A PID-based operation safety control method for the hot stamping machine roller heating system

Abstract

The present invention relates to the field of operation control technology, and specifically to a PID-based operation safety control method for a hot stamping machine roller heating system. First, a PLC program is used to receive and preprocess multi-parameter data collected by a sensor group to obtain a preprocessed multi-parameter time series vector, which is then input into a multi-source adjustment coefficient prediction module to obtain a threshold adjustment coefficient. Then, the PLC program is used to calculate a dynamic safety threshold based on the threshold adjustment coefficient and the current safety threshold, and the current safety threshold is overwritten and updated. Next, the error time series vector calculated by the PLC program, the updated PID parameters obtained by the parameter adjustment coefficient prediction module, and the dynamic safety threshold are input into a PID controller provided by the PLC to obtain a control output. The operation parameters are adjusted using the digital output port of the PLC in combination with the control output. The present invention can improve the effectiveness of the operation safety control of the hot stamping machine roller heating system.

Description

Technical Field

[0001] The present invention relates to the technical field of operation control, in particular to an operation safety control method of a hot stamping machine flower roller heating system based on PID. Background Art

[0002] The hot stamping machine flower roller heating system has complex nonlinear characteristics during actual operation, and the PID controller is a control algorithm based on a linear model. At the same time, the PID parameter settings are fixed. Therefore, the PID program control method cannot handle the dynamic control of the system operation safety parameters; in addition, the safety protection threshold and program control parameters in the actual hot stamping machine flower roller heating process will change due to environmental factors and hot stamping operation factors, resulting in incorrect judgment of the PID safety control logic; the existing technology lacks optimization and processing of program control parameters, and cannot stabilize the system's operating parameters, thereby reducing the effectiveness of the hot stamping machine flower roller heating system operation safety control.

[0003] Therefore, a PID-based operation safety control method for the hot stamping machine roller heating system is proposed. Summary of the Invention

[0004] The present invention aims to provide a PID-based operational safety control method for a hot stamping machine's roller heating system. First, a PLC program receives and preprocesses multi-parameter data collected by a sensor group to obtain a preprocessed multi-parameter time series vector, which is then input into a multi-source adjustment coefficient prediction module to obtain a threshold adjustment coefficient. Then, the PLC program calculates a dynamic safety threshold based on the threshold adjustment coefficient and the current safety threshold, overwriting and updating the current safety threshold. Next, the error time series vector calculated by the PLC program, the updated PID parameters obtained by the parameter adjustment coefficient prediction module, and the dynamic safety threshold are input into a PID controller provided by the PLC to obtain a control output. The PLC's digital output port is used in conjunction with the control output to adjust operational parameters. This invention can improve the effectiveness of operational safety control for a hot stamping machine's roller heating system.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A PID-based operation safety control method for a hot stamping machine roller heating system includes:

[0007] Acquire multi-parameter data collected by the sensor group;

[0008] Using a PLC program to preprocess the multi-parameter data according to a time dimension to obtain a preprocessed multi-parameter time series vector; storing the preprocessed multi-parameter time series vector in a data register of the PLC;

[0009] Recalling the preprocessed multi-parameter time series vector from the data register, and inputting the preprocessed multi-parameter time series vector into a multi-source adjustment coefficient prediction module to obtain a threshold adjustment coefficient;

[0010] Obtaining a dynamic safety threshold using the PLC program and according to the threshold adjustment coefficient and the current safety threshold;

[0011] Inputting the preprocessed multi-parameter time series vector and the current PID parameters into a parameter adjustment coefficient prediction module to obtain updated PID parameters; inputting the error time series vector, the updated PID parameters, and the dynamic safety threshold into a PID controller provided by a PLC to obtain a control output; wherein the error time series vector is obtained by calculating using the PLC program;

[0012] Use the digital output port of the PLC in conjunction with the control output to adjust the operating parameters.

[0013] Furthermore, the multi-parameter data includes: environmental data, hot stamping operation data, equipment status data and pattern roller status data.

[0014] Furthermore, the multi-parameter data is preprocessed according to the time dimension using a PLC program to obtain a preprocessed multi-parameter time series vector, which includes:

[0015] Performing data conversion and alignment on the multi-parameter data using a PLC program to obtain first multi-parameter data;

[0016] Using a PLC program to perform data cleaning on the first multi-parameter data to obtain second multi-parameter data;

[0017] The second multi-parameter data is normalized and vectorized using a PLC program to obtain the preprocessed multi-parameter time series vector.

[0018] Furthermore, the pre-processed multi-parameter time series vector is input into the multi-source adjustment coefficient prediction module to obtain the threshold adjustment coefficient. The process is:

[0019] dividing the preprocessed multi-parameter time series vector into a non-adjustment parameter time series vector and an adjustment parameter time series vector;

[0020] Inputting the influencing adjustment parameter time series vector and the adjustment parameter time series vector in the non-adjustment parameter time series vector into the pre-trained multi-source adjustment coefficient prediction module for processing to obtain the threshold adjustment coefficient;

[0021] Among them, the multi-source adjustment coefficient prediction module includes: an input layer, a multi-source feature extraction layer, an attention mechanism layer, an adjustment coefficient prediction layer and an output layer; the input layer is used to receive the time series vector affecting the adjustment parameter and the time series vector of the adjustment parameter; the multi-source feature extraction layer is used to extract parameter features from the time series vector; the attention mechanism layer is used to extract correlation features from the characteristics affecting the adjustment parameter and the adjustment parameter characteristics; the adjustment coefficient prediction layer is used to obtain the adjustment feature based on the correlation feature; the output layer is used to convert the adjustment feature into the threshold adjustment coefficient.

[0022] Furthermore, the process of obtaining the dynamic safety threshold using the PLC program and according to the threshold adjustment coefficient and the current safety threshold is as follows:

[0023] Using a PLC program to read the threshold adjustment coefficient of the adjustment parameter, the current safety threshold, the time sequence vector at the current moment, and the time sequence vector at the previous moment;

[0024] Using a PLC program to transform the threshold adjustment coefficient using a hyperbolic tangent function to obtain a threshold adjustment weight;

[0025] Combining the threshold adjustment weight, the current security threshold, the time series vector at the current moment, and the time series vector at the previous moment to obtain the dynamic security threshold;

[0026] The dynamic safety threshold is range-checked using a PLC program. If it does not exceed a specified range, the dynamic safety threshold is written into a PLC memory to replace the current safety threshold; otherwise, an early warning signal is issued and the current safety threshold is not updated.

[0027] Furthermore, the calculation formula of the dynamic safety threshold is:

[0028] ;

[0029] in, is the dynamic safety threshold; is the current safety threshold; adjusting a weight for the threshold; is the time series vector weight coefficient; is the Euclidean norm; and are the adjustment parameters at the current moment The time series vector and at the last moment The time series vector.

[0030] Furthermore, the preprocessed multi-parameter time series vector and the current PID parameters are input into the parameter adjustment coefficient prediction module to obtain the updated PID parameters; the error time series vector, the updated PID parameters and the dynamic safety threshold are input into the PID controller provided by the PLC to obtain the control output. The process includes:

[0031] Using a PLC program to read the parameter time series vector that can reflect the system operating state and the current PID parameters in the preprocessed multi-parameter time series vector;

[0032] Inputting the parameter time series vector and the current PID parameter into the pre-trained parameter adjustment coefficient prediction module for processing to obtain the PID parameter adjustment coefficient;

[0033] Updating the current PID parameters using the PID parameter adjustment coefficients to obtain the updated PID parameters;

[0034] Calculating the error value between the target parameter value and each dimension of the parameter time series vector using a PLC program, and forming the error time series vector from a series of the error values;

[0035] The time series error value in the error time series vector is used as an input of a PID controller, the PID controller parameters are updated using the updated PID parameters, and the dynamic safety threshold is used as a constraint condition of the PID controller to obtain the control output.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] 1. The present invention proposes a multi-source adjustment coefficient prediction method for obtaining the threshold adjustment coefficient of the adjustment parameter; the method uses a multi-source adjustment coefficient prediction module to process the time series vector of the influencing adjustment parameter and the time series vector of the adjustment parameter; by using the attention mechanism layer to extract the correlation information between multiple influencing adjustment parameter features and a single adjustment parameter feature, and then using the adjustment coefficient prediction layer to predict the adjustment coefficient according to the correlation information, the threshold adjustment coefficient of the adjustment parameter is obtained; since the influencing adjustment parameter and the adjustment parameter change with time, the threshold adjustment coefficient also changes dynamically, which can provide a data basis for the subsequent dynamic safety threshold calculation, thereby improving the effectiveness of the safety control of the hot stamping machine flower roller heating system.

[0038] 2. The present invention proposes a dynamic safety threshold for constraining the PID controller; the calculation process of the threshold combines the threshold adjustment weight of the adjustment parameter, the current safety threshold, the current time series vector and the previous time series vector; wherein, the threshold adjustment weight is obtained by performing a function transformation on the threshold adjustment coefficient output by the multi-source adjustment coefficient prediction module; the current time series vector and the previous time series vector are used to calculate the time series change of the adjustment parameter, thereby compensating and correcting the dynamic safety threshold; the dynamic safety threshold will be dynamically adjusted according to the actual operation of the system, thereby improving the effectiveness of the safety control of the hot stamping machine flower roller heating system.

[0039] 3. The present invention proposes an adaptive PID control method for improving the system operation control accuracy; the method first inputs the parameter time series vector reflecting the system operation status and the current PID parameters into the parameter adjustment coefficient prediction module to obtain the PID parameter adjustment coefficient; then, the PID parameter adjustment coefficient is used to update the current PID parameters to obtain the updated PID parameters; by calling the PID controller provided by the PLC and combining the time series error value, the updated PID parameters and the dynamic safety threshold, the system operation is effectively controlled, thereby improving the effectiveness of the operation safety control of the hot stamping machine flower roller heating system. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a flow chart of a PID-based safety control method for a hot stamping machine roller heating system according to the present invention;

[0041] Figure 2 Schematic diagram of the structure of the multi-source adjustment coefficient prediction module of the present invention;

[0042] Figure 3 Schematic diagram of the control output acquisition process of the present invention. DETAILED DESCRIPTION

[0043] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0044] See also Figures 1 to 3 The present invention provides a PID-based method for controlling the safe operation of a hot stamping machine roller heating system. The technical solution is as follows:

[0045] Example 1

[0046] In order to ensure the stable and safe operation of the hot stamping machine roller system, a hot stamping processing company used a PID-based hot stamping machine roller heating system operation safety control method proposed in this invention. This method combines a PLC program with PID control. In this invention, the PLC, as a programmable logic controller, mainly plays the role of data interaction, data storage, logic control, and process control.

[0047] The method flow diagram of the present invention can be referred to Figure 1 , as follows:

[0048] Acquire multi-parameter data collected by the sensor group;

[0049] Furthermore, the multi-parameter data includes: environmental data, hot stamping operation data, equipment status data and flower roller status data;

[0050] Furthermore, the environmental data includes: ambient temperature and ambient humidity; the hot stamping operation data includes: hot stamping rate, hot stamping pressure and hot stamping time; the equipment status data includes heater status data and sensor status data; the pattern roller status data includes: pattern roller temperature, pattern roller rate, pattern roller vibration and pattern roller pressure distribution;

[0051] Furthermore, the heater status data includes: heater power, heater voltage, heater current and heater operating time; the sensor status data is sensor signal strength.

[0052] By introducing multi-parameter data, a solid data foundation can be provided for the subsequent multi-source adjustment coefficient prediction and PID parameter adjustment coefficient prediction, so that the subsequent modules can fully learn the correlation between the parameter data. The use of data correlation can assist the PID controller to perform effective control, further improving the effectiveness of the safety control of the hot stamping machine roller heating system.

[0053] Using the PLC program to preprocess the multi-parameter data according to the time dimension to obtain a preprocessed multi-parameter time series vector; storing the preprocessed multi-parameter time series vector in the data register of the PLC;

[0054] Furthermore, the multi-parameter data is preprocessed according to the time dimension using the PLC program to obtain the preprocessed multi-parameter time series vector, which includes:

[0055] Performing data conversion and alignment on the multi-parameter data using a PLC program to obtain first multi-parameter data;

[0056] Using a PLC program to perform data cleaning on the first multi-parameter data to obtain second multi-parameter data;

[0057] Use the PLC program to perform data standardization and vectorization on the second multi-parameter data to obtain a preprocessed multi-parameter time series vector;

[0058] Furthermore, data conversion includes data type conversion and unit conversion; alignment refers to timestamp alignment;

[0059] Furthermore, data cleaning includes the following steps: noise elimination, outlier processing, and missing value processing; noise elimination can use a mean filter, median filter, Kalman filter, etc.; outliers are detected using statistical methods and missing values ​​are checked using timestamps, and then interpolation is used to fill in outliers and missing values;

[0060] Furthermore, the data normalization method may adopt Min-Max normalization or Z-score normalization; and then the normalized data is converted into a vector form to form a time series vector.

[0061] By performing data conversion and alignment, data cleaning, data standardization and vectorization on multi-parameter data, the comparability and time consistency of multi-parameter data can be ensured, and the quality of the data can be improved, thereby improving the accuracy of subsequent adjustment coefficient prediction and PID parameter adjustment coefficient prediction, and further improving the effectiveness of the safety control of the hot stamping machine roller heating system.

[0062] Calling the preprocessed multi-parameter time series vector from the data register, inputting the preprocessed multi-parameter time series vector into the multi-source adjustment coefficient prediction module to obtain the threshold adjustment coefficient;

[0063] Furthermore, the preprocessed multi-parameter time series vector is input into the multi-source adjustment coefficient prediction module to obtain the threshold adjustment coefficient as follows:

[0064] The preprocessed multi-parameter time series vector is divided into a non-adjustment parameter time series vector and an adjustment parameter time series vector;

[0065] Inputting the influencing adjustment parameter time series vector and the adjustment parameter time series vector in the non-adjustment parameter time series vector into the pre-trained multi-source adjustment coefficient prediction module for processing to obtain the threshold adjustment coefficient;

[0066] Furthermore, the influencing adjustment parameter is a parameter that can affect the control adjustment parameter; for example, if the adjustment parameter is a heater power parameter, the influencing adjustment parameter is a parameter affecting the heater power control, which may be: pattern roller temperature, hot stamping rate, ambient temperature, heater status data, etc.

[0067] Among them, the structure of the multi-source adjustment coefficient prediction module is as follows Figure 2As shown, it includes: an input layer, a multi-source feature extraction layer, an attention mechanism layer, an adjustment coefficient prediction layer and an output layer; the input layer is used to receive the time series vector of the influencing adjustment parameter and the time series vector of the adjustment parameter; the multi-source feature extraction layer is used to extract parameter features from the time series vector; the attention mechanism layer is used to extract correlation features from the influencing adjustment parameter features and the adjustment parameter features; the adjustment coefficient prediction layer is used to obtain the adjustment feature according to the correlation feature; the output layer is used to convert the adjustment feature into the threshold adjustment coefficient;

[0068] Furthermore, the multi-source feature extraction layer uses a one-dimensional convolution kernel to extract features from the parameter time series vector;

[0069] Furthermore, the attention mechanism layer adopts the Transformer encoding layer and uses the cross attention mechanism to learn the relationship between the features that affect the adjustment parameters and the features of the adjustment parameters;

[0070] Furthermore, the adjustment coefficient prediction layer adopts a fully connected neural network and obtains the adjustment feature based on the correlation feature prediction.

[0071] By utilizing the attention mechanism layer in the multi-source adjustment coefficient prediction module to extract the correlation information between multiple influencing adjustment parameter features and a single adjustment parameter feature, and then using the adjustment coefficient prediction layer to predict the adjustment coefficient based on the correlation information, the threshold adjustment coefficient of the adjustment parameter is obtained; the influencing adjustment parameter and the adjustment parameter change over time lead to dynamic changes in the threshold adjustment coefficient, thereby providing conditions for subsequent dynamic adjustment of system parameters.

[0072] Using the PLC program and according to the threshold adjustment coefficient and the current safety threshold, a dynamic safety threshold is obtained;

[0073] Furthermore, the process of obtaining the dynamic safety threshold using the PLC program and based on the threshold adjustment coefficient and the current safety threshold is as follows:

[0074] Use the PLC program to read the threshold adjustment coefficient of the adjustment parameter, the current safety threshold, the time series vector at the current moment, and the time series vector at the previous moment;

[0075] The threshold adjustment coefficient is transformed using the hyperbolic tangent function by using the PLC program to obtain the threshold adjustment weight;

[0076] The dynamic safety threshold is obtained by combining the threshold adjustment weight, the current safety threshold, the time series vector at the current moment, and the time series vector at the previous moment. Specifically, the Euclidean norm of the time series vector at the current moment and the time series vector at the previous moment are calculated, and the time series vector weight coefficient is set to weight them. Then, the dynamic safety threshold is obtained by summing the value obtained by weighting the current safety threshold using the threshold adjustment weight.

[0077] Use the PLC program to perform range check on the dynamic safety threshold. If it does not exceed the specified range, the dynamic safety threshold is written into the PLC memory to replace the current safety threshold; otherwise, an early warning signal is issued and the current safety threshold is not updated;

[0078] Furthermore, the hyperbolic tangent function is used to compress the range of the threshold adjustment coefficient to between (-1, 1); wherein a positive value indicates an increase in the baseline threshold, and a negative value indicates a decrease in the baseline threshold;

[0079] Furthermore, the safety threshold range of each parameter is related to the aging performance of the equipment and historical experience, and needs to be set according to actual conditions.

[0080] By using the PLC program to calculate the threshold adjustment weight and the dynamic safety threshold, and performing range verification on the dynamic safety threshold, the effectiveness of the dynamic safety threshold can be effectively improved, thereby ensuring effective numerical constraints on PID dynamic control and further improving the effectiveness of the safety control of the hot stamping machine roller heating system.

[0081] Furthermore, the calculation formula of the dynamic safety threshold is:

[0082] ;

[0083] in, is the dynamic safety threshold; is the current safety threshold; Adjust weights for thresholds; is the time series vector weight coefficient; is the Euclidean norm; and They are the adjustment parameters at the current moment The time series vector and the last moment The time series vector of

[0084] Furthermore, the range of the threshold adjustment weight is between (-1, 1);

[0085] Furthermore, the time series vector weight coefficient is set to 0.1; of course, the setting of the time series vector weight coefficient can be adjusted according to actual needs and is not unique.

[0086] In order to illustrate the dynamic safety threshold proposed in the present invention, the heater power data of the same system in different time periods are randomly taken and the dynamic safety threshold test is performed for the same adjustment parameter. The data are recorded as sample one, sample two and sample three respectively, and each group of data is separated by half a year; the data includes: the time series vector of the influencing adjustment parameter, the time series vector of the adjustment parameter, the safety threshold, the time series vector of a certain moment in the time period, and the time series vector of the previous moment. The time window is 1 minute and the sampling frequency is once every 5 seconds; the maximum power of the heater is 2kW; the time series vector of the influencing adjustment parameter and the time series vector of the adjustment parameter of each group are input into the pre-trained multi-source adjustment coefficient prediction module for processing to obtain their respective threshold adjustment coefficients; combined with the calculation formula of the dynamic safety threshold, the dynamic safety threshold test results are obtained, which can be referred to Table 1.

[0087] Table 1. Dynamic safety threshold test results

[0088]

[0089] The calculation process of the dynamic safety threshold in this embodiment combines the threshold adjustment weight of the adjustment parameter, the current safety threshold, the current time series vector and the previous time series vector; the current safety threshold is dynamically weighted using the variable threshold adjustment weight, and combined with the time series vector distance between the current moment and the previous moment, the threshold under the system operation state can be effectively adjusted.

[0090] The preprocessed multi-parameter time series vector and the current PID parameters are input into the parameter adjustment coefficient prediction module to obtain the updated PID parameters; the error time series vector, the updated PID parameters, and the dynamic safety threshold are input into the PID controller provided by the PLC to obtain the control output; wherein the error time series vector is obtained by calculating using the PLC program;

[0091] Furthermore, the process of obtaining the control output is shown as follows: Figure 3 As shown in the figure, the specific implementation process includes:

[0092] Use the PLC program to read the parameter time series vector that can reflect the system operation status and the current PID parameters in the pre-processed multi-parameter time series vector;

[0093] The parameter time series vector and the current PID parameters are input into the pre-trained parameter adjustment coefficient prediction module for processing to obtain the PID parameter adjustment coefficient;

[0094] The current PID parameters are updated using the PID parameter adjustment coefficient to obtain updated PID parameters;

[0095] The PLC program is used to calculate the error value between the target parameter value and each dimension in the parameter time series vector, and a series of error values ​​are used to form an error time series vector;

[0096] The time series error value in the error time series vector is used as the input of the PID controller, the PID controller parameters are updated by using the updated PID parameters, and the dynamic safety threshold is used as the constraint condition of the PID controller to obtain the control output;

[0097] Furthermore, the parameter adjustment coefficient prediction module has the same structure as the multi-source adjustment coefficient prediction module, except that the input data of the parameter adjustment coefficient prediction module includes PID controller parameters and parameters reflecting the system operation status, and the output data is the PID parameter adjustment coefficient;

[0098] Furthermore, parameters reflecting the system operating status include: pattern roller status data, hot stamping operation data and environmental data.

[0099] By utilizing the timing error value, updating the PID parameters and dynamic safety threshold and combining it with the PID controller provided by the PLC, the dynamic control of the system operating parameters can be effectively improved, thereby improving the effectiveness of the operational safety control of the hot stamping machine roller heating system.

[0100] Use the PLC's digital output ports in conjunction with control outputs to adjust operating parameters.

[0101] Furthermore, taking heater power as an example, the distribution of heater power is first divided into different levels, the corresponding level is selected according to the control result of the heater power, and the corresponding digital output port state is obtained by referring to the correspondence table between the digital output port and the heater power level; then, the output instruction of the PLC is used to set the digital output port to the corresponding state to realize the adjustment of the heater power parameters.

[0102] This embodiment proposes a PID-based operational safety control method for the hot stamping machine roller heating system. First, a PLC program receives and preprocesses multi-parameter data collected by a sensor group to obtain a preprocessed multi-parameter time series vector, which is then input into a multi-source adjustment coefficient prediction module to obtain a threshold adjustment coefficient. Then, the PLC program calculates a dynamic safety threshold based on the threshold adjustment coefficient and the current safety threshold, and overwrites and updates the current safety threshold. Next, the error time series vector calculated by the PLC program, the updated PID parameters obtained by the parameter adjustment coefficient prediction module, and the dynamic safety threshold are input into a PID controller provided by the PLC to obtain a control output. The operating parameters are adjusted using the PLC's digital output port in conjunction with the control output. This invention can improve the effectiveness of operational safety control for the hot stamping machine roller heating system.

[0103] Example 2

[0104] This paper proposes a PID-based method for safe operation control of the hot stamping machine's roller heating system. To further verify the effectiveness of the proposed adjustment coefficient prediction method and control scheme, we conducted method and control scheme effectiveness tests on different modules and schemes. We selected two companies, A and B, to conduct the above two sets of ablation tests.

[0105] The present invention selects the historical parameter data of processing enterprise A for the past three years as the data set of the module, among which the data from the first two years are used as the training set of the module, and the data from the third year are used as the validation set; the sampling of the module data set refers to the following rules: taking the day as the unit, randomly extract two groups of data from the morning, noon and evening respectively.

[0106] The present invention uses different adjustment coefficient prediction methods to process the data set collected from processing enterprise A; method one is the multi-source adjustment coefficient prediction module proposed by the present invention, method two is to remove the attention mechanism layer in the multi-source adjustment coefficient prediction module, and method three uses manual prediction combined with historical experience; therefore, the first two methods need to first input the training set into the module for training, and then use the pre-training module and the validation set for testing; while method three directly uses the validation set for testing; the threshold adjustment coefficient obtained by each group of methods is applied to the safety threshold at that time and combined with the equipment status data, flower roller status data and environmental data at that time for verification, and the proportion of the safety threshold adjusted by each method within a reasonable range is obtained, thereby obtaining the method effectiveness test results; the method effectiveness test results are shown in Table 2.

[0107] Table 2 Method effectiveness test results

[0108]

[0109] The results in Table 2 show that the effectiveness test results of the multi-source adjustment coefficient prediction module proposed in the present invention in terms of safety threshold adjustment are better than those of other methods. This shows that the multi-source adjustment coefficient prediction module proposed in the present invention can obtain a more reasonable adjustment coefficient, further improving the effectiveness of the safety control of the hot stamping machine roller heating system.

[0110] To further test the effectiveness of the control scheme, this example collected historical parameter data from processing company B for the past two years. Using the same data set partitioning method and data sampling rules, different control schemes were tested on the same data to obtain their respective control results.

[0111] To improve the efficiency of the scheme test, only the heater power was selected as the control output parameter. The respective power control results were then combined with the current roller status data and environmental data, and the control results of each scheme were obtained through manual verification to determine the proportion of the control results within a reasonable range.

[0112] The control schemes of each group are respectively recorded as Scheme 1, Scheme 2, Scheme 3 and Scheme 4; among them, Scheme 1 is the scheme proposed by the present invention, that is, the error time series vector calculated by the PLC program, the updated PID parameters obtained by the parameter adjustment coefficient prediction module and the dynamic safety threshold are input into the PID controller provided by the PLC to obtain the control output; Scheme 2 changes the updated PID parameters in Scheme 1 to fixed parameters; Scheme 3 changes the dynamic safety threshold in Scheme 1 to a fixed safety threshold; Scheme 4 changes both the updated PID parameters and the dynamic safety threshold in Scheme 1 to fixed values;

[0113] The results of the control scheme effectiveness test are shown in Table 3.

[0114] Table 3 Control scheme effectiveness test results

[0115]

[0116] From the results in Table 3, it can be seen that the test results of the effective proportion of the control results obtained by adopting Scheme 1, that is, the control scheme proposed in the present invention, are better than those obtained by using other schemes. This shows that combining the update of PID parameters and dynamic safety thresholds is necessary for effective control parameters, which can significantly improve the effectiveness of the safety control of the hot stamping machine roller heating system.

[0117] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A PID-based operation safety control method for a hot stamping machine roller heating system, characterized in that: include: Acquire multi-parameter data collected by the sensor group; Preprocessing the multi-parameter data according to the time dimension using a PLC program to obtain a preprocessed multi-parameter time series vector; Storing the preprocessed multi-parameter time series vector in a data register of a PLC; Recalling the preprocessed multi-parameter time series vector from the data register, and inputting the preprocessed multi-parameter time series vector into a multi-source adjustment coefficient prediction module to obtain a threshold adjustment coefficient; Obtaining a dynamic safety threshold using the PLC program and according to the threshold adjustment coefficient and the current safety threshold; The preprocessed multi-parameter time series vector and the current PID parameters are input into a parameter adjustment coefficient prediction module to obtain updated PID parameters; the error time series vector, the updated PID parameters, and the dynamic safety threshold are input into a PID controller provided by a PLC to obtain a control output; wherein, the error value of each dimension in the parameter time series vector and the target parameter value are calculated using a PLC program, and a series of the error values ​​constitute the error time series vector; Use the digital output port of the PLC in conjunction with the control output to adjust the operating parameters.

2. The PID-based hot stamping machine roller heating system operation safety control method according to claim 1 is characterized in that: The multi-parameter data includes: environmental data, hot stamping operation data, equipment status data and pattern roller status data.

3. The PID-based hot stamping machine roller heating system operation safety control method according to claim 1 is characterized in that: The process of preprocessing the multi-parameter data according to the time dimension using the PLC program to obtain the preprocessed multi-parameter time series vector includes: Performing data conversion and alignment on the multi-parameter data using a PLC program to obtain first multi-parameter data; Using a PLC program to perform data cleaning on the first multi-parameter data to obtain second multi-parameter data; The second multi-parameter data is normalized and vectorized using a PLC program to obtain the preprocessed multi-parameter time series vector.

4. The PID-based hot stamping machine roller heating system operation safety control method according to claim 1 is characterized in that: The process of inputting the preprocessed multi-parameter time series vector into the multi-source adjustment coefficient prediction module to obtain the threshold adjustment coefficient is as follows: dividing the preprocessed multi-parameter time series vector into a non-adjustment parameter time series vector and an adjustment parameter time series vector; Inputting the influencing adjustment parameter time series vector and the adjustment parameter time series vector in the non-adjustment parameter time series vector into the pre-trained multi-source adjustment coefficient prediction module for processing to obtain the threshold adjustment coefficient; Among them, the multi-source adjustment coefficient prediction module includes: an input layer, a multi-source feature extraction layer, an attention mechanism layer, an adjustment coefficient prediction layer and an output layer; the input layer is used to receive the time series vector affecting the adjustment parameter and the time series vector of the adjustment parameter; the multi-source feature extraction layer is used to extract parameter features from the time series vector; the attention mechanism layer is used to extract correlation features from the characteristics affecting the adjustment parameter and the adjustment parameter characteristics; the adjustment coefficient prediction layer is used to obtain the adjustment feature based on the correlation feature; the output layer is used to convert the adjustment feature into the threshold adjustment coefficient.

5. The PID-based operation safety control method for the hot stamping machine roller heating system according to claim 1 is characterized in that: The process of obtaining the dynamic safety threshold using the PLC program and according to the threshold adjustment coefficient and the current safety threshold is as follows: Using a PLC program to read the threshold adjustment coefficient of the adjustment parameter, the current safety threshold, the time sequence vector at the current moment, and the time sequence vector at the previous moment; Using a PLC program to transform the threshold adjustment coefficient using a hyperbolic tangent function to obtain a threshold adjustment weight; Combining the threshold adjustment weight, the current security threshold, the time series vector at the current moment, and the time series vector at the previous moment to obtain the dynamic security threshold; Performing a range check on the dynamic safety threshold using a PLC program, and if the dynamic safety threshold does not exceed a specified range, writing the dynamic safety threshold into a PLC memory to replace the current safety threshold; Otherwise, a warning signal is issued and the current safety threshold is not updated.

6. The PID-based hot stamping machine roller heating system operation safety control method according to claim 5 is characterized in that: The calculation formula of the dynamic safety threshold is: ； in, is the dynamic safety threshold; is the current safety threshold; adjusting a weight for the threshold; is the time series vector weight coefficient; is the Euclidean norm; are the adjustment parameters at the current moment The time series vector and at the last moment The time series vector.

7. The PID-based hot stamping machine roller heating system operation safety control method according to claim 1 is characterized in that: Inputting the preprocessed multi-parameter time series vector and the current PID parameters into the parameter adjustment coefficient prediction module to obtain the updated PID parameters; The process of inputting the error time series vector, the updated PID parameter and the dynamic safety threshold into a PID controller provided by a PLC to obtain a control output includes: Using a PLC program to read the parameter time series vector that can reflect the system operating state and the current PID parameters in the preprocessed multi-parameter time series vector; Inputting the parameter time series vector and the current PID parameter into the pre-trained parameter adjustment coefficient prediction module for processing to obtain the PID parameter adjustment coefficient; Updating the current PID parameters using the PID parameter adjustment coefficients to obtain the updated PID parameters; Calculating the error value between the target parameter value and each dimension of the parameter time series vector using a PLC program, and forming the error time series vector from a series of the error values; The time series error value in the error time series vector is used as an input of a PID controller, the PID controller parameters are updated using the updated PID parameters, and the dynamic safety threshold is used as a constraint condition of the PID controller to obtain the control output.

Images