Continuous hot stamping automation equipment and method

Abstract

The present application discloses a continuous hot stamping automation equipment and method thereof, which obtains the time queue of the roller pressure data of the transmission roller mechanism and the time queue of the temperature data of the silicone heating plate, and adopts the data analysis and encoding method based on deep learning to perform local time series encoding on the roller pressure data, and then performs feature prompt dynamic propagation and ideal temperature time series prediction on the encoded local time series characteristics of each roller pressure, and at the same time performs time series encoding on the temperature data, so as to intelligently obtain the transmission speed recommendation value of the transmission roller mechanism based on the time series offset characterization between the ideal silicone heating plate temperature time series characteristics and the actual silicone heating plate temperature time series characteristics. In this way, the time series changes of various parameters in the hot stamping process can be dynamically monitored, so as to adjust the transmission speed of the transmission roller mechanism more accurately and timely, ensure the consistency and stability of the hot stamping effect, and thus achieve higher precision and more intelligent hot stamping control.

Description

Technical Field

[0001] The present application relates to the field of intelligent control, and more specifically, to a continuous hot stamping automation device and method thereof. Background Art

[0002] Hot stamping is a common surface decoration process that achieves beautiful and durable decorative effects by transferring ink or coating materials from a film to the material to be hot stamped. Hot stamping is widely used in display cabinets, packaging bags, eyeglass accessories and other fields. For example, hot stamping can add exquisite patterns and texts to the surface of display cabinets; add brand logos, patterns and texts to packaging bags; add brand logos and decorative patterns to accessories such as eyeglass frames and temples, thereby beautifying the product and enhancing its market competitiveness and commercial value.

[0003] Patent CN110509663B proposes an automatic hot stamping machine and a hot stamping method thereof, which includes a workbench, a frame, a telescopic drive member, a lifting plate, a crossbeam and a transmission roller mechanism, etc. It is equipped with a heating plate rack and a silicone heating plate to ensure uniform heating of the film, and is provided with a film transmission mechanism, a feeding mechanism and a demoulding mechanism to make the process simple and easy to operate.

[0004] During the hot stamping process, the control of the transmission speed in the transmission roller mechanism is of great significance for ensuring the quality of hot stamping. The appropriate transmission speed can ensure uniform heat transfer and avoid the problem of insufficient or excessive hot stamping. In the above patent, the transmission speed is controlled by adjusting the rotation speed of the first motor through a traditional preset program or manually input parameters. However, relying on preset programs or manually input parameters, it is impossible to adjust and optimize in real time and lacks adaptive capabilities. This makes it difficult for the system to automatically adapt when facing different materials and environmental changes, and requires frequent manual intervention. In addition, the preset program or manually input parameters may lead to insufficient control and easy deviations, which will affect the hot stamping effect, especially when the temperature and pressure fluctuate greatly, and cannot be corrected in time, resulting in unstable product quality.

[0005] Therefore, an optimized continuous hot stamping automation equipment is desired. Summary of the invention

[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a continuous hot stamping automation equipment and method thereof, which obtains the time queue of the roller rolling pressure data of the transmission roller mechanism and the time queue of the temperature data of the silicone heating plate, and adopts the data analysis and encoding method based on deep learning to perform local time series encoding on the roller rolling pressure data, and then performs feature prompt dynamic propagation and ideal temperature time series prediction on the encoded local time series characteristics of each roller rolling pressure, and at the same time performs time series encoding on the temperature data, so as to intelligently obtain the transmission speed recommendation value of the transmission roller mechanism based on the time series offset characterization between the ideal silicone heating plate temperature time series characteristics and the actual silicone heating plate temperature time series characteristics. In this way, the time series changes of various parameters in the hot stamping process can be dynamically monitored, so as to adjust the transmission speed of the transmission roller mechanism more accurately and timely, ensure the consistency and stability of the hot stamping effect, and thus achieve higher precision and more intelligent hot stamping control.

[0007] According to one aspect of the present application, a continuous hot stamping automation equipment is provided, which includes: a workbench; a frame; a telescopic driving member installed on the inner side of both ends of the frame; a lifting plate arranged at the end of the telescopic driving member; a transmission roller mechanism installed on the crossbeam between the two lifting plates; a silicone heating plate on the heating plate frame arranged at the bottom of the lifting plate; a film transmission mechanism fixed on the frame; a feeding mechanism installed on the workbench; a demolding mechanism located below the silicone heating plate; it is characterized in that the continuous hot stamping automation equipment also includes: a pressure sensor installed on the transmission roller mechanism, which is used to collect a time queue of roller rolling pressure data of the transmission roller mechanism; a temperature sensor installed on the silicone heating plate, which is used to collect a time queue of temperature data of the silicone heating plate; a controller, which is used to control the transmission speed of the transmission roller mechanism based on the data collected by the pressure sensor and the temperature sensor.

[0008] According to another aspect of the present application, a continuous hot stamping automation method is provided, which includes: performing local time encoding on the time queue of the roller rolling pressure data to obtain a sequence of local time characteristics of the roller rolling pressure; performing node feature prompt dynamic propagation aggregation on the sequence of local time characteristics of the roller rolling pressure to obtain local time characteristic propagation aggregation characteristics of the roller rolling pressure; performing silicone heater plate temperature prediction on the local time characteristic propagation aggregation characteristics of the roller rolling pressure to obtain ideal silicone heater plate temperature timing characteristics; performing time encoding on the time queue of the temperature data to obtain actual silicone heater plate temperature timing characteristics; calculating the silicone heater plate temperature timing offset representation between the ideal silicone heater plate temperature timing characteristics and the actual silicone heater plate temperature timing characteristics, and obtaining a control result based on the silicone heater plate temperature timing offset representation.

[0009] Compared with the prior art, the present application provides a continuous hot stamping automation equipment and method thereof, which obtains the time queue of the roller rolling pressure data of the transmission roller mechanism and the time queue of the temperature data of the silicone heating plate, and adopts the data analysis and encoding method based on deep learning to perform local time series encoding on the roller rolling pressure data, and then performs feature prompt dynamic propagation and ideal temperature time series prediction on the encoded local time series characteristics of each roller rolling pressure, and at the same time performs time series encoding on the temperature data, so as to intelligently obtain the transmission speed recommendation value of the transmission roller mechanism based on the time series offset representation between the ideal silicone heating plate temperature time series characteristics and the actual silicone heating plate temperature time series characteristics. In this way, the time series changes of various parameters in the hot stamping process can be dynamically monitored, so as to adjust the transmission speed of the transmission roller mechanism more accurately and timely, ensure the consistency and stability of the hot stamping effect, and thus achieve higher precision and more intelligent hot stamping control. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0011] Figure 1 It is a block diagram of a continuous hot stamping automation device according to an embodiment of the present application.

[0012] Figure 2 Schematic diagram of data flow of continuous hot stamping automation equipment according to an embodiment of the present application.

[0013] Figure 3 It is a block diagram of a controller in a continuous hot stamping automation device according to an embodiment of the present application.

[0014] Figure 4 Flow chart of a continuous hot stamping automation method according to an embodiment of the present application. DETAILED DESCRIPTION

[0015] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0016] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0017] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0018] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0019] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0020] In the technical solution of the present application, a continuous hot stamping automation device is proposed. Figure 1 It is a block diagram of a continuous hot stamping automation device according to an embodiment of the present application. Figure 2 Schematic diagram of data flow of the continuous hot stamping automation equipment according to the embodiment of the present application. Figure 1 and Figure 2As shown, the continuous hot stamping automation equipment 300 according to the embodiment of the present application includes: a workbench 310; a frame 320; a telescopic driving member 330 installed on the inner side of both ends of the frame; a lifting plate 340 arranged at the end of the telescopic driving member; a transmission roller mechanism 350 installed on the crossbeam between the two lifting plates; a silicone heating plate 360 ​​arranged on the heating plate frame at the bottom of the lifting plate; a film transmission mechanism 370 fixed on the frame; a feeding mechanism 380 installed on the workbench; a demolding mechanism 390 located below the silicone heating plate; characterized in that the continuous hot stamping automation equipment also includes: a pressure sensor 400 installed on the transmission roller mechanism, which is used to collect the time queue of the roller rolling pressure data of the transmission roller mechanism; a temperature sensor 410 installed on the silicone heating plate, which is used to collect the time queue of the temperature data of the silicone heating plate; a controller 420, which is used to control the transmission speed of the transmission roller mechanism based on the data collected by the pressure sensor and the temperature sensor.

[0021] The workbench is the basic platform of the entire equipment, used to support and fix other mechanical parts. The workbench is usually made of strong materials such as steel or aluminum to ensure that it remains stable during long-term operation. The workbench provides a stable base for all mechanical parts to ensure that the equipment will not shake or move during operation. There are also mounting holes and guide rails on the workbench to fix the feeding mechanism and demolding mechanism to ensure that the material can smoothly enter and leave the hot stamping area. In addition, there are mounting holes and guide rails on the workbench to fix the control system and other sensors to ensure data acquisition and transmission of control signals.

[0022] The rack is the main frame of the equipment, which is used to connect and fix various mechanical parts. The rack is usually made of high-strength materials such as steel or aluminum alloy to ensure the solidity and durability of its structure. There are multiple mounting holes and guide rails on the rack to fix mechanical parts such as telescopic drive parts, lifting plates, beams, etc. to ensure the structural stability of the entire equipment. The rack provides installation and support for the transmission roller mechanism, film transmission mechanism, etc. to ensure that these parts can operate normally. The rack also plays a protective role to prevent the external environment from causing damage to the internal mechanical parts.

[0023] The telescopic drive is a mechanical device that can be extended and retracted in a straight line, usually driven by electricity or pneumatics. The telescopic drive is used to control the up and down movement of the lifting plate. The telescopic drive adjusts the distance between the silicone heating plate and the material by accurately controlling the up and down movement of the lifting plate to ensure uniform pressure during the hot stamping process. Before the hot stamping begins, the telescopic drive lifts the silicone heating plate to an appropriate position to avoid contact with the material. After the hot stamping is completed, the telescopic drive quickly lifts the silicone heating plate to facilitate the removal of the material.

[0024] The lifting plate is a flat plate installed at the end of the telescopic drive, used to carry the silicone heater and the heater frame. The movement of the lifting plate is controlled by the telescopic drive to achieve up and down movement. The lifting plate provides stable support for the silicone heater and the heater frame to ensure that they remain stable during the hot stamping process. By cooperating with the telescopic drive, the lifting plate realizes the precise up and down movement of the silicone heater and adjusts the hot stamping pressure. The lifting plate is provided with mounting holes for fixing the pressure sensor and the temperature sensor to ensure accurate data collection.

[0025] The transmission roller mechanism is a mechanical device for transmitting power and pressure, usually composed of a motor, a reducer, a roller, etc. The transmission roller mechanism is responsible for feeding the material from the feeding mechanism into the hot stamping area and applying pressure through the rotation of the roller. Driven by the motor and the reducer, the transmission roller mechanism realizes the rotation of the roller and feeds the material from the feeding mechanism into the hot stamping area. The pressure applied by the rotation of the roller ensures uniform contact between the material and the silicone heating plate, achieving a good hot stamping effect. A pressure sensor is installed on the transmission roller mechanism to monitor the pressure of the roller in real time to ensure stable pressure during the hot stamping process.

[0026] Silicone heating pads are flexible heating elements, usually made of silicone material with heating wires embedded inside. Silicone heating pads can heat evenly to ensure even heat transfer during the hot stamping process. Silicone heating pads ensure even heat transfer during the hot stamping process by evenly heating, avoiding problems such as insufficient or excessive hot stamping. A temperature sensor is installed on the silicone heating pad to monitor the temperature of the silicone heating pad in real time to ensure that the temperature is within the set range. By cooperating with the lifting plate, the silicone heating pad can achieve precise up and down movement to adjust the hot stamping pressure.

[0027] The film transmission mechanism is a mechanical device used to transport the hot stamping film, usually composed of a motor, a transmission belt, a guide rail, etc. The film transmission mechanism is responsible for transporting the hot stamping film from the storage area to the hot stamping area, and ensures the alignment of the hot stamping film and the material through precise positioning. Driven by the motor and the transmission belt, the film transmission mechanism transports the hot stamping film from the storage area to the hot stamping area. Through the guide rail and the positioning device, the film transmission mechanism ensures the precise alignment of the hot stamping film and the material to avoid misalignment. Through speed control, the film transmission mechanism ensures that the conveying speed of the hot stamping film is synchronized with the conveying speed of the material.

[0028] The feed mechanism is a mechanical device used to feed the material to be stamped into the stamping area, usually composed of a motor, a conveyor belt, a guide rail, etc. The feed mechanism is responsible for feeding the material from the feed port into the stamping area and ensuring the accurate position of the material through precise positioning. Driven by the motor and the conveyor belt, the feed mechanism feeds the material from the feed port into the stamping area. Through the guide rail and the positioning device, the feed mechanism ensures the accurate position of the material to avoid misalignment. Through speed control, the feed mechanism ensures that the material conveying speed is synchronized with the conveying speed of the film transmission mechanism.

[0029] The demoulding mechanism is a mechanical device used to separate the stamped material from the silicone heating plate, usually composed of a cylinder, a push rod, a guide rail, etc. The demoulding mechanism is responsible for lifting the material from the silicone heating plate after the stamping is completed to facilitate the removal of the material. Driven by the cylinder and the push rod, the demoulding mechanism lifts the stamped material from the silicone heating plate to prevent the material from adhering to the silicone heating plate. Through the guide rail and the positioning device, the demoulding mechanism ensures the smooth removal of the material to avoid damage. Through speed control, the demoulding mechanism ensures a smooth and efficient demoulding process.

[0030] The pressure sensor is an electronic device used to measure pressure. It is usually installed on the transmission roller mechanism to monitor the pressure changes between the roller and the material in real time. The pressure sensor monitors the pressure changes between the roller and the material in real time to ensure the stability of the pressure during the hot stamping process. The pressure sensor transmits the collected pressure data to the controller to adjust the transmission speed of the transmission roller mechanism. Through the analysis of pressure data, the pressure sensor can promptly detect abnormal conditions of the equipment, such as roller wear, material jamming, etc., so as to take corresponding maintenance measures.

[0031] The temperature sensor is an electronic device used to measure temperature. It is usually installed on the silicone heating plate to monitor the temperature changes of the silicone heating plate in real time. The temperature sensor monitors the temperature changes of the silicone heating plate in real time to ensure that the temperature is within the set range and avoid hot stamping quality problems caused by excessively high or low temperatures. The temperature sensor transmits the collected temperature data to the controller to adjust the heating power of the silicone heating plate. Through the analysis of temperature data, the temperature sensor can promptly detect abnormal conditions of the equipment, such as heating plate failure, temperature fluctuations, etc., so as to take corresponding maintenance measures.

[0032] The controller is the central processing unit of the equipment, usually composed of a microprocessor, memory, input and output interfaces, etc. The controller is responsible for receiving data from pressure sensors and temperature sensors, and adjusting the transmission speed of the transmission roller mechanism according to these data. The controller receives data from pressure sensors and temperature sensors, and monitors the pressure and temperature changes during the hot stamping process in real time. The controller processes and analyzes the collected data through built-in algorithms and models to generate control signals. The controller adjusts the transmission speed of the transmission roller mechanism according to the processing results to ensure the stability of pressure and temperature during the hot stamping process. By working in conjunction with sensors and mechanical components, the controller realizes intelligent control of the hot stamping process and improves the hot stamping quality and production efficiency.

[0033] That is, the material feeding mechanism feeds the material to be stamped from the feed port into the stamping area, and ensures the accurate position of the material through the guide rail and positioning device. The film transmission mechanism feeds the stamping film from the storage area into the stamping area, and ensures the precise alignment of the stamping film and the material through the guide rail and positioning device. The telescopic drive controls the up and down movement of the lifting plate, adjusts the distance between the silicone heating plate and the material, and ensures uniform pressure during the stamping process. The silicone heating plate ensures uniform heat transfer during the stamping process by uniform heating. The temperature sensor monitors the temperature of the silicone heating plate in real time to ensure that the temperature is within the set range. The pressure sensor and the temperature sensor monitor the pressure change between the roller and the material and the temperature change of the silicone heating plate in real time, and transmit the data to the controller. The controller generates a control signal based on the received data through the built-in algorithm and model, adjusts the transmission speed of the transmission roller mechanism, and ensures stable pressure and temperature during the stamping process. Under the intelligent control of the controller, the transmission roller mechanism and the silicone heating plate work together to achieve uniform stamping of the material. After the stamping is completed, the demoulding mechanism lifts the stamped material from the silicone heating plate to ensure the smooth removal of the material and avoid damage.

[0034] In particular, in a specific example of the present application, Figure 3 As shown, the controller 420 includes: a pressure encoding module 421, which is used to perform local time series encoding on the time queue of the roller rolling pressure data to obtain a sequence of local time series characteristics of the roller rolling pressure; a pressure aggregation module 422, which is used to perform node feature prompt dynamic propagation aggregation on the sequence of local time series characteristics of the roller rolling pressure to obtain local time series characteristic propagation aggregation characteristics of the roller rolling pressure; a temperature time series prediction module 423, which is used to perform silicone heating plate temperature prediction on the local time series characteristic propagation aggregation characteristics of the roller rolling pressure to obtain an ideal silicone heating plate temperature time series characteristic; a temperature encoding module 424, which is used to perform time series encoding on the time queue of the temperature data to obtain an actual silicone heating plate temperature time series characteristic; a control module 425, which is used to calculate the silicone heating plate temperature time series offset representation between the ideal silicone heating plate temperature time series characteristic and the actual silicone heating plate temperature time series characteristic, and obtain a control result based on the silicone heating plate temperature time series offset representation.

[0035] Specifically, the pressure encoding module 421 is used to perform local time series encoding on the time queue of the roller rolling pressure data to obtain a sequence of local time series features of the roller rolling pressure. Considering that the roller rolling pressure data has time series features on a local time scale, such as abnormal values ​​in a short period of time or continuous fluctuations within a local time range. And considering that by using the convolution kernel to slide on the time axis, the important features of the local area in the roller rolling pressure data can be automatically learned. This is critical for identifying short-term fluctuations or anomalies that occur during the roller rolling process. Therefore, in the technical solution of the present application, the time queue of the roller rolling pressure data is passed through a rolling pressure local time series encoder based on 1D-CNN to capture the pressure change pattern in different time windows by sliding the convolution layer of the one-dimensional convolution on the time series through a sliding window, and obtain a sequence of local time series feature vectors of the roller rolling pressure.

[0036] It is worth mentioning that the one-dimensional convolutional neural network (1D-CNN) is a deep learning architecture specifically designed to process one-dimensional signals (such as time series data, audio signals, etc.). 1D-CNN aims to automatically learn features in data by applying convolution operations to one-dimensional data. Its specific working principle is: a one-dimensional filter (or convolution kernel) is moved along the input data to generate a feature map. The filter consists of a set of weights that are optimized by the back-propagation algorithm during the training process to facilitate learning the most relevant features of the input data. The convolution operation is essentially looking for the degree of match between the input data and the filter, thereby emphasizing certain specific patterns. After convolution, a nonlinear activation function (such as ReLU) is usually applied to increase the expressive power of the model so that the model can learn more complex patterns. In order to reduce the number of parameters and control overfitting, a pooling layer is often added after the convolution layer. Common pooling operations include maximum pooling and average pooling, which reduce the size of the feature map by taking the maximum value or average value of the local area.

[0037] That is, first, the time queue of the roller pressure data is collected through the pressure sensor installed on the transmission roller mechanism. These data are usually continuous time series, indicating the changes in the roller pressure at different time points. The collected data may contain noise and outliers, so preprocessing is required. Common preprocessing steps include normalization, denoising and segmentation. Normalization normalizes the data to a fixed range (such as 0 to 1) for subsequent processing; denoising uses filters (such as low-pass filters) to remove high-frequency noise; segmentation divides the continuous time series data into subsequences of fixed length, each subsequence as an input sample.

[0038] Next, the time queue of the roller pressure data is feature extracted by constructing a 1D-CNN model. 1D-CNN is a deep learning model specifically designed to process one-dimensional time series data. Its basic structure includes multiple convolutional layers, pooling layers, and fully connected layers. Assuming that the length of each subsequence is T and the feature dimension of each time point is F, the shape of the input layer is (T, F). Define one or more one-dimensional convolution kernels (filters), each with a size of k and a step size of s. The convolution kernel slides on the time axis and performs convolution operations with the input data to generate feature maps. An activation function (such as ReLU) is usually added after the convolution layer to increase the nonlinear ability of the model. A pooling layer is added after the convolution layer. Commonly used pooling operations include maximum pooling and average pooling. Pooling operations can reduce the dimension of the feature map and help the model extract higher-level abstract features. In order to capture long-term and short-term dependencies in the time series, multiple convolutional layers and pooling layers can be stacked. Each layer extracts features at different levels. Flatten the feature map output by the last convolutional layer or pooling layer into a one-dimensional vector, then add one or more fully connected layers to further process the features and output the final feature vector.

[0039] Through the above 1D-CNN model, the time sequence of the input roller pressure data is converted into a series of feature vectors. Each feature vector represents the local time series characteristics of the roller pressure in a specific time window. These feature vectors will be used in subsequent steps such as feature aggregation and temperature prediction.

[0040] Specifically, the pressure aggregation module 422 is used to perform node feature prompt dynamic propagation aggregation on the sequence of local time series features of the roller rolling pressure to obtain the propagation aggregation features of the local time series features of the roller rolling pressure. Considering that the local time series features of the roller rolling pressure at different time points in the sequence of the local time series feature vectors of the roller rolling pressure have strong or weak connections with each other, and at the same time, the mutual dependence of different time periods has different effects on the entire time range. Based on this, in order to deeply capture the long-term and short-term dependencies in the roller rolling pressure time series data, so as to analyze the time series patterns of the roller rolling pressure at different time points, so as to better retain and transmit important feature information, and further enhance the expressiveness and contextual dependency of the features, in the technical solution of the present application, the sequence of local time series features of the roller rolling pressure is subjected to node feature prompt dynamic propagation aggregation to obtain the propagation aggregation features of the local time series features of the roller rolling pressure.

[0041] Specifically, first, the local time series feature vector of the rolling pressure of the current roller is extracted from the sequence of the local time series feature vectors of the rolling pressure of the roller; this operation is essentially to prepare for the subsequent feature propagation and update. In the local time series feature sequence distribution of the rolling pressure of the roller, each node can have different attributes or features, and these feature vectors represent the information of the node itself. Extracting the feature vector of the current node means locking a specific reference point and evaluating the relative importance of other nodes with this node as the center, which is similar to selecting a focus in the local time series feature sequence distribution of the rolling pressure of the roller, and then constructing a sequence attention perspective around this focus.

[0042] Next, the current roller pressure local time series feature vector is used as the feature prompt gate, and the feature association strength prompt value of each roller pressure local time series feature vector in the sequence of the roller pressure local time series feature vector is calculated relative to the current roller pressure local time series feature vector to obtain the sequence of the roller pressure local time series feature association strength prompt value; that is, the current roller pressure local time series feature vector is used as the feature prompt gate to calculate the roller pressure local time series feature association strength prompt value between other node feature vectors and the current node feature vector. This step is intended to measure the relationship strength between each node and provide a basis for determining which features should be propagated more strongly in the next step. It can essentially be regarded as an application of the attention mechanism, in which the attention score is determined by the feature similarity between nodes. In this way, the neighboring nodes that have a greater impact on the current node can be identified, and a higher weight can be given to the subsequent feature propagation. Furthermore, the median of the sequence of the roller pressure local time series feature association strength prompt value is extracted as the intermediate state value to obtain the roller pressure feature association strength gate threshold. Here, the median of the sequence of the correlation strength prompt values ​​of the local time series feature of the rolling pressure of the roller is used as the threshold value. One reason for choosing the median instead of the average is that the median is not sensitive to extreme values ​​and can better reflect the central trend of the sequence of the local time series feature vector of the rolling pressure of the roller. This approach helps to maintain the stability of feature propagation in the network and avoid the overall performance degradation caused by a few extreme cases.

[0043] More specifically, based on the current roller pressure local time series feature vector, the feature association strength prompt value of each roller pressure local time series feature vector in the sequence of roller pressure local time series feature vectors is calculated using the following feature association strength prompt calculation formula to obtain a sequence of roller pressure local time series feature association strength prompt values; wherein the feature association strength prompt calculation formula is: ;in, is the sequence of local time series characteristic vectors of the rolling pressure of the roller, and are the first, second, and third in the sequence of the local time series characteristic vectors of the rolling pressure of the pressure roller. and The local time series characteristic vector of the rolling pressure of the roller, It is The characteristic value of each position in the local time series characteristic vector of the rolling pressure of the pressure roller, It is The characteristic value of each position in the local time series characteristic vector of the rolling pressure of the pressure roller, represents the logarithmic function value with base 2, is the number of eigenvalues ​​in the local time series eigenvector of the rolling pressure of each roller, represents the exponential function value with the natural constant e as the base, yes The corresponding roller pressure local time series feature correlation strength prompt value, and are the first, second and third values ​​in the sequence of the local temporal characteristics of the rolling pressure of the roller. The prompt value of the correlation strength of the local time series characteristics of the rolling pressure of the roller, To get the median of a sequence, is the roller pressure feature associated strength gate threshold.

[0044] Then, the roller pressure feature correlation strength gate threshold is used as the switch threshold of the gate control unit, and the sequence of the roller pressure local time series feature vectors and the sequence of the roller pressure local time series feature correlation strength prompt values ​​are input into the gate control unit to obtain the sequence of the roller pressure local time series feature vectors after modulation. By inputting the sequence of the roller pressure local time series feature vectors and the sequence of the roller pressure local time series feature correlation strength prompt values ​​into the gate control unit, and using the roller pressure feature correlation strength gate threshold as the switch threshold of the gate control unit as the switch standard, the intensity modulation of the node feature vector is realized. This process controls the propagation of features by adjusting the weights, so that those features with higher correlation strength are more likely to be retained and passed to the subsequent layers, so that the feature information is selectively transmitted or suppressed. Finally, the sequence of the roller pressure local time series feature vectors after modulation is input into the feature dynamic propagation module based on the forward LSTM model to obtain the roller pressure local time series feature propagation aggregation feature vector ideal silicone heating plate temperature time series feature. Among them, LSTM is a special recurrent neural network (RNN) that is particularly good at handling long-term dependency problems. In the technical solution of this application, LSTM is used to capture the dependency between different time points or different neighbor nodes in the process of propagating the local temporal characteristics of the rolling pressure of the roller after modulation, so as to obtain the aggregated feature vector of the propagation of the local temporal characteristics of the rolling pressure of the roller.

[0045] More specifically, based on the roller rolling pressure feature association strength gate threshold and the roller rolling pressure local temporal feature association strength prompt value sequence, the roller rolling pressure local temporal feature vector sequence is gated modulated and dynamically aggregated using the following gated modulation dynamic aggregation formula to obtain the roller rolling pressure local temporal feature propagation aggregation feature vector as the roller rolling pressure local temporal feature propagation aggregation feature; wherein the gated modulation dynamic aggregation formula is: ;in, For gated mask operation, for The local time series feature vector of the roller pressure after weighted modulation, and are the first, second and third in the sequence of the local time series feature vector of the rolling pressure of the roller after modulation. The local time series characteristic vector of the rolling pressure of the roller after modulation, is the sequence of local time series characteristic vectors of the rolling pressure of the roller after the modulation, Encode for LSTM, The aggregated feature vector is propagated for the local time series feature of the rolling pressure of the pressure roller.

[0046] Specifically, the temperature timing prediction module 423 is used to predict the temperature of the silicone heater plate by propagating the aggregated characteristics of the local timing characteristics of the rolling pressure of the pressure roller to obtain the ideal silicone heater plate temperature timing characteristics. Considering that there is a complex interaction between the rolling pressure of the pressure roller and the temperature of the silicone heater plate, such as when the rolling pressure of the pressure roller increases, the contact between the pressure roller and the silicone heater plate becomes closer, and the heat conduction efficiency is improved. This means that more heat will be transferred from the silicone heater plate to the material to be stamped, which may cause the temperature of the silicone heater plate to drop slightly. Therefore, in order to infer the ideal silicone heater plate temperature based on the rolling pressure characteristics of the pressure roller, so as to better characterize the temperature information of the ideal silicone heater plate, in the technical solution of the present application, the local timing characteristics of the rolling pressure of the pressure roller are propagated and aggregated. The feature vector is passed through the silicone heater plate temperature predictor based on the decoder to obtain the ideal silicone heater plate temperature timing feature vector, that is, accurate temperature prediction can help the system make adjustments in advance and reduce downtime and scrap rate caused by temperature fluctuations.

[0047] Here, the decoder model is a neural network model used to recover original data or generate new data from a high-level feature vector. In the present invention, the decoder model is used to predict the ideal silicone heating plate temperature timing feature vector from the roller rolling pressure local timing feature propagation aggregate feature vector. The following are the specific steps for building a decoder model: The input of the decoder model is the roller rolling pressure local timing feature propagation aggregate feature vector, assuming that the input shape is (N, D). First, the input feature vector is expanded through one or more fully connected layers (Dense layers) to increase the dimension of the feature. For example, the input feature vector can be expanded from D dimensions to H dimensions, where H is a larger dimension. This helps the model better capture the complex relationships between features.

[0048] After the decoder model is built, it needs to be compiled and trained. When compiling, the optimizer and loss function are specified. When training, the roller rolling pressure local time series feature propagation aggregation feature vector is used as input, and the ideal silicone heater temperature time series feature vector is used as the label.

[0049] After the model training is completed, the model can be used to predict the new roller rolling pressure local time series feature propagation aggregation feature vector to obtain the ideal silicone heating plate temperature time series feature vector. These predicted temperature feature vectors will be used in the subsequent control module to adjust the transmission speed of the transmission roller mechanism to ensure the stability and consistency of the hot stamping process.

[0050] Specifically, the temperature encoding module 424 is used to perform time series encoding on the time queue of the temperature data to obtain the actual silicone heating plate temperature time series characteristics. Considering that the temperature data is constantly changing over time, and the correlation of the temperature time series characteristics between different time intervals is different, based on this, in the technical solution of the present application, the time queue of the temperature data is passed through a temperature time series encoder based on a forward LSTM model to capture and mine the temperature time series characteristics in different time periods, and obtain the actual silicone heating plate temperature time series feature vector. In other words, the forward LSTM model can effectively capture and mine the temperature time series characteristics in different time periods from the time queue of the temperature data. Through its powerful memory ability and dynamic processing mechanism, it can learn the complex patterns in the time series of temperature data, generate a representative actual silicone heating plate temperature time series feature vector, and provide strong support for subsequent analysis and prediction. It is worth mentioning that the forward LSTM model refers to an LSTM model that processes data in the natural order of the time series (from front to back). In time series analysis, the forward LSTM model is the most commonly used form because it conforms to the time flow direction of most time series data.

[0051] Specifically, the control module 425 is used to calculate the silicone heating plate temperature timing offset representation between the ideal silicone heating plate temperature timing characteristics and the actual silicone heating plate temperature timing characteristics, and obtain a control result based on the silicone heating plate temperature timing offset representation. In a specific example of the present application, first, the semantic difference information between the ideal silicone heating plate temperature timing feature vector and the actual silicone heating plate temperature timing feature vector is calculated to obtain the silicone heating plate temperature timing offset representation vector as the silicone heating plate temperature timing offset representation. Considering that the ideal silicone heating plate temperature timing feature vector represents the expected and ideal silicone heating plate temperature timing change, it reflects the expected trajectory of the silicone heating plate temperature change over time under optimal conditions. The actual silicone heating plate temperature timing feature vector represents the real data of the silicone heating plate temperature change over time during the actual operation of the system. Therefore, in order to calculate the difference between the ideal and actual temperature timing feature vectors, so as to quantify the deviation of the current system from the ideal state, so as to better characterize the timing characteristics of the temperature, in the technical solution of the present application, the semantic difference information between the ideal silicone heating plate temperature timing feature vector and the actual silicone heating plate temperature timing feature vector is calculated to obtain the silicone heating plate temperature timing offset characterization vector. In particular, in a specific embodiment of the present application, the position difference vector between the ideal silicone heating plate temperature timing feature vector and the actual silicone heating plate temperature timing feature vector can be calculated as the silicone heating plate temperature timing offset characterization vector, so that the actual temperature offset can be understood in a more fine-grained manner. Subsequently, the silicone heating plate temperature timing offset characterization vector is passed through a decoder-based hot stamping transmission speed controller to obtain a control result, and the control result is used to represent the recommended decoding value of the transmission speed of the transmission roller mechanism. The silicone heating plate temperature timing offset characterization vector is passed through a decoder-based hot stamping transmission speed controller to obtain a control result, and the control result is used to represent the recommended decoding value of the transmission speed of the transmission roller mechanism. It should be understood that there is a close correlation between the temperature of the silicone heating plate and the transmission speed of the transmission roller mechanism. For example, when the transmission speed is high, the contact time between the silicone heating plate and the material to be stamped is short, the heat conduction efficiency is reduced, and the temperature may not reach the ideal value. When the transmission speed is low, the contact time between the silicone heating plate and the material to be stamped is long, the heat conduction efficiency is improved, and the temperature may exceed the ideal value. Therefore, it is very important to reasonably control the transmission speed. By comprehensively considering the complex relationship between temperature and speed, the decoder can generate precise control instructions to ensure the stability and efficiency of the stamping process. More specifically, during the decoding process, the decoder receives the silicone heating plate temperature timing offset representation vector from the encoder, which is a high-dimensional feature representation that contains the timing characteristics of the temperature data.The decoder gradually restores the dimension of the silicone heater temperature timing offset representation vector through multiple intermediate layers, and learns the mapping relationship from the silicone heater temperature timing offset representation vector to the control signal. These intermediate layers usually contain fully connected layers or convolutional layers, and increase the expressive power of the model through nonlinear activation functions (such as ReLU). In this process, the decoder continuously optimizes its performance through a feedback mechanism to ensure that the temperature and speed of the hot stamping process are always kept within the optimal range, thereby ensuring the consistency and stability of the hot stamping effect, thereby achieving higher-precision and more intelligent hot stamping control.

[0052] In the case where the ideal silicone heater plate temperature timing feature vector and the actual silicone heater plate temperature timing feature vector respectively represent the temperature of the silicone heater plate and the rolling pressure of the transmission roller mechanism based on the timing correlation characteristics of different timing feature representation dimensions, when calculating the semantic difference information therebetween, the silicone heater plate temperature timing offset representation vector will also have feature timing expression differences based on the same source time domain distribution, thereby causing feature distribution diversity within the semantic difference expression domain. As a result, when the silicone heater plate temperature timing offset representation vector is decoded and regressed through a decoder-based hot stamping transmission speed controller for decoding probability, it will affect the accuracy of the decoding result.

[0053] Based on this, the temperature timing offset characterization vector of the silicone heating plate is decoded and regressed by the hot stamping transmission speed controller based on a decoder, and the temperature timing offset characterization vector of the silicone heating plate is first optimized. The optimization process specifically includes: determining the mean and variance of the feature set composed of the eigenvalues ​​of the temperature timing offset characterization vector of the silicone heating plate, and dividing the mean by the variance to obtain the probability weight value of the temperature timing offset of the silicone heating plate: ;in, represents the mean value of the feature set composed of the eigenvalues ​​of the temperature timing offset characterization vector of the silicone heating plate, represents the variance of the feature set composed of the eigenvalues ​​of the temperature time series offset characterization vector of the silicone heating plate, It represents the probability weight value of the temperature timing offset of the silicone heating plate.

[0054] The silicone heater temperature timing offset characterization vector is multiplied by the inverse of the maximum eigenvalue of the silicone heater temperature timing offset characterization vector to obtain the silicone heater temperature timing offset probability constraint vector: ;in, represents the temperature timing offset characterization vector of the silicone heating plate, represents the inverse of the maximum eigenvalue of the temporal offset characterization vector of the silicone heating plate temperature, is the dot product, represents the probability constraint vector of the temperature timing offset of the silicone heating plate.

[0055] The silicone heater temperature timing offset probability constraint vector point is added with the silicone heater temperature timing offset probability weight value, and the base 2 logarithmic value is calculated to obtain the silicone heater temperature timing offset information interaction vector: ;in, It's a little plus. is the logarithm with base 2, Represents the temperature timing offset information interaction vector of the silicone heating plate.

[0056] After subtracting the silicone heater plate temperature timing offset probability constraint vector from the silicone heater plate temperature timing offset probability weight value, the inverse of each eigenvalue is calculated to obtain the silicone heater plate temperature timing offset sequence constraint vector: ;in, It's a point reduction. represents the temperature timing offset sequence constraint vector of the silicone heating plate.

[0057] Finally, the silicone heater plate temperature timing offset information interaction vector is added to the silicone heater plate temperature timing offset sequence constraint vector to obtain an optimized silicone heater plate temperature timing offset characterization vector.

[0058] That is, considering that when the weight matrix during decoding probability regression acts on the silicone heating plate temperature timing offset representation vector, the distribution diversity of the silicone heating plate temperature timing offset representation vector causes the uncertainty of the weight stimulus parameters of the weight matrix, which makes the posterior probability density inference of the silicone heating plate temperature timing offset representation vector through the weight matrix missing, which will affect the accuracy of the probability regression result. The above optimization process simulates the mesoscale interaction structure under probability constraints between the eigenvalue scale and the eigenvector scale of the silicone heating plate temperature timing offset representation vector through the probability statistical characteristics of the silicone heating plate temperature timing offset representation vector, so as to construct a bidirectional hidden variable motif based on the mesoscale short sequence relative to the probability statistical value to perform mesoscale bidirectional migration, and perform posterior recovery based on the sequence constraints on the interactive information, so as to improve the convergence effect of the silicone heating plate temperature timing offset representation vector in the probability density domain, and improve the accuracy of the control result obtained by the hot stamping transmission speed controller based on the decoder.

[0059] As described above, the continuous hot stamping automation device 300 according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with a continuous hot stamping automation algorithm. In a possible implementation, the continuous hot stamping automation device 300 according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the continuous hot stamping automation device 300 can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the continuous hot stamping automation device 300 can also be one of the many hardware modules of the wireless terminal.

[0060] Alternatively, in another example, the continuous hot stamping automation device 300 and the wireless terminal may also be separate devices, and the continuous hot stamping automation device 300 may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in a predetermined data format.

[0061] Furthermore, a continuous hot stamping automation method is also provided.

[0062] Figure 4 Flow chart of the continuous hot stamping automation method according to an embodiment of the present application. Figure 4 As shown, the continuous hot stamping automation method according to the embodiment of the present application includes the steps of: S1, performing local time encoding on the time queue of the roller rolling pressure data to obtain a sequence of local time characteristics of the roller rolling pressure; S2, performing node feature prompt dynamic propagation aggregation on the sequence of local time characteristics of the roller rolling pressure to obtain the local time characteristic propagation aggregation feature of the roller rolling pressure; S3, performing silicone heater plate temperature prediction on the local time characteristic propagation aggregation feature of the roller rolling pressure to obtain the ideal silicone heater plate temperature timing characteristic; S4, performing time encoding on the time queue of the temperature data to obtain the actual silicone heater plate temperature timing characteristic; S5, calculating the silicone heater plate temperature timing offset representation between the ideal silicone heater plate temperature timing characteristic and the actual silicone heater plate temperature timing characteristic, and obtaining a control result based on the silicone heater plate temperature timing offset representation.

[0063] In summary, the continuous hot stamping automation method according to the embodiment of the present application is explained, which obtains the time queue of the roller pressure data of the transmission roller mechanism and the time queue of the temperature data of the silicone heating plate, and adopts the data analysis and encoding method based on deep learning to perform local time series encoding on the roller pressure data, and then performs feature prompt dynamic propagation and ideal temperature time series prediction on the encoded local time series characteristics of each roller pressure, and at the same time performs time series encoding on the temperature data, so as to intelligently obtain the transmission speed recommendation value of the transmission roller mechanism based on the time series offset characterization between the ideal silicone heating plate temperature time series characteristics and the actual silicone heating plate temperature time series characteristics. In this way, the time series changes of various parameters in the hot stamping process can be dynamically monitored, so as to adjust the transmission speed of the transmission roller mechanism more accurately and timely, ensure the consistency and stability of the hot stamping effect, and thus achieve higher precision and more intelligent hot stamping control.

[0064] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A continuous hot stamping automation equipment, comprising: Workbench; frame; A telescopic driving member installed on the inner side of both ends of the frame; A lifting plate disposed at an end of the telescopic driving member; A transmission roller mechanism installed on the crossbeam between the two lifting plates; a silicone heating plate arranged on a heating plate frame at the bottom of the lifting plate; and a film transmission mechanism fixed on the frame; A feeding mechanism installed on the workbench; A demoulding mechanism located below the silicone heating plate; characterized in that the continuous hot stamping automation equipment also includes: a pressure sensor installed on the transmission roller mechanism, used to collect the time queue of the roller rolling pressure data of the transmission roller mechanism; a temperature sensor installed on the silicone heating plate, used to collect the time queue of the temperature data of the silicone heating plate; a controller, used to control the transmission speed of the transmission roller mechanism based on the data collected by the pressure sensor and the temperature sensor; wherein the controller includes: a pressure encoding module, used to perform local time series encoding on the time queue of the roller rolling pressure data to obtain a sequence of local time series characteristics of the roller rolling pressure; a pressure aggregation module, A module for dynamically propagating and aggregating node features of the sequence of local timing features of the rolling pressure of the pressure roller to obtain propagation aggregation features of the local timing features of the rolling pressure of the pressure roller; a temperature timing prediction module for predicting the temperature of the silicone heater plate based on the propagation aggregation features of the local timing features of the rolling pressure of the pressure roller to obtain ideal silicone heater plate temperature timing features; a temperature encoding module for timing encoding the time queue of the temperature data to obtain actual silicone heater plate temperature timing features; a control module for calculating the silicone heater plate temperature timing offset representation between the ideal silicone heater plate temperature timing features and the actual silicone heater plate temperature timing features, and obtaining a control result based on the silicone heater plate temperature timing offset representation.

2. The continuous hot stamping automation equipment according to claim 1, characterized in that: The pressure encoding module is used to: pass the time queue of the roller rolling pressure data through a 1D-CNN-based rolling pressure local time series encoder to obtain a sequence of roller rolling pressure local time series feature vectors as a sequence of roller rolling pressure local time series features.

3. The continuous hot stamping automation equipment according to claim 2, characterized in that: The pressure aggregation module includes: a current roller pressure local time series feature vector extraction unit, which is used to extract the current roller pressure local time series feature vector from the sequence of the roller pressure local time series feature vectors; a pressure feature prompt value calculation unit, which is used to calculate the feature association strength prompt value of each roller pressure local time series feature vector in the sequence of the roller pressure local time series feature vectors based on the current roller pressure local time series feature vector to obtain a sequence of roller pressure local time series feature association strength prompt values, and calculate the intermediate state value of the sequence of roller pressure local time series feature association strength prompt values ​​to obtain a roller pressure feature association strength gate threshold; a pressure feature dynamic aggregation unit, which is used to perform gated modulation dynamic aggregation on the sequence of roller pressure local time series feature vectors based on the roller pressure feature association strength gate threshold and the sequence of roller pressure local time series feature association strength prompt values ​​to obtain a roller pressure local time series feature propagation aggregation feature vector as the roller pressure local time series feature propagation aggregation feature.

4. The continuous hot stamping automation equipment according to claim 3, characterized in that: The pressure feature prompt value calculation unit is used to: use the current roller rolling pressure local time series feature vector as the feature prompt gate, calculate the feature association strength prompt value of each roller rolling pressure local time series feature vector in the sequence of roller rolling pressure local time series feature vectors relative to the current roller rolling pressure local time series feature vector to obtain a sequence of roller rolling pressure local time series feature association strength prompt values; extract the median of the sequence of roller rolling pressure local time series feature association strength prompt values ​​as the intermediate state value to obtain the roller rolling pressure feature association strength gate threshold.

5. The continuous hot stamping automation equipment according to claim 4, characterized in that: The pressure feature dynamic aggregation unit is used to: use the roller pressure feature association strength gate threshold as the switch threshold of the gating unit, input the sequence of the roller pressure local time series feature vectors and the sequence of the roller pressure local time series feature association strength prompt values ​​into the gating unit to obtain the sequence of modulated roller pressure local time series feature vectors; input the sequence of modulated roller pressure local time series feature vectors into the feature dynamic propagation module based on the forward LSTM model to obtain the roller pressure local time series feature propagation aggregation feature vector.

6. The continuous hot stamping automation equipment according to claim 5, characterized in that: The temperature timing prediction module is used to propagate the aggregated feature vector of the local timing feature of the rolling pressure of the pressure roller through a decoder-based silicone heater temperature predictor to obtain an ideal silicone heater temperature timing feature vector as the ideal silicone heater temperature timing feature.

7. The continuous hot stamping automation equipment according to claim 6, characterized in that: The temperature encoding module is used to: pass the time queue of the temperature data through a temperature time series encoder based on a forward LSTM model to obtain an actual silicone heating plate temperature time series feature vector as the actual silicone heating plate temperature time series feature.

8. The continuous hot stamping automation equipment according to claim 7, characterized in that: The control module includes: a silicone heating plate temperature timing offset calculation unit, which is used to calculate the semantic difference information between the ideal silicone heating plate temperature timing feature vector and the actual silicone heating plate temperature timing feature vector to obtain a silicone heating plate temperature timing offset representation vector as the silicone heating plate temperature timing offset representation; a decoding unit, which is used to pass the silicone heating plate temperature timing offset representation vector through a decoder-based hot stamping transmission speed controller to obtain a control result, and the control result is used to represent a recommended decoding value of the transmission speed of the transmission roller mechanism.

9. A continuous hot stamping automation method, using the continuous hot stamping automation equipment according to claim 1, characterized in that: include: Performing local time series encoding on the time queue of the roller pressure data to obtain a sequence of local time series characteristics of the roller pressure; performing node feature prompt dynamic propagation aggregation on the sequence of local time series characteristics of the roller pressure to obtain local time series characteristic propagation aggregation characteristics of the roller pressure; performing silicone heater temperature prediction on the local time series characteristic propagation aggregation characteristics of the roller pressure to obtain ideal silicone heater temperature time series characteristics; performing time series encoding on the time queue of the temperature data to obtain actual silicone heater temperature time series characteristics; The silicone heating plate temperature timing offset characterization between the ideal silicone heating plate temperature timing characteristic and the actual silicone heating plate temperature timing characteristic is calculated, and a control result is obtained based on the silicone heating plate temperature timing offset characterization.

Images