AI-driven photocuring 3D printing fault prediction method

Abstract

The invention provides an AI-driven photocuring 3D printing fault prediction method, and relates to the technical field of 3D printing, and the method comprises the steps: collecting printing parameters in real time based on a sensor installed on photocuring 3D printing equipment, and carrying out data preprocessing; multi-dimensional features of photocuring 3D printing are extracted, time sequence features and spatial features are analyzed, and high-dimensional feature vectors of photocuring 3D printing are obtained; marking the collected historical printing data as a normal data sample and a fault data sample, and training and optimizing the constructed fault prediction model in combination with the high-dimensional feature vector; printing parameters collected in real time are detected based on the trained fault prediction model, an obtained prediction result is compared with a preset early warning threshold value, and if the prediction result exceeds the threshold value, early warning is given out according to a fault grading result. By monitoring and analyzing key parameters in the printing process in real time, potential faults are predicted, automatic adjustment is carried out when abnormity occurs, and the reliability of the printing process and the product quality are improved.

Description

Technical Field

[0001] The present invention relates to the field of 3D printing technology, and in particular to an AI-driven photocuring 3D printing fault prediction method. Background Art

[0002] Stereolithography 3D printing technology, also known as stereolithography (SLA) or digital light processing (DLP) technology, is an additive manufacturing technology based on the curing of photosensitive resin. Its basic principle is to use a light source (such as a laser or a projector) to irradiate the photosensitive resin, and solidify the resin layer by layer through a photocuring reaction, thereby gradually building a three-dimensional object. This technology is widely used in industrial manufacturing, medical treatment, jewelry, construction and other fields due to its high precision and surface smoothness.

[0003] In photocuring 3D printing, problems such as poor interlayer adhesion, dimensional deviation or surface defects often occur, resulting in unstable printing quality. Currently, the printing quality is mainly improved by adjusting printing parameters (such as light source intensity, printing speed, layer thickness, etc.). However, this requires relying on user experience and repeated trials, and it is difficult to cope with dynamically changing printing conditions. The current photocuring 3D printing method focuses on post-printing repair or parameter adjustment, and there is less research on real-time monitoring and fault prediction. Due to the multivariable and complex nature of the photocuring 3D printing process, it is difficult for existing technologies to achieve comprehensive and accurate fault prediction and immediate adjustment.

[0004] Therefore, it is urgent to develop a photocuring 3D printing fault prediction method that can monitor and predict faults in real time to improve the stability and reliability of photocuring 3D printing. Summary of the invention

[0005] In view of this, the present invention proposes an AI-driven photocuring 3D printing fault prediction method, which predicts potential faults and automatically adjusts when abnormalities occur by real-time monitoring and analysis of key parameters in the printing process, and conducts real-time analysis of key parameters such as temperature, light intensity, and material flow in the photocuring 3D printing process by building and applying a fault prediction model. The system is able to predict potential faults before printing begins and monitor key indicators during the printing process. For detected abnormalities, the system will automatically adjust printing parameters or issue an alarm, thereby improving the reliability of the printing process and product quality.

[0006] The present invention is implemented by the following technical solutions: In a first aspect, the present invention provides an AI-driven photocuring 3D printing fault prediction method, the method comprising the following steps: The sensors installed on the light-curing 3D printing equipment collect printing parameters in real time, and the collected printing parameters are pre-processed by the data processing module to eliminate noise and abnormal values; Extract the multi-dimensional features of photocuring 3D printing from the pre-processed data, and perform temporal and spatial feature analysis to obtain the high-dimensional feature vector of photocuring 3D printing; The collected historical printing data is labeled as normal data samples and fault data samples, and the constructed fault prediction model is trained and optimized by combining high-dimensional feature vectors; The printing parameters collected in real time are detected based on the trained fault prediction model, and the prediction results are compared with the preset warning threshold. If the threshold is exceeded, a warning is issued according to the fault classification result.

[0007] As a further solution of the present invention, the light-curing 3D printing device is equipped with a resin tank temperature sensor, a light source intensity sensor, a vibration sensor, a position sensor and a current sensor, which are respectively used to collect printing parameters during light-curing 3D printing: resin temperature, ultraviolet light intensity, vibration frequency, vibration amplitude, and Z-axis height.

[0008] As a further solution of the present invention, when the collected printing parameters are preprocessed, the collected printing parameters are subjected to moving average processing to determine the window size of the moving average; the average value of the moving window before and after the time is calculated for the printing parameters at each time point according to the sampling frequency of the data and the set smoothness level; smoothed time series data are obtained, and the resin temperature, ultraviolet light intensity, vibration data within the average set period of time and the position signal of the Z-axis height are smoothed.

[0009] As a further solution of the present invention, when performing data preprocessing on the collected printing parameters, it also includes eliminating spike noise through median filtering of the smoothed time series data, taking the median of the data at each time point and the forward and backward moving window by determining the window size of the median filter, removing sudden abnormal values ​​in the resin temperature, filtering out abnormal vibration peaks, and removing occasional current spikes.

[0010] As a further solution of the present invention, when performing data preprocessing on the collected printing parameters, it also includes identifying and processing abnormal values ​​in the data after median filtering, and calculating the mean and standard deviation of the data corresponding to each sensor, performing Z-score standardization on the resin temperature, ultraviolet light intensity, vibration frequency and vibration amplitude, and Z-axis height to obtain the preprocessed data.

[0011] As a further solution of the present invention, when extracting multidimensional features of photocuring 3D printing, the smoothed time series data in the preprocessed data is segmented according to the time series, and the time series features of each sensor data in each time period are calculated. The extracted time series features include the mean value, standard deviation, the maximum and minimum values ​​of the extreme value changes captured in each time period, and the skewness and kurtosis of the data morphological distribution are obtained; the time series data is converted to the frequency domain, and the spatial features are extracted. The spatial features include the frequency domain features of the data extracted by Fourier transform and the power spectral density at the same frequency.

[0012] As a further solution of the present invention, the AI-driven light-curing 3D printing fault prediction method further includes extracting spatial features based on a graph neural network, including the following steps: Collect sensor data of light-curing 3D printing and physical structure data of light-curing 3D printing equipment; The physical structure of the light-curing 3D printer is modeled as a graph, where nodes represent sensors and edges represent data flows; Map the sensor data to the corresponding graph nodes to form node feature vectors, where the feature vector of each node contains the multi-dimensional data collected by the node; Based on the graph convolutional network as the graph neural network architecture, a graph neural network model is constructed, in which the defined input layer receives node features, the hidden layer performs feature aggregation and transformation, and the output layer generates the final feature representation; The preprocessed graph data is divided into training set, validation set and test set. The graph neural network model is trained using the training set. The trained graph neural network model is applied to validate the test set graph data to extract the high-dimensional feature vector of each node. The extracted spatial features are fused with the time series features and frequency domain features to form a comprehensive feature vector, and the fused comprehensive feature vector is used to train the fault prediction model.

[0013] As a further solution of the present invention, temporal feature and spatial feature analysis are performed to obtain a high-dimensional feature vector of light-curing 3D printing, including the following steps: aligning data from different sensors according to timestamps, and feature splicing the extracted temporal features and spatial features into a high-dimensional feature vector.

[0014] As a further solution of the present invention, the constructed fault prediction model training and optimization includes the following steps: Collect historical printing data generated during the light-curing 3D printing process, including temperature, UV light intensity, vibration frequency, vibration amplitude, and Z-axis height; The historical printing data is divided into normal data samples and fault data samples according to whether the data is abnormal or faulty within the time period, and randomly divided into training sets and test sets in proportion. The selected model is trained using the high-dimensional feature vectors in the training set and the corresponding normal or faulty labels, and the hyperparameters of the model are adjusted. The model is cross-validated through the test set to obtain the trained and optimized fault prediction model.

[0015] In a second aspect, the present invention also includes a computer device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor executes the AI-driven photocuring 3D printing fault prediction method.

[0016] In a third aspect, the present invention also includes a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the AI-driven light-curing 3D printing fault prediction method.

[0017] Compared with the prior art, the AI-driven photocuring 3D printing fault prediction method provided by the present invention realizes fault prediction and early warning in the 3D printing process through real-time data collection, comprehensive feature extraction and advanced machine learning models, and has the following beneficial effects: 1. Improved printing quality and success rate, and realized real-time monitoring and dynamic adjustment.

[0018] The light-curing 3D printing fault prediction method of the present invention monitors various parameters of the printer (such as temperature, humidity, light intensity, material flow, etc.) in real time through sensors. Real-time monitoring can promptly detect abnormal situations, dynamically adjust printing parameters according to real-time data, optimize the printing process, avoid quality problems, and improve the printing success rate.

[0019] 2. Realize preventive maintenance and fault warning, facilitate preventive maintenance and reduce downtime.

[0020] The present invention avoids equipment damage and prolongs the service life of equipment by predicting failures in advance and performing maintenance and servicing in time; analyzes equipment operation data, identifies potential failure modes, reduces unexpected downtime, and improves equipment utilization and production efficiency.

[0021] 3. Improved production efficiency and realized intelligent early warning.

[0022] Through real-time data monitoring and dynamic adjustment of printing parameters, the printing process is optimized, production efficiency is improved, dependence on manual monitoring and intervention is reduced, and the level of production automation is improved; multi-dimensional data is collected through a variety of sensors, and pre-processing and feature extraction are performed to fully utilize the data to improve prediction accuracy; combined with time series characteristics, frequency domain characteristics and spatial characteristics, various data in the printing process are comprehensively analyzed to improve the accuracy of fault detection. It is suitable for different types of light-curing 3D printing equipment and various printing materials, and has broad application prospects.

[0023] In summary, the AI-driven photocuring 3D printing fault prediction method of the present invention can not only significantly improve printing quality and efficiency, but also effectively reduce costs, extend equipment life, and provide users with a better user experience and higher production capacity, with significant technical and economic benefits.

[0024] These and other aspects of the present invention will be more clearly understood in the following description of the embodiments. It should be understood that the above general description and the following detailed description are only exemplary and explanatory and cannot limit the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or related technologies, the following briefly introduces the drawings required for use in the exemplary embodiments or related technical descriptions. The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation to the present invention. In the drawings: Figure 1 This is a flow chart of the AI-driven photocuring 3D printing fault prediction method according to an embodiment of the present invention.

[0026] Figure 2 This is a flowchart of spatial feature extraction based on graph neural network in the AI-driven photocuring 3D printing fault prediction method in an embodiment of the present invention.

[0027] Figure 3 The present invention provides a flowchart of training and optimizing a fault prediction model constructed in the AI-driven photocuring 3D printing fault prediction method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0029] In some of the processes described in the specification and claims of the present invention and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0030] The following will be combined with the drawings in the exemplary embodiments of the present invention to clearly and completely describe the technical solutions in the exemplary embodiments of the present invention. Obviously, the exemplary embodiments described are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0031] The technical solution of the present invention is further described below in conjunction with specific embodiments: See also Figure 1 As shown, Figure 1 A flowchart of an AI-driven photocuring 3D printing fault prediction method provided by the present invention. An AI-driven photocuring 3D printing fault prediction method provided in one embodiment of the present invention comprises the following steps: Step S10: collecting printing parameters in real time based on sensors installed on the light-curing 3D printing device, and performing data preprocessing on the collected printing parameters through a data processing module to eliminate noise and abnormal values.

[0032] In this step, the light-curing 3D printing device is equipped with a resin tank temperature sensor, a light source intensity sensor, a vibration sensor, a position sensor and a current sensor, which are used to collect printing parameters during light-curing 3D printing: resin temperature, ultraviolet light intensity, vibration frequency, vibration amplitude, and Z-axis height.

[0033] In this embodiment, when the collected printing parameters are preprocessed, the collected printing parameters are subjected to moving average processing to determine the window size of the moving average; the average value of the moving window before and after the time is calculated for the printing parameters at each time point according to the sampling frequency of the data and the set smoothness level; smoothed time series data are obtained, and the resin temperature, ultraviolet light intensity, vibration data within the average set period of time, and the position signal of the Z-axis height are smoothed.

[0034] Among them, the data preprocessing of the collected printing parameters also includes eliminating spike noise through median filtering of smoothed time series data, taking the median of the data at each time point and the forward and backward moving window by determining the window size of the median filter, removing sudden abnormal values ​​in the resin temperature, filtering out abnormal vibration peaks, and removing occasional current spikes.

[0035] Among them, the data preprocessing of the collected printing parameters also includes identifying and processing abnormal values ​​in the data after median filtering. By calculating the mean and standard deviation of the data corresponding to each sensor, the resin temperature, ultraviolet light intensity, vibration frequency and vibration amplitude, and Z-axis height are Z-score standardized to obtain the preprocessed data.

[0036] In the above steps, the AI-driven photocuring 3D printing fault prediction method collects data through a variety of sensors and uses data preprocessing technology to eliminate noise and outliers, ensuring the high accuracy of the acquired printing parameter data and providing a reliable data basis for subsequent fault prediction.

[0037] Among them, when performing moving average processing, the moving average processing formula used is:

[0038] In the formula, is the moving average at time point t; is the window size of the determined moving average; The first Collect data at each time point.

[0039] When the median filter eliminates the spike noise, the calculation formula of the median filter value at time point t is:

[0040] in, is the median filter value at time point t; is the determined window size of the median filter; For the Collect data at each time point.

[0041] For example, assume that during a certain light-curing 3D printing process, the sensor collects the following data (partial data per unit time): Resin temperature (°C): 30, 31, 32, 35, 30, 29, 28, 27, 30, 29; UV light intensity (mW / cm²): 10, 11, 10, 12, 11, 10, 9, 8, 10, 11; Vibration frequency (Hz): 50, 51, 50, 52, 53, 54, 55, 56, 55, 54; Vibration amplitude (mm): 0.5, 0.6, 0.7, 0.8, 0.7, 0.6, 0.5, 0.4, 0.5, 0.6; Z-axis height (mm): 10, 10.1, 10.2, 10.3, 10.2, 10.1, 10, 9.9, 10, 10.1; Then the data is preprocessed, first of all, moving average processing is performed, where: Assuming window size =3, then the moving average of the resin temperature at the fifth time point is:

[0042] Then perform median filtering, where: Assuming window size =3, then the median filter value of the ultraviolet light intensity at the fifth time point is:

[0043] Finally, Z-score standardization is performed, where: Assuming that the mean value of the resin temperature μ=30°C and the standard deviation σ=2°C, the Z-score standardized value of the resin temperature at the fifth time point is:

[0044] In this embodiment, through the steps of sensor data collection, moving average processing, median filtering and Z-score standardization, noise and outliers can be effectively eliminated and data quality can be improved. The pre-processed data provides a reliable input basis for the AI ​​system, thereby improving the accuracy and timeliness of light-curing 3D printing fault prediction, and ultimately improving the printing quality and equipment operation stability.

[0045] Step S20: extracting multi-dimensional features of photocuring 3D printing from the pre-processed data, and performing temporal feature and spatial feature analysis to obtain a high-dimensional feature vector of photocuring 3D printing.

[0046] In this step, when extracting the multidimensional features of light-curing 3D printing, the smoothed time series data in the preprocessed data is segmented according to the time series, and the time series features of each sensor data in each time period are calculated. The extracted time series features include the mean value, standard deviation, the maximum and minimum values ​​of the extreme value changes in each time period, and the skewness and kurtosis of the data morphological distribution; the time series data is converted to the frequency domain to extract spatial features. The spatial features include the frequency domain features of the data extracted by Fourier transform and the power spectral density at the same frequency.

[0047] In this embodiment, temporal features and spatial features are analyzed to obtain a high-dimensional feature vector of light-curing 3D printing, including the following steps: aligning data from different sensors by timestamps, and splicing the extracted temporal features and spatial features into a high-dimensional feature vector.

[0048] Among them, see Figure 2 As shown, the AI-driven light-curing 3D printing fault prediction method also includes spatial feature extraction based on a graph neural network, including the following steps: Step S201, collecting sensor data of light-curing 3D printing and physical structure data of light-curing 3D printing equipment; Step S202: Modeling the physical structure of the light-curing 3D printer as a graph, where nodes represent sensors and edges represent data flows; Step S203: Map the sensor data to the corresponding graph nodes to form node feature vectors, wherein the feature vector of each node contains the multi-dimensional data collected by the node; Step S204: Based on the graph convolutional network as the graph neural network architecture, a graph neural network model is constructed, wherein the defined input layer receives node features, the hidden layer performs feature aggregation and transformation, and the output layer generates the final feature representation; Step S205: divide the preprocessed graph data into a training set, a validation set, and a test set, use the training set to train the graph neural network model, apply the trained graph neural network model to validate the test set graph data, and extract the high-dimensional feature vector of each node; Step S206: Fuse the extracted spatial features with the time series features and the frequency domain features to form a comprehensive feature vector, and use the fused comprehensive feature vector to train the fault prediction model.

[0049] In this embodiment, the AI-driven photocuring 3D printing fault prediction method further constructs a high-dimensional feature vector by extracting the temporal features and spatial features from the preprocessed data, and extracts the spatial features based on the graph neural network (GNN). Finally, the temporal features, spatial features and frequency domain features are fused to form a comprehensive feature vector for training and verification of the fault prediction model.

[0050] For example, when obtaining a high-dimensional feature vector of photocuring 3D printing, it is assumed that the following sensor data are collected during the photocuring 3D printing process: Resin temperature: 30℃, 31℃, 32℃, 33℃, 31℃, 30℃; Vibration frequency:50Hz,51Hz,50Hz,52Hz,51Hz,50Hz.

[0051] Then, when performing time series feature extraction, the mean and standard deviation of the resin temperature are calculated, where: The average value of resin temperature Mean = (30 + 31 + 32 + 33 + 31 + 30) ÷ 6 ≈ 31 ° C; Standard deviation of resin temperature Std= .

[0052] Then the time series features, frequency domain features and feature vectors based on graph neural networks are concatenated to form a comprehensive feature vector for fault prediction.

[0053] By combining time series features, frequency domain features and spatial features, especially introducing spatial structure modeling of graph neural networks, this AI-driven photocuring 3D printing fault prediction method can accurately capture the multi-dimensional features of the equipment operation process and improve the ability to predict potential faults. Through these complex feature extraction and fusion, the high-dimensional comprehensive feature vector finally generated provides richer information for the fault prediction model, improves the accuracy, robustness and timeliness of the prediction model, and ensures the stability and reliability of the printing process.

[0054] Step S30: annotate the collected historical printing data into normal data samples and fault data samples, and train and optimize the constructed fault prediction model in combination with the high-dimensional feature vector.

[0055] In this step, see Figure 3 As shown, the training and optimization of the constructed fault prediction model includes the following steps: Step S301, collecting historical printing data generated during the light-curing 3D printing process, including temperature, ultraviolet light intensity, vibration frequency, vibration amplitude and Z-axis height; Step S302: dividing the historical printing data into normal data samples and faulty data samples according to whether the data is abnormal or faulty within a time period, and randomly dividing them into a training set and a test set in proportion; Step S303: Use the high-dimensional feature vectors in the training set and the corresponding normal or faulty labels to train the selected model, adjust the hyperparameters of the model, and perform cross-validation through the test set to obtain a trained and optimized fault prediction model.

[0056] In this step, historical data generated by each sensor during the light-curing 3D printing process is collected, covering the following main data: Temperature: Temperature data inside the resin or equipment.

[0057] UV Intensity: The intensity of the UV light used in the curing process.

[0058] Vibration frequency and vibration amplitude: detect vibration data during equipment operation.

[0059] Z-axis height: The Z-axis movement of the device.

[0060] When annotating data, the data is annotated as normal samples and fault samples based on the abnormal conditions of the printing process in the historical data. Failure criteria can include equipment downtime, printing failure, or parameters exceeding the safe range. The data is randomly divided into training sets and test sets in proportion (80% for training and 20% for testing). When training the model, the support vector machine (SVM) classifier is used to optimize the target, and the random forest is used to adjust the hyperparameters of the model. The accuracy rate is used as the evaluation indicator.

[0061] Assume that the historical printing data of the light-curing 3D printer is as follows: Temperature: [30℃, 32℃, 33℃, 34℃, 32℃] UV light intensity: [100W / m2, 105W / m2, 108W / m2, 102W / m2, 110W / m2] Vibration frequency: [50Hz, 51Hz, 52Hz, 50Hz, 53Hz] The above-mentioned historical printing data are marked as normal data or fault data according to whether a fault occurs.

[0062] Then, the time series features (mean and standard deviation of temperature), frequency domain features (power spectrum density of vibration frequency) and spatial features are combined to form a comprehensive high-dimensional feature vector: Eigenvector 1: [31.5, 2.5, 105, 50.5, ...]; Eigenvector 2: [32, 1.8, 108, 51, ...]; These high-dimensional feature vectors are used together with the corresponding labels (normal or faulty) to train the model and optimize it through cross-validation. In the prediction of photocuring 3D printing faults, data preprocessing and high-dimensional feature extraction provide rich information for the model. By using classification models such as SVM, random forest, GBDT, etc., combined with appropriate hyperparameter adjustment and cross-validation, the model can effectively identify potential faults and improve the reliability and stability of printing equipment.

[0063] Step S40: Detect the real-time collected printing parameters based on the trained fault prediction model, compare the obtained prediction result with the preset warning threshold, and if it exceeds the threshold, issue a warning according to the fault classification result.

[0064] In this step, the real-time collected data is input into the trained fault prediction model. Through the learned features and patterns, a prediction value is generated to indicate whether the equipment is in a normal or faulty state. As the probability of a certain fault, a warning threshold is set as the critical point for "normal" or "faulty" judgment. For example, if the predicted fault probability value P (fault|X) exceeds a certain set threshold T, it indicates that the equipment is at risk of failure. If the fault probability output by the model exceeds the threshold, the warning mechanism is triggered. For example, when the predicted value exceeds the minor fault threshold but is lower than the major fault threshold, it can be judged as a moderate fault and a warning of the corresponding level is issued.

[0065] For example, based on the size of the predicted value, you can set a multi-level warning, where different fault classification thresholds are defined: , , .

[0066] Minor faults: ; Moderate failure: ; Serious failure: .

[0067] For example, assume that during the light-curing 3D printing process, the system collects the following real-time data: Temperature: 40℃; UV light intensity: 120W / m²; Vibration frequency: 55Hz; Z-axis height: 0.4mm.

[0068] The data is input into a trained random forest fault prediction model, and the model outputs a fault prediction probability of 0.8; If the system sets the warning threshold value as T=0.75, and the predicted value exceeds the preset threshold value, the system will judge it as a moderate fault and issue a warning to prompt the operator to check the equipment.

[0069] If the failure prediction value is 0.90, it exceeds the threshold of severe failure. =0.85, the system will shut down immediately and sound an alarm, requiring the equipment to be checked and repaired.

[0070] The AI-driven photocuring 3D printing fault prediction method of the present invention combines historical data with real-time collected data and uses a trained model to monitor the equipment operation status in real time. By comparing with the warning threshold, the system can issue a warning before the fault occurs, thereby reducing the risk of printing failure. The fault classification and automatic response mechanism make the system more intelligent and can provide effective solutions for different fault levels, greatly improving the reliability and production efficiency of the equipment.

[0071] It should be understood that, although described in a certain order, these steps are not necessarily performed in sequence in the above order. Unless there is clear explanation in this article, the execution of these steps does not have strict order restriction, and these steps can be performed in other orders. Moreover, a part of the steps of the present embodiment may include a plurality of steps or a plurality of stages, and these steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a part of the steps or stages in other steps or other steps.

[0072] In one embodiment, a computer device is also provided in an embodiment of the present invention, comprising at least one processor and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor performs the steps of the AI-driven photocuring 3D printing fault prediction method.

[0073] In one embodiment, the present invention further provides a computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable the computer to execute the steps of the AI-driven photocuring 3D printing fault prediction method.

[0074] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be realized by instructing the relevant hardware through a computer program represented by computer instructions, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory.

[0075] Non-volatile memory may include read-only memory, magnetic tape, floppy disk, flash memory or optical storage, etc. Volatile memory may include random access memory or external cache memory. As an illustration and not limitation, RAM may be in various forms, such as static random access memory or dynamic random access memory, etc.

[0076] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An AI-driven light-curing 3D printing fault prediction method, characterized in that: The method comprises the following steps: The sensors installed on the light-curing 3D printing equipment collect printing parameters in real time, and the collected printing parameters are pre-processed by the data processing module to eliminate noise and abnormal values; Extract the multi-dimensional features of photocuring 3D printing from the pre-processed data, and perform temporal and spatial feature analysis to obtain the high-dimensional feature vector of photocuring 3D printing; The collected historical printing data is labeled as normal data samples and fault data samples, and the constructed fault prediction model is trained and optimized by combining high-dimensional feature vectors; The printing parameters collected in real time are detected based on the trained fault prediction model, and the prediction results are compared with the preset warning threshold. If the threshold is exceeded, a warning is issued according to the fault classification result.

2. The AI-driven light-curing 3D printing fault prediction method according to claim 1, characterized in that: The light-curing 3D printing device is equipped with a resin tank temperature sensor, a light source intensity sensor, a vibration sensor, a position sensor and a current sensor, which are respectively used to collect printing parameters during light-curing 3D printing: resin temperature, ultraviolet light intensity, vibration frequency, vibration amplitude, and Z-axis height.

3. The AI-driven light-curing 3D printing fault prediction method according to claim 2, characterized in that: The data preprocessing of the collected printing parameters also includes eliminating spike noise through median filtering of the smoothed time series data. By determining the window size of the median filter, the median of the data at each time point and the moving window before and after is taken to remove sudden abnormal values ​​in the resin temperature, filter out abnormal vibration peaks, and remove occasional current spikes.

4. The AI-driven light-curing 3D printing fault prediction method according to claim 3, characterized in that: The data preprocessing of the collected printing parameters also includes identifying and processing outliers in the median filtered data. By calculating the mean and standard deviation of the data corresponding to each sensor, the resin temperature, ultraviolet light intensity, vibration frequency and vibration amplitude, and Z-axis height are Z-score standardized to obtain the preprocessed data.

5. The AI-driven light-curing 3D printing fault prediction method according to claim 4, characterized in that: When extracting the multidimensional features of light-curing 3D printing, the smoothed time series data in the preprocessed data is segmented according to the time series, and the time series features of each sensor data in each time period are calculated. The extracted time series features include the mean value, standard deviation, the maximum and minimum values ​​of the extreme value changes in each time period, and the skewness and kurtosis of the data morphological distribution; the time series data is converted to the frequency domain to extract spatial features. The spatial features include the frequency domain features of the data extracted by Fourier transform and the power spectral density at the same frequency.

6. The AI-driven light-curing 3D printing fault prediction method according to claim 5, characterized in that: The AI-driven light-curing 3D printing fault prediction method further includes performing spatial feature extraction based on a graph neural network, including the following steps: Collect sensor data of light-curing 3D printing and physical structure data of light-curing 3D printing equipment; The physical structure of the light-curing 3D printer is modeled as a graph, where nodes represent sensors and edges represent data flows; Map the sensor data to the corresponding graph nodes to form node feature vectors, where the feature vector of each node contains the multi-dimensional data collected by the node; Based on the graph convolutional network as the graph neural network architecture, a graph neural network model is constructed, in which the defined input layer receives node features, the hidden layer performs feature aggregation and transformation, and the output layer generates the final feature representation; The preprocessed graph data is divided into training set, validation set and test set. The graph neural network model is trained using the training set. The trained graph neural network model is applied to validate the test set graph data to extract the high-dimensional feature vector of each node. The extracted spatial features are fused with the time series features and frequency domain features to form a comprehensive feature vector, and the fused comprehensive feature vector is used to train the fault prediction model.

7. The AI-driven light-curing 3D printing fault prediction method according to claim 6, characterized in that: The temporal feature and spatial feature analysis are performed to obtain a high-dimensional feature vector of light-curing 3D printing, including the following steps: aligning data from different sensors according to timestamps, and feature splicing the extracted temporal features and spatial features into a high-dimensional feature vector.

8. The AI-driven light-curing 3D printing fault prediction method according to claim 7, characterized in that: The training and optimization of the fault prediction model include the following steps: Collect historical printing data generated during the light-curing 3D printing process, including temperature, UV light intensity, vibration frequency, vibration amplitude, and Z-axis height; The historical printing data is divided into normal data samples and fault data samples according to whether the data is abnormal or faulty within the time period, and randomly divided into training sets and test sets in proportion. The selected model is trained using the high-dimensional feature vectors in the training set and the corresponding normal or faulty labels, and the hyperparameters of the model are adjusted. The model is cross-validated through the test set to obtain the trained and optimized fault prediction model.

9. A computer device, characterized in that: include: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor executes the AI-driven photocuring 3D printing fault prediction method described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the AI-driven photocuring 3D printing fault prediction method according to any one of claims 1 to 8.

Images