A temperature control-based gilding process optimization method

Abstract

The application discloses a temperature control-based gilding process optimization method, and relates to the technical field of gilding processes.The method comprises the following steps: collecting real-time temperatures of multiple hot-pressing areas of a gilding device, obtaining multi-area temperature field distribution data and extracting temperature field characteristics; performing multi-scale heat conduction dynamic evolution trend analysis according to the temperature field characteristics, obtaining heat conduction dynamic evolution trend data, and combining with gilding plate thermal physical characteristic parameters to simulate gilding plate thermal inertia response, thereby obtaining thermal inertia response characteristic data; predicting temperature deviation trends according to the characteristic data, and determining a thermal compensation direction; quantifying a thermal compensation coefficient in combination with the compensation direction and the deviation prediction, thereby generating a feedforward compensation control signal; comparing the measured temperature with a target value to obtain a real-time deviation, thereby generating a feedback compensation control signal; time-aligning and superimposing the feedforward and feedback signals to synthesize a composite driving signal, and outputting the composite driving signal to a heating execution end, thereby realizing real-time optimization and regulation of the temperature.The application has the effects of improving gilding quality and production efficiency.

Description

Technical Field

[0001] This application relates to the field of hot stamping technology, and in particular to a method for optimizing hot stamping process based on temperature control. Background Technology

[0002] Hot stamping is widely used in packaging printing, label manufacturing, and other fields. Its core principle is to heat the hot stamping plate, melting the hot melt adhesive layer on the electroplated aluminum foil and transferring the metal foil to the substrate. The stamping temperature is a crucial parameter determining the quality of the hot stamping. Too low a temperature will result in insufficient melting of the hot melt adhesive, preventing proper stamping; too high a temperature will cause oxidation, discoloration, edge smudging, or even burn-through of the substrate. Therefore, precise control of the hot stamping plate temperature directly determines the yield and consistency of the hot stamped products.

[0003] It is worth noting that in high-speed continuous hot stamping production, the heat source of the hot stamping plate is not solely the active heating provided by the heating element. During each stamping cycle, the high-speed contact friction between the hot stamping plate and the substrate / electroplated aluminum generates continuous frictional heat accumulation. Simultaneously, as a metal component with a large heat capacity and thermal resistance, the hot stamping plate exhibits significant thermal inertia in its temperature changes. Each work cycle alternates between contact heat transfer cooling and intermittent heating heating, causing its instantaneous temperature to fluctuate periodically. When the stamping speed increases or the continuous running time extends, the frictional heat accumulates and accumulates, causing the actual temperature of the hot stamping plate to slowly drift from the set value. Operators often face the dilemma that the set temperature appears stable while the actual contact temperature deviates significantly, forcing them to reduce the stamping speed to accommodate the temperature control system's responsiveness, severely restricting production efficiency.

[0004] However, existing hot stamping equipment generally relies on PID feedback control for temperature control. Its post-deviation elimination mechanism means the controller only responds after the temperature has deviated from the target value. This inherent delay between deviation generation and control effectiveness makes it unsuitable for proactive prediction and compensation due to thermal inertia and heat accumulation. While some solutions incorporate empirical corrections such as temperature compensation during startup, the compensation amount is a fixed empirical constant, unable to adapt to dynamic conditions like changes in stamping speed, accumulated runtime, and ambient temperature fluctuations, resulting in low compensation accuracy. Furthermore, in existing solutions, temperature, speed, and pressure are independently controlled by different subsystems, lacking a collaborative optimization mechanism targeting the actual temperature at the moment of stamping contact. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this application provides an optimized hot stamping process based on temperature control.

[0006] This application provides a temperature-controlled optimization method for hot stamping processes, comprising the following steps: Step S1: Real-time temperature acquisition is performed on multiple hot pressing areas of the hot stamping equipment using a temperature sensor array to obtain multi-area temperature field distribution data; feature extraction is performed on the multi-area temperature field distribution data to obtain temperature field feature data; Step S2: Perform multi-scale dynamic evolution trend analysis of heat conduction based on temperature field characteristic data to obtain dynamic evolution trend data of heat conduction; obtain the thermophysical property parameters of the hot stamping plate, and simulate the thermal inertial response of the hot stamping plate in combination with the dynamic evolution trend data of heat conduction to obtain thermal inertial response characteristic data. Step S3: Based on the thermal inertia response characteristic data, predict the temperature deviation trend of the hot stamping process to obtain temperature deviation prediction data; determine the heat compensation direction of the hot stamping plate according to the temperature deviation prediction data to obtain heat compensation direction determination data. Step S4: Quantize the thermal compensation coefficient based on the thermal compensation direction determination data and temperature deviation prediction data to obtain the thermal compensation coefficient; generate a feedforward compensation control signal based on the thermal compensation coefficient; Step S5: Based on the comparison between the multi-region temperature field distribution data and the preset target temperature data, obtain the real-time temperature deviation value, and generate a feedback compensation control signal based on the real-time temperature deviation value; perform time-series registration and superposition synthesis of the feedforward compensation control signal and the feedback compensation control signal to generate a composite drive signal; output the composite drive signal to the heating execution end to realize real-time optimization and control of the hot stamping plate temperature.

[0007] In summary, this application includes at least one of the following beneficial technical effects: 1. Existing hot stamping equipment often uses fixed empirical constants for temperature control compensation, which cannot adaptively adjust to dynamic conditions such as changes in stamping speed, accumulated runtime, and ambient temperature fluctuations, resulting in low compensation accuracy. This application employs an empirical mode decomposition method to adaptively decompose the temperature time-series signal during the temperature field feature extraction stage. This method decomposes the signal entirely based on its local characteristics without pre-setting basis functions, making it particularly suitable for processing measured engineering signals with nonlinear and non-stationary characteristics, such as the temperature signal during hot stamping. The Hilbert-Huang transform is used to characterize the time-varying distribution of temperature fluctuation energy on the time-frequency plane, extracting nonlinear features such as the energy proportion of each intrinsic mode function, the instantaneous frequency mean, the slope of the residual trend term, and the spatial correlation coefficient matrix. This provides an accurate data foundation for analyzing the dynamic evolution trend of heat conduction and simulating thermal inertia response. In the temperature deviation trend prediction stage, an autoregressive integral moving average model is used to model the historical time series of temperature deviations. When the actual temperature deviation direction is detected to be inconsistent with the predicted deviation direction, the model parameters are immediately updated online, enabling the prediction model to quickly track sudden disturbances such as abrupt changes in substrate material and sudden changes in ambient temperature. The above mechanism enables the feedforward compensation to always dynamically adapt to changes in actual operating conditions, significantly improving the robustness and adaptability of temperature control.

[0008] 2. In existing solutions, temperature, speed, and pressure are independently controlled by different subsystems, lacking a collaborative optimization mechanism targeting the actual temperature at the instant of hot stamping contact. This application arranges a temperature sensor array according to the spatial distribution of the hot-pressing areas, ensuring that each functional area of ​​the plate is covered by sensor nodes, and performs spatial interpolation on each node to obtain a continuous temperature field estimate. During the quantification stage of the thermal compensation coefficient, a thermal inertia correction factor is introduced to dynamically correct the differences in the equivalent thermal inertia response time constant of each hot-pressing area. Areas with higher thermal inertia have their compensation coefficients appropriately increased to compensate for thermal inertia attenuation. Simultaneously, during the feedback compensation stage, the proportional, integral, and derivative coefficients are tuned according to the thermal inertia response time constant of each hot-pressing area. Areas with higher inertia have their proportional and derivative coefficients appropriately decreased and their integral coefficients appropriately increased to avoid oscillation overshoot caused by response hysteresis. Through the above multi-area differentiated compensation strategy, the temperature of each hot-pressing area of ​​the hot stamping plate remains consistently stable near the set value, effectively ensuring the consistency and stability of the hot stamping process quality. Attached Figure Description

[0009] To more clearly illustrate the technical solutions of the embodiments of the present invention, 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a schematic diagram of the steps of an optimized hot stamping process based on temperature control according to this application.

[0011] Figure 2 This is a data flow diagram of a temperature-controlled hot stamping process optimization method according to this application.

[0012] Figure 3 This is a schematic diagram of the system architecture of a temperature-controlled hot stamping process optimization method proposed in this application. Detailed Implementation

[0013] The following description, in conjunction with the implementation of the present invention, is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention, and all such modifications and additions should fall within the protection scope of the present invention.

[0014] Example

[0015] This application applies to real-time optimization control of hot stamping plate temperature in hot stamping processes, especially in high-speed continuous hot stamping production where the actual temperature slowly drifts from the set value and the instantaneous temperature fluctuates periodically due to the continuous accumulation of frictional heat and the thermal inertia of the hot stamping plate itself. Its core idea is not simply to rely on PID feedback control for post-deviation elimination, but rather to simultaneously collect multi-region temperature field distribution data of multiple hot-pressing areas through a temperature sensor array during the hot stamping process. First, temperature field features are extracted and multi-scale dynamic evolution trend analysis of heat conduction is performed. After obtaining the dynamic evolution trend data of heat conduction, thermal inertia response simulation is further performed using the thermophysical characteristic parameters of the hot stamping plate to obtain thermal inertia response characteristic data. Subsequently, temperature deviation trend prediction and thermal compensation direction determination are performed based on the thermal inertia response characteristic data, thereby generating a feedforward compensation control signal. Simultaneously, the real-time collected temperature value is compared with the preset target temperature to obtain the real-time temperature deviation value and generate a feedback compensation control signal. The feedforward compensation control signal and the feedback compensation control signal are time-registered and superimposed before being output to the heating actuator. With this treatment, even if the hot stamping speed is increased or the continuous running time is extended, leading to increased heat accumulation and temperature response lag due to thermal inertia, the heating actuator can still obtain a composite drive signal that takes into account both advance prediction and real-time correction, reducing temperature overshoot, fluctuation and lag, and improving hot stamping quality and production efficiency.

[0016] This application discloses an optimization method for hot stamping process based on temperature control.

[0017] Reference Figure 1-3 A temperature-controlled hot stamping process optimization method includes the following steps: The temperature of multiple hot pressing areas of the hot stamping equipment is collected in real time by a temperature sensor array to obtain multi-area temperature field distribution data; feature extraction is performed on the multi-area temperature field distribution data to obtain temperature field feature data. Furthermore, the temperature of multiple hot-pressing areas of the hot stamping equipment is collected in real time using a temperature sensor array to obtain multi-region temperature field distribution data; feature extraction is performed on the multi-region temperature field distribution data to obtain temperature field feature data, including the following steps: The temperature of multiple hot pressing areas of the hot stamping equipment is collected in real time by a temperature sensor array to obtain temperature field distribution data of multiple areas. The multi-region temperature field distribution data is cleaned and spatially interpolated to obtain standard temperature field distribution data. Temperature field temporal fluctuation analysis was performed on the standard temperature field distribution data to obtain temperature field temporal fluctuation data. The time-frequency domain nonlinear features of the temperature field fluctuation data are extracted to obtain the temperature field feature data.

[0018] In one specific implementation, the core physical process of hot stamping is the transfer of electroplated aluminum foil to the substrate surface under high temperature and high pressure using a hot stamping plate. Temperature is the primary process parameter determining the transfer quality. The uniformity and stability of the hot stamping plate surface temperature directly affect the hot stamping effect: excessively high temperatures will cause the foil layer to soften excessively, resulting in spread of the transfer area and causing smearing defects; excessively low temperatures will cause the hot melt adhesive layer between the foil layer and the substrate to fail to activate sufficiently, resulting in missed stamping or insufficient adhesion defects. Therefore, real-time and comprehensive acquisition of the temperature of multiple hot-pressing areas of the hot stamping plate is the data foundation for the entire optimization method.

[0019] In actual operation, the surface of the hot stamping plate in a hot stamping machine is not uniformly heated. Due to differences in the layout of heating elements, the thermal conductivity of the plate material, and heat dissipation conditions, significant temperature gradients exist between different areas of the plate. If temperature sensors are only placed at a single point or a few points, the collected temperature information cannot reflect the overall temperature distribution of the plate. When overheating or underheating occurs in a local area of ​​the plate, the single-point acquisition method cannot detect it in time. Therefore, a temperature sensor array is used to collect real-time, comprehensive data across multiple hot-pressing areas of the hot stamping machine. The sensor array is arranged according to the spatial distribution of the hot-pressing areas to ensure that sensor nodes cover all functional areas of the plate.

[0020] The surface of the hot stamping stencil is divided into M hot-pressing areas according to functional regions. The value of M is determined based on the size of the hot stamping stencil, typically ranging from 6 to 24. For a common hot stamping stencil with a size of 300mm × 400mm, its surface can be divided into 12 hot-pressing areas (3 rows × 4 columns), each area approximately 100mm × 100mm in size. For a large hot stamping stencil with a size of 500mm × 700mm, it can be divided into 24 hot-pressing areas (4 rows × 6 columns). The hot-pressing areas are evenly distributed along the length and width of the hot stamping stencil surface.

[0021] At the geometric center of each hot-pressed area, a blind hole with a diameter of 3mm to 5mm is made. The depth of the blind hole is 1 / 3 to 1 / 2 of the thickness of the hot stamping plate. The temperature sensor probe is embedded in the blind hole, and the gap between the probe and the wall of the blind hole is filled with thermal grease to ensure good thermal contact. The temperature sensor is a type K thermocouple or a PT100 platinum resistance temperature sensor, and its signal line is led out along the pre-set wiring groove on the back of the hot stamping plate.

[0022] The sensor array is arranged according to the following principles: one sensor node is placed at each of the four corner areas of the hot stamping plate, one sensor node is placed at the geometric center of the plate, and the remaining sensor nodes are evenly interpolated between the corner nodes and the center node. This arrangement ensures that all functional areas of the plate are covered by sensor nodes, while also acquiring spatial distribution information of the plate's temperature field with a relatively small number of sensors. The center-to-center distance between adjacent sensor nodes does not exceed 150mm to ensure the accuracy of the temperature field estimation after spatial interpolation.

[0023] For hot stamping plates with partitioned heating tubes, the division of each hot pressing area corresponds one-to-one with the partitioned arrangement of the heating tubes, and at least one sensor node is arranged in the area covered by each heating tube.

[0024] Each node uses a uniform sampling period. Synchronous data acquisition is performed. At the k-th sampling time... The raw temperature reading collected by the m-th sensor node is denoted as . The raw temperature readings of all M nodes are arranged in both spatial and temporal dimensions to form multi-regional temperature field distribution data. .

[0025] Because sensor nodes may be affected by factors such as electromagnetic interference, mechanical vibration, or aging under actual operating conditions. Outliers, missing values, or short-term noise inevitably exist in the raw data. If feature extraction is performed directly on the raw data, these outliers will be mixed into the subsequent analysis, causing bias. Therefore, the temperature field distribution data of multiple regions is first cleaned.

[0026] Data cleaning employs an anomaly detection method based on the moving median. This is used for the time-series data of the m-th node. , with length as Sliding window calculation of median within the window Median of absolute deviation When the deviation of a certain sampling point exceeds When this happens, the sampling point is marked as an outlier and removed. This represents the anomaly detection sensitivity coefficient. Compared to the 3σ criterion based on the mean and standard deviation, the detection method based on the absolute deviation of the median is more robust to extreme outliers. It avoids the misclassification of a large amount of normal data as anomalies due to a single extreme point raising the mean, making it more suitable for real-world scenarios involving occasional spikes in the temperature signal during hot stamping. For rejected outliers, linear interpolation of the preceding and following normal sampling points is used to fill in the gaps, ensuring the continuity of the time-series data.

[0027] After completing the node-by-node time-series cleaning, the insufficient spatial resolution caused by the sparse arrangement of sensor nodes needs to be addressed. The temperature distribution of each area on the hot stamping plate is spatially continuous, while the sensor nodes only cover a limited number of discrete locations, and the temperature of the areas between sensor nodes on the plate cannot be directly obtained. Therefore, spatial interpolation is performed on the cleaned temperature values ​​of each node at the current moment to obtain an estimate of the continuous temperature field on the plate. Considering that the temperature field of the hot stamping plate is a deterministic physical field governed by the heat conduction equation, and that the number of effective sensor nodes on the plate is limited, it is difficult to meet the requirement of Kriging interpolation for semi-variogram statistical fitting of a large number of sample point pairs. Therefore, this embodiment uses the radial basis function (RBF) interpolation method for spatial interpolation.

[0028] RBF interpolation does not rely on large-sample statistical estimation and can construct a stable, smooth, and oscillating continuous temperature field distribution even for sparse discrete nodes. Furthermore, it boasts high computational efficiency, meeting the response time requirements for real-time temperature control in the hot stamping process. Specifically, the radial basis function used in this embodiment is an inverse polyquadratic function: ; in, The Euclidean distance between the interpolation point and the known node is given. To determine shape parameters based on node density, and to ensure the interpolation matrix has a good condition number and the interpolation surface is smooth and free from overfitting, this embodiment... The distance is determined based on the average spatial distance d between each sensor node. Let the spatial distance between the i-th and j-th sensor nodes on the hot stamping plate be d. Then the average distance between all adjacent node pairs The shape parameters are calculated from the actual coordinates of each node, according to... Values.

[0029] Inverse quadratic functions possess the excellent property of nonsingular interpolation matrices, are insensitive to the number of nodes, and can generate temperature field surfaces with good smoothness, which highly matches the physical characteristics of continuous and smooth transition of the temperature field in hot stamping plates. After spatial interpolation, the number of spatial nodes obtained on the plate's spatial grid is... Standard temperature field distribution data .

[0030] Standard temperature field distribution data This describes the absolute distribution of the temperature field. However, for subsequent analysis, the relative fluctuations of the temperature field over time—that is, the deviations of the temperature in each hot-pressing region from its set value and their temporal evolution—are more valuable than the absolute temperature values ​​themselves. Therefore, [the following text is incomplete and requires further context]. Perform time-series fluctuation analysis of the temperature field.

[0031] Let the first The temperature setting value for each grid node is... The temperature deviation of this node at the k-th sampling time is defined as: ; Time series of temperature deviations at each node { Perform time-series fluctuation analysis and calculate the fluctuations of each node in continuous time. Fluctuation statistics within each sampling time point, including the mean deviation. Deviation and standard deviation and deviation range The three statistics characterize the temperature fluctuations of each hot-pressing region from three dimensions: the level of deviation center, the degree of dispersion, and the amplitude of change, respectively, and together constitute the time-series fluctuation data of the temperature field. .

[0032] Temperature field time series fluctuation data Statistical descriptions of temperature fluctuations at each node are provided, but time-domain statistics alone are insufficient to fully describe the dynamic evolution of the temperature field. Therefore, [further details are needed]. Nonlinear features in the time and frequency domains are extracted to capture the dynamic features of temperature fluctuations in both the time and frequency domains.

[0033] Empirical Mode Decomposition (EMD) was chosen for time-frequency domain feature extraction. Compared to Short-Time Fourier Transform (SFT), which requires a pre-defined fixed window length, and Wavelet Transform, which requires pre-selected wavelet basis functions, EMD performs adaptive decomposition entirely based on the signal's local features, without the need for pre-defined basis functions. This makes it particularly suitable for processing measured engineering signals with nonlinear and non-stationary characteristics, such as temperature signals from hot stamping processes. The EMD method extracts the time series temperature deviations from each node... The adaptive decomposition consists of several intrinsic mode function (IMF) components. With residual trend term .

[0034] The sieving stopping criterion for empirical mode decomposition adopts the standard deviation criterion. Specifically, for the j-th sieving iteration, the standard deviation value between two adjacent sieving results is calculated. : ; in, These are the candidate components obtained after the (k-1)th and kth sieving processes, respectively. When the SD value falls between 0.2 and 0.3, the current sieving iteration is considered to have met the stopping condition, and the current iteration is... The IMF component is output as an intrinsic mode function. The selection of the above threshold range can ensure that the obtained IMF component has a stable instantaneous frequency physical meaning, while avoiding excessive screening that would cause the component to lose its physical authenticity.

[0035] Specifically, the number n of the Intrinsic Mode Function (IMF) components is not preset, but adaptively determined by the signal's local time-scale characteristics. Each successful extraction of an IMF component... Then, subtract this component from the original signal to obtain the residual signal. and to Determine the extreme point: when the residual signal When the number of extreme points is less than two, it is determined that the residual signal no longer has a clear oscillation mode, and the decomposition process terminates. The residual signal at this point is the final residual trend term. The total number of IMF components obtained from the decomposition is n.

[0036] The IMF components obtained from the above adaptive decomposition Arranged from highest to lowest frequency, each component has a distinct physical meaning. Specifically, the high-frequency IMF components... The instantaneous frequency of the temperature pulse response of each hot-pressing zone at the moment of a single pressing is consistent with the hot stamping frequency, and the amplitude reflects the intensity of frictional heat input during a single pressing; the mid-frequency IMF component to Corresponding to the heat dissipation process within adjacent pressing intervals and the interactive modulation effect of heat accumulation and release across multiple pressing cycles, its instantaneous frequency is lower than the hot stamping frequency, and the amplitude variation rate reflects the stability of heat dissipation conditions; low-frequency IMF component Corresponding to the slow drift trend of the plate temperature during continuous production, its instantaneous frequency is close to zero, and the amplitude change reflects the net increase of the heat accumulation effect. The final residual trend term r(t) characterizes the long-term monotonic evolution baseline of the temperature deviation, and its slope The sign and magnitude of the value directly reflect the net balance between heat accumulation and heat dissipation during continuous production.

[0037] Among them, each IMF component Representing temperature fluctuation patterns at different time scales, arranged from high to low frequency, residual trend term This represents the long-term trend component of the temperature deviation. The Hilbert-Huang transform is further calculated for each IMF component to obtain its instantaneous frequency. With instantaneous amplitude This allows us to characterize the time-varying distribution of temperature fluctuation energy on the time-frequency plane.

[0038] The following nonlinear characteristic quantities are extracted from each IMF component and its Hilbert spectrum to form the temperature field characteristic data. Energy percentage of each IMF component Instantaneous frequency mean of dominant IMF components The slope of the residual trend term and the spatial correlation coefficient matrix of IMF components between nodes Temperature field characteristic data It integrates the characteristics of three dimensions: time-domain statistics, frequency-domain energy, and time-frequency coupling, and fully describes the dynamic characteristics of the temperature field during hot stamping.

[0039] Multi-scale dynamic evolution trend analysis of heat conduction is performed based on temperature field characteristic data to obtain dynamic evolution trend data of heat conduction; thermophysical property parameters of hot stamping plate are obtained, and thermal inertial response of hot stamping plate is simulated in combination with dynamic evolution trend data of heat conduction to obtain thermal inertial response characteristic data. Furthermore, based on the temperature field characteristic data, a multi-scale dynamic evolution trend analysis of heat conduction is performed to obtain dynamic evolution trend data of heat conduction. The thermophysical property parameters of the hot stamping plate are obtained, and the thermal inertial response of the thermophysical property parameters of the hot stamping plate is simulated based on the temperature field characteristic data and the dynamic evolution trend data of heat conduction to obtain thermal inertial response characteristic data.

[0040] Furthermore, based on the temperature field characteristic data and the heat conduction dynamic evolution trend data, thermal stress coupling fluctuation response calculation is performed to obtain thermal stress coupling fluctuation response data; The initial heat capacity and thermal resistance parameters of the hot stamping plate are extracted to obtain the initial heat capacity and thermal resistance parameter data. Based on the thermal stress coupling fluctuation response data, the initial thermal capacity and thermal resistance parameter data are corrected for thermal inertial response characteristics to obtain thermal inertial response characteristic data.

[0041] In one specific implementation, during continuous production, the hot stamping plate does not immediately return to its initial temperature after each pressing action. The heat accumulated during the previous pressing process is slowly dissipated to the surrounding medium through heat conduction, convection, and radiation. However, before the next pressing action begins, the residual heat on the plate surface is often not completely dissipated, resulting in a cumulative heat accumulation phenomenon between successive pressing actions, known as the heat accumulation effect. The existence of this heat accumulation effect causes the actual working temperature of the hot stamping plate to gradually increase with the number of pressing actions during continuous production. If this is not quantified and compensated for, the plate temperature will deviate from the set value, leading to a systematic deterioration in hot stamping quality as production batches increase.

[0042] From temperature field characteristic data The first step is to analyze the dynamic evolution trend of heat conduction at multiple scales. The heat conduction process inside the hot stamping plate exhibits multi-level characteristics on a time scale: at the millisecond time scale of a single press, it mainly manifests as contact heat conduction between the plate and the substrate; at the second time scale of the interval between adjacent presses, it mainly manifests as heat dissipation from the plate to the surrounding air and mechanical structure; at the minute time scale of continuous production, it mainly manifests as the net heat accumulation trend formed by the repeated superposition of the above two processes. The dynamic characteristics of heat conduction at these three time scales are superimposed and jointly determine the evolution law of plate temperature.

[0043] In view of the above-mentioned multi-scale characteristics, the temperature field characteristic data The dynamic evolution of heat conduction was analyzed for the IMF components at each time scale. At the millisecond scale, the temperature step amplitude at the compression moment was extracted from the high-frequency IMF components. This describes the instantaneous heat input intensity of a single compression; on a second-scale, it extracts the temperature decay within adjacent compression intervals from the mid-frequency IMF component. This describes the rate of heat dissipation on the printing plate during the pressing interval; at a minute-level scale, the temperature drift rate of the residual trend term is extracted. This is used to describe the long-term accumulation trend of plate temperature during continuous production. , and Combined into dynamic evolution trend data of heat conduction .

[0044] Thermal inertia refers to the characteristic that, due to the inherent heat capacity of the hot stamping plate, the plate's temperature response to changes in external heating power exhibits a time lag and amplitude attenuation relative to the change in heating power. The existence of thermal inertia means that even if a compensation control signal is applied to the heating actuator at the current moment, the actual change in plate temperature will only be fully reflected after several control cycles. If the compensation control does not consider the influence of thermal inertia, problems of overshoot or undershoot will occur.

[0045] Thermal inertial response simulation first requires obtaining the thermophysical properties of the hot stamping plate, including the specific heat capacity of the plate. Plate density Plate thickness h and plate thermal conductivity .

[0046] Specifically, for hot stamping plates made of steel, the values ​​of their thermophysical properties are as follows: specific heat capacity Approximately 460 Up to 500 ,density Approximately to thermal conductivity Approximately to The thickness h of the hot stamping plate is determined according to the actual hot stamping plate specifications, and is usually 10mm to 30mm. For hot stamping plates made of zinc or copper, the above example parameters can be replaced with the corresponding specific heat capacity, density, and thermal conductivity values ​​in the material handbook, and the thermal inertia response can be estimated using the same method.

[0047] The thermal inertial response simulation uses a lumped heat capacity model, which treats each hot-pressed region as an equivalent heat capacity. With a thermal resistance A series of first-order heat networks, in which For the region volume, the response of the plate temperature to a step input of heating power is expressed in first-order exponential form, with a response time constant of . Compared to the finite element thermal analysis method, the lumped heat capacity model requires far less computation and can complete the simulation calculation in real time within each control cycle, meeting the real-time requirements of online control. Compared to purely empirical formulas, the lumped heat capacity model has clear physical meaning, and the parameters can be identified through measured data, making the simulation results more interpretable.

[0048] However, during continuous production, the contact thermal resistance between the printing plate and the substrate... It is not constant, but rather changes non-linearly with variations in plate temperature, pressing pressure, and substrate material. Directly using nominal thermal resistance parameters for simulation will introduce systematic errors. Therefore, a non-linear correction for the contact thermal resistance is necessary. This is based on temperature field characteristic data. Extract the interfacial heat flux density distribution data at the pressing time of each hot-pressing region. Based on Fourier's law of heat conduction, and determined by interfacial heat flux density The equivalent contact thermal resistance at the current moment is calculated by inversely calculating the measured temperature difference. The revised Substituting into the lumped heat capacity model, we obtain the equivalent thermal inertia response time constant of each hot-pressed region at the current moment. This time constant is the core characteristic data characterizing the thermal inertia response of the hot stamping plate, quantitatively describing the degree of lag in the response of the plate surface temperature to the heating control signal and the duration of the transition process in each area. The time constant of each area... Along with its corresponding heat capacity With thermal resistance Together, they constitute complete thermal inertial response characteristic data.

[0049] Based on the thermal inertia response characteristic data, the temperature deviation trend of the hot stamping process is predicted to obtain temperature deviation prediction data; based on the temperature deviation prediction data, the heat compensation direction of the hot stamping plate is determined to obtain heat compensation direction determination data.

[0050] Furthermore, based on the thermal inertia response characteristic data, a temperature deviation trend analysis of the hot stamping process is performed to obtain temperature deviation trend data; The temperature deviation trend data is subjected to predictive extrapolation to obtain predicted temperature deviation data; Obtain the actual temperature deviation direction of the current hot stamping plate, compare the consistency of the actual temperature deviation direction with the predicted deviation direction in the temperature deviation prediction data, and generate heat compensation direction determination data.

[0051] Furthermore, based on the temperature deviation trend data, the temperature deviation in subsequent control cycles is recursively predicted to obtain a multi-step predicted deviation sequence. Statistical features are extracted from the multi-step prediction deviation sequence to extract the deviation peak, deviation mean, and deviation duration interval, and temperature deviation prediction data is generated based on the deviation peak, deviation mean, and deviation duration interval.

[0052] In one specific implementation, thermal inertial response characteristic data The current thermal response dynamics of the hot stamping plate are described, but the current thermal inertia parameters alone are insufficient to directly guide compensation control. The control system needs to know how the plate temperature will deviate from the set value, the magnitude of the deviation, and whether the deviation will be too high or too low over several future control cycles. Only by reliably predicting the future evolution trend of the temperature deviation can targeted feedforward compensation be generated in advance, achieving proactive suppression of the temperature deviation rather than reactive tracking. Therefore, the purpose of step S3 is based on... Predict future trends in temperature deviation and determine the direction of thermal compensation accordingly.

[0053] It should be noted that the temperature deviation trend prediction in this application is not a fitting prediction of the residual deviation after closed-loop control, but rather based on the dynamic evolution trend data of heat conduction. The inherent temperature drift rate and heat dissipation rate extracted from the model are used to extrapolate and estimate the natural evolution trend of temperature deviation without feedforward compensation. Historical temperature deviation time-series data are preferentially selected for model initialization, prioritizing data from the initial startup phase before feedforward compensation is implemented or the calibration phase where the feedforward compensation is set to zero, to avoid interference from control actions in identifying the disturbance evolution pattern. During operation, the predicted deviation sign output by the model is used to determine the direction of temperature deviation that will naturally occur in the future due to physical heat accumulation, rather than a repetitive description of the current closed-loop response.

[0054] from The equivalent thermal inertia response time constant of each hot-pressed region at the current moment is read from the middle. Heat capacity With thermal resistance Combined with temperature field characteristic data The provided current temperature deviation values ​​for each region And its recent trends, and the rate of change of temperature deviation in each region. The estimation is performed, and the rate of change reflects whether the temperature deviation is widening or converging at the current moment, and the extent of this widening or convergence. Specifically, The difference estimation is performed by comparing the temperature deviation sequence between the current time and the previous few sampling times, and then combined with... Thermal inertia correction is applied to the difference results to eliminate the influence of thermal inertia hysteresis on the rate estimation, resulting in an estimated value of the deviation change rate after inertia correction.

[0055] Temperature deviation values ​​of each region and the rate of change of deviation Organized according to both spatial and temporal dimensions, and incorporating dynamic evolution trend data of heat conduction. Provided long-term drift rate With heat dissipation rate The current evolution trend of temperature deviation in each region is comprehensively assessed, including whether the deviation is in a continuous accumulation trend of increasing, whether it has shown signs of decline after reaching a peak, or whether it is in a transitional response process caused by external heating control intervention. The above assessment results constitute temperature deviation trend data. .

[0056] In obtaining temperature deviation trend data Subsequently, it is necessary to quantitatively predict the temperature deviation of each hot-pressing region within multiple subsequent control cycles. The prediction extrapolation employs a time-series modeling method, first analyzing the historical time series of temperature deviations for each region. Time series modeling is performed to establish a parameterized model that can describe the evolution of the deviation over time, thus obtaining the parameter set of the temperature deviation time series model. .

[0057] The time series modeling uses the Autoregressive Integral Moving Average (ARIMA) model. Compared to simple linear extrapolation methods, the ARIMA model can simultaneously capture the autocorrelation and random disturbance components of the temperature deviation time series, thus making more reliable multi-step predictions for non-stationary, drifting temperature deviation time series. Compared to methods such as state-space models, the ARIMA model has a smaller computational burden for parameter identification and can complete online updates within each control cycle, meeting the response time requirements of real-time control.

[0058] Specifically, the order determination and parameter identification of the ARIMA model are carried out according to the following steps: First, the historical time series of temperature deviations in each hot-pressing zone were analyzed. Perform an ADF test. The null hypothesis of the ADF test is that the series has a unit root. If the p-value obtained from the test is less than the significance level of 0.05, the null hypothesis is rejected, and the series is determined to be stationary. In this case, the difference order d = 0. If the p-value is greater than or equal to 0.05, the series is first-order differencing is performed, and the ADF test is performed again on the differised series until the series is stationary. The difference order at this point is the integration order d. Due to the physical characteristics of temperature deviation time series, the value of d is usually 0 or 1, and rarely exceeds 1.

[0059] After determining the difference order d, the autocorrelation function (ACF) and partial autocorrelation function (PACF) of the stationary sequence are calculated. By observing the truncation and tailing characteristics of ACF and PACF, the candidate value ranges of p and q are preliminarily determined: if PACF truncates after lag p and ACF tails, it is preliminarily determined to be an AR(p) model; if ACF truncates after lag q and PACF tails, it is preliminarily determined to be an MA(q) model; if both ACF and PACF exhibit tailing, it is determined to be an ARMA(p,q) model.

[0060] To further refine the determination of the order, the Akaike Information Criterion (AIC) was used for quantitative optimization. Specifically, the search range for both p and q was set to 0 to... ,in The value is determined based on the length N of the historical data, and is usually taken as [value missing]. Alternatively, a fixed value of 5 can be preset. Iterate through all (p,q) combinations within this range and calculate the AIC value for each combination: ; in This represents the maximum likelihood estimate of the model at a given order. The order combination (p,q) with the smallest AIC value is selected as the optimal order of the model. When the difference between the minimum and second smallest AIC values ​​is less than 2, combinations with lower orders are preferred to reduce the risk of overfitting.

[0061] After determining the optimal order (p,q,d), the unknown parameters of the ARIMA model are identified using the maximum likelihood estimation method. For the m-th hot-pressed region, the model expression is: ; Where B is the lag operator, It is an autoregressive polynomial. It is a moving average polynomial. The sequence is white noise. The parameters to be identified include autoregressive coefficients. , ,…, and moving average coefficient , ,…, There are a total of p+q parameters. Maximum likelihood estimation maximizes the likelihood function given the observed data. To obtain the optimal estimate of the parameters, the iterative solution employs the Newton-Raphson algorithm or the BFGS quasi-Newton algorithm. The convergence condition is set as the parameter change between two consecutive iterations being less than 1 / 2. .

[0062] After completing parameter identification, the model fitting residual series White noise testing was performed using the Ljung-BoxQ statistic. ; Where N is the data length, L is the maximum lag period, and is taken as... , For the residual sequence in lag The autocorrelation coefficient of the order. At a significance level of 0.05, if the QQ value is less than the critical value of the chi-square distribution. If the residual sequence is deemed white noise, the model passes the diagnostic test and can be used for subsequent multi-step predictions. If it fails the test, the order optimization step is repeated, re-identifying parameters and testing them again in the next best order combination with the best AIC value, until it passes. If no model passes the diagnostic test after traversing all candidate order combinations, the current model parameters are kept unchanged, and the temperature deviation prediction data for that region is marked as low confidence. The corresponding thermal compensation coefficient is multiplied by a decay factor between 0 and 1, such as 0.5, to reduce the dependence on feedforward compensation.

[0063] Based on the parameter set of the time series model for temperature deviation The temperature deviation of the m-th region within the subsequent H control cycles is recursively predicted to obtain a multi-step prediction deviation sequence. The prediction step size H is based on the current hot stamping frequency and the response time constant of the heating execution end. The prediction step size must be dynamically determined and must at least cover the time span from the current moment until the compensation control signal fully applies to the plate temperature. This is to ensure that the prediction time domain is sufficient to support the advance calculation of the feedforward compensation amount.

[0064] Statistical features were extracted from the multi-step prediction bias sequences of each region, and the peak bias was extracted. Mean deviation and the duration of the deviation These three factors together constitute a quantitative description of the temperature deviation in the m-th region. The predicted statistical characteristics of all M hot-pressing regions are then summarized to form complete temperature deviation prediction data. .

[0065] The purpose of determining the direction of thermal compensation is to analyze temperature deviation prediction data. Based on this, the direction of thermal compensation required for each hot-pressing zone is further determined, i.e., whether to apply increased or decreased thermal compensation to the heating actuator under the current operating conditions. Correctly determining the direction of thermal compensation is a key prerequisite for avoiding reverse superposition of compensations and preventing temperature overshoot.

[0066] Obtain the actual temperature deviation direction of each hot-pressing area of ​​the current hot stamping plate. Compare the actual temperature deviation direction with the predicted temperature deviation data. The consistency of the prediction deviation direction of the multi-step prediction deviation sequence in each region is compared: if the actual temperature deviation direction is consistent with the prediction deviation direction, it indicates that the current temperature deviation is developing along the prediction trajectory, the prediction result is reliable, and the mean of the prediction deviation is directly used. The symbol serves as the compensating polarity identifier for this region. , The corresponding temperature is too high, so heat reduction compensation needs to be applied. (Compensation polarity indicator) ; The corresponding temperature is too low, so heat compensation needs to be applied. (Compensation polarity indicator) If the actual temperature deviation direction is inconsistent with the predicted deviation direction, it indicates that there is a sudden disturbance that has not been captured by the prediction model at the current moment, such as a sudden change in the substrate material or a sudden change in ambient temperature. In this case, the compensation polarity label should be re-determined based on the actual temperature deviation direction. Simultaneously triggering the time series model parameter set Online updates are performed to correct the predictive model so that it keeps up with the latest operating conditions. The compensation polarity of each hot-pressing zone is identified. ( The data is compiled to form the thermal compensation direction determination data. .

[0067] The thermal compensation coefficient is quantized based on the thermal compensation direction determination data and temperature deviation prediction data to obtain the thermal compensation coefficient; a feedforward compensation control signal is generated based on the thermal compensation coefficient.

[0068] Furthermore, the compensation polarity identifier in the thermal compensation direction determination data and the amplitude characteristic quantity in the temperature deviation prediction data are obtained; Based on the compensation polarity identifier and the amplitude characteristic quantity, the expected adjustment amount and adjustment direction of the heating power are calculated, and the feedforward compensation amount is generated. Obtain the real-time hot stamping frequency of the hot stamping plate, and perform timing advance phase correction on the feedforward compensation amount based on the real-time hot stamping frequency of the hot stamping plate to obtain the timing advance correction amount. The timing advance correction amount is mapped to the amplitude of the drive signal at the heating actuator to generate a feedforward compensation control signal.

[0069] In one specific implementation, thermal compensation direction determination data The direction in which compensation needs to be applied to each hot-pressing zone has been determined, and the temperature deviation prediction data has been obtained. It provides quantitative information on the magnitude of future temperature deviations in each region. Based on this, the magnitude and timing of the compensation amount are further determined, and finally converted into a feedforward compensation control signal that can be directly responded to by the heating actuator.

[0070] The fundamental difference between feedforward compensation and feedback compensation lies in their timing and magnitude. Feedback compensation relies on responding after the temperature deviation has occurred, inherently lagging behind the disturbance's emergence. Feedforward compensation, on the other hand, applies compensation control based on predictions of future temperature deviations before they fully materialize, suppressing the disturbance before the plate temperature fully detects it. Feedforward compensation can suppress predictive biases that feedback compensation cannot overcome, but the compensation amount must precisely match the actual requirements in both amplitude and timing dimensions; otherwise, it will introduce new overshoot or undercompensation problems. Therefore, the quantification of the thermal compensation coefficient must consider both the compensation amplitude and timing dimensions simultaneously.

[0071] thermal compensation coefficient This is the core conversion parameter connecting the temperature deviation prediction result and the driving quantity at the heating execution end. Its physical meaning is the ratio between the heating power adjustment required to offset the predicted temperature deviation and the magnitude of the predicted deviation. For the m-th hot-pressing region, the quantification of the thermal compensation coefficient utilizes the compensation polarity identifier. Temperature deviation prediction data The amplitude characteristic quantity in the middle is used.

[0072] From temperature deviation prediction data Extract the compensation polarity identifier of the m-th region. Prediction bias peak Mean deviation from prediction The mean of the prediction deviation As the main amplitude basis for quantifying the thermal compensation coefficient, it is used to predict the peak deviation. As a constraint on the upper limit of compensation. Thermal compensation coefficient. Compensation polarity identifier The sign is determined by the mean of the prediction deviation. The absolute value multiplied by the power-temperature conversion factor of the heating actuator. Determine the amplitude, where The steady-state temperature change of the plate caused by a unit change in heating power in region m is characterized by a calibration experiment that determines this beforehand.

[0073] Based on this, a thermal inertia correction factor is introduced. Dynamic corrections are performed. This is due to the equivalent thermal inertia response time constant of each region. Unlike regions with low thermal inertia, regions with higher thermal inertia experience a significantly reduced transient temperature rise within a limited heating window due to the buffering effect of their thermal capacity. This means that the temperature increase achieved within the same heating window is lower. To achieve synchronized temperature rise responses across all regions at the end of the heating window, regions with higher thermal inertia require a larger compensation coefficient to compensate for this transient amplitude reduction.

[0074] The correction method is as follows: based on the current With reference thermal inertia time constant The ratio of the transient amplitude compensation factor is used as the transient amplitude compensation factor. By scaling up proportionally, the final thermal compensation coefficient after transient amplitude correction is obtained. .

[0075] After obtaining the set of thermal compensation coefficients, they need to be further converted into feedforward compensation control signals that can be directly received by the heating actuator. This process includes two key steps: timing advance phase correction and drive signal amplitude mapping.

[0076] The necessity of timing advance phase correction stems from the existence of thermal inertia effects. If at the current moment... The output feedforward compensation control signal needs to be in the future. If the plate temperature can only fully respond at a certain time, then the control signal must be sent in advance. Output to ensure that the actual response of the plate temperature occurs precisely at the peak of the deviation predicted by the temperature deviation prediction data. Compensation was completed previously. Timing lead time. Based on the current hot stamping frequency Equivalent thermal inertial response time constant of each region Together, these factors determine that the higher the hot stamping frequency and the shorter the interval between adjacent presses, the narrower the time window for the compensation control signal response, and the greater the required lead time; thermal inertia time constant The larger the value, the slower the plate temperature response to the control signal, and the greater the required lead time. Based on the current... and Calculate the timing advance correction amount for each region. The thermal compensation coefficient The corresponding compensation action is shifted forward at any time. The compensation amount after timing advance correction is obtained. This ensures that the compensation action is applied before the peak of the temperature deviation occurs.

[0077] After completing the timing correction, advance the timing correction amount. This is mapped to the amplitude of the drive signal at the heating actuator. The heating actuator is typically a thyristor power regulator or a PWM-driven heating element. Its input is a drive signal in the form of voltage or duty cycle, rather than a direct power value. Therefore, a pre-calibrated drive signal-power mapping relationship is needed to map the signal to the heating actuator. The corresponding heating power adjustment is converted into the corresponding drive signal amplitude. The amplitude of the drive signals for all M hot-pressing zones will be... ( Arrange the heating execution channels according to the corresponding regions to generate feedforward compensation control signals. .

[0078] The system acquires the real-time temperature deviation value and generates a feedback compensation control signal based on the real-time temperature deviation value. The system performs time-series registration and superposition of the feedforward compensation control signal and the feedback compensation control signal to generate a composite drive signal. The composite drive signal is output to the heating actuator to achieve real-time optimization and control of the hot stamping plate temperature.

[0079] Furthermore, the real-time temperature deviation value is obtained, and the feedback control increment calculated based on the real-time temperature deviation value is used to generate a feedback compensation control signal. The feedforward compensation control signal and the feedback compensation control signal are time-registered according to a unified time base to obtain the time-registered feedforward signal and the time-registered feedback signal. The amplitude of the timing registration feedforward signal and the timing registration feedback signal are superimposed and synthesized to generate the initial composite driving signal; Obtain the output limiting value of the heating actuator, and compare the initial composite drive signal with the output limiting value to determine whether the limit is exceeded. If the limit is not exceeded, the initial composite drive signal is used as the composite drive signal output. If the limit is exceeded, the initial composite drive signal is truncated using the output limiting value to generate a limited composite drive signal. The composite drive signal is output to the heating actuator to achieve real-time optimization and control of the hot stamping plate temperature.

[0080] In one specific implementation, the feedforward compensation control signal The feedforward compensation is an open-loop control quantity generated in advance based on temperature deviation prediction data. Its compensation effect depends on the accuracy of the prediction model and the precision of the thermal inertia parameters. However, in actual production, factors such as batch differences in substrate materials, sudden changes in environmental temperature and humidity, and the aging and wear of the hot stamping plate itself can introduce random disturbances that the prediction model cannot fully cover, resulting in residual deviations in the feedforward compensation. Therefore, relying solely on feedforward compensation cannot guarantee that the plate temperature will converge to the set value under any operating condition. Feedback compensation based on real-time temperature measurement must be introduced to form a closed loop to correct the residual error of feedforward compensation in real time. Step S5 precisely integrates feedforward compensation and feedback compensation in both timing and amplitude dimensions to generate a composite driving signal that can simultaneously take into account both advance suppression and real-time correction capabilities. This is the execution closed-loop link of the entire temperature optimization and control method.

[0081] Real-time temperature deviation value Defined as the difference between the measured temperature and the set value of each hot-pressing region at the current sampling time, i.e. ,in Data from standard temperature field distribution The corresponding value at the current moment. This is the temperature setpoint for region m. Real-time temperature deviation value. It reflects the actual residual deviation of the screen temperature after the feedforward compensation is executed at the current moment, and is the direct input basis for feedback compensation.

[0082] The feedback compensation control signal is generated using an incremental PID control algorithm. Compared to positional PID, incremental PID outputs the increment of the control quantity relative to the previous moment in each iteration. Rather than the absolute value of the control quantity, this characteristic has a significant advantage when used in conjunction with the feedforward-feedback composite architecture of this application: its incremental output form is easy to integrate with the feedforward compensation control signal generated in step S44. It achieves shock-free superposition and can more naturally suppress integral saturation in subsequent amplitude limiting processing.

[0083] The core technological contribution of this application in the feedback control stage lies in: using the equivalent thermal inertia response characteristic data of each hot-pressed region obtained online in step S2. By directly embedding the PID parameter tuning criteria, deep coupling between control parameters and real-time operating conditions is achieved. Specifically, the PID coefficients for each region include the proportional coefficient. Integral coefficient With differential coefficients It is not set as a fixed constant, but rather based on the initial heat capacity and thermal resistance parameters and the equivalent thermal inertia response time constant of each region. Perform separate adjustments.

[0084] Differential tuning follows the following coherent rules: for the thermal inertia response time constant For larger areas, the proportional coefficient should be appropriately reduced. With differential coefficients This is to avoid differential oscillations and proportional overshoot caused by the feedback quantity lagging behind the actual temperature change; at the same time, the integral coefficient should be appropriately increased. This enhances the ability to eliminate low-frequency, slow drift deviations, ensuring that the region converges to the set value during long-term operation. Conversely, for the thermal inertia response time constant... For smaller regions, the proportional and derivative coefficients can be appropriately increased to enhance the rapid response capability of feedback control.

[0085] This differentiated tuning scheme works in synergy with the feedforward compensation coefficient quantization strategy in step S4: feedforward compensation is responsible for predictive amplitude compensation, while feedback compensation is based on... The response sensitivity is adaptively adjusted, and together the two construct a two-dimensional thermal inertia suppression mechanism of feedforward amplitude compensation and feedback sensitivity adaptation.

[0086] Feedforward compensation control signal In step S4, timing advance phase correction is performed, and the output time of each region is relative to the current sampling time. It has been done in advance Feedback compensation control signal Based on the measured deviation at the current moment It is generated in real time, and its output time is the current sampling time. The two are not naturally aligned on the time axis. If the amplitudes are directly superimposed, the feedforward and feedback components of each region will act on the heating actuator at different times. The superposition effect of the two will deviate from the expected composite compensation design goal.

[0087] Therefore, it is necessary to adjust the feedforward compensation control signal. With feedback compensation control signal Timing registration is performed using a unified time reference. Specifically, it is based on the current sampling time. To unify the time base, The output times of each region that have undergone timing advance correction are mapped back to Using the relative time axis as a reference, a timing registration feedforward signal is generated. Feedback compensation control signal It is itself Generated as a reference, it requires no offset and can be directly used as a timing registration feedback signal. After timing registration is completed, and The timing alignment of control quantities in each region was achieved on a unified timeline, providing a correct timing basis for subsequent amplitude superposition and synthesis.

[0088] After time registration and The amplitudes of each hot-pressing region are superimposed one by one to obtain the initial composite driving signal for each region. The feedforward component undertakes the main task of advance compensation, while the feedback component undertakes the task of real-time correction of residual deviation. The composite drive signal resulting from their superposition can theoretically balance the advance and accuracy of compensation.

[0089] However, under certain extreme conditions, such as the rapid heating phase when the plate temperature is far below the set value in the early stages of production, or when a sudden change in the substrate material causes a large instantaneous thermal disturbance, the feedforward component and the feedback component may superimpose in the same direction, causing the initial composite drive signal to... Exceeding the physical output limit of the heating actuator If an over-limit signal is applied to the heating actuator without restriction, it will cause overload damage to the actuator or a sharp overshoot of the plate temperature. Therefore, the initial composite drive signal needs to be limited: obtain the output limiting value of each channel of the heating actuator. ,Will and The comparison is performed. If the amplitude limit is not exceeded, the initial composite drive signal is directly used as the final composite drive signal output; if the amplitude limit is exceeded, then... The initial composite drive signal is truncated to generate a limited composite drive signal. .

[0090] It is worth noting that when the clipping occurs, the integral term of the feedback compensation will experience integral saturation due to the truncation of the executed quantity. That is, the integral term continues to accumulate, but the executed quantity has reached its limit. Once the clipping is lifted, the accumulated integral term will cause a significant overshoot in the control quantity. Therefore, when the clipping is detected, the integral accumulation of the incremental PID controller is simultaneously paused, and the process continues until... The integral action resumes only after the value falls back to within the limit, effectively preventing overshoot caused by integral saturation.

[0091] The composite drive signal after limiting all M hot-pressing zones Arranged according to the heating execution channels corresponding to each region, the final composite drive signal is formed. The output is sent to the heating execution end, driving the heating elements of each hot pressing area to adjust the heating according to the power level specified by the composite drive signal. This enables real-time optimization control of the temperature of each hot pressing area of ​​the hot stamping plate during continuous production, ensuring that the plate temperature of each area remains stable near the set value, thereby guaranteeing the consistency and stability of the hot stamping process quality.

[0092] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention, they should all fall within the protection scope of the present invention.

[0093] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0094] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A method for optimizing hot stamping process based on temperature control, characterized in that, Includes the following steps: Step S1: Real-time temperature acquisition is performed on multiple hot pressing areas of the hot stamping equipment using a temperature sensor array to obtain multi-area temperature field distribution data; feature extraction is performed on the multi-area temperature field distribution data to obtain temperature field feature data; Step S2: Perform multi-scale dynamic evolution trend analysis of heat conduction based on temperature field characteristic data to obtain dynamic evolution trend data of heat conduction; obtain the thermophysical property parameters of the hot stamping plate, and simulate the thermal inertial response of the hot stamping plate in combination with the dynamic evolution trend data of heat conduction to obtain thermal inertial response characteristic data. Step S3: Based on the thermal inertia response characteristic data, predict the temperature deviation trend of the hot stamping process to obtain temperature deviation prediction data; Based on the temperature deviation prediction data, the hot stamping plate heat compensation direction is determined to obtain heat compensation direction determination data. Step S4: Quantize the thermal compensation coefficient based on the thermal compensation direction determination data and temperature deviation prediction data to obtain the thermal compensation coefficient; generate a feedforward compensation control signal based on the thermal compensation coefficient; Step S5: Based on the comparison between the multi-region temperature field distribution data and the preset target temperature data, obtain the real-time temperature deviation value, and generate a feedback compensation control signal based on the real-time temperature deviation value; perform time-series registration and superposition synthesis of the feedforward compensation control signal and the feedback compensation control signal to generate a composite drive signal; output the composite drive signal to the heating execution end to realize real-time optimization and control of the hot stamping plate temperature.

2. The hot stamping process optimization method based on temperature control according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Real-time temperature acquisition is performed on multiple hot pressing areas of the hot stamping equipment using a temperature sensor array to obtain multi-area temperature field distribution data; Step S12: Perform data cleaning and spatial interpolation on the multi-region temperature field distribution data to obtain standard temperature field distribution data; Step S13: Perform time-series fluctuation analysis on the standard temperature field distribution data to obtain time-series fluctuation data of the temperature field; Step S14: Extract time-frequency domain nonlinear features from the temperature field time-series fluctuation data to obtain temperature field feature data.

3. The hot stamping process optimization method based on temperature control according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: Perform multi-scale dynamic evolution trend analysis of heat conduction based on the temperature field characteristic data to obtain dynamic evolution trend data of heat conduction; Step S22: Obtain the thermophysical property parameters of the hot stamping plate, and perform thermal inertial response simulation on the thermophysical property parameters of the hot stamping plate according to the temperature field characteristic data and the dynamic evolution trend data of heat conduction, to obtain thermal inertial response characteristic data.

4. The hot stamping process optimization method based on temperature control according to claim 3, characterized in that, Step S22 includes the following steps: Step S221: Calculate the thermal stress coupling fluctuation response based on the temperature field characteristic data and the heat conduction dynamic evolution trend data to obtain thermal stress coupling fluctuation response data; Step S222: Extract the initial heat capacity and thermal resistance parameters of the hot stamping plate to obtain the initial heat capacity and thermal resistance parameter data; Step S223: Based on the thermal stress coupling fluctuation response data, the initial thermal capacity and thermal resistance parameter data are corrected for thermal inertia response characteristics to obtain thermal inertia response characteristic data.

5. The hot stamping process optimization method based on temperature control according to claim 1, characterized in that, Step S3 includes the following steps: Step S31: Based on the thermal inertia response characteristic data, perform temperature deviation trend analysis of the hot stamping process to obtain temperature deviation trend data; Step S32: Perform predictive extrapolation on the temperature deviation trend data to obtain predicted temperature deviation data; Step S33: Obtain the actual temperature deviation direction of the current hot stamping plate, compare the actual temperature deviation direction with the predicted deviation direction in the temperature deviation prediction data, and generate heat compensation direction determination data.

6. The hot stamping process optimization method based on temperature control according to claim 5, characterized in that, Step S32 includes the following steps: Step S321: Based on the temperature deviation trend data, recursively predict the temperature deviation in the subsequent control cycle to obtain a multi-step predicted deviation sequence; Step S322: Statistical feature extraction is performed on the multi-step prediction deviation sequence to extract the deviation peak value, deviation mean value and deviation duration interval, and temperature deviation prediction data is generated based on the deviation peak value, deviation mean value and deviation duration interval.

7. The hot stamping process optimization method based on temperature control according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: Obtain the compensation polarity identifier in the thermal compensation direction determination data and the amplitude characteristic quantity in the temperature deviation prediction data; Step S42: Calculate the expected adjustment amount and adjustment direction of the heating power based on the compensation polarity identifier and the amplitude characteristic quantity, and generate the feedforward compensation amount; Step S43: Obtain the real-time hot stamping frequency of the hot stamping plate, and perform timing advance phase correction on the feedforward compensation amount according to the real-time hot stamping frequency of the hot stamping plate to obtain the timing advance correction amount. Step S44: Map the timing advance correction amount to the driving signal amplitude of the heating actuator to generate a feedforward compensation control signal.

8. The hot stamping process optimization method based on temperature control according to claim 1, characterized in that, Step S5 includes the following steps: Step S51: Obtain the real-time temperature deviation value, and generate a feedback compensation control signal based on the feedback control increment calculated from the real-time temperature deviation value. Step S52: Perform timing registration processing on the feedforward compensation control signal and the feedback compensation control signal according to a unified time base to obtain the timing registration feedforward signal and the timing registration feedback signal; Step S53: The timing registration feedforward signal and the timing registration feedback signal are superimposed and synthesized to generate the initial composite driving signal; Step S54: Obtain the output limiting value of the heating actuator, and compare the initial composite drive signal with the output limiting value to determine: if it does not exceed the limit, the initial composite drive signal is used as the composite drive signal output; if it exceeds the limit, the initial composite drive signal is truncated using the output limiting value to generate a limited composite drive signal. Step S55: Output the composite drive signal to the heating actuator to achieve real-time optimization and control of the hot stamping plate temperature.

Images