Printing success rate judgment method and system based on AI

Abstract

The invention discloses an AI-based printing success rate judgment method and system. The method comprises the following steps: firstly, collecting real-time data of a printer; then preprocessing the real-time data of the printer to generate feature data used for judging the printing success rate; constructing an initial multi-dimensional feature prediction model for judging a printing success rate, and training the initial multi-dimensional feature prediction model by using historical printing job data to obtain a trained target multi-dimensional feature prediction model; and finally, inputting the feature data into a trained target multi-dimensional feature prediction model, and outputting a success rate prediction result of the printing job. By providing an AI-based target multi-dimensional feature prediction model, the pre-judgment capability of printing faults is improved, printing parameters can be adjusted in real time, the printing process can be optimized, and the printing failure rate can be reduced.

Description

Technical Field

[0001] The present invention belongs to the field of printing technology, and in particular to an AI-based printing success rate judgment method and system. Background Art

[0002] In the current field of 3D printing and other printing technologies, the printing success rate is an important indicator to measure the quality and efficiency of printing jobs. However, due to the influence of various factors, such as the operating status of the printer, the physical and chemical properties of the printing materials, and the environmental conditions of the printing job, the printing success rate is often difficult to accurately predict. Traditional methods for judging the printing success rate mostly rely on the operator's experience or simple threshold setting. These methods are not only inefficient, but also cannot guarantee accuracy, which has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] The purpose of the present invention is to provide an AI-based printing success rate judgment method and system to solve the deficiencies in the prior art. It improves the ability to predict printing failures by providing an AI-based target multidimensional feature prediction model, can adjust printing parameters in real time, optimize the printing process, and reduce the printing failure rate.

[0004] An embodiment of the present application provides a method for determining a printing success rate based on AI, the method comprising: Collecting real-time data of the printer; wherein the real-time data of the printer includes operation data of the printer, characteristic data of the printing material and data of printing environment factors; Perform preprocessing of the real-time data of the printer to generate characteristic data for judging the printing success rate; Constructing an initial multidimensional feature prediction model for determining a printing success rate, and training the initial multidimensional feature prediction model using historical printing job data to obtain a trained target multidimensional feature prediction model; The feature data is input into a trained target multi-dimensional feature prediction model, and a prediction result of the success rate of the printing job is output.

[0005] Optionally, the operating data includes printing speed, printing resolution, laser power, scanning speed, layer thickness and filling mode; the characteristic data includes the performance, purity and particle size distribution of the printing material; and the environmental factor data includes temperature and humidity data.

[0006] Optionally, before collecting the real-time data of the printer, the method further includes: Acquire printer operation standard data, and parse out several indicator data affecting printing success rate in the printer operation standard data; The variable parameters in each of the index data affecting the printing success rate are subjected to principal component analysis to obtain a weight value corresponding to each of the index data affecting the printing success rate.

[0007] Optionally, collecting real-time data of the printer includes: By utilizing a data collection system based on the Internet of Things and combining the weight values ​​corresponding to the indicator data affecting the printing success rate, real-time data of the printer is obtained; wherein the data collection system based on the Internet of Things includes: an operation risk warning unit, a printing material identification unit, and a printing environment monitoring unit.

[0008] Optionally, before constructing an initial multidimensional feature prediction model for determining a printing success rate and training the initial multidimensional feature prediction model using historical printing job data to obtain a trained target multidimensional feature prediction model, the method further includes: Acquire a total set of historical printing job data, wherein the total set of historical printing job data includes operation data of the printer under various printing scenarios, characteristic data of printing materials, and printing environment factor data; Using a clustering algorithm, the total set of historical printing job data is divided into historical printing job data subsets corresponding to a plurality of printing scenarios; A number of sampling values ​​are randomly selected from the historical printing job data subsets corresponding to the multiple printing scenarios to generate the historical printing job data.

[0009] Optionally, the initial multi-dimensional feature prediction model includes: an operation data processing unit, a printing material analysis unit and an environmental factor interference unit, wherein the operation data processing unit is constructed based on a multi-layer perceptron model, the printing material analysis unit is constructed based on a recursive neural network and a variant long short-term memory network model, and the environmental factor interference unit is constructed based on a deep belief neural network model; The using the historical printing job data to train the initial multidimensional feature prediction model to obtain a trained target multidimensional feature prediction model includes: The operation data processing unit and the printing material analysis unit are integrated through a multi-layer perception algorithm to obtain a two-dimensional hybrid model, and the historical printing job data are trained using the two-dimensional hybrid model to obtain an initial printing success rate matrix; Verifying the initial printing success rate matrix using the environmental factor interference model and the historical printing job data, and determining a target printing success rate matrix according to the verification result; Based on the target printing success rate matrix and the preset loss function, a target multi-dimensional feature prediction model is determined.

[0010] Optionally, after inputting the feature data into a trained target prediction model and outputting a prediction result of the success rate of the printing job, the method further includes: Based on the prediction result of the success rate of the printing job, adjustment suggestions are fed back to the user; the adjustment suggestions at least include adjusting the printing settings, changing the printing materials and / or improving the printing environment.

[0011] Another embodiment of the present application provides a printing success rate judgment system based on AI, the system comprising: A collection module, used to collect real-time data of the printer; wherein the real-time data of the printer includes operation data of the printer, characteristic data of printing materials and data of printing environment factors; An execution module, used for performing preprocessing of the real-time data of the printer and generating characteristic data for judging the printing success rate; A construction module, used to construct an initial multidimensional feature prediction model for determining a printing success rate, and to train the initial multidimensional feature prediction model using historical printing job data to obtain a trained target multidimensional feature prediction model; The output module is used to input the feature data into the trained target multi-dimensional feature prediction model and output the success rate prediction result of the printing job.

[0012] Yet another embodiment of the present application provides a storage medium, wherein the storage medium stores a computer program, wherein the computer program is configured to implement any of the above methods when executed.

[0013] Another embodiment of the present application provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to implement any of the above methods.

[0014] Compared with the prior art, the present invention first collects the real-time data of the printer; then performs preprocessing on the real-time data of the printer to generate feature data for judging the printing success rate; constructs an initial multidimensional feature prediction model for judging the printing success rate, and uses historical printing job data to train the initial multidimensional feature prediction model to obtain a trained target multidimensional feature prediction model; finally, the feature data is input into the trained target multidimensional feature prediction model to output the prediction result of the success rate of the printing job. It improves the ability to predict printing failures by providing an AI-based target multidimensional feature prediction model, can adjust printing parameters in real time, optimize the printing process, and reduce the printing failure rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1A hardware structure block diagram of a computer terminal for an AI-based printing success rate determination method provided in an embodiment of the present invention; Figure 2 A flowchart of a method for determining a printing success rate based on AI provided by an embodiment of the present invention; Figure 3 A schematic diagram of the structure of a printing success rate judgment system based on AI provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0016] The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, but should not be construed as limiting the present invention.

[0017] The embodiment of the present invention first provides a printing success rate judgment method based on AI, which can be applied to electronic devices, such as computer terminals, specifically ordinary computers, quantum computers, etc.

[0018] The following describes it in detail by taking running on a computer terminal as an example. Figure 1 The hardware structure block diagram of a computer terminal for a printing success rate determination method based on AI provided by an embodiment of the present invention. Figure 1 As shown, the computer terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data. Optionally, the computer terminal may also include a transmission device 106 for communication functions and an input and output device 108. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above-mentioned computer terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0019] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / modules corresponding to the AI-based printing success rate judgment method in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, implementing the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories can be connected to the computer terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0020] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of a computer terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0021] See also Figure 2 , Figure 2 A flowchart of a method for determining a printing success rate based on AI provided by an embodiment of the present invention may include the following steps: S201: Collecting real-time data of the printer; wherein the real-time data of the printer includes operation data of the printer, characteristic data of printing materials, and data of printing environment factors.

[0022] Specifically, the operating data includes printing speed, printing resolution, laser power, scanning speed, layer thickness and filling mode; the characteristic data includes the performance, purity and particle size distribution of the printing material; and the environmental factor data includes temperature and humidity data.

[0023] Among them, print speed is used to monitor the speed at which the printer moves when performing a printing task. Print resolution is used to record the number of dots per inch (DPI) of the print output. High resolution usually means better image quality and detail performance. For laser printers, laser power is used to monitor the power intensity of the laser beam, which determines the adhesion and density of the toner or carbon powder printed on the paper. Scan speed is used to record the speed when scanning documents or images. For 3D printers, layer thickness is used to measure the height of each printed layer. Layer thickness affects the accuracy and surface quality of 3D prints. Fill mode is used to record the filling method inside the print, such as grid, honeycomb or solid filling. Print material performance is used to evaluate the physical and chemical properties of the printed material, such as weather resistance, heat resistance, chemical corrosion resistance, etc. Purity is used to measure the content of impurities or contaminants in the material. High-purity materials usually provide better printing effects and longer service life. Particle size distribution is used to analyze the size and distribution of particles in the printed material, which is particularly important for powdered printing materials, such as metal powder for 3D printing, because it affects the printing accuracy and interlayer bonding. Temperature is used to monitor the temperature of the environment where the printer is located. Changes in temperature may affect the performance of printed materials and the operating stability of the printer. Humidity is used to record the humidity level in the environment. High humidity may cause problems such as paper deformation, ink diffusion or print head clogging.

[0024] First, most modern printers are equipped with built-in data logging and monitoring functions. You can view the printer's status, print tasks, ink levels, paper usage, etc. through the printer's management interface or supporting software. Secondly, for printers that use specific types of paper or consumables, you can identify the type of material used through the printer's consumables management system or driver. Or use professional color measurement instruments and color management software, such as X-Rite's i1Pro2 spectrophotometer and ProfileMaker or i1Profiler color management software, to measure and record the color characteristic data of printed materials. These data include data in color spaces such as RGB, CMYK, and colorimetric values ​​such as Lab*. At the same time, you can also establish and maintain a printed material database to record the characteristic data and usage of different materials so that you can quickly find and compare when needed.

[0025] In order to obtain printing environment factor data, you can install temperature and humidity sensors, dust sensors and other equipment in the application environment where the printer is located to monitor the temperature, humidity, air quality and other parameters of the environment in real time. Or use an environmental data acquisition system, such as a client and server architecture system, to collect and process these environmental parameter data. The client is responsible for data collection and wireless transmission, while the server is responsible for data storage and concurrent services. The collected environmental factor data is correlated with data such as print quality and printer failure rate to analyze the impact of environmental factors on printing effects, and optimize the printing environment and printing parameters accordingly.

[0026] In summary, real-time data collection of printers can be achieved by using a combination of tools and methods such as the printer's built-in functions, professional color measurement instruments, and color management software. At the same time, it is also necessary to establish and maintain a printing material database and use an environmental data collection system to monitor and analyze environmental factor data in real time. These measures will help to fully understand the operating status of the printer, the characteristics of the printing materials, and the impact of printing environmental factors on the printing effect, thereby optimizing the printing process and improving the printing quality.

[0027] In an optional implementation, before collecting the real-time data of the printer, the method may further include: 1. Obtain printer operation standard data, and parse out several indicator data affecting the printing success rate in the printer operation standard data; 2. Performing principal component analysis on the variable parameters in each of the index data affecting the printing success rate to obtain the weight value corresponding to each of the index data affecting the printing success rate.

[0028] Specifically, first, the printer's operating standard data can be obtained from the printer manufacturer or related databases. These data usually include the printer's design parameters, recommended working conditions, performance specifications, etc. These data are the basis for evaluating printer performance and predicting printing success rate.

[0029] After obtaining the printer's operating standard data, these data can be analyzed to extract several key indicator data that affect the printing success rate. These indicator data may include the printer's physical parameters (such as printing speed, resolution, etc.), material parameters (such as the performance and purity of the printing material, etc.) and environmental factors (such as temperature, humidity, etc.).

[0030] Next, principal component analysis (PCA) is performed on the variable parameters in the analyzed indicator data that affect the printing success rate. Principal component analysis is a commonly used data dimensionality reduction technique that can convert high-dimensional data into low-dimensional data through linear transformation while retaining as much information as possible in the original data.

[0031] In principal component analysis, the variance contribution rate of each indicator data can be calculated, that is, the proportion of the indicator data in the total variance. This proportion can be regarded as the weight value of the indicator data's influence on the printing success rate. Through principal component analysis, the weight value corresponding to each indicator data affecting the printing success rate can be obtained, and these weight values ​​will be used for subsequent feature data generation and model training.

[0032] In summary, steps such as obtaining printer operation standard data, analyzing indicator data that affect printing success rate, and performing principal component analysis to determine weight values ​​are indispensable preliminary work in the AI-based printing success rate judgment method. They provide an important foundation for subsequent data collection, feature generation, and model training.

[0033] In an optional implementation, the collecting of real-time data of the printer may include: By utilizing a data collection system based on the Internet of Things and combining the weight values ​​corresponding to the indicator data affecting the printing success rate, real-time data of the printer is obtained; wherein the data collection system based on the Internet of Things includes: an operation risk warning unit, a printing material identification unit, and a printing environment monitoring unit.

[0034] Specifically, the operation risk warning unit is one of the core parts of the data acquisition system. It is mainly responsible for monitoring the operating status of the printer and predicting and identifying potential failures or risks by real-time monitoring of key parameters. The operation risk warning unit can integrate a variety of sensors, such as temperature sensors, pressure sensors, vibration sensors, etc. These sensors can collect various operating data of the printer in real time. The collected data will be transmitted to the data processing center in real time and quickly analyzed through preset algorithms and models. Once the data exceeds the normal range or reaches the preset threshold, the system will immediately trigger the early warning mechanism and send an alarm message to the management personnel. In addition to the early warning function, the operation risk early warning unit can also identify the specific fault type through data analysis and locate the specific location of the fault, providing strong support for subsequent repairs and maintenance.

[0035] The printing material identification unit is mainly responsible for identifying and distinguishing different types of printing materials as well as the batch and quality of the materials, thereby ensuring that the printer can print with the correct materials.

[0036] The printing material identification unit can use advanced material identification technologies, such as RFID (radio frequency identification), barcode scanning or image recognition. These technologies can accurately and quickly identify the type, batch and quality information of the printing material. In order to achieve accurate identification of different materials, the printing material identification unit can also establish a huge material database. The database contains detailed information on various materials, such as physical properties, chemical properties, usage conditions, etc. By comparing the information in the database, the system can determine whether the currently used materials meet the printing requirements. In addition to the identification function, the printing material identification unit can also effectively manage materials. For example, it can record the inventory status and usage records of materials, etc., to provide data support for subsequent procurement and inventory management.

[0037] The printing environment monitoring unit is mainly responsible for monitoring various parameters of the printing environment, such as temperature and humidity, to ensure that the printing process is carried out under optimal conditions. The printing environment monitoring unit can be equipped with a variety of environmental sensors, such as temperature sensors, humidity sensors, etc. These sensors can collect data about the printing environment in real time and transmit it to the data processing center for analysis. Based on the collected environmental data, the system can automatically adjust the settings of the printer to adapt to the current environmental conditions. For example, in an environment with high humidity, the system may automatically adjust the humidity of the printing material or increase the drying time to prevent the material from being damp and affecting the printing quality. In addition to the real-time adjustment function, the unit can also generate environmental monitoring reports. The report contains information such as historical data and change trends of the printing environment, providing managers with a comprehensive analysis of environmental conditions.

[0038] In summary, the various units in the IoT-based data acquisition system work together to achieve comprehensive monitoring and optimization of the printer's operating status. The application of these units not only improves the printing success rate, but also reduces the failure rate, providing strong data support for the maintenance and care of the printer.

[0039] S202: Execute preprocessing of the real-time data of the printer to generate characteristic data for determining the printing success rate.

[0040] Specifically, in order to generate characteristic data for determining the printing success rate, the printer operation data, the characteristic data of the printing material, and the printing environment factor data may be comprehensively preprocessed.

[0041] First, perform data cleaning and normalization of printer operation data, printing material characteristic data, and printing environment factor data, including deleting missing key information, such as print task ID, print start / end time records, or filling them in according to business logic. Ensure that all timestamp formats are consistent and value types are correct, such as converting temperature and humidity into a unified unit of measurement. Identify and handle obviously unreasonable values, such as negative print pages, excessively high ambient temperature, etc.

[0042] Secondly, feature extraction and conversion operations include printer operation features, which may include print task duration, print speed stability, error rate, etc. Print material features may include paper type coding, such as using label coding to convert paper type into a numerical value; printing paper weight grade, dividing grades according to paper weight; using binary classification features to perform compatibility identification of printing ink / toner, etc.

[0043] For example, the stability index can be measured by calculating the standard deviation of the printing speed:

[0044] in, Indicates the printing speed at a preset time point. Indicates the average printing speed, Indicates the number of data points to print.

[0045] The printing environment characteristics may include an environment temperature range, such as dividing the temperature into different intervals; a humidity level, such as dividing the levels according to humidity, and the like.

[0046] You can use correlation analysis methods, such as Pearson correlation coefficient, to analyze the correlation between features and printing success rate. Or you can use machine learning models, such as random forests, logistic regression, etc. to evaluate the importance of the above features. If there are too many features, you can use dimensionality reduction techniques such as linear discriminant analysis (LDA) to reduce the number of features while maintaining the main information of the feature data.

[0047] S203: constructing an initial multidimensional feature prediction model for determining a printing success rate, and using historical printing job data to train the initial multidimensional feature prediction model to obtain a trained target multidimensional feature prediction model.

[0048] Specifically, before constructing an initial multidimensional feature prediction model for determining a printing success rate and training the initial multidimensional feature prediction model using historical printing job data to obtain a trained target multidimensional feature prediction model, the method may further include: Step 1: Acquire a total set of historical printing job data, wherein the total set of historical printing job data includes operation data of the printer under various printing scenarios, characteristic data of printing materials, and printing environment factor data.

[0049] Step 2: using a clustering algorithm to divide the total set of historical printing job data into historical printing job data subsets corresponding to a variety of printing scenarios.

[0050] Step 3: randomly selecting a number of sampling values ​​from the historical printing job data subset corresponding to the multiple printing scenarios to generate historical printing job data.

[0051] Specifically, to obtain the total set of historical printing job data, comprehensive historical printing job data can be collected from multiple sources, such as printer logs, material databases, environmental monitoring systems, etc. The data should cover printer operation data under various printing scenarios, characteristic data of printing materials, and printing environmental factor data. Then, the collected data is preliminarily cleaned to remove invalid, duplicate or outlier values. Format, encode or convert the data as needed to ensure that it is suitable for subsequent analysis. Select a suitable clustering algorithm, such as K-means, hierarchical clustering, DBSCAN, etc., and divide the total set of historical printing job data into multiple data subsets according to the characteristics of the data.

[0052] Exemplarily, cluster analysis is performed in combination with the print success rate pointing index in the total set of historical print job data, and the total set of historical print job data is divided into ontological feature distributions of multiple data subsets according to the characteristics of the data:

[0053] in, It is an indicator of the printing success rate in the total set of historical printing job data. Represents the similarity of fuzzy cluster centers, uses vector quantization decomposition method for data fusion and balanced configuration, and obtains the variable transfer function of the ontology feature distribution of the data subset:

[0054] in, Represents the quantized decomposition function.

[0055] In summary, the vector quantization decomposition method is used to realize the division of the total set of historical printing job data and determine the variable transfer function of the ontological feature distribution of the data subset, which will provide strong support for subsequent data analysis and decision support.

[0056] Each data subset represents a specific printing scenario, which may vary based on factors such as printer configuration, material type, environmental conditions, etc. Randomly select sampling values ​​from the data subset, and for each divided data subset, randomly select a certain number of samples as sampling values ​​for training data. The sampling process should ensure the representativeness of the data, that is, the important features and information in each subset can be fully reflected. Finally, the sampling values ​​selected from each data subset are merged to form a historical printing job dataset for training the model. This dataset should contain data under a variety of printing scenarios to simulate the diversity in the actual printing environment. In the generated historical printing job dataset, select features that have a significant impact on the printing success rate, and divide the dataset into training set, validation set, and test set for subsequent model training, validation, and testing.

[0057] In an optional embodiment, the initial multidimensional feature prediction model includes: an operating data processing unit, a printing material analysis unit and an environmental factor interference unit, the operating data processing unit is constructed based on a multi-layer perceptron model, the printing material analysis unit is constructed based on a recursive neural network and a variant long short-term memory network model, and the environmental factor interference unit is constructed based on a deep belief neural network model.

[0058] Specifically, a multilayer perceptron (MLP) is a widely used artificial neural network model, whose structure includes an input layer, one or more hidden layers, and an output layer. MLP has shown excellent performance in classification, regression, pattern recognition and other fields with its powerful nonlinear mapping ability and flexible structural design.

[0059] In the operation data processing unit, MLP can learn the complex features in the operation data of the equipment or system through the connections between its multiple layers of neurons. These features can include the operation status and performance parameters of the equipment. The output layer of MLP can generate prediction results of the equipment or system status, such as fault warning, performance evaluation, etc.

[0060] Recurrent neural networks (RNN) and their variants, long short-term memory networks (LSTM), excel in processing sequence data. They have a memory function that allows them to retain previous information and apply it to current decisions. In the printing material analysis unit, RNN / LSTM can learn the laws of how the properties of printing materials change over time. These laws can include the evolution trends of the material's strength, toughness, weather resistance, etc. over time. By inputting the historical performance data of the printing material, RNN / LSTM can predict the material performance at future moments and provide a reference for material selection for printing jobs.

[0061] For example, the number of input layer neurons based on the recurrent neural network and variant long short-term memory network model can be set to , the number of neurons in the evaluation layer is set to , the number of neurons in the output layer is set to , the parameter weights between the input layer and the evaluation layer are set as , the parameter weights between the evaluation layer and the output layer are set as , the learning rate is set to , the correlation strength function between the pooling layer nodes and the output layer nodes based on the recurrent neural network and variant long short-term memory network model is:

[0062] The input data of the evaluation layer neurons in the forward propagation process is:

[0063] The output data of the output layer neurons are:

[0064] If there is an error signal in the output prediction result of the output layer, the specific formula for adjusting the parameter weights of the evaluation layer and the output layer is:

[0065] in, Indicates the reverse error value.

[0066] A deep belief neural network (DBN) is a neural network model composed of multiple layers of restricted Boltzmann machines (RBM). DBN can learn deep feature representations in data by training layer by layer. In the environmental factor interference unit, DBN can process complex and changeable environmental factor data. These factors may include temperature, humidity, etc. DBN can learn the interaction between these factors and predict their potential impact on device or system performance. By inputting real-time environmental factor data, DBN can output the predicted results of device or system performance, providing a basis for environmental adjustment of printing jobs.

[0067] In summary, each unit of the initial multi-dimensional feature prediction model selected a different neural network model for construction. The operation data processing unit selected the MLP model to process the operation data of the equipment or system; the printing material analysis unit selected the RNN / LSTM model to process the performance data of the printing material; the environmental factor interference unit selected the DBN model to process the complex and changeable environmental factor data. These selections are intended to give full play to the advantages of each neural network model and achieve accurate prediction results.

[0068] In an optional implementation, the using the historical printing job data to train the initial multidimensional feature prediction model to obtain a trained target multidimensional feature prediction model may include: a. By using a multi-layer perception algorithm, the operation data processing unit and the printing material analysis unit are integrated to obtain a two-dimensional hybrid model, and the historical printing job data is trained using the two-dimensional hybrid model to obtain an initial printing success rate matrix; b. Using the environmental factor interference model and the historical print job data to verify the initial print success rate matrix, and determining the target print success rate matrix according to the verification results; c. Based on the target printing success rate matrix and the preset loss function, determine the target multi-dimensional feature prediction model.

[0069] First, through the multi-layer perception algorithm, the operation data processing unit and the printing material analysis unit are integrated to form a two-dimensional hybrid model. This model can simultaneously process the operation data of the equipment or system and the performance data of the printing material, so as to more comprehensively capture the factors affecting the printing success rate.

[0070] Next, this two-dimensional hybrid model is used to train historical print job data. The historical print job data contains printer operation data under various printing scenarios, characteristic data of printing materials, and known printing success rate information. Through training, the model can learn the correlation between these data and generate an initial printing success rate matrix. After obtaining the initial printing success rate matrix, the initial matrix is ​​verified using the environmental factor interference model and historical printing job data. The environmental factor interference model can handle complex and changeable environmental factor data and predict their impact on the printing success rate. The verification process includes inputting the environmental factor data into the environmental factor interference model to obtain the predicted value of the impact of environmental factors on the printing success rate. Then, this predicted value is combined with the initial printing success rate matrix, and the success rate value in the matrix is ​​adjusted to reflect the actual impact of environmental factors.

[0071] According to the verification results, the target printing success rate matrix is ​​determined. The target printing success rate matrix more accurately reflects the printing success rate under various printing scenarios, taking into account the combined effects of equipment operation, printing materials and environmental factors.

[0072] Finally, based on the target printing success rate matrix and the preset loss function, the target multi-dimensional feature prediction model is determined. Among them, the loss function is a function that measures the difference between the model's prediction results and the actual results. By minimizing the loss function, the parameters and structure of the model can be adjusted to make its prediction results more accurate. During the training process, optimization algorithms such as gradient descent can be used to iteratively update the parameters of the model until the loss function reaches a smaller value or converges. At this point, the model has learned the complex relationship between the multiple dimensional features that affect the printing success rate, and can predict the printing success rate based on the input feature data.

[0073] It should be noted that the print success rate matrix is ​​a multidimensional array used to represent the success rate of print jobs under different conditions. Each row and column of the matrix represents different influencing factors, such as printing materials, printer models, printing environment, etc., and each element in the matrix represents the print success rate under specific conditions.

[0074] In summary, by integrating the operation data processing unit, the printing material analysis unit and the environmental factor interference model and using the historical printing job data for training, an accurate and reliable target multi-dimensional feature prediction model can be obtained. This model can provide strong support for the prediction of the success rate of printing jobs.

[0075] S204: Input the feature data into the trained target multi-dimensional feature prediction model, and output the prediction result of the success rate of the printing job.

[0076] Specifically, the sorted feature data is input into the trained target multi-dimensional feature prediction model. Ensure that the dimension and order of the input data are consistent with the data used in model training to avoid model processing errors. After receiving the input data, the model will perform calculations according to its internal structure and parameters. For example, in the combination of models such as multi-layer perceptron (MLP), recurrent neural network (RNN / LSTM) and deep belief neural network (DBN), the model will comprehensively consider the influence of multiple dimensions such as equipment operation data, printing material information and environmental factors. Through the internal calculation of the model, a predicted value representing the success rate of the print job is finally output. After the model calculation is completed, the prediction result of the success rate of the print job is output. This prediction result is usually a value between 0 and 1, indicating the probability of success of the print job. According to actual needs, this prediction result can be presented in the form of percentage, chart or other intuitive ways for users to better understand and use.

[0077] In an optional implementation, after inputting the feature data into the trained target prediction model and outputting the prediction result of the success rate of the printing job, the method may further include: Based on the prediction result of the success rate of the printing job, adjustment suggestions are fed back to the user; the adjustment suggestions at least include adjusting the printing settings, changing the printing materials and / or improving the printing environment.

[0078] Specifically, the prediction results of the success rate of the print job are first analyzed, which can include understanding the specific value of the predicted success rate and its comparison with the historical success rate. If the predicted success rate is lower than a certain threshold, for example, lower than 90%, it may be necessary to further explore the cause and generate corresponding adjustment suggestions.

[0079] Combined with the feature dimensions considered by the target multidimensional feature prediction model, such as equipment operation data, printing material information, environmental factors, etc., analyze which factors may lead to a decrease in the printing success rate. Through the weight distribution or feature importance analysis within the model, several key factors that have the greatest impact on the printing success rate can be identified. Based on the identified key factors, specific adjustment suggestions are generated. These suggestions may include: adjusting print settings, such as adjusting the print speed, print quality, print mode, etc. to adapt to different printing materials and environmental conditions. Or changing printing materials, such as choosing a paper type, ink type, etc. that is more suitable for the current printer model and printing environment. And improving the printing environment, such as adjusting the temperature, humidity, and other conditions of the printing room to reduce the negative impact on the print job.

[0080] Compared with the prior art, the present invention first collects the real-time data of the printer; then performs preprocessing on the real-time data of the printer to generate feature data for judging the printing success rate; constructs an initial multidimensional feature prediction model for judging the printing success rate, and uses historical printing job data to train the initial multidimensional feature prediction model to obtain a trained target multidimensional feature prediction model; finally, the feature data is input into the trained target multidimensional feature prediction model to output the prediction result of the success rate of the printing job. It improves the ability to predict printing failures by providing an AI-based target multidimensional feature prediction model, can adjust printing parameters in real time, optimize the printing process, and reduce the printing failure rate.

[0081] Another embodiment of the present application provides a printing success rate judgment system based on AI, such as Figure 3 The structure diagram of a printing success rate judgment system based on AI is shown, and the system includes: The acquisition module 301 is used to acquire real-time data of the printer; wherein the real-time data of the printer includes operation data of the printer, characteristic data of the printing material and data of printing environment factors; An execution module 302 is used to perform preprocessing of the real-time data of the printer and generate characteristic data for determining the printing success rate; A construction module 303 is used to construct an initial multidimensional feature prediction model for determining a printing success rate, and train the initial multidimensional feature prediction model using historical printing job data to obtain a trained target multidimensional feature prediction model; The output module 304 is used to input the feature data into the trained target multi-dimensional feature prediction model and output the prediction result of the success rate of the printing job.

[0082] Compared with the prior art, the present invention first collects the real-time data of the printer; then performs preprocessing on the real-time data of the printer to generate feature data for judging the printing success rate; constructs an initial multidimensional feature prediction model for judging the printing success rate, and uses historical printing job data to train the initial multidimensional feature prediction model to obtain a trained target multidimensional feature prediction model; finally, the feature data is input into the trained target multidimensional feature prediction model to output the prediction result of the success rate of the printing job. It improves the ability to predict printing failures by providing an AI-based target multidimensional feature prediction model, can adjust printing parameters in real time, optimize the printing process, and reduce the printing failure rate.

[0083] An embodiment of the present invention further provides a storage medium, in which a computer program is stored, wherein the computer program is configured to implement the steps in the above method embodiment when running.

[0084] Specifically, in this embodiment, the above storage medium may be configured to store a computer program for performing the following steps: S201: Collecting real-time data of the printer; wherein the real-time data of the printer includes operation data of the printer, characteristic data of printing materials, and data of printing environment factors; S202: performing preprocessing on the real-time data of the printer to generate characteristic data for determining the printing success rate; S203: constructing an initial multidimensional feature prediction model for determining a printing success rate, and training the initial multidimensional feature prediction model using historical printing job data to obtain a trained target multidimensional feature prediction model; S204: Input the feature data into the trained target multi-dimensional feature prediction model, and output the prediction result of the success rate of the printing job.

[0085] Specifically, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store computer programs.

[0086] Compared with the prior art, the present invention first collects the real-time data of the printer; then performs preprocessing on the real-time data of the printer to generate feature data for judging the printing success rate; constructs an initial multidimensional feature prediction model for judging the printing success rate, and uses historical printing job data to train the initial multidimensional feature prediction model to obtain a trained target multidimensional feature prediction model; finally, the feature data is input into the trained target multidimensional feature prediction model to output the prediction result of the success rate of the printing job. It improves the ability to predict printing failures by providing an AI-based target multidimensional feature prediction model, can adjust printing parameters in real time, optimize the printing process, and reduce the printing failure rate.

[0087] An embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in the above method embodiment.

[0088] Specifically, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0089] Specifically, in this embodiment, the processor may be configured to perform the following steps through a computer program: S201: Collecting real-time data of the printer; wherein the real-time data of the printer includes operation data of the printer, characteristic data of printing materials, and data of printing environment factors; S202: performing preprocessing on the real-time data of the printer to generate characteristic data for determining the printing success rate; S203: constructing an initial multidimensional feature prediction model for determining a printing success rate, and training the initial multidimensional feature prediction model using historical printing job data to obtain a trained target multidimensional feature prediction model; S204: Input the feature data into the trained target multi-dimensional feature prediction model, and output the prediction result of the success rate of the printing job.

[0090] Compared with the prior art, the present invention first collects the real-time data of the printer; then performs preprocessing on the real-time data of the printer to generate feature data for judging the printing success rate; constructs an initial multidimensional feature prediction model for judging the printing success rate, and uses historical printing job data to train the initial multidimensional feature prediction model to obtain a trained target multidimensional feature prediction model; finally, the feature data is input into the trained target multidimensional feature prediction model to output the prediction result of the success rate of the printing job. It improves the ability to predict printing failures by providing an AI-based target multidimensional feature prediction model, can adjust printing parameters in real time, optimize the printing process, and reduce the printing failure rate.

[0091] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0092] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0093] In the several embodiments provided by the present invention, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the above units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0094] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0095] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0096] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a memory, including several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the above-mentioned methods of each embodiment of the present invention. The aforementioned memory includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0097] The embodiments of the present invention are described in detail above. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A printing success rate judgment method based on AI, characterized in that: The method comprises: Collecting real-time data of the printer; wherein the real-time data of the printer includes operation data of the printer, characteristic data of the printing material and data of printing environment factors; Perform preprocessing of the real-time data of the printer to generate characteristic data for judging the printing success rate; Constructing an initial multidimensional feature prediction model for determining a printing success rate, and training the initial multidimensional feature prediction model using historical printing job data to obtain a trained target multidimensional feature prediction model; The feature data is input into a trained target multi-dimensional feature prediction model, and a prediction result of the success rate of the printing job is output.

2. The method according to claim 1, characterized in that The operation data includes printing speed, printing resolution, laser power, scanning speed, layer thickness and filling mode; the characteristic data includes the performance, purity and particle size distribution of the printing material; and the environmental factor data includes temperature and humidity data.

3. The method according to claim 2, characterized in that Before collecting the real-time data of the printer, the method further includes: Acquire printer operation standard data, and parse out several indicator data affecting printing success rate in the printer operation standard data; The variable parameters in each of the index data affecting the printing success rate are subjected to principal component analysis to obtain a weight value corresponding to each of the index data affecting the printing success rate.

4. The method according to claim 3, characterized in that The real-time data of the printer is collected, including: By utilizing a data collection system based on the Internet of Things and combining the weight values ​​corresponding to the indicator data affecting the printing success rate, real-time data of the printer is obtained; wherein the data collection system based on the Internet of Things includes: an operation risk warning unit, a printing material identification unit, and a printing environment monitoring unit.

5. The method according to claim 4, characterized in that Before constructing an initial multidimensional feature prediction model for determining a printing success rate and training the initial multidimensional feature prediction model using historical printing job data to obtain a trained target multidimensional feature prediction model, the method further includes: Acquire a total set of historical printing job data, wherein the total set of historical printing job data includes operation data of the printer under various printing scenarios, characteristic data of printing materials, and printing environment factor data; Using a clustering algorithm, the total set of historical printing job data is divided into historical printing job data subsets corresponding to a plurality of printing scenarios; A number of sampling values ​​are randomly selected from the historical printing job data subsets corresponding to the multiple printing scenarios to generate the historical printing job data.

6. The method according to claim 5, characterized in that The initial multi-dimensional feature prediction model includes: an operation data processing unit, a printing material analysis unit and an environmental factor interference unit, wherein the operation data processing unit is constructed based on a multi-layer perceptron model, the printing material analysis unit is constructed based on a recursive neural network and a variant long short-term memory network model, and the environmental factor interference unit is constructed based on a deep belief neural network model; The using the historical printing job data to train the initial multidimensional feature prediction model to obtain a trained target multidimensional feature prediction model includes: The operation data processing unit and the printing material analysis unit are integrated through a multi-layer perception algorithm to obtain a two-dimensional hybrid model, and the historical printing job data are trained using the two-dimensional hybrid model to obtain an initial printing success rate matrix; Using the environmental factor interference model and the historical printing job data to verify the initial printing success rate matrix, and determining a target printing success rate matrix according to the verification result; Based on the target printing success rate matrix and the preset loss function, a target multi-dimensional feature prediction model is determined.

7. The method according to claim 6, characterized in that After inputting the feature data into the trained target prediction model and outputting the prediction result of the success rate of the printing job, the method further includes: Based on the prediction result of the success rate of the printing job, adjustment suggestions are fed back to the user; the adjustment suggestions at least include adjusting the printing settings, changing the printing materials and / or improving the printing environment.

8. A printing success rate judgment system based on AI, characterized in that: The system comprises: A collection module, used to collect real-time data of the printer; wherein the real-time data of the printer includes operation data of the printer, characteristic data of printing materials and data of printing environment factors; An execution module, used for performing preprocessing of the real-time data of the printer and generating characteristic data for judging the printing success rate; A construction module, used to construct an initial multidimensional feature prediction model for determining a printing success rate, and to train the initial multidimensional feature prediction model using historical printing job data to obtain a trained target multidimensional feature prediction model; The output module is used to input the feature data into the trained target multi-dimensional feature prediction model and output the success rate prediction result of the printing job.

9. A storage medium, characterized in that: The storage medium stores a computer program, wherein the computer program is configured to implement the method according to any one of claims 1 to 7 when executed.

10. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the computer program to implement the method according to any one of claims 1 to 7.

Images