connect the upper die loss analysis channel and the lower die loss analysis channel, and to make stamping die fault decisions based on dynamic evolution partitioning in conjunction with the stamping control parameters.

[0030] The fault diagnosis and repair module 16 is the core component of the stamping die fault diagnosis and repair system. This module primarily connects to the upper and lower die wear analysis channels and, in conjunction with stamping control parameters, performs fault diagnosis on the internal components of the stamping die, providing corresponding repair suggestions. The fault diagnosis and repair module 16 first connects to the upper and lower die wear analysis channels, receiving and analyzing the die wear data transmitted from these channels in real time. During fault diagnosis, the module incorporates current stamping control parameters, such as pressure, stamping path, and stamping speed, to gain a more comprehensive understanding of the die's working status and potential problems. Based on the received die wear data and stamping control parameters, the module uses advanced fault diagnosis algorithms to accurately diagnose potential faults within the die. The diagnosis results clearly indicate the type, location, and severity of the fault, providing a basis for subsequent repair decisions. Based on the fault diagnosis results, the module generates corresponding repair suggestions, including adjusting stamping control parameters, replacing die components, and optimizing the die structure, aiming to eliminate the fault, restore the die to its normal working state, and prevent similar faults from recurring. The fault diagnosis and repair module 16, by connecting the upper and lower die wear analysis channels and combining them with stamping control parameters, enables accurate diagnosis and repair decisions for internal faults in stamping dies. This module not only improves the accuracy and efficiency of fault diagnosis but also provides strong support for die maintenance and optimization, helping to extend die life and improve production efficiency.

[0031] Furthermore, the system is used to perform the following steps: Connect to the stamping die database, and use the model parameters of the stamping die as constraints to obtain stamping die failure instances; based on the failure duration corresponding to the stamping die failure instances, divide them according to a preset service life to obtain stamping die failure instances that have exceeded their service life and stamping die failure instances that have not exceeded their service life; use the stamping die failure instances that have exceeded their service life as positive labels and the stamping die failure instances that have not exceeded their service life as negative labels, and perform feature label iterative training by referring to the working twin model of the stamping die.

[0032] The weak point coordinate positioning module 13 establishes a connection with the stamping die database to ensure successful data retrieval. Based on specific requirements, the model parameters of the stamping die are set as constraints for querying the database. Using the set model parameters as query conditions, corresponding stamping die failure instances are retrieved from the database. These instances should include detailed information such as failure time, failure cause, and failure location. Failure duration information, i.e., the time from when the die is put into use to when it fails, is extracted from the retrieved failure instances. Based on the characteristics of the stamping die and empirical data, a reasonable preset service life is set. Failure instances whose failure duration exceeds the preset service life are classified as overdue failure instances and marked positively. Failure instances whose failure duration does not exceed the preset service life are classified as within-limit failure instances and marked negatively. Based on the design parameters and working conditions of the stamping die, a working twin model of the stamping die is constructed. This model can simulate the working state and performance of the die in the actual working environment. For overdue failure instances associated with positive markings, their performance in the working twin model is analyzed, and failure-related features such as stress distribution, temperature changes, and wear degree are extracted. For negatively labeled failure instances that have not yet exceeded their expiration date, their performance in the working twin model is analyzed, and relevant features are extracted. Using the extracted features and labeled failure instance data, a fault diagnosis model is constructed or optimized. This model can be based on machine learning, deep learning, or other artificial intelligence technologies. The labeled failure instance data is divided into training and validation sets. The fault diagnosis model is trained using the training set data, and the trained model is evaluated using the validation set data, calculating metrics such as accuracy, recall, and F1 score. Based on the evaluation results, model parameters or optimization algorithms are adjusted for the next round of iterative training. The above steps are repeated until the model performance meets the preset requirements or the maximum number of iterations is reached. Through this process, the weakness coordinate localization module 13 can perform iterative training of feature labels based on failure instance data from the stamping die database, combined with the stamping die working twin model. This data- and model-based approach can continuously optimize the performance of the fault diagnosis model, improving the accuracy and efficiency of stamping die fault diagnosis.

[0033] like Figure 2 As shown, the system is further configured to perform the following steps: Based on the failure example of the stamping die that did not exceed its time limit, the failure time is located; based on the failure time, abnormal features are extracted and key parameter sequences are collected; first-order difference calculation is performed through the key parameter sequences to analyze weak degradation signals, and the negative markers include weak degradation signals.

[0034] All stamping die failure instances that have not exceeded their service life are screened out. For each instance, the specific failure time is determined. This is typically based on monitoring data or log records from the die's usage. For each failure instance, data within a specific time period, such as one hour before failure, is extracted centered on the failure time. Features potentially related to the failure, such as temperature, pressure, vibration, and noise, are selected from the extracted data. Anomalous features compared to normal conditions are extracted using statistical methods, machine learning algorithms, or deep learning models. Key parameters closely related to the failure are selected from all monitored parameters. For each failure instance, time-series data of these key parameters before and after the failure time are collected. First-order differencing is performed on the collected key parameter sequences, i.e., the difference between parameter values ​​at adjacent time points is calculated. By analyzing the differencing signal, weak degradation signals are identified. These signals may indicate a gradual decline in die performance or the emergence of potential problems. Identified weak degradation signals are marked as part of a negative label, as they represent performance degradation or potential problems occurring before the die reaches its preset service life. Data labeled with weak degradation signals is used to update or optimize fault diagnosis models, improving their ability to identify early degradation and potential problems. This process allows for more effective extraction and analysis of weak degradation signals from stamping die failure instances that have not yet exceeded their service life. These signals are crucial for preventing potential problems and extending die life. Furthermore, incorporating weak degradation signals as part of a negative label can further enhance the accuracy and robustness of the fault diagnosis model.

[0035] Furthermore, the system is used to perform the following steps: First-order difference calculations are performed using the key parameter sequence to output first-order difference results, which include the average, standard deviation, maximum, and minimum values ​​corresponding to continuous differences. Periodic characteristics are analyzed based on the first-order difference results to identify anomalies. Based on the anomalies, in-depth analysis is performed using the key parameter sequence to determine early failure indicators. The weak degradation signal is used to trigger the mining of associated features of the early failure indicators.

[0036] First-order difference calculation involves calculating the difference between each data point in the key parameter sequence and the previous data point; these differences constitute the first-order difference sequence. The average value is the sum of all differences in the first-order difference sequence, reflecting the average rate of parameter change. The standard deviation measures the dispersion of the differences in the first-order difference sequence, i.e., the fluctuation of the rate of change. The maximum value, the point with the largest rate of change in the difference sequence, may represent a sharp change in the parameter value. The minimum value, the point with the smallest rate of change in the difference sequence, may indicate that the parameter value is relatively stable or changes slowly. It is important to check for periodic patterns in the first-order difference sequence. Periodic patterns may represent the inherent cycle of mold operation, including machine operation cycles and processing cycles. These periodic features can be identified through Fourier analysis, autocorrelation functions, or other periodicity detection methods. Points in the first-order difference sequence that significantly differ from periodic patterns or the overall trend are considered outliers. These outliers may indicate sudden changes in mold performance or potential problems. When outliers are found, the original key parameter sequence is traced back to examine the actual parameter values ​​corresponding to these outliers. Analyze the changing trends of these parameter values, their correlation with other parameters, and whether they exhibit any patterns or trends. Combine outlier analysis and in-depth analysis of key parameter sequences to identify early signs or characteristics related to mold failure. These early signs may include specific patterns of parameter value changes, the rate at which specific thresholds are reached, and specific relationships with other parameters. Weak degradation signals refer to those signals that are not easily observed directly but have a long-term cumulative impact on mold performance. Once early failure signs are identified, these features can be used to trigger feature mining of weak degradation signals. Other parameters or characteristics highly correlated with weak degradation signals can be identified, thereby predicting and preventing mold failure earlier. The goal of the entire process is to identify signs of mold performance degradation in advance through meticulous data analysis and feature mining, enabling timely maintenance and repair measures to extend mold life and reduce the risk of failure in production.

[0037] Furthermore, the system is used to perform the following steps: The equipment operation mode of the punch press is obtained, including servo drive mode, frequency conversion drive mode, pressure closed-loop control mode, and multi-level pressure control mode. Based on the equipment operation mode, a joint analysis is performed on the failure examples of the punching die that have not exceeded their expiration period, and a risk warning is issued.

[0038] The process involves acquiring the operating modes of the punch press, including: servo drive mode (using a servo motor for precise position and speed control); frequency converter drive mode (adjusting motor speed via a frequency converter for stepless speed regulation); pressure closed-loop control mode (adjusting parameters based on real-time pressure data from sensors to maintain preset pressure values); and multi-level pressure control mode (setting multiple pressure levels during stamping to meet the needs of different workpieces or processes). This information can typically be obtained by consulting the punch press's technical documentation, communicating with the manufacturer, or extracting it directly from the equipment's control system. Next, the operating modes are jointly analyzed with examples of stamping die failures within their expiration dates. All data related to failures within their expiration dates is collected, including the operating mode at the time of failure, key parameter sequences, and operation logs. The operating modes in the failure examples are analyzed to identify specific operating modes or combinations of modes associated with the failure mode. The changing trends of key parameters under specific operating modes are studied to identify parameter anomalies or patterns related to the failure. Based on the results of the joint analysis, the risk level of stamping die failure under different operating modes is assessed. A risk warning system can be established based on the results of the joint analysis to issue timely warnings when potential risks arise. By following the above steps, the equipment operation mode of the punch press can be effectively obtained, and combined with the failure examples of punching dies that have not exceeded their time limit for joint analysis, so as to achieve accurate risk warning and fault prevention.

[0039] Furthermore, the system is used to perform the following steps: Based on the equipment operation mode, and combined with the failure instances of the stamping dies that have not exceeded their expiration dates, a joint analysis is performed. The system includes: performing stability analysis on the equipment operation mode based on the proportion of the failure instances of the stamping dies that have not exceeded their expiration dates, and obtaining a stability index; calculating the pressure switching smoothness based on the equipment operation mode; and performing feature matching based on the early failure symptoms, combined with the stability index and the pressure switching smoothness, to obtain the joint analysis results.

[0040] Based on the proportion of stamping die failures within their expiration dates, a stability analysis is performed on the equipment operation modes. All failure instances of stamping dies within their expiration dates are collected, and the equipment operation mode and related parameters at the time of their occurrence are recorded. The proportion of failure instances within their expiration dates under each equipment operation mode is calculated, and a stability index is assigned to each operation mode based on the proportion of failure instances. For example, operation modes with a lower proportion of failure instances can obtain a higher stability index. Pressure switching smoothness is calculated based on the equipment operation modes. Data related to pressure switching, such as switching time and pressure changes during the switching process, are extracted from the operation data. Algorithms such as moving average and Fourier analysis are used to calculate the smoothness of pressure switching. A high smoothness indicates stable pressure changes, which may mean more stable equipment operation. Feature matching is performed based on previously identified early failure indicators, extracting features related to early failure indicators from the operation data. The extracted features are matched with known early failure indicator features to identify potential risks. The results of the stability index, pressure switching smoothness, and early failure indicator feature matching are combined to obtain joint analysis results. A comprehensive evaluation report is formed by integrating the results of the stability index, pressure switching smoothness, and early failure indicator feature matching. Based on the assessment report, risk levels are assigned to different equipment operating modes and potential risks, triggering corresponding early warning mechanisms. According to the analysis results, parameters for stability indices, smoothness calculation methods, and failure early warning feature matching are adjusted to improve the accuracy and reliability of the analysis. As new failure instances and data accumulate, the system's analysis model and algorithms are updated to adapt to the constantly changing operating environment. Through these methods, a comprehensive stability analysis and risk warning system for the operating modes of punch presses can be implemented, thereby improving the stability and safety of equipment operation.

[0041] Furthermore, the system is used to perform the following steps: Obtain stamping die repair records, which include the number of repairs, repair period, repair location, repair method, repair material, and repair reason. After the repair is completed, evaluate the restoration degree of each stamping die repair task in the stamping die repair record and obtain a restoration degree score. Based on the restoration degree score, mark the stamping die repair record.

[0042] Obtain stamping die repair records, which include: repair count (total number of repairs); repair period (specific date or time period for each repair); repair location (specific part or area on the die that was repaired); repair method (technique or method used for repair, common methods include welding, grinding, and component replacement); repair material (material type used for repair, common materials include welding rods, alloys, and plastics); and repair reason (description of the problem or fault that caused the die to require repair). After collecting the repair records, the degree of restoration for each repair task needs to be evaluated, and a clear scoring standard needs to be defined to quantify the degree of restoration. The scoring can be based on multiple factors, such as the lifespan of the repaired die, the precision of the repaired die, and the performance stability of the repaired die. For example, the above conditions can be weighted to obtain a score, and the degree of restoration can be scored based on the score. Each repair task is scored according to the scoring standard, and further, the score is determined by an expert system in conjunction with actual operating data, including production efficiency and scrap rate. Based on the degree of restoration score, the stamping die repair records can be marked. The marking can be a simple classification, such as good, average, or poor, or it can be a specific score value. Markings can help users quickly understand the repair history and current status of a mold.

[0043] Furthermore, the system is used to perform the following steps: The mloptimizer database is invoked, and each stamping die repair task in the stamping die repair record is combined to optimize the repair decision. The repair decision is optimized with the goal of maximizing the degree of restoration score.

[0044] Repair records of stamping dies are extracted from the mloptimizer database, including the number of repairs, repair period, repair location, repair method, repair material, repair reason, and corresponding restoration score. Based on the repair records, features related to repair decisions, such as repair location, repair method, and repair material, are extracted, encoded, and transformed for subsequent processing. The objective function of the optimization problem is defined: maximizing the restoration score. The decision variables for the optimization problem are determined, including selectable options such as repair method and repair material. Constraints for the optimization problem are defined based on the actual situation, such as repair cost and repair time. An optimizer suitable for the optimization problem is selected, and its parameters, such as learning rate and number of iterations, are configured according to the characteristics of the problem. A prediction model is trained using historical repair records as training data to predict the restoration score under different repair decisions. The prediction model is combined with the optimizer, and through an iterative optimization process, the repair decision that maximizes the restoration score is found. The performance of the optimized repair decision is evaluated using a validation set or test set, and the restoration scores before and after optimization are compared to verify the effectiveness of the optimizer. By following the steps above, the mloptimizer database can be accessed and combined with stamping die repair records to optimize repair decisions, aiming to maximize the restoration score. This can improve the efficiency and quality of die repair, and reduce production costs and failure rates.

[0045] For specific embodiments of the stamping die fault decision system based on dynamic evolution partitioning, please refer to the embodiments above, which will not be repeated here. The above modules can be embedded in hardware or independent of the processor in a computer device, or stored in software in the memory of a computer device, so that the processor can call and execute the operations corresponding to each module.

[0046] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0047] The embodiments described above are merely examples of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. 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 protection scope of this application.

Claims

1. A stamping die fault decision system based on dynamic evolution partitioning, characterized in that, The system includes: The stamping control parameter acquisition module is used to connect to the punch press and acquire stamping control parameters, including pressure, stamping path, and stamping speed. The multi-source monitoring data acquisition module is used to monitor the stamping die in real time based on the multi-source monitoring equipment and acquire multi-source monitoring data, including temperature data, pressure data, and vibration amplitude. The weak point coordinate localization module is used to obtain failure instances of stamping dies, establish a working twin model of stamping dies, and locate the coordinates of U weak points; The dynamic evolution partition setting module is used to set up U stamping loss micro-regions based on the U weak point coordinates, and associate and map the temperature change, pressure distribution and vibration frequency corresponding to the multi-source monitoring data to set up U dynamic evolution partitions. The mold loss analysis channel setting module is used to set the upper mold loss analysis channel based on the U dynamic evolution partitions, with the protruding top of the upper mold punch as the center; and to set the lower mold loss analysis channel based on the U dynamic evolution partitions, with the concave bottom of the lower mold cavity as the center. The fault diagnosis and repair module is used to connect the upper die loss analysis channel and the lower die loss analysis channel, and, in conjunction with the stamping control parameters, to make stamping die fault decisions based on dynamic evolution partitioning.

2. The stamping die fault decision system based on dynamic evolution partitioning as described in claim 1, characterized in that, The system acquires examples of stamping die failures, establishes a working twin model of the stamping die, and locates the coordinates of U weak points. The system then performs the following steps: Connect to the stamping die database and use the model parameters of the stamping die as constraints to obtain stamping die failure instances; Based on the failure duration corresponding to the stamping die failure instances, and divided by a preset service life, we obtain stamping die failure instances that have exceeded the service life and stamping die failure instances that have not exceeded the service life. Using the failure instances of stamping dies that have exceeded their expiration date as positive labels and the failure instances of stamping dies that have not exceeded their expiration date as negative labels, feature labeling is iteratively trained by referring to the working twin model of the stamping die.

3. The stamping die fault decision system based on dynamic evolution partitioning as described in claim 2, characterized in that, Using the example of a stamping die failure that did not exceed its expiration date as a negative marker, the system is used to perform the following steps, including: Based on the example of stamping die failure before its expiration date, the time of failure is determined. Based on the failure time, abnormal features are extracted, and key parameter sequences are collected; The weak degradation signal is analyzed by performing first-order difference calculations on the key parameter sequence, and the negative label includes the weak degradation signal.

4. The stamping die fault decision system based on dynamic evolution partitioning as described in claim 3, characterized in that, The system performs first-order difference calculations using the key parameter sequence to analyze weak degradation signals. The system is used to execute the following steps: First-order difference calculation is performed using the key parameter sequence to output first-order difference results, which include the average, standard deviation, maximum, and minimum values ​​corresponding to continuous differences. Based on the first-order difference results, periodic characteristics are analyzed to identify outliers; Based on the anomalies, in-depth analysis is performed using the key parameter sequences to determine early failure symptoms. The weak degradation signals are used to trigger the mining of associated features of the early failure symptoms.

5. The stamping die fault decision system based on dynamic evolution partitioning as described in claim 4, characterized in that, The equipment operation mode of the punch press is obtained, including servo drive mode, frequency conversion drive mode, pressure closed-loop control mode, and multi-level pressure control mode. Based on the equipment operation mode, a joint analysis is conducted on the failure examples of stamping dies that have not exceeded their expiration dates, and a risk warning is issued.

6. The stamping die fault decision system based on dynamic evolution partitioning as described in claim 5, characterized in that, Based on the equipment operating mode, and combined with the failure examples of stamping dies that have not exceeded their expiration dates, the system performs the following steps: Based on the failure instances of the stamping dies that have not exceeded their expiration date and the proportion of such failure instances, a stability analysis is performed on the equipment operation mode to obtain a stability index. Calculate the pressure switching smoothness based on the device's operating mode; Feature matching is performed based on the aforementioned early failure symptoms, and combined with the stability index and pressure switching smoothness to obtain joint analysis results.

7. The stamping die fault decision system based on dynamic evolution partitioning as described in claim 1, characterized in that, Obtain stamping die repair records, which include the number of repairs, repair period, repair location, repair method, repair material, and repair reason; After the repair is completed, the degree of restoration of each stamping die repair task in the stamping die repair record is evaluated, and a restoration degree score is obtained. The stamping die repair record is marked based on the restoration degree score.

8. The stamping die fault decision system based on dynamic evolution partitioning as described in claim 7, characterized in that, The system is used to perform the following steps, including: The mloptimizer database is invoked, and each stamping die repair task in the stamping die repair record is combined to optimize the repair decision. The repair decision is optimized with the goal of maximizing the degree of restoration score.

Images