Online production control management system for printing equipment

Abstract

The invention discloses an online production control management system for printing equipment, and relates to the technical field of automatic control. Collecting an equipment operation parameter set during the operation of the target printing equipment through a preset time period; preprocessing the equipment operation parameter set to obtain an effective operation parameter set; performing quality detection on the target product to obtain a quality detection result; whether the quality detection result of the target product meets production requirements or not is judged, and if the quality detection result of the target product does not meet the production requirements, the quality detection result, the production requirements and the effective operation parameter set are substituted into a preset model to obtain updated operation parameters of the target printing equipment; and controlling and managing the target printing equipment according to the updated operation parameters. After the operation parameter set is preprocessed, in combination with the quality detection result and the production requirement, the updated operation parameters are calculated by using the preset model, so that the printing equipment is controlled to guarantee the stable product quality, and the online production efficiency of the printing equipment is improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of automatic control, and particularly relates to a printing equipment online production control management system. BACKGROUND

[0002] In the field of printing production, stable operation of printing equipment and product quality control are key links to guarantee production efficiency and economic benefits. Traditional printing equipment management relies on manual experience for parameter adjustment and quality control. An operator judges the quality of printed matter according to visual observation or simple detection tools, and manually adjusts operating parameters such as printing speed, pressure and drying temperature accordingly. This method is not only inefficient, but also highly dependent on the professional experience of the operator, making it difficult to achieve precise and consistent quality control of printed matter. At the same time, manual adjustment often has a lag, and cannot respond to dynamic changes in the operating state of the equipment in real time, which can easily lead to the production of a large number of defective products, causing waste of raw materials and an increase in production costs.

[0003] Patent CN112356579A discloses a printing machine control method and device. A control board acquires overprint error information. The control board acquires overprint error information, wherein the overprint error information is used to indicate the corresponding relationship between the running speed of the mobile platform and the overprint error. The overprint error is the error between the actual position of the color mark sensor and the target position of the color mark sensor. The control board determines the running speed of the mobile platform. The control board adjusts the color mark sensor according to the overprint error information and the running speed of the mobile platform, so that the actual position of the color mark sensor and the target position of the color mark sensor are aligned.

[0004] However, the quality detection of the prior art only relies on the color mark sensor, which is single in means and cannot achieve precise and comprehensive quality evaluation. This bottleneck leads to a lag in equipment parameter adjustment, making it difficult to quickly respond to flexible and variable production tasks, ultimately resulting in low online production efficiency of the printing equipment. SUMMARY

[0005] The purpose of the present application is to solve the problem that the quality detection of the prior art only relies on the color mark sensor, which is single in means and cannot achieve precise and comprehensive quality evaluation, ultimately resulting in low online production efficiency of the printing equipment. A printing equipment online production control management system is proposed.

[0006] The present application proposes a printing equipment online production control management system, which comprises:

[0007] A device operating parameter acquisition module is configured to acquire a set of device operating parameters of a target printing equipment during operation through a preset time period. The device operating parameters include printing speed, pressure and drying temperature.

[0008] A data preprocessing module is configured to preprocess the set of device operation parameters to obtain an effective set of operation parameters.

[0009] A product quality detection module is configured to detect the quality of a target product to obtain a quality detection result. The target product is an object product printed under the device operation parameters. The quality detection result includes excellent products, qualified products, and defective products.

[0010] An updated operation parameter determination module is configured to determine whether the quality detection result of the target product meets production requirements. If the quality detection result does not meet the production requirements, the quality detection result, the production requirements, and the effective set of operation parameters are substituted into a preset model to obtain updated operation parameters of the target printing device. The production requirements are that the quality detection result is an excellent product or a qualified product.

[0011] A printing device control management module is configured to control and manage the target printing device according to the updated operation parameters.

[0012] Optionally, the data preprocessing module includes:

[0013] A mean and variance calculation module is configured to calculate the mean and variance of each type of data in the set of device operation parameters to obtain mean operation parameters and variance operation parameters.

[0014] A dynamic abnormal data marking module is configured to mark the device operation parameters at a time point as dynamic abnormal data if the device operation parameters at the time point exceed three times the variance operation parameters.

[0015] A behavior operation recognition module is configured to recognize device operation logs, update the dynamic abnormal data according to the device operation logs and the mean operation parameters to obtain dynamic normal data.

[0016] A dynamic abnormal data replacement module is configured to replace the dynamic abnormal data in each type of data in the set of device operation parameters with the dynamic normal data to obtain the effective set of operation parameters.

[0017] Optionally, the behavior operation recognition module includes:

[0018] A behavior operation determination module is configured to determine whether there is a behavior operation according to the device operation logs.

[0019] A first dynamic normal data generation module is configured to update the dynamic abnormal data according to the mean operation parameters to obtain dynamic normal data if there is a behavior operation.

[0020] A second dynamic normal data generation module is configured to replace the dynamic abnormal data according to the mean operation parameters to obtain dynamic normal data if there is no behavior operation.

[0021] Optionally, the product quality detection module comprises:

[0022] a printing feature extraction module, configured to collect image data of the target product, and perform feature extraction on the image data to obtain a printing feature;

[0023] an enhanced printing feature extraction module, configured to perform feature enhancement on the printing feature to obtain an enhanced printing feature;

[0024] a quality detection result determination module, configured to substitute the enhanced printing feature into a detection network to obtain a quality detection result.

[0025] Optionally, the printing feature extraction module comprises:

[0026] a multi-channel feature extraction module, configured to substitute the image data into a CBS module to obtain a first feature, and perform channel splitting on the first feature to obtain a first channel feature, a second channel feature, a third channel feature and a fourth channel feature; the first channel feature has a channel number that is half of the first feature; the second channel feature has a channel number that is half of the first channel feature; the third channel feature has a channel number that is half of the second channel feature; and the fourth channel feature has a channel number that is the same as that of the third channel feature;

[0027] a spliced channel feature extraction module, configured to splice the first channel feature, the second channel feature and the third channel feature with the fourth channel feature after substituting them into a channel feature extraction module to obtain a spliced channel feature;

[0028] a printing feature generation module, configured to substitute the spliced channel feature into a CBS module to obtain a printing feature;

[0029] the working principle of the channel feature extraction module is as follows:

[0030] after performing 3x3 convolution on an input feature, substitute the input feature into a progressive diversity function module to obtain an enhanced input feature;

[0031] after performing element addition on the input feature and the enhanced input feature, perform 1x1 convolution to obtain an output channel feature.

[0032] Optionally, the enhanced printing feature extraction module comprises:

[0033] a nonlinear enhanced feature extraction module, configured to substitute the printing feature into a GELU activation function after performing 1x1 convolution on the printing feature to obtain a nonlinear enhanced feature;

[0034] a semantic perception feature generation module, configured to substitute the nonlinear enhanced feature into a region interaction auxiliary reconstruction unit and a global semantic perception unit in sequence to obtain a semantic perception feature;

[0035] An enhanced printing feature generation module is configured to multiply the printing feature and the semantic perception feature and perform 1x1 convolution to obtain an enhanced printing feature.

[0036] Optionally, the quality detection result determination module comprises:

[0037] A height width feature extraction module is configured to perform height feature extraction and width feature extraction on the enhanced printing feature to obtain a height feature and a width feature.

[0038] A first activation feature generation module is configured to multiply the height feature and the width feature, perform 1x1 convolution, and obtain a first activation feature through a Sigmoid function.

[0039] A second activation feature generation module is configured to fuse the first activation feature and the enhanced printing feature, perform 1x1 convolution, and obtain a second activation feature through a Softmax function.

[0040] A target printing feature generation module is configured to fuse the second activation feature and the enhanced printing feature, perform 1x1 convolution, and obtain a target printing feature.

[0041] A quality detection result generation module is configured to search a preset database according to the target printing feature to obtain a quality detection result. The preset database stores printing features and quality grades of different printing templates under different device running parameters.

[0042] Optionally, the update running parameter determination module comprises:

[0043] An original running parameter set generation module is configured to search the preset database according to the target printing feature corresponding to the quality detection result, search device running parameters corresponding to similar printing features according to a preset similarity, and obtain an original running parameter set.

[0044] A target population generation module is configured to generate an initial population according to the original running parameter set, perform iterative optimization on the initial population through a preset genetic algorithm, and obtain a target population.

[0045] A fitness set generation module is configured to perform fitness evaluation on each chromosome in the target population through a target model to obtain a fitness set.

[0046] An updated target population generation module is configured to remove chromosomes with fitness values less than a preset threshold from the target population to obtain an updated target population.

[0047] The quality detection result generation module is used to substitute the updated target population as the initial population into the target population generation module, and repeat the target population generation module to the updated target population generation module until the preset termination condition is met to obtain the updated target population; the updated target population is the updated running parameters.

[0048] Optionally, the fitness set generation module includes:

[0049] The historical printing data acquisition module is used to acquire historical printing data; the historical printing data includes historical printing operation parameters and corresponding historical printing characteristics;

[0050] The model update module is used to input the historical printing data into a preset model to obtain model parameters, and update the preset model according to the model parameters to obtain the target model.

[0051] Optionally, the system may also include an alarm module:

[0052] The alarm module is used to issue an alarm and send relevant information to maintenance personnel if the quality inspection result of the target product after control and management is a defective product.

[0053] The beneficial effects of this invention are:

[0054] This invention proposes an online production control and management system for printing equipment. The system collects the operating parameters of the printing equipment at preset intervals via an equipment operating parameter acquisition module. After data preprocessing, and combined with the quality inspection results of the printed products by the product quality inspection module, an updated operating parameter determination module calculates updated operating parameters using a preset model when products fail to meet production requirements. Finally, the printing equipment control and management module controls the equipment accordingly. This approach allows for precise monitoring of equipment operation, timely detection and resolution of quality issues, and dynamic parameter adjustment to ensure stable product quality, thereby improving the online production efficiency of the printing equipment. Attached Figure Description

[0055] The invention will now be further described with reference to the accompanying drawings.

[0056] Figure 1 A framework diagram of an online production control and management system for printing equipment provided in an embodiment of the present invention;

[0057] Figure 2 This is a flowchart illustrating a printing feature extraction module provided in an embodiment of the present invention. Detailed Implementation

[0058] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0059] 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.

[0060] This invention provides an online production control and management system for printing equipment. See also... Figure 1 , Figure 1 This is a framework diagram of an online production control and management system for printing equipment provided in an embodiment of the present invention. The method includes the following steps:

[0061] The equipment operation parameter acquisition module is used to acquire a set of equipment operation parameters of the target printing equipment during operation at a preset time period; the equipment operation parameters include printing speed, pressure and drying temperature;

[0062] The data preprocessing module is used to preprocess the set of operating parameters of the equipment to obtain a valid set of operating parameters;

[0063] The product quality inspection module is used to perform quality inspection on the target product and obtain the quality inspection results; the target product is the product to be printed under the operating parameters of the equipment; the quality inspection results include excellent products, qualified products, and defective products.

[0064] The update operation parameter determination module is used to determine whether the quality inspection result of the target product meets the production requirements. If it does not meet the production requirements, the quality inspection result, the production requirements, and the set of valid operation parameters are substituted into a preset model to obtain the updated operation parameters of the target printing equipment. The production requirements are that the quality inspection result is an excellent product or a qualified product.

[0065] The printing equipment control and management module is used to control and manage the target printing equipment according to the updated operating parameters.

[0066] Based on the online production control and management system for printing equipment provided in this embodiment of the invention, the system collects the operating parameters of the printing equipment at a preset cycle through the equipment operating parameter acquisition module. After data preprocessing by the data preprocessing module, and combined with the quality inspection results of the printed products by the product quality inspection module, the updated operating parameter determination module can calculate updated operating parameters using a preset model when the products do not meet production requirements. Finally, the printing equipment control and management module controls the equipment accordingly. This approach enables precise monitoring of the equipment's operating status, timely detection and resolution of quality problems, and ensures stable product quality through dynamic parameter adjustment, thereby improving the online production efficiency of the printing equipment.

[0067] In one implementation, the equipment operation parameter acquisition module collects key parameters such as printing speed, pressure, and drying temperature according to a preset cycle, which can comprehensively and accurately record the equipment operation status and provide a rich and accurate data foundation for subsequent analysis and optimization. The preset cycle is determined by the technicians, and multiple sampling points are generated within the preset cycle. The spacing between each sampling point is the same and is determined by the technicians.

[0068] In one implementation, the product quality inspection module can promptly inspect the target product to determine whether it is an excellent, qualified, or defective product, allowing producers to quickly understand the product quality status and facilitate timely problem detection. When the product is inspected as excellent or qualified, no action is taken. The update operating parameter determination module compares the quality inspection results with production requirements; if they are not met, it calculates updated operating parameters using a preset model. This dynamic adjustment mechanism can quickly respond to quality issues, ensuring that production always moves towards meeting quality requirements. Because printing equipment experiences wear and tear during historical use, this mechanism avoids direct control based on historical data, improving the accuracy of printing equipment control.

[0069] In one embodiment, the data preprocessing module includes:

[0070] The mean and variance calculation module is used to calculate the mean and variance of each type of data in the equipment operating parameter set to obtain the mean operating parameters and variance operating parameters.

[0071] The dynamic abnormal data marking module is used to record the equipment operating parameters at a certain moment as dynamic abnormal data if the operating parameters at that moment exceed three times the variance of the operating parameters.

[0072] The behavior operation recognition module is used to identify the equipment operation logs and update the dynamic abnormal data based on the equipment operation logs and the average operation parameters to obtain dynamic normal data.

[0073] The dynamic abnormal data replacement module is used to replace the dynamic abnormal data in each type of data in the device operating parameter set with dynamic normal data to obtain a valid operating parameter set.

[0074] In one implementation, the mean and variance calculation module calculates the mean and variance of the operating parameters for each type of equipment to obtain the mean operating parameters and the variance operating parameters. The mean reflects the central tendency of the data, and the variance reflects the dispersion of the data. These two indicators provide key reference benchmarks for subsequent analysis and help to more accurately grasp the normal range of equipment operation.

[0075] In one implementation, the dynamic anomaly data marking module, based on the three-fold variance principle, can quickly and accurately identify dynamic anomaly data in the equipment operating parameters; the anomaly detection method based on statistical methods avoids the subjectivity and arbitrariness of human judgment, and improves the accuracy and reliability of anomaly data identification.

[0076] In one implementation, the equipment operation log records the operation information of the printing equipment during online production. By analyzing the equipment operation log and combining it with the average operating parameters, the dynamic abnormal data is updated to obtain dynamic normal data. This fully utilizes the historical information of equipment operation and the parameter characteristics under normal conditions, making the corrected data more consistent with the actual operating conditions of the equipment and avoiding information loss caused by simply removing abnormal data.

[0077] In one embodiment, the behavior recognition module includes:

[0078] The behavior operation judgment module is used to determine whether a behavior operation exists based on the device operation log;

[0079] The first dynamic normal data generation module is used to update the dynamic abnormal data according to the average running parameters to obtain dynamic normal data if there is a behavioral operation.

[0080] The second dynamic normal data generation module is used to replace the dynamic abnormal data with dynamic normal data based on the mean running parameters if no action operation exists.

[0081] In one implementation, the first dynamic normal data generation module updates the dynamic abnormal data based on the average operating parameters. Since the behavioral operation may be a purposeful equipment adjustment, the update operation will combine the average parameters of normal operation, so that the corrected data not only takes into account the impact of the behavioral operation, but also conforms to the characteristics of normal equipment operation, thereby obtaining dynamic normal data that is more in line with the actual situation. Specifically, updating the dynamic abnormal data based on the average operating parameters to obtain dynamic normal data involves averaging the average operating parameters and the dynamic abnormal data as the dynamic normal data.

[0082] In one implementation, the second dynamic normal data generation module directly replaces the dynamic abnormal data based on the average operating parameters. Under normal operating conditions, the normal operating parameters of the equipment should fluctuate around the average. Replacing the abnormal data with the average operating parameters can effectively eliminate abnormal values, making the data closer to the real state and ensuring the accuracy and reliability of the data.

[0083] In one embodiment, the product quality inspection module includes:

[0084] The printing feature extraction module is used to collect image data of the target product and extract printing features from the image data.

[0085] The enhanced printing feature extraction module is used to enhance printing features to obtain enhanced printing features;

[0086] The quality inspection result determination module is used to input enhanced printing features into the inspection network to obtain the quality inspection result.

[0087] In one implementation, quality inspection is performed by extracting and enhancing printing features, independent of the shape, size, or material of a specific product. Therefore, it is applicable to various types of printed products, exhibiting broad applicability and versatility. In actual printing production, various complex printing scenarios may be encountered, such as multi-color printing and special effects printing. The printing feature extraction module and the enhanced printing feature extraction module can effectively process image data under these complex conditions, extract representative printing features, and perform reasonable enhancements, enabling the inspection system to accurately perform quality inspection in different printing scenarios.

[0088] In one embodiment, see Figure 2 , Figure 2 A flowchart illustrating a printing feature extraction module provided in an embodiment of the present invention includes:

[0089] The multi-channel feature extraction module is used to input image data into the CBS module to obtain the first feature, and to split the first feature into channels to obtain the first channel feature, the second channel feature, the third channel feature, and the fourth channel feature; the number of channels of the first channel feature is half that of the first feature; the number of channels of the second channel feature is half that of the first channel feature; the number of channels of the third channel feature is half that of the second channel feature; and the number of channels of the fourth channel feature is the same as that of the third channel feature.

[0090] The splicing channel feature extraction module is used to substitute the first channel feature, the second channel feature, and the third channel feature into the channel feature extraction module respectively, and then splice them with the fourth channel feature to obtain the spliced ​​channel feature;

[0091] The printing feature generation module is used to substitute the splicing channel features into the CBS module to obtain printing features;

[0092] The working principle of the channel feature extraction module is as follows:

[0093] The input features are then subjected to a 3×3 convolution and fed into the progressive diversity function module to obtain enhanced input features;

[0094] The output channel features are obtained by summing the input features and the enhanced input features element by element and then performing a 1×1 convolution.

[0095] In one implementation, after substituting the image data into the CBS module to obtain the first feature, channel splitting is performed to obtain four channel features with different numbers of channels. The multi-channel splitting method can capture image features from different scales. The first channel feature has a large number of channels, which can retain more original information and include the macroscopic features of the image. The number of channels in subsequent channel features gradually decreases, which can focus more on the local detail features of the image. Through this multi-scale feature extraction, feature information of the printed image at different levels can be comprehensively obtained, providing rich basis for accurate judgment of printing quality.

[0096] In one implementation, printed products may have various types of defects, such as large areas of uneven color (macro-defects) and tiny missing strokes in text (micro-defects); multi-channel feature extraction can simultaneously capture these defect features at different scales, improving the detection capability for various defects.

[0097] In one implementation, the spliced ​​channel features have stronger expressive power and can better reflect the true characteristics of the printed product; the enhanced feature expressive power helps the subsequent quality inspection module to more accurately distinguish between excellent products, qualified products and defective products, thereby improving the accuracy of quality inspection.

[0098] In one implementation, performing 3×3 convolution on the input features can effectively extract local features of the image. The 3×3 convolution kernel can capture the relationships and patterns between pixels in the local region. The progressive diversity function module can further enhance the convolutional features. By progressively diversifying and optimizing the features, important feature information is highlighted, noise and interference are suppressed, and the enhanced input features can more accurately reflect the key features of the printed image.

[0099] In one implementation, the CBS module consists of a convolutional layer (Conv), a batch normalization layer (BN), and a SiLU activation function layer, which is an existing module. The progressive diversity function module is a technique for feature enhancement through staged processing. Its core principle is based on multi-scale feature fusion and progressive optimization. This module splits the input features into multiple sub-features, gradually enhances key information through step-by-step processing, and finally fuses them into more robust output features, which is an existing technology.

[0100] In one embodiment, the enhanced printing feature extraction module includes:

[0101] The non-linear enhancement feature extraction module is used to perform a 1×1 convolution on the printing features and then substitute them into the GELU activation function to obtain the non-linear enhancement features.

[0102] The semantic-aware feature generation module is used to sequentially substitute the nonlinear enhanced features into the regional interaction-assisted reconstruction unit and the global semantic-aware unit to obtain semantic-aware features.

[0103] The enhanced printing feature generation module is used to multiply the printing features and semantic-aware features and then perform a 1×1 convolution to obtain enhanced printing features.

[0104] In one implementation, the 1×1 convolution can adjust the channel dimension, while the GELU (Gaussian Error Linear Unit) activation function enables the model to capture complex patterns in printing features through a smooth nonlinear transformation; GELU is more stable than ReLU and can prevent the model from being overly sensitive to noise when the amount of data is small, thus improving generalization ability.

[0105] In one implementation, a regional interactive assisted reconstruction unit is used to further refine features based on global semantic information, suppress background noise, and highlight defect details. The regional interactive assisted reconstruction unit first calculates the mean and variance of each channel of the input feature T. It generates gate weights (W) through group normalization and Sigmoid activation and generates binary masks W1 (rich in information) and W2 (less information) based on a threshold (0.5). T is multiplied by W1 and W2 respectively to obtain two parts of features T1 and T2. Then, T1 and T2 are multiplied to obtain spatial refinement features S. S is divided into high-level information (Sup) and low-level details (Slow) according to the channel allocation rate (0.5). Group convolution (GWC, group size g=2) and point convolution (PWC) are applied to Sup in sequence to obtain the first spatial refinement feature S1. 1×1 PWC is applied to Slow to generate the second spatial refinement feature S2. S1 and S2 are fused to obtain the output feature.

[0106] In one implementation, a global semantic awareness unit is used to capture global contextual information of the input features and enhance the semantic correlation between the target and its surrounding environment. The input features are subjected to depthwise separable convolution to generate an initial feature Y1. Y1 is divided into four sub-features along the channel dimension, with each sub-feature having the same number of channels. 1×k and k×1 deep strip-shaped large kernel convolutions are applied to each sub-feature in sequence. The four sub-features are then output and fused through a convolutional layer to generate feature Y2. The average of Y1 and Y2 is used to obtain the output feature.

[0107] In one embodiment, the quality inspection result determination module includes:

[0108] The height and width feature extraction module is used to extract height and width features from the enhanced printing features to obtain height and width features.

[0109] The first activation feature generation module is used to multiply the height feature and the width feature, perform a 1×1 convolution, and obtain the first activation feature through the Sigmoid function;

[0110] The second activation feature generation module is used to fuse the first activation feature and the enhanced printing feature, perform a 1×1 convolution, and obtain the second activation feature through the Softmax function.

[0111] The target printing feature generation module is used to fuse the second activation feature and the enhanced printing feature and then perform a 1×1 convolution to obtain the target printing feature;

[0112] The quality inspection result generation module is used to find the quality inspection result from the preset database based on the target printing characteristics; the preset database stores the printing characteristics and quality grades of different printing templates under different equipment operating parameters.

[0113] In one implementation, the height and width feature extraction modules extract features in the height and width directions of the enhanced printing features, which can deeply explore the key information of the printing features from different dimensions. The directional extraction method is more comprehensive than the single-dimensional feature extraction, which helps to capture the subtle differences and changes in the height and width of the printed matter, such as the thickness of the printed lines and the width-to-height ratio of the pattern, providing a rich feature basis for subsequent accurate quality assessment.

[0114] In one implementation, the first activation feature generation module multiplies the height and width features, performs a 1×1 convolution, and then processes the result using the Sigmoid function. The multiplication operation enables interaction between features in different directions, uncovering their potential relationships. The 1×1 convolution can linearly combine features and reduce channel dimensionality without changing the feature space size, further fusing and refining feature information. The Sigmoid function maps the output value to the (0, 1) interval, serving to filter and emphasize important features, allowing the first activation feature to highlight feature combinations that have a critical impact on printing quality.

[0115] In one implementation, the second activation feature generation module fuses the first activation feature with the enhanced printing feature, performs a 1×1 convolution, and then processes it using the Softmax function. This fusion method combines the first activation feature, which has been preliminarily selected and emphasized, with the original enhanced printing feature, making full use of information at different levels. The 1×1 convolution further optimizes and adjusts the fused feature, while the Softmax function transforms the output into a probability distribution, enabling the second activation feature to more precisely represent the relative importance of different features for printing quality assessment, providing strong support for generating accurate target printing features.

[0116] In one implementation, the target printing feature generation module merges the second activation feature and the enhanced printing feature again and performs a 1×1 convolution, further integrating the feature information after multi-stage processing. Through this multiple fusion and convolution operation, a more accurate and comprehensive target printing feature can be gradually constructed. This feature can better reflect the actual quality of the printed matter and provide a reliable basis for subsequent quality inspection.

[0117] In one embodiment, the update runtime parameter determination module includes:

[0118] The original operating parameter set generation module is used to search a preset database based on the target printing features corresponding to the quality inspection results, and to search for the equipment operating parameters corresponding to similar printing features based on the preset similarity to obtain the original operating parameter set.

[0119] The target population generation module is used to generate an initial population based on the original set of operating parameters, and to obtain the target population by iteratively optimizing the initial population through a preset genetic algorithm.

[0120] The fitness set generation module is used to evaluate the fitness of each chromosome in the target population using the target model to obtain a fitness set.

[0121] The target population generation module is updated to remove chromosomes with fitness values ​​less than a preset threshold from the target population to obtain an updated target population.

[0122] The quality detection result generation module is used to substitute the updated target population as the initial population into the target population generation module, repeat the target population generation module to the updated target population generation module, until the preset termination condition is met to obtain the updated target population; the updated target population is used to update the running parameters.

[0123] In one implementation, the original operating parameter set generation module searches for the equipment operating parameters corresponding to similar printing features in a preset database based on the target printing features corresponding to the quality inspection results and a preset similarity, thereby obtaining the original operating parameter set. This allows for the quick and accurate location of original parameters that are similar to the current printing quality status from a large amount of historical data, providing a reliable starting point for subsequent parameter optimization and avoiding the inefficiency and poor results caused by blindly adjusting parameters.

[0124] In one implementation, the target population generation module generates an initial population based on the original set of operating parameters, and iteratively optimizes the initial population using a preset genetic algorithm (such as a classical genetic algorithm or an existing genetic algorithm) to obtain the target population. As a global optimization algorithm, the genetic algorithm has powerful search capabilities, enabling extensive exploration within the parameter space and avoiding getting trapped in local optima. By simulating natural selection and genetic mechanisms, superior parameter individuals are continuously screened and combined, gradually approaching the optimal parameter combination, thereby achieving efficient global optimization of the equipment's operating parameters.

[0125] In one implementation, the preset similarity is determined by a technician, and when searching a preset database, the preset similarity is calculated as the cosine similarity.

[0126] In one implementation, the quality inspection result generation module updates the target population as the initial population and substitutes it back into the target population generation module for iterative optimization until a preset termination condition is met (the preset number of iterations, to be determined by technical personnel). This dynamic iterative optimization process can continuously adjust and improve parameters based on the results of each iteration, so that the equipment operating parameters gradually adapt to different printing needs and production environments, thereby achieving continuous improvement and enhancement of printing quality.

[0127] In one implementation, the fitness set generation module evaluates the fitness of each chromosome (i.e., different parameter combinations) in the target population using the target model to obtain a fitness set. Fitness evaluation quantifies the degree to which each parameter combination improves print quality, providing a scientific basis for parameter selection and optimization. This approach allows for a clear understanding of the strengths and weaknesses of each parameter combination, enabling targeted selection and adjustment.

[0128] In one implementation, the target population generation module removes chromosomes in the target population with fitness values ​​lower than a preset threshold (determined by technicians) to obtain an updated target population. This step can promptly eliminate parameter combinations that do not effectively improve print quality, ensuring that subsequent optimization processes always move towards improving print quality, thus improving optimization efficiency and effectiveness.

[0129] In one embodiment, the fitness set generation module includes:

[0130] The historical printing data acquisition module is used to acquire historical printing data; historical printing data includes historical printing operation parameters and corresponding historical printing characteristics;

[0131] The model update module is used to input historical printing data into a preset model to obtain model parameters, and then update the preset model according to the model parameters to obtain the target model.

[0132] In one implementation, the historical printing data acquisition module can collect a large amount of historical printing operation parameters and corresponding historical printing characteristics. These data cover a variety of information at different times, for different printing tasks, and under different equipment conditions, providing rich materials for subsequent analysis and modeling.

[0133] In one implementation, the printing production environment is constantly changing. New printing materials, equipment aging, process improvements, and other factors may affect the relationship between equipment operating parameters and printing quality. The model update module inputs historical printing data into the preset model and updates the preset model according to the obtained model parameters, so that the target model can adapt to these changes in a timely manner.

[0134] In one implementation, the preset model is a multilayer perceptron (MLP), including GMLP, ECgMLP, etc.

[0135] In one implementation, as historical printing data accumulates, the model update module can continuously optimize and improve the target model; more data means that the model can learn more complex patterns and relationships, thereby improving the model's accuracy and generalization ability.

[0136] In one embodiment, the system further includes an alarm module:

[0137] The alarm module is used to issue an alarm and send relevant information to maintenance personnel if the quality inspection of the target product after control and management results in a defective product.

[0138] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. An online production control and management system for printing equipment, characterized in that, The system includes: The equipment operation parameter acquisition module is used to acquire a set of equipment operation parameters of the target printing equipment during operation at a preset time period; the equipment operation parameters include printing speed, pressure and drying temperature; The data preprocessing module is used to preprocess the set of operating parameters of the equipment to obtain a valid set of operating parameters; The product quality inspection module is used to perform quality inspection on the target product and obtain the quality inspection results; the target product is the product to be printed under the operating parameters of the equipment; the quality inspection results include excellent products, qualified products, and defective products. The update operation parameter determination module is used to determine whether the quality inspection result of the target product meets the production requirements. If it does not meet the production requirements, the quality inspection result, the production requirements, and the set of valid operation parameters are substituted into a preset model to obtain the updated operation parameters of the target printing equipment. The production requirements are that the quality inspection result is an excellent product or a qualified product. The printing equipment control and management module is used to control and manage the target printing equipment according to the updated operating parameters.

2. The online production control and management system for printing equipment according to claim 1, characterized in that, The data preprocessing module includes: The mean and variance calculation module is used to calculate the mean and variance of each type of data in the set of equipment operating parameters to obtain the mean operating parameters and the variance operating parameters. The dynamic abnormal data marking module is used to record the equipment operating parameters at a certain moment as dynamic abnormal data if the operating parameters at that moment exceed three times the variance of the operating parameters. The behavior operation recognition module is used to recognize the device operation log, and update the dynamic abnormal data according to the device operation log and the average operation parameters to obtain dynamic normal data. The dynamic abnormal data replacement module is used to replace the dynamic abnormal data in each type of data in the device operating parameter set with dynamic normal data to obtain a valid operating parameter set.

3. The online production control and management system for printing equipment according to claim 2, characterized in that, The behavior recognition module includes: The behavior operation judgment module is used to determine whether a behavior operation exists based on the device operation log; The first dynamic normal data generation module is used to update the dynamic abnormal data according to the average running parameters to obtain dynamic normal data if there is a behavioral operation. The second dynamic normal data generation module is used to replace the dynamic abnormal data with dynamic normal data based on the mean running parameters if no action operation exists.

4. The online production control and management system for printing equipment according to claim 1, characterized in that, The product quality inspection module includes: The printing feature extraction module is used to collect image data of the target product and extract features from the image data to obtain printing features; An enhanced printing feature extraction module is used to enhance the printing features to obtain enhanced printing features; The quality inspection result determination module is used to input enhanced printing features into the inspection network to obtain the quality inspection result.

5. The online production control and management system for printing equipment according to claim 4, characterized in that, The printing feature extraction module includes: A multi-channel feature extraction module is used to input the image data into a CBS module to obtain a first feature, and to perform channel splitting on the first feature to obtain a first channel feature, a second channel feature, a third channel feature, and a fourth channel feature; the number of channels of the first channel feature is half that of the first feature; the number of channels of the second channel feature is half that of the first channel feature; the number of channels of the third channel feature is half that of the second channel feature; and the number of channels of the fourth channel feature is the same as that of the third channel feature. The splicing channel feature extraction module is used to substitute the first channel feature, the second channel feature, and the third channel feature into the channel feature extraction module respectively, and then splice them with the fourth channel feature to obtain the spliced ​​channel feature; A printing feature generation module is used to substitute the splicing channel features into the CBS module to obtain printing features; The working principle of the channel feature extraction module is as follows: The input features are then subjected to a 3×3 convolution and fed into the progressive diversity function module to obtain enhanced input features; The input features and the enhanced input features are element-wise summed and then subjected to a 1×1 convolution to obtain the output channel features.

6. The online production control and management system for printing equipment according to claim 4, characterized in that, The enhanced printing feature extraction module includes: The nonlinear enhancement feature extraction module is used to perform a 1×1 convolution on the printing features and then substitute them into the GELU activation function to obtain nonlinear enhancement features. The semantic-aware feature generation module is used to sequentially substitute the nonlinear enhanced features into the regional interaction-assisted reconstruction unit and the global semantic-aware unit to obtain semantic-aware features. An enhanced printing feature generation module is used to multiply the printing features and the semantic-aware features and then perform a 1×1 convolution to obtain enhanced printing features.

7. The online production control and management system for printing equipment according to claim 4, characterized in that, The quality inspection result determination module includes: The height and width feature extraction module is used to extract height and width features from the enhanced printing features to obtain height and width features; The first activation feature generation module is used to multiply the height feature and the width feature, perform a 1×1 convolution, and obtain the first activation feature through the Sigmoid function; The second activation feature generation module is used to fuse the first activation feature and the enhanced printing feature, perform a 1×1 convolution, and obtain the second activation feature through the Softmax function; The target printing feature generation module is used to fuse the second activation feature and the enhanced printing feature and then perform a 1×1 convolution to obtain the target printing feature; The quality inspection result generation module is used to search a preset database to obtain the quality inspection result based on the target printing features; the preset database stores the printing features and quality grades of different printing templates under different equipment operating parameters.

8. The online production control and management system for printing equipment according to claim 7, characterized in that, The update operation parameter determination module includes: The original operating parameter set generation module is used to search the preset database based on the target printing features corresponding to the quality inspection results, and to search the equipment operating parameters corresponding to similar printing features based on the preset similarity to obtain the original operating parameter set. The target population generation module is used to generate an initial population based on the original set of operating parameters, and to obtain the target population by iteratively optimizing the initial population through a preset genetic algorithm. The fitness set generation module is used to evaluate the fitness of each chromosome in the target population using the target model to obtain a fitness set. The target population generation module is used to remove chromosomes with fitness values ​​less than a preset threshold from the target population to obtain an updated target population. The quality detection result generation module is used to substitute the updated target population as the initial population into the target population generation module, and repeat the target population generation module to the updated target population generation module until the preset termination condition is met to obtain the updated target population; the updated target population is the updated running parameters.

9. The online production control and management system for printing equipment according to claim 8, characterized in that, The fitness set generation module includes: The historical printing data acquisition module is used to acquire historical printing data; the historical printing data includes historical printing operation parameters and corresponding historical printing characteristics; The model update module is used to input the historical printing data into a preset model to obtain model parameters, and update the preset model according to the model parameters to obtain the target model.

10. The online production control and management system for printing equipment according to claim 1, characterized in that, The system also includes an alarm module: The alarm module is used to issue an alarm and send relevant information to maintenance personnel if the quality inspection result of the target product after control and management is a defective product.

Images