A servo press control method and system

Abstract

The application discloses a servo press control method and system, and relates to the technical field of servo press control. The method comprises the following steps: collecting working condition data in real time through a multi-modal sensing module and performing filtering processing; calculating core characteristic parameters in real time based on the filtered data, including displacement error and body vibration amplitude; setting a preset trigger threshold and performing real-time comparison; when the core characteristic parameters exceed the threshold, starting an online intelligent optimization algorithm, taking minimizing the displacement error and the body vibration amplitude as the target, and iteratively optimizing the low-order coefficient of the seven polynomial stamping curve; in each interpolation period, the controller calculates the instruction displacement value according to the optimized coefficient and outputs; and storing the optimal parameter group meeting the performance index into a process database. Through multi-modal sensing, online intelligent optimization and autonomous learning, the application realizes autonomous sensing, adaptive optimization and parameter reuse of the servo press, and improves the stamping precision and stability.

Description

Technical Field

[0001] This application relates to the field of servo press control technology, and in particular to a servo press control method and system. Background Technology

[0002] As a core piece of equipment in metal forming industrial machinery, servo presses have become key equipment in the field of high-end precision manufacturing due to their advantages of high precision, high flexibility, and high efficiency. With the rapid development of industries such as new energy and electronics, higher requirements have been placed on the stamping accuracy, high-speed performance, and adaptability of servo presses. Servo presses of various specifications are widely used in industrial production, and the traditional control methods of servo presses can no longer meet the universal adaptation needs of various scenarios.

[0003] The existing control method for servo presses has obvious defects: First, it lacks autonomous sensing capabilities and cannot capture information such as changes in material properties, die wear, and abnormal vibration during the stamping process in real time. It can only passively execute preset stamping curves. When material batches are changed or the die experiences slight wear, problems such as decreased stamping accuracy, increased vibration, and workpiece scrap are likely to occur. Moreover, the control methods are often designed for specific specifications and cannot be universally adapted to servo presses with different strokes and strokes.

[0004] Secondly, the stamping curve parameters rely on manual adjustment, which is cumbersome and time-consuming. Moreover, human experience makes it difficult to achieve a balance between multiple objectives, and it is impossible to meet the requirements of minimizing dynamics, maximizing accuracy, and achieving the required efficiency.

[0005] Third, they lack self-learning ability. Every time materials or working conditions are changed, parameters need to be readjusted, making it impossible to reuse historical experience, resulting in low production efficiency and high trial and error costs.

[0006] Fourth, when using seventh-order polynomials and other high-order curves for stamping planning, Runge phenomenon is prone to occur. Traditional numerical solution methods are difficult to obtain the optimal coefficients within the feasible time frame of the project, which affects the stability of stamping.

[0007] Embodied intelligence, as an important branch of artificial intelligence, emphasizes that intelligent agents acquire information, learn knowledge, and complete tasks through interaction with their environment via their physical bodies. Its core characteristics include physical embodiment, real-time closed-loop feedback, contextualized learning, and proactive exploration, providing a novel approach to solving the control pain points of servo presses. However, the application of embodied intelligence technology in the industrial machine tool field is still in its early stages. There is currently no mature technical solution that deeply integrates embodied intelligence with the seventh-order polynomial stamping curve and intelligent optimization algorithms of servo presses to achieve autonomous perception, adaptive optimization, and autonomous learning. Summary of the Invention

[0008] To address the aforementioned issues, this application provides a servo press control method and system. This method utilizes a multimodal sensing module to collect real-time operating data, employs an online intelligent optimization algorithm to dynamically fine-tune the coefficients of the seventh-order polynomial stamping curve, and leverages a process database to achieve autonomous learning and parameter matching. This enables the servo press to transition from passive execution to autonomous decision-making, thereby improving stamping accuracy and stability.

[0009] The first technical solution adopted in this application is: providing a servo press control method, including: The working condition data of the servo press stamping process is collected in real time by the multimodal sensing module, and the working condition data is filtered to obtain filtered working condition data. Based on the filtered working condition data, core characteristic parameters for characterizing the stamping process state are calculated in real time. These core characteristic parameters include displacement error and machine body vibration amplitude. Set a preset trigger threshold for trigger optimization, and compare the core feature parameters with the preset trigger threshold in real time; In response to determining that at least one of the core feature parameters exceeds the preset trigger threshold, an online intelligent optimization algorithm is configured and started to minimize the displacement error and the vibration amplitude of the fuselage as the optimization target. The low-order coefficients of the preset seventh-order polynomial stamping curve used to generate the slider command displacement are iteratively optimized to solve and generate a set of optimized coefficients. In each interpolation cycle, a controller calculates the command displacement value at the current moment based on the optimized coefficients and outputs it to the servo drive mechanism; The optimized coefficients corresponding to the stamping results that satisfy the preset performance indicators with the core feature parameters are stored as the optimal parameter set in a process database.

[0010] Furthermore, the multimodal sensing module includes: a grating ruler for acquiring the actual displacement of the slider, a force sensor for acquiring the actual impact force, a vibration sensor for acquiring the vibration acceleration of the machine body, and an encoder for acquiring the actual speed of the slider.

[0011] Furthermore, the preset trigger threshold includes at least one of the following: displacement error threshold, vibration acceleration amplitude threshold, impact force fluctuation threshold, and material stiffness change threshold.

[0012] Furthermore, the online intelligent optimization algorithm uses a fitness function to evaluate the quality of the low-order term coefficients, and the evaluation dimensions of the fitness function include: The peak value of the absolute value of the jerk during operation and the displacement error during operation are used to calculate the weighted sum of the evaluation dimensions to obtain the objective function value.

[0013] Furthermore, the fitness function is also configured with a constraint penalty term; the triggering logic for the constraint penalty term is as follows: During the iterative optimization process, if the calculated value of velocity, acceleration, or jerk at any moment exceeds the preset physical limit of the device, the objective function value is increased to eliminate coefficient solutions that do not meet the physical constraints.

[0014] Furthermore, the iterative optimization process constrains the coefficients of the higher-order terms of the seventh-order polynomial stamping curve to remain unchanged, and the solution is performed only within the search space composed of the coefficients of the lower-order terms.

[0015] Furthermore, the iterative optimization adopts a limited update strategy, that is, the coefficient generated in the new round of iteration is the sum of the coefficient of the previous round and a correction amount, and the magnitude of the correction amount is limited to a preset percentage of the coefficient of the previous round.

[0016] Furthermore, the online intelligent optimization algorithm is a particle swarm optimization algorithm, and its configuration parameters include population size, number of iterations, inertia weight, individual learning factor and swarm learning factor.

[0017] Furthermore, the controller also executes real-time safety monitoring logic: The controller compares the operating data with their respective preset safety constraints in real time; when any data exceeds the corresponding preset safety constraint, the controller performs an emergency shutdown and alarm action.

[0018] The second technical solution adopted in this application is: a servo press control system is provided, which can realize the control method as described in any of the preceding claims, including: A multimodal sensing module is used to collect and filter operating condition data in real time; An intelligent decision control module is configured as follows: Calculate the core feature parameters based on the aforementioned operating condition data; In response to the core feature parameters exceeding the preset trigger threshold, an online intelligent optimization algorithm is initiated to iteratively optimize the lower-order coefficients of a seventh-order polynomial stamping curve, thereby generating a set of optimized coefficients. In each interpolation cycle, the command displacement value is calculated based on the optimized coefficients; A servo execution module is used to receive the instruction displacement value and drive the slider to move; A process database module is used to store the optimized coefficients.

[0019] Due to the adoption of the above technical solution, this application has at least one of the following beneficial effects compared with the prior art:

[0020] 1. By collecting working condition data of the entire stamping process in real time through the multimodal sensing module, the servo press has autonomous sensing capabilities and can capture working condition information such as changes in material properties, die wear, and abnormal vibration in real time, providing a data foundation for adaptive control.

[0021] 2. Configure an online intelligent optimization algorithm. When the core feature parameters exceed the preset trigger threshold, the optimization will be automatically started. The lower-order coefficients of the seventh polynomial stamping curve will be iteratively optimized. It can adapt to changes in working conditions without manual intervention, which greatly reduces the cost of manual debugging.

[0022] 3. The seventh-order polynomial stamping curve is adopted to achieve continuous and abrupt changes in displacement, speed, acceleration, and jerk throughout the entire stroke. The Runge phenomenon is suppressed by online fine-tuning of the coefficients of the lower-order terms. At the same time, the constraint penalty term is used to ensure that the speed, acceleration, and jerk do not exceed the physical limits of the equipment, thus balancing stamping accuracy and equipment safety.

[0023] 4. Build a process database to enable self-learning. Store the optimal parameter set that meets the performance indicators into the database. When producing similar workpieces in the future, the database can automatically load the historical optimal coefficients, realize one-click start-up and use, reduce trial and error costs and improve production efficiency. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 A schematic flowchart illustrating a servo press control method provided in an embodiment of this application; Figure 2 This is a schematic diagram of the entire stamping stroke provided in an embodiment of this application; Figure 3 A schematic diagram of the framework of a servo press control system provided in an embodiment of this application. Detailed Implementation

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It is understood that the specific embodiments described herein are only for explaining this application and not for limiting it. Furthermore, it should be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all structures. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0026] The terms "first," "second," etc., used in this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0027] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0028] Existing servo press control methods lack autonomous sensing capabilities, failing to capture real-time information on material property changes, die wear, and abnormal vibrations during the stamping process. They can only passively execute preset stamping curves; the stamping curve parameters rely on manual adjustment, lacking autonomous learning capabilities and unable to reuse historical experience. To address these issues, this application provides a servo press control method and system that uses a multimodal sensing module to collect real-time operating data, dynamically fine-tunes the seventh-order polynomial stamping curve coefficients using an online intelligent optimization algorithm, and achieves autonomous learning and parameter matching through a process database. Figure 1 As shown, Figure 1 This is a flowchart illustrating a servo press control method provided in an embodiment of this application. The method includes the following steps: The multimodal sensing module collects real-time working condition data of the entire stamping process of the servo press and filters the working condition data to obtain filtered working condition data. The multimodal sensing module includes: a grating ruler for collecting the actual displacement of the slider, a force sensor for collecting the actual stamping force, a vibration sensor for collecting the vibration acceleration of the machine body, and an encoder for collecting the actual speed of the slider.

[0029] In this embodiment, the multimodal sensing module includes a high-precision grating ruler, a piezoelectric force sensor, a triaxial vibration sensor, and a servo encoder. The installation positions and adaptation principles of each sensor are as follows: A high-precision grating ruler is mounted on the side of the slider, with a measuring range matched to the slider stroke of the servo press. It has a resolution ≥0.1μm and a sampling rate of 1kHz, and is used to acquire the actual displacement of the slider. To ensure the accuracy of displacement sensing, the selection principle for the grating ruler is as follows: the measuring range must cover the maximum slider stroke of the servo press, the resolution must be no less than 0.1μm to meet the positioning requirements of precision stamping, and the sampling rate of 1kHz must match the 1ms interrupt cycle of the PLC.

[0030] The piezoelectric force sensor is installed between the lower end face of the slider and the die. Its measuring range matches the rated punching force of the servo press, with a response time of <1ms and a sampling rate of 1kHz, and is used to collect the actual punching force. It can sense the deformation characteristics of materials and the stress state of molds; piezoelectric force sensors have the characteristics of fast response speed and high measurement accuracy, and are suitable for measuring dynamic impact forces during stamping.

[0031] The triaxial vibration sensor is mounted on the upper part of the fuselage column, with a frequency range of 0~10kHz and a sampling rate of 1kHz, and is used to collect the amplitude of fuselage vibration acceleration. It monitors the operating status of the equipment and the impact of stamping. The triaxial vibration sensor can simultaneously collect vibration acceleration in the X, Y, and Z directions, and the upper part of the machine body column can effectively sense the impact vibration during the stamping process.

[0032] The servo encoder, mounted at the rear of the servo motor, is a ≥23-bit absolute encoder with a sampling rate of 1kHz, used to acquire the actual speed of the slider. This enables real-time speed feedback. The absolute encoder retains position information even after a power outage, eliminating the need for re-zeroing and improving system usability and reliability.

[0033] The PLC synchronously acquires raw data from the four sensors with a 1ms interrupt cycle. Due to impact interference during the stamping process, the raw data exhibits spike noise; therefore, a filtering algorithm is used for data preprocessing. Preferably, a first-order hysteresis filtering algorithm is employed, with the following filtering formula: .

[0034] in, The raw data collected, The filtered data, and These represent the current time and the previous time, respectively. These are the filter coefficients, ranging from 0 to 1. The first-order lag filter algorithm is simple to calculate and has a fast response time, making it suitable for real-time data processing within a 1ms interrupt period. The values ​​of the filter coefficients need to be calibrated according to the actual operating conditions. The larger the value, the stronger the filtering effect, but the slower the response to data changes; The smaller the value, the faster the response but the weaker the filtering effect. As a preferred option, A value of 0.3 to 0.5 can be used to effectively eliminate spike noise and ensure data stability.

[0035] Based on the filtered working condition data, the core characteristic parameters used to characterize the stamping process state are calculated in real time. The core characteristic parameters include displacement error and machine vibration amplitude. A preset trigger threshold is set for trigger optimization, and the core characteristic parameters are compared with the preset trigger threshold in real time.

[0036] The PLC calculates core feature parameters every 1ms, transforming raw data into effective information for decision-making. The calculation cycle for feature extraction is consistent with the data acquisition cycle (1ms), ensuring that the feature parameters can reflect the dynamic changes during the stamping process in real time. The specific extracted feature parameters include: Displacement error : ,in The command displacement is calculated using a seventh-degree polynomial. The actual displacement collected by the grating ruler reflects the accuracy of the slider movement. The sign of the displacement error indicates whether the actual displacement leads or lags behind the commanded displacement, and the absolute value reflects the level of accuracy.

[0037] The vibration amplitude of the machine body is the absolute value of the vibration acceleration amplitude, which is used to reflect the stability of the equipment operation. The larger the vibration value, the greater the impact during the stamping process. Since the vibration acceleration has physical meaning in both positive and negative directions, taking the absolute value can intuitively reflect the vibration intensity.

[0038] Material stiffness: The first derivative of the impact force with respect to the displacement is used to identify batch differences and changes in material properties. Fluctuations in material stiffness can directly reflect changes in material hardness and thickness. When material batches are changed, the stiffness value will change significantly.

[0039] Actual jerk : The first derivative of acceleration with respect to time is used to reflect the smoothness of the stamping curve; sudden changes in acceleration can lead to increased equipment vibration and affect stamping stability.

[0040] In response to the determination that at least one of the core feature parameters exceeds a preset trigger threshold, an online intelligent optimization algorithm is configured and started. With minimizing displacement error and body vibration amplitude as the optimization objectives, the lower-order coefficients of the preset seventh-order polynomial stamping curve used to generate slider command displacement are iteratively optimized to solve for and generate a set of optimized coefficients.

[0041] The preset trigger thresholds include at least one of the following: displacement error threshold, vibration acceleration amplitude threshold, impact force fluctuation threshold, and material stiffness change threshold. That is, when any extracted feature parameter meets the corresponding trigger condition, the online intelligent optimization algorithm is activated. By setting trigger thresholds, frequent optimization activation under normal operating conditions can be avoided, saving computational resources and ensuring real-time control. It should be noted that this application does not limit the specific range of threshold selection; it can be adaptively adjusted based on different actual operating conditions without any limitation.

[0042] The expression for the seventh-order polynomial stamping curve is:

[0043] in, Let be the slider displacement at time t, where t is the running time within the segment. to The coefficients are undetermined. The first, second, and third derivatives of the seventh-order polynomial stamping curve are successively calculated to obtain the corresponding velocity curves. Acceleration curve and acceleration curve They are respectively:

[0044]

[0045]

[0046]

[0047] The seventh-order polynomial has eight coefficients, which can simultaneously constrain the displacement, velocity, acceleration and jerk at the start and end points of the segment, achieving continuous S / V / A / J without abrupt changes throughout the entire stroke.

[0048] In an optional embodiment, the iterative optimization process constrains the coefficients of the higher-order terms of the seventh-order polynomial stamping curve to remain unchanged, and the solution is performed only in the search space composed of the coefficients of the lower-order terms. Fixing the coefficients of the higher-order terms can effectively suppress the Runge phenomenon of the seventh-order polynomial, that is, the severe oscillation at the end of the curve caused by small fluctuations in the coefficients, and ensure the smoothness of the curve. Fine-tuning only the coefficients of the lower-order terms can achieve smooth adjustment of speed and acceleration without affecting the overall smoothness of the curve.

[0049] In an optional embodiment, the iterative optimization employs a limited-amplitude update strategy, whereby the coefficients generated in the new iteration are the sum of the coefficients from the previous iteration and a correction amount, and the magnitude of the correction amount is limited to a preset percentage of the coefficients from the previous iteration; in this embodiment, the magnitude of the correction amount is limited to ±5% of the coefficients from the previous iteration. Limited-amplitude updates ensure that the curve has no abrupt changes and only a small portion is adjusted each time, avoiding shocks and vibrations caused by coefficient jumps; it should be noted that in other embodiments, the magnitude of the correction amount can be selected differently, without any limitation.

[0050] In an optional embodiment, the online intelligent optimization algorithm uses a fitness function to evaluate the quality of low-order term coefficients. The evaluation dimensions of the fitness function include: The peak value of the absolute value of the jerk during operation and the displacement error during operation are used to calculate the objective function value by weighted summation of the evaluation dimensions.

[0051] The fitness function is as follows:

[0052] in, The fitness value is the smaller the value, the better the combination of coefficients. The maximum jerk within the corresponding segment is assigned a weight of 0.6 to prioritize minimizing vibration. This represents the absolute value of the displacement error, with a weight of 0.3, to ensure that the accuracy meets the standards. For a single stamping cycle, the weight is 0.1 to ensure that stamping efficiency is not reduced; This is a penalty item.

[0053] The fitness function is also configured with constraint penalty terms; the triggering logic for the constraint penalty terms is as follows: During the iterative optimization process, if the calculated value of velocity, acceleration, or jerk at any moment exceeds the preset physical limit of the device, the objective function value is increased to eliminate coefficient solutions that do not meet the physical constraints.

[0054] Specifically, when the velocity V exceeds the device's preset constraints, the acceleration A exceeds the device's preset constraints, or the jerk J exceeds the device's preset constraints, ,otherwise = 0. By setting a large penalty coefficient (such as 10000), the fitness value of any solution that violates the physical constraints is increased sharply, thus being naturally eliminated during the iteration process, ensuring that the final output is a feasible solution that fully meets the physical limits of the device.

[0055] The online intelligent optimization algorithm is a particle swarm optimization algorithm, and its configuration parameters include population size, number of iterations, inertia weight, individual learning factor and swarm learning factor.

[0056] The online intelligent optimization algorithm is Particle Swarm Optimization (PSO), with the following configuration parameters: population size 50-60, number of iterations 100-120, inertia weight W = 0.6-0.7, individual learning factor C1 = 1.5-1.8, and swarm learning factor C2 = 1.5-1.8. PSO has the advantages of fast convergence speed, few parameters, and ease of implementation, making it suitable for online real-time optimization scenarios.

[0057] coefficients of a seventh-order polynomial to The range of values ​​for are respectively , , , , to Solving directly from boundary constraints: , , , ,in , , , The starting displacement, velocity, acceleration, and jerk of the corresponding stamping segment.

[0058] As another optional implementation, the online intelligent optimization algorithm can also be a multi-objective collaborative optimization algorithm based on NSGA-III, and its optimization steps are as follows: Construct a multi-objective optimization function, which is decomposed into two independent optimization objectives: Objective 1 is to minimize the maximum accelerometer value within a segment. Target 2 is displacement error ; The penalty function method is used to handle equipment constraints related to velocity, acceleration, and jerk, and solutions that violate the constraints are removed from the non-dominated solution set. Set the population size to 50-60, the number of iterations to 100-120, and the number of reference points to match the optimization objective dimension; Pareto optimal solution sets for the two optimization objectives are obtained iteratively by non-dominated sorting, population selection, and crossover mutation operations. For different stamping processes, three types of recommended optimal solutions are output: Precision mode prioritizes the optimal solution of objective 2, high-speed mode prioritizes the optimal solution of objective 1, and balance mode selects the solution whose Pareto leading edge is closest to the ideal point.

[0059] The entire stamping stroke is divided into a rapid approach section, a material contact section, a forming section, and a return section, such as... Figure 2 As shown, Figure 2 This is a schematic diagram of the entire stamping stroke provided in one embodiment of this application. The specific parameters of each segment are set according to the slide stroke, number of strokes, and stamping process requirements of the servo press. The total duration of each segment is consistent with a single stamping cycle. Taking a servo press with a slide stroke of 200mm and a stroke rate of 200 strokes / min as an example, the single stamping cycle is 0.3s. The displacement range of the rapid approach segment is 0~195mm, and the duration is 0.12s; the displacement range of the material contact segment is 195~198mm, and the duration is 0.03s; the displacement range of the forming segment is 198~200mm, and the duration is 0.02s; the displacement range of the return segment is 200~0mm, and the duration is 0.13s.

[0060] The rapid approach phase is the process of the slider descending from its upper end point to a preset position above the material. Its displacement range is determined by the slider stroke and the material's placement position, and its duration is allocated based on the number of strokes and the total cycle. The starting state is denoted as (…). , , , The final state is denoted as ( , , , ).

[0061] The material contact section is the process where the slider decelerates from a preset position above the material to the material contact position. Its displacement range is determined based on the material thickness and contact accuracy requirements, and its duration is set according to the deceleration smoothness requirements. The starting state is consistent with the ending state of the rapid approach section, and the ending state is denoted as (…). , , , ).

[0062] The forming section is the process of the slider pressing from the material contact position to the bottom dead center. Its displacement range is the distance from the material contact position to the bottom dead center. The duration is set according to the forming process requirements. The starting state is consistent with the ending state of the material contact section. The ending state is denoted as ( , , , ).

[0063] The return stroke is the process of the slider returning from the bottom dead center to the top end point. Its displacement range is the distance between the bottom dead center and the top end point. The duration is allocated according to the number of strokes and the total cycle. The starting state is denoted as ( , , , The final state is denoted as ( , , , ).

[0064] Within each interpolation cycle, a controller calculates the command displacement value at the current moment based on the optimized coefficients and outputs it to the servo drive mechanism; the PLC executes the real-time closed-loop control process with a 1ms interrupt cycle. The PLC reads the current stamping segment and its corresponding seventh-order polynomial coefficients in real time, reads the running time t within the segment (reset to 0 at the beginning of the segment and incremented by 0.001s every 1ms), and substitutes it into the seventh-order polynomial expression to calculate the command displacement at the current moment. Then read the actual displacement of the slider. Calculate displacement error , shift the command The command is sent to the servo execution module, and the servo driver drives the servo motor to run according to the command, causing the slider to move along the optimized curve.

[0065] The PLC monitors the segment running time t in real time. When t reaches the duration of the current segment, the segment running time t=0 is reset and the process switches to the next segment. At the same time, it ensures that the displacement, velocity, acceleration, and jerk at the end of the previous segment are completely equal to the corresponding parameters at the beginning of the next segment, so as to achieve a smooth connection of the entire stroke without impact or sudden changes.

[0066] The optimized coefficients corresponding to the stamping results that satisfy the preset performance indicators with the core feature parameters are stored as the optimal parameter set in a process database; this is explained in detail below: The process database module automatically stores a set of optimal data every few stamping cycles, including material type, stamping force, vibration value, displacement error, coefficients after seventh-order polynomial optimization, and stamping result (OK / NG), forming a complete process data archive.

[0067] When producing the same type of material again, the intelligent decision control module automatically retrieves the process database and loads the historical best coefficients, eliminating the need for manual readjustment and enabling one-click start-up.

[0068] When stamping defects occur, the process database automatically records the curve parameters, operating data, and causes of defects under abnormal conditions. The intelligent decision control module automatically avoids this parameter combination during subsequent optimization processes to prevent the same error from recurring and achieves autonomous iterative learning.

[0069] In an optional embodiment, the controller also performs real-time safety monitoring logic: The controller compares the operating data with the corresponding preset safety constraints in real time; when any data exceeds the corresponding preset safety constraint, the controller performs an emergency stop and alarm action.

[0070] The intelligent decision control module monitors operating data in real time. When the following abnormal conditions occur, it immediately sends an emergency stop command and triggers alarms for over-limit travel, overload, and excessive vibration: The slider displacement exceeds the preset stroke range of the equipment; The velocity V, acceleration A, or jerk J exceeds the preset constraints of the equipment; Abnormal overload of punching force (exceeding 90% of the equipment's rated punching force); Sensor malfunction (data loss or abnormal fluctuations).

[0071] This application also provides a servo press control system capable of implementing the control methods of any of the above embodiments; such as Figure 3 As shown, Figure 3 A schematic diagram of the framework of a servo press control system provided in an embodiment of this application includes: A multimodal sensing module is used to collect and filter operating condition data in real time; An intelligent decision control module is configured as follows: Calculate core feature parameters based on operating condition data; In response to the core feature parameters exceeding the preset trigger threshold, an online intelligent optimization algorithm is initiated to iteratively optimize the coefficients of the lower-order terms of a seventh-order polynomial stamping curve, thereby generating a set of optimized coefficients. In each interpolation cycle, the command displacement value is calculated based on the optimized coefficients; A servo execution module is used to receive instruction displacement values ​​and drive the slider movement; A process database module is used to store the optimized coefficients.

[0072] In the several embodiments provided in this application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0073] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0074] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0075] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A servo press control method, characterized in that, include: The working condition data of the servo press stamping process is collected in real time by the multimodal sensing module, and the working condition data is filtered to obtain filtered working condition data. Based on the filtered working condition data, core characteristic parameters for characterizing the stamping process state are calculated in real time. These core characteristic parameters include displacement error and machine body vibration amplitude. A preset trigger threshold is set for triggering optimization, and the core characteristic parameters are compared with the preset trigger threshold in real time. In response to determining that at least one of the core feature parameters exceeds the preset trigger threshold, an online intelligent optimization algorithm is configured and started to minimize the displacement error and the vibration amplitude of the fuselage as the optimization target. The low-order coefficients of the preset seventh-order polynomial stamping curve used to generate the slider command displacement are iteratively optimized to solve and generate a set of optimized coefficients. In each interpolation cycle, a controller calculates the command displacement value at the current moment based on the optimized coefficients and outputs it to the servo drive mechanism; The optimized coefficients corresponding to the stamping results that satisfy the preset performance indicators with the core feature parameters are stored as the optimal parameter set in a process database.

2. The control method according to claim 1, characterized in that, The multimodal sensing module includes: a grating ruler for acquiring the actual displacement of the slider, a force sensor for acquiring the actual impact force, a vibration sensor for acquiring the vibration acceleration of the machine body, and an encoder for acquiring the actual speed of the slider.

3. The control method according to claim 1, characterized in that, The preset trigger thresholds include at least one of the following: displacement error threshold, vibration acceleration amplitude threshold, impact force fluctuation threshold, and material stiffness change threshold.

4. The control method according to claim 1, characterized in that, The online intelligent optimization algorithm uses a fitness function to evaluate the quality of the low-order term coefficients. The evaluation dimensions of the fitness function include: The peak value of the absolute value of the jerk during operation and the displacement error during operation are used to calculate the weighted sum of the evaluation dimensions to obtain the objective function value.

5. The control method according to claim 4, characterized in that, The fitness function is also configured with a constraint penalty term; the triggering logic for the constraint penalty term is as follows: During the iterative optimization process, if the calculated value of velocity, acceleration, or jerk at any moment exceeds the preset physical limit of the device, the objective function value is increased to eliminate coefficient solutions that do not meet the physical constraints.

6. The control method according to claim 1, characterized in that, The iterative optimization process constrains the coefficients of the higher-order terms of the seventh-order polynomial stamping curve to remain unchanged, and solves the problem only within the search space composed of the coefficients of the lower-order terms.

7. The control method according to claim 1, characterized in that, The iterative optimization adopts a limited update strategy, that is, the coefficient generated in the new round of iteration is the sum of the coefficient of the previous round and a correction amount, and the magnitude of the correction amount is limited to a preset percentage of the coefficient of the previous round.

8. The control method according to claim 1, characterized in that, The online intelligent optimization algorithm is a particle swarm optimization algorithm, and its configuration parameters include population size, number of iterations, inertia weight, individual learning factor and swarm learning factor.

9. The control method according to claim 1, characterized in that, The controller also performs real-time safety monitoring logic: The controller compares the operating data with their respective preset safety constraints in real time; when any data exceeds the corresponding preset safety constraint, the controller performs an emergency shutdown and alarm action.

10. A servo press control system, capable of implementing the control method as described in any one of claims 1-9, characterized in that, include: A multimodal sensing module is used to collect and filter operating condition data in real time; An intelligent decision control module is configured as follows: Calculate the core feature parameters based on the aforementioned operating condition data; In response to the core feature parameters exceeding the preset trigger threshold, an online intelligent optimization algorithm is initiated to iteratively optimize the lower-order coefficients of a seventh-order polynomial stamping curve, thereby generating a set of optimized coefficients. In each interpolation cycle, the command displacement value is calculated based on the optimized coefficients; A servo execution module is used to receive the instruction displacement value and drive the slider to move; A process database module is used to store the optimized coefficients.

Images