Printing pressure self-adaptive regulation and control system and method based on historical work order big data

Abstract

The invention discloses a printing pressure self-adaptive regulation and control system and method based on historical work order big data, and belongs to the technical field of printing pressure regulation. The invention discloses a printing pressure self-adaptive regulation and control system and method based on historical work order big data. The method comprises the following steps: collecting historical work order big data; extracting work order features; constructing a pressure regulation and control model; collecting current work order real-time data; screening historical reference work orders; obtaining a final pressure regulation and control parameter; executing pressure regulation and control; the system comprises a historical data acquisition module, a feature extraction module, a model construction module, a real-time data acquisition module, a working condition matching module, a pressure regulation and control module and an instruction execution module. The problems that historical work order big data analysis is not available, optimal pressure parameters cannot be reused, and the pre-judgment capability of data support is lacked in the prior art are solved. The printing pressure is accurately controlled through a three-level regulation and control mechanism based on historical work order big data, the quality and efficiency are improved, the cost is reduced, and the stability of printed products is improved.

Description

Technical Field

[0001] This invention relates to the field of printing pressure regulation technology, specifically to an adaptive printing pressure control system and method based on historical work order big data. Background Technology

[0002] Printing pressure is one of the core process parameters in printing. Its magnitude directly determines key quality indicators such as ink uniformity, image clarity, and registration accuracy of printed materials. It also affects the lifespan of printing plates, the consumption of printing consumables, and the operational stability of printing equipment. As the printing industry develops towards higher precision, higher speed, and personalized customization, increasingly higher demands are being placed on the precision of printing pressure control and response speed.

[0003] Chinese Patent Publication No. CN118838436A discloses an automatic printing pressure adjustment control system and method for a satellite-type flexographic printing press. The system includes a programmable logic controller (PLC). The PLC is connected via an industrial Ethernet bus to a central roller servo unit, a printing roller servo unit, and an anilox roller servo unit, all connected to the central roller. The PLC is also connected via the industrial Ethernet bus to a printing roller pressure adjustment servo unit that drives the printing roller to feed towards the central roller, and another servo unit that drives the anilox roller to feed towards the printing roller. By controlling the rotational speed of the central roller, the printing speed is changed. Furthermore, by controlling the servo motors on both sides of the printing roller and the anilox roller, the center distance between the printing roller and the central roller, and between the printing roller and the anilox roller, is changed, thus achieving automatic adjustment of the printing pressure.

[0004] In actual use, the control logic of the aforementioned patent only revolves around the mechanical linkage of the roller spacing and rotation speed of the equipment, without incorporating historical work order big data for analysis. As a result, it is impossible to refer to the optimal pressure parameters at different rotation speeds in similar historical work orders, and the control relies entirely on real-time mechanical status, lacking data-supported predictive capabilities. Therefore, it does not meet the current needs. To address this, we propose a printing pressure adaptive control system and method based on historical work order big data. Summary of the Invention

[0005] The purpose of this invention is to provide a printing pressure adaptive control system and method based on historical work order big data. By constructing a historical work order big data pressure control model, collecting full-dimensional data to extract features, setting a three-level control mechanism, predicting the optimal pressure parameters in advance and dynamically fine-tuning them, quickly matching similar work orders to correct parameters, accurately controlling pressure, improving quality and efficiency, reducing costs, reducing material waste and reliance on manual adjustments, and improving the stability and consistency of printed product quality, thus solving the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a printing pressure adaptive control method based on historical work order big data, comprising the following steps: S1: Collect all dimensions of historical printing work orders, and obtain a standardized historical work order dataset after preprocessing; S2: Extract features from the standardized historical work order dataset to obtain the feature vectors of the historical work orders; S3: Using the feature vector of historical work orders as input and the corresponding optimal pressure control parameters as output, the initial model is continuously iterated and optimized to obtain the optimized pressure control model. S4: Collect real-time working condition data of the current printing work order, extract features from the real-time working condition data of the current work order, and obtain the real-time feature vector of the current work order. S5: Calculate the cosine similarity between the real-time feature vector and the historical work order feature vector, and select the N historical reference work orders with the highest similarity. S6: Utilize the optimized pressure control model to output the initial pressure control parameters, correct the initial parameters, and dynamically fine-tune the corrected parameters to obtain the final pressure control parameters. S7: Convert the final pressure control parameters into control commands and send them to the actuator to drive the actuator to adjust the printing pressure.

[0007] Preferably, the full-dimensional data of the historical work orders includes basic work order information, material characteristic data, process parameter data, equipment operation data, printing quality inspection data, and pressure control parameter data.

[0008] Preferably, the feature extraction process in S2 specifically includes: The paper's basis weight, thickness, smoothness, stiffness, and moisture content are normalized and then used as paper characteristics. After normalizing the viscosity, fineness, and drying speed of the ink, it is used as the ink characteristics. The paper characteristics and ink characteristics are combined to form a material characteristic sub-vector. The printing speed, anilox roller line count, ink supply, and drying temperature are normalized and combined with the coding value of the printing plate type to form a process feature sub-vector. The drum speed, drum temperature, bearing temperature, and motor current are normalized, and the equipment running time and the time of the most recent maintenance are converted into equipment aging coefficients and combined to form equipment status feature sub-vectors. The quality requirement levels are converted into numerical codes, and combined with preset thresholds for ink uniformity, image clarity, and registration accuracy to form a quality requirement feature sub-vector. By concatenating the material feature sub-vectors, process feature sub-vectors, equipment status feature sub-vectors, and quality requirement feature sub-vectors, the feature vectors of historical work orders are obtained.

[0009] Preferably, the model training process in S3 specifically includes: The standardized historical work order dataset was divided into a training set and a validation set in a 7:3 ratio. Set initial values ​​for hyperparameters of the gradient boosting tree algorithm, such as the number of decision trees, tree depth, learning rate, and minimum number of sample splits. The historical work order feature vectors of the training set are input into the gradient boosting tree model, and the model is trained with the corresponding optimal pressure control parameters as the target values. The initial model obtained from training was validated using a validation set, and the mean square error between the predicted pressure control parameters and the actual optimal pressure control parameters was calculated. The hyperparameters were iteratively optimized using a grid search method until the mean square error reached a preset threshold, thus obtaining the initial pressure control model. At preset time intervals, newly added standardized historical work order data are added to the training set to retrain the initial pressure control model and continuously optimize the model to obtain an optimized pressure control model.

[0010] Preferably, the similarity filtering process in S5 specifically includes: The cosine similarity algorithm is used to calculate the similarity between the real-time feature vector of the current work order and the feature vector of each historical work order; The calculated similarity scores are sorted in descending order, and the top N most similar historical work orders are selected as historical reference work orders.

[0011] Preferably, the parameter calculation and fine-tuning process in S6 specifically includes: The real-time feature vector of the current work order is input into the optimized pressure control model to obtain the initial pressure control parameters predicted by the model. Calculate the weighted average of the optimal pressure control parameters for N historical reference work orders, with the weights being the similarity between each historical reference work order and the current work order; The initial pressure control parameters are combined with the weighted average value to obtain the corrected pressure control parameters; Calculate the deviation between real-time quality monitoring data and preset quality standards. If the deviation is less than or equal to the preset allowable deviation, keep the corrected parameters unchanged. If the deviation is greater than the preset allowable deviation, the corrected pressure control parameters are dynamically fine-tuned based on the direction and magnitude of the deviation and the adjustment experience of similar historical deviations, to obtain the final pressure control parameters. The finalized pressure control parameters are converted into electrical signal commands that the actuator can recognize. These electrical signal commands include the pressure adjustment amount, adjustment speed, and stabilization time.

[0012] The printing pressure adaptive control system based on historical work order big data is applied in the printing pressure adaptive control method based on historical work order big data, including: The historical data acquisition module is used to collect full-dimensional data corresponding to historical printing work orders, and to preprocess the collected raw data to obtain a standardized historical work order dataset. The feature extraction module is used to extract features from the various dimensions of the standardized historical work order dataset to obtain the feature vector of the historical work orders. The model building module is used to train the model using the gradient boosting tree algorithm to obtain an initial pressure regulation model, and to continuously iterate and optimize the initial pressure regulation model to obtain an optimized pressure regulation model. The real-time data acquisition module is used to collect real-time working condition data of the current printing work order and extract features from the collected real-time working condition data to obtain the real-time feature vector of the current work order. The work condition matching module is used to perform similarity matching between the real-time feature vector of the current work order and the feature vector of historical work orders, and to filter out several historical reference work orders with the highest similarity. The pressure control module is used to obtain initial pressure control parameters, correct the initial pressure control parameters to obtain final pressure control parameters, dynamically fine-tune the final pressure control parameters, and generate pressure control commands. The instruction execution module is used to receive pressure control instructions and drive the pressure adjustment components of the printing equipment to perform precise actions, thereby achieving adaptive adjustment of printing pressure.

[0013] Preferably, the pressure regulation module specifically includes: The initial pressure control parameter acquisition unit is used to input the real-time feature vector of the current work order into the optimized pressure control model to obtain the initial pressure control parameters predicted by the model. The initial parameter correction unit is used to calculate the weighted average of the optimal pressure control parameters of N historical reference work orders, and to fuse the initial pressure control parameters with the weighted average to obtain the corrected pressure control parameters. The dynamic fine-tuning unit is used to calculate the deviation between real-time quality monitoring data and preset quality standards. Based on the direction and magnitude of the deviation, and combined with the adjustment experience of similar historical deviations, the corrected pressure control parameters are dynamically fine-tuned to obtain the final pressure control parameters. The instruction generation unit is used to dynamically fine-tune the final pressure control parameters based on the deviation between real-time quality monitoring data and preset quality standards, and generate pressure control instructions.

[0014] Preferably, the final determined pressure control parameters are converted into electrical signal commands that the actuator can recognize, including: Obtain the safety operating standards of the implementing agency; A safety analysis is performed on the final pressure control parameters according to the safe operation standards to determine whether the final pressure control parameters meet the safe operation standards and to obtain the safety analysis results. Based on the safety analysis results, the final pressure control parameters are verified for compliance. When the safety analysis results are found to be inconsistent with the safe operation standards, the final pressure control parameters are verified in conjunction with the safe operation standards to determine the compliance data of the final pressure control parameters and obtain the final pressure control verification parameters. By using the physical quantity mapping relationship, the final pressure control parameter or the final pressure control verification parameter is converted into the corresponding physical control quantity of the actuator to obtain the target physical quantity. The interface type of the actuator is combined with the target physical quantity to encapsulate the electrical signal instruction, thus obtaining the first target electrical signal instruction; While the executing agency responds to the first target electrical signal command, it performs stability monitoring and statistics, and combines the stability time for maintenance control.

[0015] Preferably, the real-time feature vector of the current work order is matched with the feature vectors of historical work orders based on similarity, including: Dimensional analysis and alignment are performed on the real-time feature vector of the current work order and the feature vector of the historical work orders to obtain the real-time feature processing vector of the current work order and the feature processing vector of the historical work orders. Timestamp analysis is performed on the feature processing vectors of historical work orders, and attenuation factor analysis is calculated using the following formula: ; In the above formula, For historical work orders The attenuation factor, The attenuation coefficient is... The time of the current work order. For historical work orders Time, The decay period; The similarity calculation formula is used to calculate the similarity between the real-time feature processing vector of the current work order and the feature processing vector of historical work orders, and a preliminary similarity calculation value is obtained. The initial similarity calculation value is revised by combining it with the corresponding decay factor to obtain the final similarity between the current work order and historical work orders.

[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention constructs a pressure control model based on historical work order big data. At the start of each work order, it can predict the optimal pressure control parameters in advance by combining real-time working condition data, achieving pre-setting of printing pressure. Dynamic fine-tuning based on real-time quality data further ensures the stability of pressure during the printing process, effectively reducing defective printed materials in the transition stage, minimizing material waste, and improving production efficiency. Employing a three-level control mechanism of model prediction, historical data correction, and real-time dynamic fine-tuning, it achieves continuous, smooth, and precise adjustment of printing pressure. Model prediction and historical data correction ensure the accuracy of initial pressure parameters, while real-time dynamic fine-tuning can quickly respond to minor fluctuations during the printing process, significantly improving the quality stability and consistency of printed materials.

[0017] 2. This invention collects comprehensive data from historical work orders, extracting features to form feature vectors containing multiple sub-vectors. This comprehensively reflects various factors affecting printing pressure and their coupling relationships. The pressure control model is trained based on these multi-dimensional features, enabling it to output pressure control parameters highly matched to the current working conditions. This solves the problem of insufficient control accuracy caused by existing technologies that only consider a single factor. Even when printing materials or process parameters change, it can still quickly output reasonable pressure control parameters, ensuring stable printing quality. Through the collection and preprocessing of historical work order big data, a complete historical work order database is constructed, and a working condition matching module enables rapid matching of the current work order with similar historical work orders. When a new work order is started, the optimal pressure control parameters from similar historical work orders can be directly used to correct the model's predicted values, eliminating the need for extensive manual adjustments and significantly shortening production preparation time while improving production efficiency. Simultaneously, this mechanism reduces reliance on operator experience, avoids quality problems caused by manual adjustments, and lowers production management costs. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the printing pressure adaptive control method based on historical work order big data of the present invention; Figure 2 This is a schematic diagram of the printing pressure adaptive control system based on historical work order big data of the present invention. Figure 3 This is a schematic diagram of the parameter calculation and fine-tuning process of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] To address the issues of existing technologies relying on the mechanical linkage between roller spacing and rotation speed, lacking historical work order big data analysis, being unable to reuse optimal pressure parameters, and lacking data-supported predictive capabilities, please refer to [the relevant documentation / reference]. Figures 1-3 This embodiment provides the following technical solution: The adaptive control method for printing pressure based on historical work order big data includes the following steps: S1: Collect all dimensions of historical printing work orders, and obtain a standardized historical work order dataset after preprocessing; S2: Extract features from the standardized historical work order dataset to obtain the feature vectors of the historical work orders. The feature vectors include material feature sub-vectors, process feature sub-vectors, equipment status feature sub-vectors, and quality requirement feature sub-vectors. S3: Using the feature vectors of historical work orders as input and the corresponding optimal pressure control parameters as output, the gradient boosting tree algorithm is used to train the initial pressure control model; and the initial model is continuously iterated and optimized based on the newly added historical work order data to obtain the optimized pressure control model. S4: Collect real-time working condition data of the current printing work order, extract features from the real-time working condition data of the current work order, and obtain the real-time feature vector of the current work order. S5: Calculate the cosine similarity between the real-time feature vector and the historical work order feature vector, and select the N historical reference work orders with the highest similarity. S6: Input the real-time feature vector of the current work order into the optimized pressure control model to obtain the initial pressure control parameters. Combine the optimal pressure control parameters of historical reference work orders to correct the initial parameters. Based on the deviation between the real-time quality detection data and the preset quality standard, dynamically fine-tune the corrected parameters to obtain the final pressure control parameters. S7: Convert the final pressure control parameters into control commands and send them to the actuator to drive the actuator to adjust the printing pressure.

[0021] Preferably, the comprehensive data of the historical work orders includes basic work order information, material characteristic data, process parameter data, equipment operation data, printing quality inspection data, and pressure control parameter data. The basic work order information includes work order number, production time, product name, production batch, and quality requirement level. Material characteristic data includes paper weight, thickness, smoothness, stiffness, and moisture content; ink viscosity, fineness, drying speed, and color type. Process parameter data includes printing speed, printing plate type, anilox roller line count, ink supply, and drying temperature. Equipment operation data includes printing press cylinder speed, cylinder temperature, bearing temperature, motor current, equipment operating time, and the time of the most recent maintenance. Printing quality inspection data includes ink uniformity, image clarity, registration accuracy, dot gain, and printed defect rate. Pressure control parameter data includes the pressure value between the printing plate cylinder and the blanket cylinder, the pressure value between the blanket cylinder and the impression cylinder, the pressure adjustment step size, and the pressure stabilization time.

[0022] Preferably, the feature extraction process in S2 specifically includes: The paper's basis weight, thickness, smoothness, stiffness, and moisture content are normalized and then used as paper characteristics. After normalizing the viscosity, fineness, and drying speed of the ink, it is used as the ink characteristics. The paper characteristics and ink characteristics are combined to form a material characteristic sub-vector. The printing speed, anilox roller line count, ink supply, and drying temperature are normalized and combined with the coding value of the printing plate type to form a process feature sub-vector. The drum speed, drum temperature, bearing temperature, and motor current are normalized, and the equipment running time and the time of the most recent maintenance are converted into equipment aging coefficients and combined to form equipment status feature sub-vectors. The quality requirement levels are converted into numerical codes, and combined with preset thresholds for ink uniformity, image clarity, and registration accuracy to form a quality requirement feature sub-vector. By concatenating the material feature sub-vectors, process feature sub-vectors, equipment status feature sub-vectors, and quality requirement feature sub-vectors, the feature vectors of historical work orders are obtained.

[0023] Preferably, the model training process in S3 specifically includes: The standardized historical work order dataset was divided into a training set and a validation set in a 7:3 ratio. Set initial values ​​for hyperparameters of the gradient boosting tree algorithm, such as the number of decision trees, tree depth, learning rate, and minimum number of sample splits. The feature vectors of historical work orders in the training set are input into the gradient boosting tree model, and the model is trained with the corresponding optimal pressure control parameter as the target value. The optimal pressure control parameter is the pressure control parameter when all printing quality inspection data in the corresponding historical work order meet the preset quality standards and the scrap rate of printed products is the lowest. The initial model obtained from training was validated using a validation set, and the mean square error between the predicted pressure control parameters and the actual optimal pressure control parameters was calculated. The hyperparameters were iteratively optimized using a grid search method until the mean square error reached a preset threshold, thus obtaining the initial pressure control model. At preset time intervals, newly added standardized historical work order data are added to the training set to retrain the initial pressure control model. The model is then continuously optimized to obtain an optimized pressure control model, thereby improving the model's generalization ability and prediction accuracy.

[0024] Preferably, the similarity filtering process in S5 specifically includes: The cosine similarity algorithm is used to calculate the similarity between the real-time feature vector of the current work order and the feature vector of each historical work order. The specific calculation is as follows: the cosine similarity is equal to the dot product of the current feature vector and the historical feature vector divided by the product of the magnitude of the current feature vector and the magnitude of the historical feature vector. The calculated similarity scores are sorted in descending order, and the top N historical work orders with the highest similarity scores are selected as historical reference work orders. The value of N ranges from 3 to 5 and can be adjusted according to the size of the historical work order data.

[0025] Preferably, the parameter calculation and fine-tuning process in S6 specifically includes: The real-time feature vector of the current work order is input into the optimized pressure control model to obtain the initial pressure control parameters predicted by the model. Calculate the weighted average of the optimal pressure control parameters for N historical reference work orders, with the weights being the similarity between each historical reference work order and the current work order; The initial pressure control parameters and the weighted average value are combined in a ratio of 60%-70%:30%-40% to obtain the corrected pressure control parameters; Calculate the deviation between real-time quality monitoring data and preset quality standards. If the deviation is less than or equal to the preset allowable deviation, keep the corrected parameters unchanged. If the deviation is greater than the preset allowable deviation, the corrected pressure control parameters are dynamically fine-tuned according to the direction and magnitude of the deviation, combined with the adjustment experience of similar historical deviations, to obtain the final pressure control parameters. The fine-tuning amount is positively correlated with the magnitude of the deviation; the larger the deviation, the larger the fine-tuning amount. At the same time, the adjustment direction and fine-tuning amplitude are determined by referring to the pressure adjustment experience under similar historical deviation conditions. The finalized pressure control parameters are converted into electrical signal commands that the actuator can recognize. These electrical signal commands include the pressure adjustment amount, adjustment speed, and stabilization time.

[0026] The printing pressure adaptive control system based on historical work order big data is applied in the printing pressure adaptive control method based on historical work order big data, including: The historical data acquisition module is used to collect full-dimensional data corresponding to historical printing work orders. The collected raw data is cleaned, deduplicated, missing value imputation, outlier removal and data standardization processing to obtain a standardized historical work order dataset. The feature extraction module is used to extract features from the various dimensions of the standardized historical work order dataset to obtain the feature vector of the historical work orders. The model building module is used to train the model using the gradient boosting tree algorithm to obtain an initial pressure regulation model, and to continuously iterate and optimize the initial pressure regulation model to obtain an optimized pressure regulation model. The real-time data acquisition module is used to collect real-time operating data of the current printing work order. The real-time operating data includes real-time material characteristic data, real-time process parameter data, real-time equipment operation data, and real-time quality inspection data. The module also performs feature extraction on the collected real-time operating data to obtain the real-time feature vector of the current work order. The work condition matching module is used to perform similarity matching between the real-time feature vector of the current work order and the feature vector of historical work orders, and to filter out several historical reference work orders with the highest similarity. The pressure control module is used to obtain initial pressure control parameters, correct the initial pressure control parameters to obtain final pressure control parameters, dynamically fine-tune the final pressure control parameters, and generate pressure control commands. The instruction execution module is used to receive pressure control instructions and drive the pressure adjustment components of the printing equipment to perform precise actions, thereby achieving adaptive adjustment of printing pressure.

[0027] Preferably, the pressure regulation module specifically includes: The initial pressure control parameter acquisition unit is used to input the real-time feature vector of the current work order into the optimized pressure control model to obtain the initial pressure control parameters predicted by the model. The initial parameter correction unit is used to calculate the weighted average of the optimal pressure control parameters of N historical reference work orders, and to fuse the initial pressure control parameters with the weighted average to obtain the corrected pressure control parameters. The dynamic fine-tuning unit is used to calculate the deviation between real-time quality monitoring data and preset quality standards. Based on the direction and magnitude of the deviation, and combined with the adjustment experience of similar historical deviations, the corrected pressure control parameters are dynamically fine-tuned to obtain the final pressure control parameters. The instruction generation unit is used to dynamically fine-tune the final pressure control parameters based on the deviation between real-time quality monitoring data and preset quality standards, and generate pressure control instructions.

[0028] Preferably, the final determined pressure control parameters are converted into electrical signal commands that the actuator can recognize, including: Obtain the safety operating standards for the actuator; these standards include: maximum permissible pressure value, minimum permissible pressure value, and maximum permissible adjustment speed. A safety analysis is performed on the final pressure control parameters according to the safe operation standards to determine whether the final pressure control parameters meet the safe operation standards and to obtain the safety analysis results. Based on the safety analysis results, the final pressure control parameters are verified for compliance. When the safety analysis results are found to be inconsistent with the safe operation standards, the final pressure control parameters are verified in conjunction with the safe operation standards to determine the compliance data of the final pressure control parameters and obtain the final pressure control verification parameters. By using the physical quantity mapping relationship, the final pressure control parameter or the final pressure control verification parameter is converted into the corresponding physical control quantity of the actuator to obtain the target physical quantity. The interface type of the actuator is combined with the target physical quantity to encapsulate the electrical signal instruction, thus obtaining the first target electrical signal instruction; While the executing agency responds to the first target electrical signal command, it performs stability monitoring and statistics, and combines the stability time for maintenance control.

[0029] In the above technical solution, when the safety analysis result meets the safety operation standards, the final pressure control parameters are directly converted into the corresponding physical control quantities of the actuator through the physical quantity mapping relationship.

[0030] In the above technical solution, the physical quantity mapping relationship refers to a pre-determined pressure-displacement conversion function and a pressure-current lookup table. When converting the final pressure control parameter or final pressure control verification parameter into the corresponding physical control quantity of the actuator through the physical quantity mapping relationship, the conversion is performed according to the control module. When the control mode is position control mode, the target control displacement is determined by the pressure adjustment amount in the final pressure control parameter or final pressure control verification parameter through the pressure-displacement conversion function. Analysis and calculation of target control displacement ,in, This refers to the pressure adjustment amount in the final pressure control parameter or the final pressure control verification parameter. This is the pressure-displacement conversion function. Then, based on the target control displacement, it passes through... The analysis and conversion yield the physical control quantity (number of pulses) of the servo motor actuator. ,in, To control the displacement of the target, For the lead screw, For encoder resolution. When the control mode is force control mode, the target control quantity is determined by the pressure adjustment amount in the final pressure control parameter or the final pressure control verification parameter of the pressure-current lookup table, i.e., by... Analyze and calculate the target control power. ,in, This refers to the pressure adjustment amount in the final pressure control parameter or the final pressure control verification parameter. This is a pressure-current correlation function. Then, based on the target control quantity, the power is controlled through... Analysis and conversion yield the target control power. ,in, This represents the zero-point offset of the current signal, using a live zero-point design; here, its value is set to 4. For the maximum range of the actuator, The effective signal span is determined by the difference between the maximum and minimum current values.

[0031] In the above technical solution, when encapsulating electrical signal commands by combining the interface type of the actuator with the target physical quantity, the interface type of the actuator is identified. The interface types include analog interfaces, pulse interfaces, and industrial bus interfaces. Different signal encapsulation processes are performed according to the interface type. When the interface type is an analog interface, the target physical quantity is converted into a corresponding analog voltage or current signal through a D / A converter. When the interface type is a pulse interface, a pulse sequence command packet is generated based on the target physical quantity. The pulse sequence command packet includes target position fields, velocity fields, acceleration fields, deceleration fields, etc. When the interface type is an industrial bus interface, data frames are encapsulated for the target physical quantity according to a preset industrial bus protocol format. The data frames include control words, target position, contour velocity, contour acceleration, etc.

[0032] In the above technical solution, when the actuator responds to the command of the first target electrical signal while performing stability monitoring and statistics, and combining the stability time for maintenance and control, it includes: While the actuator responds to the first target electrical signal command, it monitors the actual pressure value and determines whether a steady state has been reached based on the actual pressure value. When a steady state is reached, timing is started according to the stabilization time. During the timing process, the deviation between the actual pressure value and the target pressure value is analyzed and calculated in real time. When the deviation exceeds 50% but does not exceed 100% of the analysis and judgment threshold, [the system]... PID fine-tuning is performed, where, This is the adjustment amount for PID fine-tuning. For the pre-tuned PID parameters, This is the current deviation data, and this is the cumulative deviation data. The current rate of change of deviation is used as the time until the actual pressure value meets the analysis and judgment threshold and reaches a stable time, thus completing the maintenance control.

[0033] The aforementioned technical solution ensures that the electrical signal commands identifiable by the actuator are compatible with its operating standards by acquiring the safety operating standards. This prevents actuator malfunctions or safety hazards caused by non-compliance with these standards, thus guaranteeing the normal operation of the actuator. Furthermore, by encapsulating electrical signal commands based on the actuator's interface type and target physical quantities, and employing refined processing for different interfaces, the solution ensures that electrical signal commands can be accurately recognized and executed by various types of actuators. This allows for effective electrical signal command responses in both traditional analog signal control environments and modern industrial bus communication environments, improving adaptability and versatility, and ensuring that electrical signal commands are effectively received and responded to by the actuator. Moreover, maintenance and control ensure that the actuator remains in a safe and stable operating state during pressure regulation, preventing safety hazards caused by excessive pressure fluctuations. This, in turn, guarantees printing stability, precise pressure control, improved quality and efficiency, reduced costs, decreased material waste and reliance on manual adjustments, and enhanced print quality stability and consistency.

[0034] Preferably, the real-time feature vector of the current work order is matched with the feature vectors of historical work orders based on similarity, including: Dimensional analysis and alignment are performed on the real-time feature vector of the current work order and the feature vector of the historical work orders to obtain the real-time feature processing vector of the current work order and the feature processing vector of the historical work orders. Timestamp analysis is performed on the feature processing vectors of historical work orders, and attenuation factor analysis is calculated using the following formula: ; In the above formula, For historical work orders The attenuation factor, The attenuation coefficient is... The time of the current work order. For historical work orders Time, The decay period; The similarity calculation formula is used to calculate the similarity between the real-time feature processing vector of the current work order and the feature processing vector of historical work orders, and a preliminary similarity calculation value is obtained. The initial similarity calculation value is revised by combining it with the corresponding decay factor to obtain the final similarity between the current work order and historical work orders.

[0035] In the above technical solution, the attenuation coefficient The value range is (0,1), usually 0.5, but can be adjusted appropriately according to actual needs.

[0036] In the above technical solution, the attenuation period The value range is 90-180 days.

[0037] In the above technical solution, when revising the preliminary similarity calculation value in combination with the corresponding attenuation factor, the final similarity calculation value is obtained by multiplying the attenuation factor corresponding to the current work order by the preliminary similarity calculation value, that is, the final similarity between the current work order and the historical work orders is obtained.

[0038] The above technical solution ensures that the feature vectors are consistent by performing dimensional analysis and alignment between the real-time feature vector of the current work order and the feature vectors of historical work orders. This provides a guarantee for subsequent similarity analysis calculations and avoids similarity errors caused by different feature vector dimensions. Moreover, by introducing a decay factor, the impact of time on work order similarity is considered, improving the sensitivity to time. As time goes by, the relevance and reference value of historical work orders gradually decrease, making the similarity calculation better adapt to the ever-changing business environment and more in line with the actual situation. This ensures the accuracy and timeliness of the similarity matching results, and allows the selection of the most similar historical reference work orders to provide more reliable and valuable data support for optimization and adjustment.

[0039] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0040] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

Claims

1. A method for adaptive control of printing pressure based on historical work order big data, characterized in that, Includes the following steps: S1: Collect all dimensions of historical printing work orders, and obtain a standardized historical work order dataset after preprocessing; S2: Extract features from the standardized historical work order dataset to obtain the feature vectors of the historical work orders; S3: Using the feature vector of historical work orders as input and the corresponding optimal pressure control parameters as output, the initial model is continuously iterated and optimized to obtain the optimized pressure control model. S4: Collect real-time working condition data of the current printing work order, extract features from the real-time working condition data of the current work order, and obtain the real-time feature vector of the current work order. S5: Calculate the cosine similarity between the real-time feature vector and the historical work order feature vector, and select the N historical reference work orders with the highest similarity. S6: Utilize the optimized pressure control model to output the initial pressure control parameters, correct the initial parameters, and dynamically fine-tune the corrected parameters to obtain the final pressure control parameters. S7: Convert the final pressure control parameters into control commands and send them to the actuator to drive the actuator to adjust the printing pressure.

2. The adaptive printing pressure control method based on historical work order big data according to claim 1, characterized in that, The comprehensive data of the historical work orders includes basic work order information, material characteristic data, process parameter data, equipment operation data, printing quality inspection data, and pressure control parameter data.

3. The adaptive printing pressure control method based on historical work order big data according to claim 1, characterized in that, The feature extraction process in S2 specifically includes: The paper's basis weight, thickness, smoothness, stiffness, and moisture content are normalized and then used as paper characteristics. After normalizing the viscosity, fineness, and drying speed of the ink, it is used as the ink characteristics. The paper characteristics and ink characteristics are combined to form a material characteristic sub-vector. The printing speed, anilox roller line count, ink supply, and drying temperature are normalized and combined with the coding value of the printing plate type to form a process feature sub-vector. The drum speed, drum temperature, bearing temperature, and motor current are normalized, and the equipment running time and the time of the most recent maintenance are converted into equipment aging coefficients and combined to form equipment status feature sub-vectors. The quality requirement levels are converted into numerical codes, and combined with preset thresholds for ink uniformity, image clarity, and registration accuracy to form a quality requirement feature sub-vector. By concatenating the material feature sub-vectors, process feature sub-vectors, equipment status feature sub-vectors, and quality requirement feature sub-vectors, the feature vectors of historical work orders are obtained.

4. The adaptive printing pressure control method based on historical work order big data according to claim 1, characterized in that, The model training process in S3 specifically includes: The standardized historical work order dataset was divided into a training set and a validation set in a 7:3 ratio. Set initial values ​​for hyperparameters of the gradient boosting tree algorithm, such as the number of decision trees, tree depth, learning rate, and minimum number of sample splits. The historical work order feature vectors of the training set are input into the gradient boosting tree model, and the model is trained with the corresponding optimal pressure control parameters as the target values. The initial model obtained from training was validated using a validation set, and the mean square error between the predicted pressure control parameters and the actual optimal pressure control parameters was calculated. The hyperparameters were iteratively optimized using a grid search method until the mean square error reached a preset threshold, thus obtaining the initial pressure control model. At preset time intervals, newly added standardized historical work order data are added to the training set to retrain the initial pressure control model and continuously optimize the model to obtain an optimized pressure control model.

5. The printing pressure adaptive control method based on historical work order big data according to claim 1, characterized in that, The similarity filtering process in S5 specifically includes: The cosine similarity algorithm is used to calculate the similarity between the real-time feature vector of the current work order and the feature vector of each historical work order; The calculated similarity scores are sorted in descending order, and the top N most similar historical work orders are selected as historical reference work orders.

6. The printing pressure adaptive control method based on historical work order big data according to claim 1, characterized in that, The parameter calculation and fine-tuning process in S6 specifically includes: The real-time feature vector of the current work order is input into the optimized pressure control model to obtain the initial pressure control parameters predicted by the model. Calculate the weighted average of the optimal pressure control parameters for N historical reference work orders, with the weights being the similarity between each historical reference work order and the current work order; The initial pressure control parameters are combined with the weighted average value to obtain the corrected pressure control parameters; Calculate the deviation between real-time quality monitoring data and preset quality standards. If the deviation is less than or equal to the preset allowable deviation, keep the corrected parameters unchanged. If the deviation is greater than the preset allowable deviation, the corrected pressure control parameters are dynamically fine-tuned based on the direction and magnitude of the deviation and the adjustment experience of similar historical deviations, to obtain the final pressure control parameters. The finalized pressure control parameters are converted into electrical signal commands that the actuator can recognize. These electrical signal commands include the pressure adjustment amount, adjustment speed, and stabilization time.

7. A printing pressure adaptive control system based on historical work order big data, applied in the printing pressure adaptive control method based on historical work order big data as described in claim 6, characterized in that, include: The historical data acquisition module is used to collect full-dimensional data corresponding to historical printing work orders, and to preprocess the collected raw data to obtain a standardized historical work order dataset. The feature extraction module is used to extract features from the various dimensions of the standardized historical work order dataset to obtain the feature vector of the historical work orders. The model building module is used to train the model using the gradient boosting tree algorithm to obtain an initial pressure regulation model, and to continuously iterate and optimize the initial pressure regulation model to obtain an optimized pressure regulation model. The real-time data acquisition module is used to collect real-time working condition data of the current printing work order and extract features from the collected real-time working condition data to obtain the real-time feature vector of the current work order. The work condition matching module is used to perform similarity matching between the real-time feature vector of the current work order and the feature vector of historical work orders, and to filter out several historical reference work orders with the highest similarity. The pressure control module is used to obtain initial pressure control parameters, correct the initial pressure control parameters to obtain final pressure control parameters, dynamically fine-tune the final pressure control parameters, and generate pressure control commands. The instruction execution module is used to receive pressure control instructions and drive the pressure adjustment components of the printing equipment to perform precise actions, thereby achieving adaptive adjustment of printing pressure.

8. The printing pressure adaptive control system based on historical work order big data according to claim 7, characterized in that, The pressure regulation module specifically includes: The initial pressure control parameter acquisition unit is used to input the real-time feature vector of the current work order into the optimized pressure control model to obtain the initial pressure control parameters predicted by the model. The initial parameter correction unit is used to calculate the weighted average of the optimal pressure control parameters of N historical reference work orders, and to fuse the initial pressure control parameters with the weighted average to obtain the corrected pressure control parameters. The dynamic fine-tuning unit is used to calculate the deviation between real-time quality monitoring data and preset quality standards. Based on the direction and magnitude of the deviation, and combined with the adjustment experience of similar historical deviations, the corrected pressure control parameters are dynamically fine-tuned to obtain the final pressure control parameters. The instruction generation unit is used to dynamically fine-tune the final pressure control parameters based on the deviation between real-time quality monitoring data and preset quality standards, and generate pressure control instructions.

9. The printing pressure adaptive control system based on historical work order big data according to claim 6, characterized in that, The finalized pressure control parameters are converted into electrical signal commands that the actuator can recognize, including: Obtain the safety operating standards of the implementing agency; A safety analysis is performed on the final pressure control parameters according to the safe operation standards to determine whether the final pressure control parameters meet the safe operation standards and to obtain the safety analysis results. Based on the safety analysis results, the final pressure control parameters are verified for compliance. When the safety analysis results do not meet the safety operation standards, the final pressure control parameters are verified in conjunction with the safety operation standards to determine the compliance data of the final pressure control parameters and obtain the final pressure control verification parameters. By using the physical quantity mapping relationship, the final pressure control parameter or the final pressure control verification parameter is converted into the corresponding physical control quantity of the actuator to obtain the target physical quantity. The interface type of the actuator is combined with the target physical quantity to encapsulate the electrical signal instruction, thus obtaining the first target electrical signal instruction; While the executing agency responds to the first target electrical signal command, it performs stability monitoring and statistics, and combines the stability time for maintenance control.

10. The printing pressure adaptive control system based on historical work order big data according to claim 7, characterized in that, Perform similarity matching between the real-time feature vector of the current work order and the feature vectors of historical work orders, including: Dimensional analysis and alignment are performed on the real-time feature vector of the current work order and the feature vector of the historical work orders to obtain the real-time feature processing vector of the current work order and the feature processing vector of the historical work orders. Timestamp analysis is performed on the feature processing vectors of historical work orders, and attenuation factor analysis is calculated using the following formula: ; In the above formula, For historical work orders The attenuation factor, The attenuation coefficient is... The time of the current work order. For historical work orders Time, The decay period; The similarity calculation formula is used to calculate the similarity between the real-time feature processing vector of the current work order and the feature processing vector of historical work orders, and a preliminary similarity calculation value is obtained. The initial similarity calculation value is revised by combining it with the corresponding decay factor to obtain the final similarity between the current work order and historical work orders.

Images