A control method applied to self-piercing riveting

Abstract

The application discloses a control method applied to self-piercing riveting, and relates to the technical field of self-piercing riveting control. The method realizes real-time acquisition of riveting material data through a multi-source data fusion interface, captures a high dynamic deformation process of a riveting head, extracts deformation rate and rebound value characteristic values, and constructs a multi-dimensional process parameter set. The hardness, thickness and plastic dissipation characteristics of the material are comprehensively analyzed, and a deformation resistance capability evaluation function is established. By introducing stamping force fluctuation characteristics, the nonlinear relationship between the stamping force and the rivet elastic deformation is analyzed, a hybrid compensation model is constructed, and a riveting dynamic coupling compensation vector is output. Through an adaptive optimization algorithm, a compensation weight matrix is generated, and the stamping force and the rivet displacement are dynamically adjusted in combination with an anti-saturation control strategy. Gain parameters are dynamically adjusted through multi-scale filtering and Lyapunov stability criterion, and compensation weight parameter iterative updating is triggered when an error is exceeded, so that the control effect is optimized. The riveting precision and stability are improved, and the service life of the equipment is prolonged.

Description

Technical Field

[0001] This invention relates to the field of self-piercing riveting control technology, specifically a control method applied to self-piercing riveting. Background Technology

[0002] Self-piercing riveting, as a highly efficient and precise joining technology, is increasingly widely used in modern manufacturing, especially in industries such as automotive, aerospace, and electronics. This technology directly rivets metal materials together using impact force, eliminating the need for heating and filler materials required in traditional welding processes. It offers advantages such as high efficiency, environmental friendliness, and low cost. With the continuous improvement of industrial automation, the requirements for the precision and reliability of the riveting process are becoming increasingly stringent. Therefore, how to adjust key parameters such as impact force and riveting displacement in real time and accurately to ensure riveting quality has become one of the key technologies for improving the application effect of self-piercing riveting technology.

[0003] Currently, traditional self-piercing riveting control methods largely rely on the monitoring results of a single sensor, such as force or displacement sensors, to monitor the riveting process in real time. Especially when dealing with complex materials and deformation processes, a single control strategy is insufficient to meet the precise adjustment requirements of multi-dimensional process parameters. Furthermore, existing methods often fail to achieve efficient and accurate compensation control when faced with significant differences in material hardness, thickness, and plasticity, leading to fluctuations in riveting quality. In addition, existing control strategies fail to fully consider the nonlinear relationship between punching force and rivet elastic deformation, neglecting the combination of dynamic compensation and anti-saturation control. This results in the risk of insufficient or excessive punching force in practical applications, leading to substandard joint strength or deformation failure. Therefore, a more intelligent, adaptable, and dynamically adjustable self-piercing riveting control method is needed to improve riveting accuracy and reliability. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a control method for self-piercing riveting, which solves the problems mentioned in the background.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a control method for self-piercing riveting, comprising the following steps: S1. Real-time acquisition of hardness, thickness, and punching force time-series data of the riveting material through a multi-source data fusion interface, capturing the high-dynamic deformation process of the riveting joint, extracting deformation rate and springback characteristic values, and constructing a multi-dimensional process parameter set; S2. Based on the multi-dimensional process parameter set, comprehensive analysis of material hardness, thickness, and plastic dissipation characteristics, establishment of an anti-deformation capability evaluation function, quantification of the influence of different materials on the stopping distance, and analysis of the nonlinear relationship between punching force and rivet elastic deformation, combined with punching force fluctuation characteristics, constructing a hybrid... S3. Based on the riveting dynamic coupling compensation vector and the process adaptability analysis results, the compensation weight matrix is ​​obtained through an adaptive optimization algorithm, and the riveting punching force and rivet displacement are dynamically adjusted in combination with the anti-saturation control strategy to generate an anti-saturation compensation instruction set; S4. Based on the anti-saturation compensation instruction set, the riveting punching force and rivet displacement are dynamically corrected, the real-time displacement deviation trend component is extracted through multi-scale filtering, and the compensation strength is optimized in combination with the anti-saturation compensation instruction set and the Lyapunov stability criterion is used to dynamically adjust the gain parameters of the control system. If the deviation exceeds the limit continuously, the iterative update of the compensation weight parameters is triggered.

[0006] Furthermore, the specific process of capturing the high dynamic deformation process of the rivet joint, extracting the deformation rate and springback feature values, and constructing a multi-dimensional process parameter set is as follows: continuously capturing high-deformation time-series images of the rivet joint using a high frame rate vision sensor, calculating the instantaneous change gradient of the deformation rate using the optical flow method, and extracting the springback feature using an edge detection algorithm; and fusing material hardness, thickness, and stamping pressure time-series data to construct a multi-dimensional process parameter set that includes dynamic deformation features, material properties, and process states.

[0007] Furthermore, the specific process of establishing the deformation resistance evaluation function by comprehensively analyzing the material hardness, thickness, and plastic dissipation characteristics is as follows: Based on the interaction between material hardness and thickness, a nonlinear deformation resistance evaluation function is constructed, and the influence of different materials on the stopping distance is quantified by introducing the material plastic dissipation rate; combined with experimental calibration data, a dynamic mapping relationship is generated to characterize the differentiated compensation requirements of high-hardness thin plates and low-hardness thick plates.

[0008] Furthermore, combining the characteristics of stamping pressure fluctuations, the nonlinear relationship between stamping pressure and rivet elastic deformation is analyzed, a hybrid compensation model is constructed, and the specific process of outputting the riveting dynamic coupling compensation vector is as follows: the stamping pressure time series data is modeled in stages, and hybrid compensation functions are constructed for the pressure increase, pressure stabilization, and pressure decrease stages respectively to analyze the hysteresis correlation between rivet elastic deformation and pressure fluctuation; by introducing a dynamic hysteresis factor to correct the deviation, a dynamic coupling compensation vector containing pressure deformation compensation and material resistance compensation is output.

[0009] Furthermore, based on the riveting dynamic compensation vector and process adaptability analysis results, the specific process of obtaining the compensation weight matrix through the adaptive optimization algorithm is as follows: Construct a multi-objective optimization problem with the optimization objectives of minimizing head height deviation, maximizing compensation stability, and minimizing equipment life attenuation rate. Combine the weight benchmark parameters in the historical process database, solve the global optimal weight combination through the gradient projection algorithm to generate the compensation weight matrix.

[0010] Furthermore, the specific process of dynamically adjusting the riveting punching force and rivet displacement in conjunction with the anti-saturation control strategy to generate the anti-saturation compensation instruction set is as follows: Based on the compensation weight matrix and real-time process status, an anti-saturation control strategy is designed, and the upper and lower thresholds of punching force and displacement compensation are constrained by the amplitude limiting function; combined with the dynamic boundary expansion mechanism, the safety boundary is adaptively adjusted in the case of material mutation or equipment aging, and the anti-saturation compensation instruction set is generated.

[0011] Furthermore, based on the anti-saturation compensation instruction set, the specific process for dynamically correcting the riveting punching force and rivet displacement is as follows: the anti-saturation compensation instruction set is converted into an execution end control signal, which drives the servo system through a feedback fusion mechanism; the feedforward channel injects the compensation reference value, and the feedback channel dynamically corrects the execution amount according to the real-time displacement deviation, thereby achieving high-precision dynamic correction of the punching force and displacement.

[0012] Furthermore, the specific process of extracting the real-time displacement deviation trend component through multi-scale filtering, and optimizing the compensation strength and dynamically adjusting the gain parameters of the control system by combining the anti-saturation compensation instruction set and the Lyapunov stability criterion is as follows: the low-frequency trend component of the displacement deviation is separated from the high-frequency noise by wavelet transform, and the compensation strength is optimized by combining the anti-saturation instruction set; the stability criterion is constructed based on the Lyapunov function, and the proportional-integral gain parameter is dynamically adjusted to ensure the global asymptotic stability of the system under disturbance.

[0013] Furthermore, if there are consecutive out-of-tolerance events, the specific process of triggering the iterative update of the compensation weight parameters is as follows: Set an out-of-tolerance trigger threshold. When an out-of-tolerance event is detected, start the incremental parameter update mechanism: correct the compensation weight matrix through recursive least squares method, and feed the updated weights back to the hybrid compensation model.

[0014] The present invention has the following beneficial effects:

[0015] (1) A control method for self-piercing riveting, which collects hardness, thickness and punching force time-series data of riveting materials in real time through a multi-source data fusion interface, can capture the high dynamic deformation process of the riveting joint and extract the deformation rate and springback characteristic values, thereby constructing a multi-dimensional process parameter set. This process effectively realizes the accurate perception and analysis of riveting materials and processes, thus providing accurate basic data for subsequent compensation models. By comprehensively analyzing the material hardness, thickness and plastic dissipation characteristics, an anti-deformation capability evaluation function is established, which can quantify the influence of different materials on the stopping distance, and combined with the punching force fluctuation characteristics, analyze the nonlinear relationship between punching force and rivet elastic deformation. It can dynamically adjust for different material properties, greatly improving the accuracy and stability of the riveting process.

[0016] (2) A control method for self-piercing riveting, which combines a dynamic compensation model and an adaptive optimization algorithm with an anti-saturation control strategy, can achieve dynamic adjustment of riveting punching force and rivet displacement. This method can generate an anti-saturation compensation instruction set and extract the real-time displacement deviation trend component through multi-scale filtering to achieve refined control of the riveting process. In addition, by optimizing the compensation force and Lyapunov stability criterion, the gain parameters of the control system can be dynamically adjusted, thereby effectively avoiding system deviation or instability and ensuring high precision and high stability of the riveting process. If continuous deviation occurs, the iterative update of the compensation weight parameters can be triggered in time to further enhance the adaptability and robustness of the system.

[0017] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0018] Figure 1 This is a flowchart of a control method for self-piercing riveting according to the present invention. Detailed Implementation

[0019] This application provides a control method for self-piercing riveting, solving the problem of unstable riveting accuracy caused by differences in material properties, fluctuations in punching pressure, and the nonlinear relationship of riveting joint deformation during the self-piercing riveting process. Traditional riveting methods often rely on fixed parameters, making it difficult to cope with the influence of material hardness, thickness variations, and punching pressure fluctuations on the riveting effect, leading to unstable riveting quality or failures. The method of this invention, combining a multi-dimensional process parameter set, a dynamic compensation model, and an adaptive optimization algorithm, can adjust the punching pressure and displacement during the riveting process in real time, achieving dynamic control and compensation of the riveting process. This significantly improves riveting accuracy and stability, avoiding quality problems caused by process fluctuations or material inhomogeneity.

[0020] The overall concept of the solution in this application embodiment is as follows:

[0021] A control method for self-piercing riveting includes the following steps: S1. Real-time acquisition of hardness, thickness and punching force time data of riveting material through a multi-source data fusion interface, capturing the high dynamic deformation process of the riveting joint, extracting deformation rate and springback characteristic values, and constructing a multi-dimensional process parameter set.

[0022] Based on a multi-dimensional set of process parameters, the material hardness, thickness and plastic dissipation characteristics are comprehensively analyzed to establish a deformation resistance evaluation function, quantify the influence of different materials on the stopping distance, and combine the stamping pressure fluctuation characteristics to analyze the nonlinear relationship between stamping pressure and rivet elastic deformation. A hybrid compensation model is constructed to output the riveting dynamic coupling compensation vector.

[0023] Based on the riveting dynamic compensation vector and process adaptability analysis results, the compensation weight matrix is ​​obtained through an adaptive optimization algorithm, and the riveting punching force and rivet displacement are dynamically adjusted in combination with the anti-saturation control strategy to generate an anti-saturation compensation instruction set.

[0024] Based on the anti-saturation compensation instruction set, the riveting punching force and rivet displacement are dynamically corrected. The real-time displacement deviation trend component is extracted through multi-scale filtering. The compensation strength is optimized by combining the anti-saturation compensation instruction set and the Lyapunov stability criterion to dynamically adjust the gain parameters of the control system. If the deviation exceeds the limit continuously, the iterative update of the compensation weight parameters is triggered.

[0025] Please see Figure 1 This invention provides a technical solution: a control method for self-piercing riveting, comprising the following steps: S1. Real-time acquisition of hardness, thickness, and punching force time-series data of the riveting material through a multi-source data fusion interface, capturing the high-dynamic deformation process of the riveting joint, extracting deformation rate and springback characteristic values, and constructing a multi-dimensional process parameter set; S2. Based on the multi-dimensional process parameter set, comprehensively analyzing the material hardness, thickness, and plastic dissipation characteristics, establishing a deformation resistance evaluation function, quantifying the influence of different materials on the stopping distance, and combining the punching force fluctuation characteristics to analyze the nonlinear relationship between punching force and rivet elastic deformation, and constructing a hybrid compensation model. S3. Based on the riveting dynamic coupling compensation vector and the process adaptability analysis results, obtain the compensation weight matrix through an adaptive optimization algorithm, and dynamically adjust the riveting punching force and rivet displacement in combination with the anti-saturation control strategy to generate an anti-saturation compensation instruction set; S4. Based on the anti-saturation compensation instruction set, dynamically correct the riveting punching force and rivet displacement, extract the real-time displacement deviation trend component through multi-scale filtering, and dynamically adjust the gain parameter of the control system by optimizing the compensation strength and Lyapunov stability criterion in combination with the anti-saturation compensation instruction set. If the error exceeds the tolerance continuously, trigger the iterative update of the compensation weight parameter.

[0026] In this implementation plan, step S1 involves real-time acquisition of the hardness, thickness, and stamping pressure time-series data of the riveting material via a multi-source data fusion interface. This step uses a multi-source data fusion interface (i.e., multiple sensors or data sources) to collect key parameters related to the riveting process in real time: the hardness (material's resistance to deformation), thickness (material's geometric properties), and stamping pressure (the pressure applied during riveting) and time-series data (i.e., records of pressure changes over time). These data reflect the dynamic changes during the riveting process. The high-dynamic deformation process of the riveted joint is captured, and the deformation rate and springback characteristic values ​​are extracted. During riveting, the material deforms. The deformation rate refers to the speed of material deformation, and the springback refers to the degree to which the material recovers its shape after riveting. By capturing these data in real time, the deformation process of the riveted joint under high dynamic pressure can be analyzed. A multi-dimensional process parameter set is constructed: multiple parameters such as hardness, thickness, stamping pressure, deformation rate, and springback are integrated to form a multi-dimensional process parameter set. This parameter set provides complete process data support for subsequent steps, helping to formulate more precise control strategies. Step S2: Comprehensive Analysis of Material Hardness, Thickness, and Plastic Dissipation Characteristics: The deformation resistance of a material is affected by factors such as hardness, thickness, and plastic dissipation (i.e., the energy consumed by the material during stress). By comprehensively analyzing these factors, we can understand the performance of different materials during the riveting process. Establish a deformation resistance evaluation function to quantify the impact of different materials on the stopping distance: Based on the above analysis, establish a deformation resistance evaluation function. This function can quantify the deformation resistance of different materials, that is, determine the ease with which the material deforms during the riveting process, thus affecting the decision on the "stopping distance" during riveting. The stopping distance refers to the displacement of the rivet joint at the end of riveting, involving precise control and compensation. Combined with the characteristics of punching pressure fluctuations, analyze the nonlinear relationship between punching pressure and the elastic deformation of the rivet rod, and construct a hybrid compensation model: Punching pressure and the elastic deformation of the rivet rod are interrelated, but their relationship is nonlinear. This means that the change in punching pressure is not entirely proportional to the deformation of the rivet rod. By analyzing these fluctuation characteristics, construct a hybrid compensation model (including different types of compensation functions) to cope with the nonlinear effects of punching pressure fluctuations and deformation. Output riveting dynamic coupling compensation vector: The model outputs a dynamic coupling compensation vector, which contains compensation information for multiple process parameters (such as pressure, deformation, etc.) to adjust the riveting process in real time and ensure riveting quality. Step S3: Based on the riveting dynamic compensation vector and process adaptability analysis results: Combining the compensation vector from step S2, further analyze the process adaptability, i.e., the degree of adaptability of the compensation strategy under different material and process conditions. Obtain the compensation weight matrix through an adaptive optimization algorithm: The adaptive optimization algorithm dynamically adjusts the compensation strategy based on real-time data, generating a compensation weight matrix. The compensation weight matrix is ​​a tool used to adjust the proportions of various process parameters (such as punching force, rivet displacement, etc.).Dynamic adjustment of riveting punching force and rivet displacement is achieved by combining an anti-saturation control strategy: In some cases, the punching force or displacement may reach the saturation point of the system, meaning the control system can no longer effectively adjust. In this case, an anti-saturation control strategy is used to ensure that the system is not limited by saturation, and can still maintain stability and effectively control the punching force and rivet displacement. An anti-saturation compensation instruction set is generated: Based on the above adjustments, an anti-saturation compensation instruction set is generated, indicating how to dynamically adjust the punching force and displacement to avoid saturation during riveting and ensure riveting quality. Step S4: Dynamic correction of riveting punching force and rivet displacement is performed according to the anti-saturation compensation instruction set: Using the previously generated anti-saturation compensation instruction set, the punching force and rivet displacement during the riveting process are corrected in real time. The purpose of dynamic correction is to correct possible errors and ensure that the riveting quality is always at its best. Real-time displacement deviation trend components are extracted through multi-scale filtering: Multi-scale filtering technology is used to extract the trend components in the real-time displacement data. Through this process, errors from noise and interference can be smoothed out, and the true deviation trend can be extracted. Combining anti-saturation compensation instruction set to optimize compensation intensity and Lyapunov stability criterion: The compensation intensity of the control system is optimized using the anti-saturation compensation instruction set. The Lyapunov stability criterion is a mathematical method used to determine whether a system is stable. In this step, the gain parameter is dynamically adjusted to ensure that the system maintains stability while compensating. If continuous deviations occur, the compensation weight parameter is iteratively updated: If the system experiences continuous deviations (i.e., exceeding the allowable error range), the compensation weight parameter is updated. By iteratively updating the compensation parameter, the riveting process can be continuously optimized, avoiding repeated errors. Deformation resistance assessment function: A mathematical function used to quantify the material's deformation resistance, which can predict the stopping distance during the riveting process based on the material's hardness, thickness, and plastic dissipation characteristics. Nonlinear relationship: The relationship between punching force and rivet deformation is not a simple linear relationship; it may involve complex dynamic effects such as hysteresis and elastic deformation. Hybrid compensation model: A composite model combining different compensation strategies (such as exponential and logarithmic compensation functions) to compensate for fluctuations in punching force and deformation. Adaptive optimization algorithm: An algorithm that dynamically adjusts control parameters based on real-time feedback data to optimize control system performance. Anti-saturation control strategy: A strategy to prevent control systems from failing due to reaching saturation (e.g., when the control signal can no longer be adjusted), ensuring the system always operates effectively. Lyapunov stability criterion: A mathematical tool for analyzing system stability, determining whether a control system can remain stable under different conditions.

[0027] Specifically, the process of capturing the high dynamic deformation process of the rivet joint, extracting the deformation rate and springback feature values, and constructing a multi-dimensional process parameter set is as follows: continuously capturing high-deformation time-series images of the rivet joint using a high-frame-rate visual sensor, calculating the instantaneous change gradient of the deformation rate using the optical flow method, and extracting the springback feature using an edge detection algorithm; and fusing material hardness, thickness, and stamping pressure time-series data to construct a multi-dimensional process parameter set that includes dynamic deformation features, material properties, and process states.

[0028] In this implementation, a high frame rate visual sensor is used to capture the deformation process of the riveting joint: by using a high frame rate visual sensor, the high dynamic deformation time-series images of the riveting joint during the riveting process can be captured in real time. Since the riveting process is usually very rapid and has strong dynamic changes, a high frame rate sensor can provide continuous, high-resolution image data for accurate monitoring of the deformation process. Instantaneous gradient calculation of deformation rate: The captured image sequence is processed using optical flow. Optical flow is an image processing technique used to estimate the temporal motion of objects or regions in an image. In this method, the deformation rate of the riveting joint region is calculated using optical flow, and the instantaneous gradient of the deformation rate, i.e., the rate of change of deformation, is obtained. This gradient reveals the instantaneous deformation process experienced by the riveting joint during riveting, providing accurate data for subsequent compensation and adjustment. Edge detection algorithm extracts springback features: During the riveting process, the springback amount is an important feature that directly affects the final riveting quality. Edge detection algorithms (such as Canny edge detection) can accurately extract the springback feature from the high frame rate images. These characteristic values ​​reflect the degree of recovery of the riveted joint after deformation during the riveting process, i.e., the springback. Multi-source data is integrated: in addition to deformation rate and springback, material hardness, thickness, and stamping pressure time-series data are also collected in real time. Combining this information provides a comprehensive reflection of various process states during riveting. This data is then fused to construct a multi-dimensional set of process parameters, including dynamic deformation characteristics (such as deformation rate and springback), material properties (such as hardness and thickness), and process states (such as stamping pressure). This parameter set provides the necessary foundational data support for subsequent compensation models and process optimization.

[0029] Specifically, the process of establishing a deformation resistance evaluation function by comprehensively analyzing material hardness, thickness, and plastic dissipation characteristics is as follows: Based on the interaction between material hardness and thickness, a nonlinear deformation resistance evaluation function is constructed, and the influence of different materials on the stopping distance is quantified by introducing the material plastic dissipation rate; combined with experimental calibration data, a dynamic mapping relationship is generated to characterize the differentiated compensation requirements of high-hardness thin plates and low-hardness thick plates.

[0030] In this implementation scheme, a nonlinear deformation resistance evaluation function is constructed based on the interaction between material hardness and thickness: the material's deformation resistance is closely related to its hardness and thickness. Hardness typically affects the compressive strength of a material, while thickness determines the deformation amplitude when the material is subjected to external force. To quantify the interaction between these two factors, we construct a nonlinear evaluation function to represent the material's deformation resistance under specific conditions. The formula is as follows: ;in: The material's resistance to deformation indicates its ability to resist deformation during the riveting process. The hardness of a material affects its compressive strength. The thickness of the material affects its ability to withstand stress during the riveting process. : A constant representing the overall effect of material hardness and thickness on its resistance to deformation. , : An index to be determined, reflecting the degree to which hardness and thickness affect the resistance to deformation. These parameters are typically determined through experimental or theoretical analysis. The influence of different materials on the stopping distance is quantified by introducing the material's plastic dissipation rate: During plastic deformation, materials consume a certain amount of energy; this process is called plastic dissipation. The magnitude of plastic dissipation is related to factors such as the material's hardness, thickness, and impact force. In the free steel ball process, the material's plastic dissipation capacity directly affects the stopping distance during deformation. Therefore, it is necessary to introduce the material's plastic dissipation rate to quantify the performance differences of different materials in the steel ball process. The formula is as follows: ;in: Plastic dissipation rate: This represents the energy consumed by a material per unit time. : Energy consumed during plastic deformation. It is equal to the integral of the product of stress and strain in the material during the plastic stage. t: Duration of the plastic deformation process, reflecting the time taken for the deformation to complete. The integral of the product of stress and strain during the plastic deformation stage represents the energy consumed during plastic deformation. A dynamic mapping relationship is generated by combining experimental calibration data to characterize the differentiated compensation requirements between high-hardness thin plates and low-hardness thick plates: To achieve accurate compensation, the differences in compensation requirements between high-hardness thin plates and low-hardness thick plates need to be calibrated based on experimental data. By generating a dynamic mapping relationship from experimental data, the corresponding compensation amount can be calculated based on different material hardness and thickness parameters. This helps to dynamically adjust the compensation strategy during riveting, thereby ensuring riveting quality and efficiency. The formula is expressed as follows: ;in: Compensation amount indicates the stopping distance or impact force that needs to be adjusted. The material's resistance to deformation. Plastic dissipation rate of the material. The hardness of a material. The thickness of the material. Using the above formula, the compensation amount can be dynamically adjusted according to the hardness, thickness, and plasticity of different materials, thereby meeting the riveting requirements of different materials.

[0031] Specifically, based on the characteristics of stamping pressure fluctuations, the nonlinear relationship between stamping pressure and rivet elastic deformation is analyzed, a hybrid compensation model is constructed, and the dynamic coupling compensation vector of riveting is output. The specific process is as follows: the stamping pressure time series data is modeled in stages, and hybrid compensation functions are constructed for the pressure increase, pressure stabilization, and pressure decrease stages respectively to analyze the hysteresis correlation between rivet elastic deformation and pressure fluctuations; by introducing a dynamic hysteresis factor to correct the deviation, a dynamic coupling compensation vector containing pressure deformation compensation and material resistance compensation is output.

[0032] In this implementation plan, the core of this process is to analyze and model the stamping pressure time series data, and to resolve the nonlinear relationship between stamping pressure fluctuations and rivet elastic deformation, ultimately constructing a hybrid compensation model to output a riveting dynamic coupling compensation vector. This process includes three main steps: staged modeling, hysteresis correlation analysis, and dynamic hysteresis correction. Staged modeling of the stamping pressure time series data: Since the stamping process includes three different stages—pressure increase, pressure stabilization, and pressure decrease—a compensation function needs to be constructed for each stage. This is because the stamping pressure variation characteristics of each stage differ from the physical behavior during the riveting process. For example, the pressure gradually increases during the pressure increase stage, remains stable during the pressure stabilization stage, and gradually decreases during the pressure decrease stage. Pressure increase stage: During this stage, the stamping pressure rises, the rivet begins to deform and enters the plastic zone. We need to construct a compensation model that reflects the elastic deformation of the rivet when the pressure increases. Pressure stabilization stage: During this stage, the stamping pressure stabilizes and reaches its maximum value, and the elastic deformation of the rivet tends to stabilize, but is affected by fluctuations. Compensation in this stage needs to consider the correlation between small pressure fluctuations and rivet deformation. Pressure Reduction Phase: As the pressure decreases, the elastic deformation of the rivet recovers, resulting in a certain amount of springback. Compensation in this phase needs to consider the relationship between the springback effect and the pressure reduction. Analysis of the Hysteresis Correlation between Rivet Elastic Deformation and Pressure Fluctuations: In actual self-piercing riveting, the punching force and the elastic deformation of the rivet are not instantaneously synchronized, but exhibit a hysteresis effect. Specifically, the deformation of the rivet often lags behind the change in punching force, resulting in a hysteresis response of pressure fluctuations to the rivet's elastic deformation. To better handle this hysteresis effect, we introduce a dynamic hysteresis factor, which can correct the deviation caused by the hysteresis. The hysteresis effect generally manifests as: when the pressure increases, the deformation response of the rivet lags; when the pressure decreases, the springback of the rivet also exhibits a hysteresis effect. Correction of Deviations by Introducing the Dynamic Hysteresis Factor: The dynamic hysteresis factor can make real-time corrections at each stage of the punching force, correcting the deformation deviation caused by the hysteresis effect. By introducing the dynamic hysteresis factor, we can more accurately adjust the compensation model, making the compensation vector more consistent with the actual riveting process. Phased Modeling and Compensation Functions: Compensation functions are constructed separately for the boost, stabilization, and buck phases. , and Boost stage compensation function: in: Impact pressure. : A constant that reflects the effect of pressure on deformation. The index reflects the sensitivity of the willow stalk to elastic deformation during the pressurization stage. The constant relating the rate of pressure change to the deformation response time. Time. Voltage stabilization stage compensation function: ;in: : Constant, compensation coefficient during the voltage stabilization stage. : Index, the sensitivity between pressure and deformation during the stabilization phase. Frequency reflects the effect of minute fluctuations in impact force on deformation. Compensation function during the pressure reduction stage: ;in: : Constant, the influence coefficient of compensation during the pressure reduction stage. The index represents the impact of pressure during the depressurization phase. The attenuation coefficient reflects the impact of the rebound speed during the pressure reduction phase. Hysteresis correlation and the introduction of a dynamic hysteresis factor: Considering the hysteresis relationship between pressure and the elastic deformation of the willow stalk, we introduce a dynamic hysteresis factor. Correcting errors caused by lag: ;in: Deformation deviation at the current moment. The dynamic hysteresis factor changes over time to correct for hysteresis errors. Delay time Reference impact force at that time. The hysteresis time constant reflects the time delay between pressure fluctuations and deformation response. Output of the dynamic coupling compensation vector: By combining the compensation functions of each stage with the hysteresis correction, we obtain the final dynamic coupling compensation vector C, which includes both pressure deformation compensation and material resistance compensation. ;in: The dynamic coupling compensation vector reflects the comprehensive compensation effect during the riveting process. Each compensation function reflects the compensation amount during the voltage boost, voltage stabilization, and voltage reduction stages, respectively. This is a lag correction factor, used for dynamic correction and compensation.

[0033] Specifically, based on the riveting dynamic compensation vector and the results of process adaptability analysis, the specific process of obtaining the compensation weight matrix through an adaptive optimization algorithm is as follows: Construct a multi-objective optimization problem with the optimization objectives of minimizing head height deviation, maximizing compensation stability, and minimizing equipment life attenuation rate. Combine the weight benchmark parameters in the historical process database, solve the global optimal weight combination through the gradient projection algorithm, and generate the compensation weight matrix.

[0034] In this implementation plan, the process of obtaining the compensation weight matrix through an adaptive optimization algorithm, based on the riveting dynamic compensation vector and the results of process adaptability analysis, mainly revolves around the construction and solution of a multi-objective optimization problem. The core of this process is to use an adaptive optimization algorithm (such as gradient projection) to weigh different optimization objectives and solve for the globally optimal combination of compensation weights to ensure the efficiency, stability, and durability of the riveting process. Constructing the multi-objective optimization problem: The optimization objectives include three key aspects: Minimizing head height deviation: During riveting, the head height deviation of the rivet directly affects the quality and structural strength of the connector. Therefore, reducing the rivet head height deviation is an important optimization objective. Maximizing compensation stability: The stability of the compensation system is an important indicator for ensuring riveting quality. High stability means that the compensation strategy can maintain consistent effectiveness under different working conditions, avoiding large fluctuations and unstable compensation behavior. Minimizing equipment lifespan attenuation rate: The lifespan of the riveting equipment is an important factor affecting long-term production costs. Reasonable compensation weights can reduce the load on the equipment, thereby extending its service life. Combining weight benchmark parameters from the historical process database: The data in the historical process database provides a large number of riveting process parameters and results, which can be used to provide weight benchmark parameters for the current optimization problem. These benchmark parameters reflect empirical data on compensation weights under different process conditions. Gradient projection algorithm for finding the globally optimal weight combination: To optimize the compensation effect during the riveting process, the gradient projection algorithm is used to find the optimal compensation weight combination. This algorithm guides the search direction by calculating the gradient of the objective function and restricts the solution to an effective solution space through a projection step. This ensures the feasibility of the solution and the optimization effect. Generation of the compensation weight matrix: Based on the above solution process, the final compensation weight matrix reflects the contribution of different process parameters (such as pressure, displacement, etc.) to each optimization objective, enabling dynamic adjustment of the compensation strategy. Formula for the multi-objective optimization problem: The objective function of the optimization problem can be expressed as: ;in: : Compensation weight matrix, which contains the weight combination of all compensation parameters. : Head height deviation function, representing the head height deviation under compensation weight. : The compensation stability function represents the stability measure of the compensation system. : Lifetime decay rate function, representing the rate at which the equipment's lifespan decays. Weighting coefficients are used to balance the influence between different objectives. Gradient projection method: The gradient projection method calculates the gradient of the objective function and projects it into the constraint space to obtain the optimal solution. The specific steps are as follows: Gradient calculation: Where J(w) represents the objective function, This is the gradient of the objective function. Projection steps: ;in: : The updated compensation weight matrix. Step size parameter determines the size of the update step. The gradient of the objective function. Generation of the compensation weight matrix: Through the optimization algorithm described above, a compensation weight matrix W can be generated, which contains the optimal compensation strategy for each optimization objective. The compensation weight matrix will guide real-time compensation adjustments in subsequent processes.

[0035] Specifically, the process of dynamically adjusting the riveting punching force and rivet displacement in conjunction with the anti-saturation control strategy to generate the anti-saturation compensation instruction set is as follows: Based on the compensation weight matrix and real-time process status, an anti-saturation control strategy is designed, and the upper and lower thresholds of punching force and displacement compensation are constrained by the amplitude limiting function; combined with the dynamic boundary expansion mechanism, the safety boundary is adaptively adjusted in the case of material mutation or equipment aging, and the anti-saturation compensation instruction set is generated.

[0036] In this implementation scheme, the compensation weight matrix and real-time process status are used: By combining the aforementioned compensation weight matrix (including optimized combinations of factors such as punching pressure, displacement, and head height deviation) with real-time process status (such as pressure, displacement, riveting speed, and riveting time), a suitable compensation strategy can be provided for each riveting stage. The real-time process status changes continuously based on current production conditions (such as equipment performance, material properties, and operating environment). An anti-saturation control strategy is designed: To prevent punching pressure and displacement from reaching the physical or safety limits of the equipment, an anti-saturation control strategy is adopted. This strategy constrains the compensation amounts of punching pressure and displacement through a limiting function to prevent excessive compensation and avoid equipment overload or inaccurate deformation. The limiting function effectively limits the compensation amounts of punching pressure P and displacement d, ensuring that they do not exceed preset maximum or minimum thresholds. Assuming the threshold is... and Therefore, the limiting function can be expressed as: ; ;in: and These are the compensated impact force and displacement, respectively. and It is the safe upper limit for impact force and displacement. This means that x is restricted to the interval [c, d], and values ​​outside the interval are restricted to the nearest boundary value. Dynamic boundary expansion mechanism: In abnormal situations such as abrupt material changes (e.g., changes in material hardness or uneven thickness) or equipment aging (e.g., pressure sensor errors or mechanical wear), the compensation mechanism needs to adaptively adjust to ensure the safety and stability of the riveting process. The dynamic boundary expansion mechanism adjusts the safety boundary in a timely manner by monitoring changes in process conditions and material properties. Expansion mechanism: Assuming changes in material hardness or equipment condition, the system needs to dynamically adjust the safety range of punching force and displacement. New thresholds can be automatically derived through adaptive algorithms (e.g., models based on historical data or real-time feedback). For example, when material hardness increases, the upper limit of punching force and displacement can be appropriately increased.

[0037] This extension layer mechanism can be described as follows: ; ;in: and It is a dynamic adjustment factor that changes with time t, representing a correction coefficient based on changes in material properties or equipment aging. and These are the adjusted punching force and displacement safety thresholds. Generating a saturation compensation instruction set: Based on the above strategies, a final compensation instruction set is generated through disturbance rejection and control strategies. These instruction sets include compensation amounts adjusted according to real-time process conditions, ensuring that the compensation amount does not exceed the safety threshold at each stage, and considering the effects of material mutations and equipment aging. Compensation instruction set: Each compensation instruction set includes the punching force compensation amount. Displacement compensation amount And the corresponding time and operating conditions. Generate corresponding compensation instructions based on the compensation needs at different stages.

[0038] Specifically, the process of dynamically correcting the riveting punching force and rivet displacement based on the anti-saturation compensation instruction set is as follows: the anti-saturation compensation instruction set is converted into an execution end control signal, and the servo system is driven through a feedback fusion mechanism; the feedforward channel injects the compensation reference value, and the feedback channel dynamically corrects the execution amount according to the real-time displacement deviation, thereby achieving high-precision dynamic correction of the punching force and displacement.

[0039] In this implementation scheme, the anti-saturation compensation instruction set is converted into execution-end control signals. The anti-saturation compensation instruction set includes compensation amounts calculated based on factors such as process conditions, material properties, and equipment status. These compensation amounts include punching force compensation and displacement compensation, used to guide the dynamic adjustment of the servo system. Conversion process: The instruction set is converted into specific control signals, typically current, speed, or position control signals, through a certain algorithm or conversion function. For example, punching force compensation may be converted into the output current of the servo motor, while displacement compensation may be converted into displacement control instructions. These converted signals are sent to the execution end (such as the servo system). Feedback fusion mechanism drives the servo system: The feedback fusion mechanism combines real-time feedback data from sensors (such as pressure sensors and displacement sensors) to correct the punching force and displacement executed by the system. This feedback data includes the error between the actual state and the expected target. The feedback fusion mechanism integrates the error information with the original compensation instructions, adjusting the control signals at the execution end to ensure greater accuracy with each adjustment. Feedback mechanism: For example, if a deviation in punching force or displacement is detected, the feedback signal is transmitted to the control system, which dynamically adjusts the compensation amount based on this error. Commonly used control algorithms, such as Proportional-Integral-Derivative (PID) control and fuzzy control, can correct errors in real time. The feedforward channel injects pre-calculated compensation reference values. These reference values ​​are determined in previous steps based on factors such as process parameters, material properties, and equipment status. Before the compensation process begins, the feedforward control system directly injects these compensation reference values ​​into the control signal, providing the system with an initial compensation amount. The function of feedforward compensation is to provide the system with an advance compensation signal, avoiding large errors caused by real-time feedback delays. The feedforward compensation amount is usually an ideal compensation value obtained through modeling under known or expected operating conditions. The feedback channel dynamically corrects the execution amount based on real-time displacement deviation: After the riveting process begins, the real-time displacement sensor monitors the actual displacement of the rivet rod and compares it with the target displacement value in real time, generating a displacement deviation. The feedback channel adjusts the execution amount based on this deviation to ensure that the riveting displacement reaches the predetermined target. Real-time correction: For example, if the actual displacement deviation of the rivet rod is negative (i.e., insufficient displacement) during riveting, the feedback channel will increase the compensation signal to adjust the execution amount of the riveting equipment. Conversely, if the displacement deviation is positive, the system will reduce the compensation amount. Achieving high-precision dynamic correction of punching force and displacement: Through the combined action of the above feedforward and feedback mechanisms, the compensation amount of punching force and displacement can be dynamically adjusted at every moment during the riveting process. This dynamic correction ensures that punching force and displacement maintain high precision under various operating environments and process conditions, thereby improving the quality and accuracy of riveting.Dynamic correction effect: Through continuous feedback and feedforward mechanisms, the system can always maintain accurate compensation under the influence of factors such as pressure fluctuations, displacement changes, and material property fluctuations, so that the impact force and displacement do not overshoot or lag, and achieve the expected goal.

[0040] Specifically, the process of extracting the real-time displacement deviation trend component through multi-scale filtering and optimizing the compensation strength by combining the anti-saturation compensation instruction set with the Lyapunov stability criterion to dynamically adjust the gain parameters of the control system is as follows: the low-frequency trend component of the displacement deviation is separated from the high-frequency noise by wavelet transform, and the compensation strength is optimized by combining the anti-saturation instruction set; the stability criterion is constructed based on the Lyapunov function, and the proportional-integral gain parameter is dynamically adjusted to ensure the global asymptotic stability of the system under disturbance.

[0041] In this implementation plan, the specific process steps are as follows: Wavelet transform separates the low-frequency trend component and high-frequency noise of the displacement deviation: During the steel welding process, the real-time displacement deviation typically contains a low-frequency trend component (representing the system's stability or long-term variation trend) and a high-frequency noise component (caused by external disturbances or sensor errors). To more accurately capture the dynamic behavior of the system, wavelet transform is used to decompose the displacement deviation signal into low-frequency trend and high-frequency noise. Wavelet transform: Wavelet transform can effectively separate different frequency components in a high signal-to-noise ratio signal. The low-frequency component reflects the overall trend of the system (such as compensation requirements), while the high-frequency component represents short-period fluctuations and noise. Through wavelet transform, the low-frequency trend component of the displacement deviation can be extracted, which is particularly important for the optimization and adjustment of the compensation system. Extracting the low-frequency trend: The low-frequency trend component can help determine the steady-state condition of the current steel welding process, thus providing a basis for the optimization of the anti-saturation compensation instruction set. High-frequency noise can be suppressed to avoid interfering with the accuracy of the compensation system. Integrating the compensation intensity with anti-saturation compensation instructions: Through the anti-saturation compensation instruction set, the compensation system will adjust the compensation amount (such as punching force and displacement compensation amount) according to different process parameters. By combining the low-frequency trend components extracted by wavelet transform, the compensation intensity is optimized to cope with changes in actual working conditions. Optimizing the compensation intensity: Based on the trend components of the displacement deviation, it is possible to analyze whether the current processing is in a stable range, and then adjust the compensation intensity accordingly. If the system tends to saturate, the compensation system will adjust the amplitude of the compensation signal through anti-saturation control to avoid instability caused by over-compensation. Constructing a stability criterion based on the Lyapunov function: To ensure the stability of the control system under anti-motion conditions, the Lyapunov stability criterion can be used to determine the global asymptotic stability of the system. The Lyapunov function is a mathematical tool used to analyze the stability of force systems. By constructing the Lyapunov function, the stability of the system can be quantitatively evaluated, and theoretical support can be provided for the dynamic adjustment of the gain parameter. Construction of the Lyapunov function: The Lyapunov function is generally represented as V(x), where x is the state variable of the system. This function satisfies the conditions V(x)>0 and V'(x)≤0, indicating that the system is asymptotically stable. By evaluating the change in system energy through the Lyapunov function, it can be ensured that the system can return to a steady state under disturbance. Dynamic adjustment of proportional-integral gain parameters: In a control system, a proportional-integral (PI) controller is used to regulate the system output. This is achieved by dynamically adjusting the proportional gain (PI). ) and integral gain ( The PI gain parameter can be adjusted to optimize the system's response performance. Combined with the Lyapunov stability criterion, the PI gain can be adjusted in real time to ensure the system remains stable under disturbances and avoid oscillations or overreactions. Adjusting the gain parameter: proportional gain. The degree of response of the control system to errors, integral gain This is used to eliminate long-term deviations in the system. The Lyapunov function is used to determine the system's stability. When the system faces disturbances, these two gain parameters can be automatically adjusted to restore the system to a stable state. Ensuring global asymptotic stability: Throughout the process, the Lyapunov stability criterion ensures that the system can asymptotically stabilize and eventually return to the predetermined target under the presence of disturbances. The strategy of dynamically adjusting the PI gain parameters will be adjusted in real time according to the stability criterion, thereby ensuring that the impulse force, displacement, and other compensation quantities maintain accuracy and stability throughout the process. Global asymptotic stability: By adjusting the gain parameters and optimizing the compensation amount in combination with the anti-saturation compensation instruction set, the system can recover to a stable operating state and maintain optimal dynamic performance regardless of any disturbance or process change.

[0042] Specifically, if there are consecutive out-of-tolerance events, the specific process of triggering the iterative update of the compensation weight parameters is as follows: Set an out-of-tolerance trigger threshold. When an out-of-tolerance event is detected, start the incremental parameter update mechanism: correct the compensation weight matrix by recursive least squares method, and feed the updated weights back to the hybrid compensation model.

[0043] In this implementation scheme, an out-of-tolerance trigger threshold is set: an out-of-tolerance threshold (usually a deviation tolerance) is set to monitor the punching force, displacement, or other key process parameters during the riveting process. When a deviation of a parameter exceeds this threshold, the system determines it as an out-of-tolerance event and triggers the update process of the compensation weight parameters. An incremental parameter update mechanism is initiated: This mechanism means that after an out-of-tolerance event is detected, the system will gradually adjust the compensation weights according to the magnitude and trend of the deviation. This differs from the traditional full update mechanism. The incremental update mechanism avoids making a large adjustment to the entire model at once, thereby reducing the risk of over-correction and improving the system's adaptability and flexibility. Incremental updates calculate the compensation weight correction amount at the current moment and gradually adjust the weight parameters. The compensation weight matrix is ​​corrected using the recursive least squares method: The recursive least squares method is an adaptive filtering algorithm commonly used for parameter estimation in dynamic systems. This method allows for continuous updating of the compensation weight matrix based on real-time data. In practical applications, the RLS method adjusts the compensation weight matrix based on the data and errors of each measurement, making it more consistent with the recursive least squares formula for the current process state: The weight matrix is ​​set as W(k), where k represents time. The compensation weight matrix is ​​updated through a recursive calculation formula: ;in, The weight update amount is calculated using the following formula: ;here: Let be the covariance matrix. This represents the prediction error. This represents the update amount at the current moment. The updated weights are fed back to the hybrid compensation model: After updating the compensation weight matrix, these corrected parameters are fed back into the hybrid compensation model. The hybrid compensation model provides multi-dimensional compensation based on different process parameters (such as punching force, displacement, material properties, etc.), ensuring the real-time adaptability of the compensation system. Feedback process: The corrected weight parameters are applied to each stage of the compensation model, enabling the compensation effect to match the new process and equipment states in real time, maintaining system accuracy and stability.

[0044] In summary, this application has at least the following effects:

[0045] A control method for self-piercing riveting is proposed. Through multi-source data fusion interfaces and real-time data acquisition, it accurately captures key process parameters such as material hardness, thickness, and punching force during the riveting process, providing a comprehensive set of process parameters to ensure real-time monitoring and control of dynamic changes during the riveting operation. Combining the material's hardness, thickness, and plasticity characteristics, a deformation resistance evaluation function is established. Through the output of a dynamic compensation vector, precise compensation of punching force and rivet displacement is achieved, effectively solving dynamic errors in the riveting process caused by material properties or punching force fluctuations. An adaptive optimization algorithm generates a compensation weight matrix, and combined with an anti-saturation control strategy, dynamically adjusts the riveting punching force and rivet displacement. This adaptively responds to changes in different materials, thicknesses, or pressure fluctuations, improving process adaptability and ensuring riveting accuracy. By extracting real-time displacement deviation trends through multi-scale filtering and dynamically adjusting gain parameters, combined with the Lyapunov stability criterion, the control system maintains global stability under different disturbances, avoiding system instability or over-compensation, and improving the accuracy and stability of the riveting process. When consecutive deviations occur, the compensation weight parameters are iteratively updated. Through a recursive update mechanism, the compensation strategy is continuously optimized, further enhancing the system's adaptability under complex working conditions and ensuring that all parameters remain under optimal control throughout the riveting process. This method reduces equipment load and wear caused by inaccurate control by dynamically adjusting the control system's compensation strategy, thereby extending equipment lifespan and ensuring a more reliable riveting process.

[0046] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0047] This invention is described with reference to flowchart illustrations and / or block diagrams of systems, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0048] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0049] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0050] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0051] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A control method for self-piercing riveting, characterized in that, Includes the following steps: S1. Real-time acquisition of hardness, thickness and punching time data of riveting material through multi-source data fusion interface, capture of high dynamic deformation process of riveting joint, extraction of deformation rate and springback characteristic values, and construction of multi-dimensional process parameter set; S2. Based on the multi-dimensional process parameter set, comprehensively analyze the material hardness, thickness and plastic dissipation characteristics, establish a deformation resistance evaluation function, quantify the influence of different materials on the stopping distance, combine the stamping pressure fluctuation characteristics, analyze the nonlinear relationship between stamping pressure and rivet elastic deformation, construct a hybrid compensation model, and output the riveting dynamic coupling compensation vector. S3. Based on the riveting dynamic coupling compensation vector and process adaptability analysis results, the compensation weight matrix is ​​obtained through an adaptive optimization algorithm, and the riveting punching force and rivet displacement are dynamically adjusted in combination with the anti-saturation control strategy to generate an anti-saturation compensation instruction set. S4. Based on the anti-saturation compensation instruction set, the riveting punching force and rivet displacement are dynamically corrected. The real-time displacement deviation trend component is extracted through multi-scale filtering. The compensation strength is optimized by combining the anti-saturation compensation instruction set and the Lyapunov stability criterion to dynamically adjust the gain parameters of the control system. If the deviation exceeds the limit continuously, the iterative update of the compensation weight parameters is triggered. The specific process of capturing the high-dynamic deformation process of riveted joints, extracting the deformation rate and springback feature values, and constructing a multi-dimensional process parameter set is as follows: The high-deformation time-series images of the rivet joint are continuously captured by a high-frame-rate visual sensor. The instantaneous change gradient of the deformation rate is calculated by optical flow method, and the springback feature is extracted by edge detection algorithm. By integrating material hardness, thickness, and stamping time-series data, a multi-dimensional set of process parameters is constructed, which includes dynamic deformation characteristics, material properties, and process conditions. The specific process for establishing a deformation resistance evaluation function based on a comprehensive analysis of material hardness, thickness, and plastic dissipation characteristics is as follows: Based on the interaction between material hardness and thickness, a nonlinear deformation resistance evaluation function is constructed, and the influence of different materials on stopping distance is quantified by introducing the material plasticity dissipation rate. By combining experimental calibration data, a dynamic mapping relationship is generated to characterize the differentiated compensation requirements between high-hardness thin plates and low-hardness thick plates. Based on the characteristics of stamping force fluctuation, the nonlinear relationship between stamping force and rivet elastic deformation is analyzed, a hybrid compensation model is constructed, and the specific process of outputting the riveting dynamic coupling compensation vector is as follows: The time series data of the counter-pressure were modeled in stages, and hybrid compensation functions were constructed for the pressure increase, pressure stabilization and pressure decrease stages respectively to analyze the hysteresis relationship between the elastic deformation of the rivet and the pressure fluctuation. By introducing a dynamic hysteresis factor to correct the deviation, a dynamic coupled compensation vector containing pressure deformation compensation and material resistance compensation is output. Based on the riveting dynamic coupling compensation vector and process adaptability analysis results, the specific process of obtaining the compensation weight matrix through an adaptive optimization algorithm is as follows: A multi-objective optimization problem is constructed, with the optimization objectives being minimizing head height deviation, maximizing compensation stability, and minimizing equipment lifespan decay rate. Combining the weight benchmark parameters in the historical process database, the global optimal weight combination is solved through the gradient projection algorithm to generate the compensation weight matrix.

2. The control method for self-piercing riveting according to claim 1, characterized in that: The specific process of dynamically adjusting the riveting punching force and rivet displacement by combining the anti-saturation control strategy and generating the anti-saturation compensation instruction set is as follows: Based on the compensation weight matrix and real-time process status, an anti-saturation control strategy is designed, and the upper and lower thresholds of the punching force and displacement compensation are constrained by the amplitude limiting function. By combining a dynamic boundary expansion mechanism, the safety boundary is adaptively adjusted in scenarios of material mutation or equipment aging, generating an anti-saturation compensation instruction set.

3. The control method for self-piercing riveting according to claim 2, characterized in that: The specific process for dynamically correcting the riveting punching force and rivet displacement based on the anti-saturation compensation instruction set is as follows: The anti-saturation compensation instruction set is converted into execution-end control signals, which drive the servo system through a feedback fusion mechanism. The feedforward channel injects a compensation reference value, and the feedback channel dynamically corrects the execution amount based on the real-time displacement deviation, thereby achieving high-precision dynamic correction of the punching force and displacement.

4. The control method for self-piercing riveting according to claim 3, characterized in that: The specific process of extracting real-time displacement deviation trend components through multi-scale filtering, and then optimizing the compensation strength and dynamically adjusting the gain parameters of the control system by combining the anti-saturation compensation instruction set with the Lyapunov stability criterion is as follows: The low-frequency trend component of the displacement deviation is separated from the high-frequency noise by wavelet transform, and the compensation intensity is optimized by combining the anti-saturation instruction set. A stability criterion is constructed based on the Lyapunov function, and the proportional-integral gain parameter is dynamically adjusted to ensure the global asymptotic stability of the system under disturbance.

5. The control method for self-piercing riveting according to claim 4, characterized in that: If the error exceeds the limit consecutively, the specific process of triggering the iterative update of the compensation weight parameters is as follows: Set an out-of-tolerance trigger threshold. When an out-of-tolerance event is detected, start an incremental parameter update mechanism: correct the compensation weight matrix by recursive least squares method, and feed the updated weights back to the hybrid compensation model.

Images