3D printing wire production process monitoring data processing method and system based on Internet of Things

Abstract

The invention provides a 3D printing wire production process monitoring data processing method and system based on the Internet of Things, and relates to the technical field of printing wire production data processing, and the method comprises the steps: carrying out the quality interference monitoring of 3D printing wire data, whether the 3D printing wire feature loss condition and the qualification condition of the 3D printing wire feature dynamically captured by the model are analyzed or not is judged based on the output monitoring result; if yes, after analysis is finished, printing wire characteristic quality decision feedback is conducted, and if not, printing wire data source distortion optimization and printing wire data characteristic optimization are executed. According to the invention, the processing accuracy of the 3D printing wire production data can be improved, and the problem of low processing accuracy of the 3D printing wire production data in the prior art can be solved.

Description

Technical Field

[0001] This invention relates to the field of filament production data processing technology, and in particular to a method and system for monitoring and processing 3D printing filament production process data based on the Internet of Things. Background Technology

[0002] 3D printing filament is a building material produced using the fused deposition modeling (FDM) process, bridging digital models into tangible entities. The production of 3D printing filament is a continuous process integrating polymer plasticization, precision mechanics, and closed-loop control. Its core process begins with raw material feeding and preparation, where the main plastic granules, such as PLA (Polylactic Acid), are thoroughly mixed with the required masterbatch or modifying materials in a dry environment. Subsequently, raw material extrusion and cooling are performed. The raw material mixture is fed into a single-screw extruder for melting and homogenization. The fully plasticized and homogenized melt is pushed by the screw into a high-precision extrusion die for precise shaping. The high-temperature filament extruded from the die immediately enters a three-meter dual-stage heated water bath for initial warming and then a four-meter cooling water bath, thoroughly and uniformly cooling the filament to room temperature. Next, traction and stretching are performed. Based on the four-wheel traction machine located behind the cooling water tank, which provides stable and slip-free traction, the wire is pulled out of the mold at a uniform speed. The diameter gauge can measure the real-time diameter of the wire. The wire can be transmitted through the horizontal tension storage rack. The wire reaches the single-reel servo winding machine, which neatly and evenly winds the wire onto the reel. Before packaging, a final quality inspection is carried out, including diameter recheck, roundness test and printing performance test. Finally, it is sealed in moisture-proof packaging to complete the entire production process.

[0003] To achieve online monitoring of the 3D printing filament production process, improve production efficiency, reduce defect rates, and ensure product quality, it is necessary to monitor production quality based on 3D printing filament production data. Existing production monitoring data processing methods mainly integrate multiple sensors, wireless communication technologies, and data analysis techniques to monitor and analyze the 3D printing filament production process in real time. Firstly, sensors are deployed at key stages of the 3D printing filament production line (such as raw material feeding, raw material extrusion, cooling and shaping, and traction stretching) to collect multi-dimensional filament production data such as temperature, pressure, and tension. Then, this data is transmitted in real time to the cloud or local server using IoT technology. Big data analytics and machine learning algorithms are used to clean, process, and model the data. Specifically, preprocessing methods such as moving average, Kalman filtering, or wavelet denoising are used. Machine learning and deep learning algorithms, such as regression analysis and clustering, are used to build a quality prediction model, aiming to reveal the complex nonlinear relationship between process parameters (such as extrusion temperature and the average screw speed of six types of feeding screws) and the final filament quality (such as diameter consistency, roundness, and mechanical properties). Finally, potential anomalies or non-conforming conditions in the production process are identified, and key parameters such as traction speed, cooling air volume, or extrusion temperature are adjusted online through closed-loop feedback. In addition, data security and privacy protection are considered during the processing, and encryption technology is used to ensure the security of production data during transmission.

[0004] The above-mentioned technology has at least the following technical problems: In the production of 3D printing filaments, severe noise interference may occur under complex working conditions (such as cooling water flow fluctuations and mechanical vibrations), and dynamic changes in process parameters and quality indicators (such as the dynamic impact of changes in melt rheological properties caused by batch replacement of raw materials). This leads to disorder and distortion in the 3D printing filament data source itself. For example, effective features in the raw signals collected by sensors may be submerged by noise, resulting in disorder and distortion at the 3D printing filament data level. Furthermore, the correlation between the 3D printing filament data source and features may be weakened, manifesting as structural changes between 3D printing filament features due to changes in working conditions (such as correlation drift). Existing technologies typically use fixed thresholds to remove outliers (such as using a moving average algorithm and setting the wire diameter deviation between adjacent sampling points to >0.07). The use of outlier removal (mm) and static modeling has limitations. It may lead to the loss of key 3D printing filament feature information (such as pressure fluctuation details characterizing melt viscoelasticity changes, temperature gradient characteristics reflecting cooling crystallization dynamics, and high-frequency tension harmonic components indicating diameter perturbations) and the inability of the quality prediction model to accurately capture dynamic nonlinearity. This can result in deviations in the quality prediction model output. Using incorrect model outputs for state assessment may misjudge the stability of the production process, ultimately leading to low accuracy in 3D printing filament production data processing. Summary of the Invention

[0005] Therefore, embodiments of the present invention provide a method and system for monitoring and processing 3D printing filament production process data based on the Internet of Things, which can improve the accuracy of 3D printing filament production data processing.

[0006] The technical solution of this invention is implemented as follows: This invention provides a data processing method for monitoring the production process of 3D printing filament based on the Internet of Things (IoT), including: monitoring the quality interference of 3D printing filament data, and determining whether to analyze the loss of 3D printing filament features and the passability of the model's dynamic capture of 3D printing filament features based on the output monitoring results; if so, after the analysis, performing printing filament feature quality decision feedback to evaluate the predicted passability of 3D printing filament production quality; if not, performing printing filament data source distortion optimization and printing filament data feature optimization, and after the optimization is qualified, analyzing the loss of 3D printing filament features and the passability of the model's dynamic capture of 3D printing filament features.

[0007] This application also provides an IoT-based 3D printing filament production process monitoring data processing system. This system employs an IoT-based 3D printing filament production process monitoring data processing method. The system includes: a filament data quality interference monitoring module, a feature quality decision feedback module, and a filament data optimization module. The filament data quality interference monitoring module monitors the quality interference of 3D printing filament data and, based on the output monitoring results, determines whether to analyze the loss of 3D printing filament features and the passability of the model's dynamic capture of 3D printing filament features. If the feature quality decision feedback is performed, the module performs filament feature quality decision feedback after the analysis to evaluate the predicted passability of 3D printing filament production quality. If the feature quality decision feedback is not performed, the module performs filament data source distortion optimization and filament data feature optimization, and, after successful optimization, analyzes the loss of 3D printing filament features and the passability of the model's dynamic capture of 3D printing filament features.

[0008] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: 1. By monitoring the quality interference of 3D printing filament data and analyzing the output monitoring results to determine whether to analyze the loss of 3D printing filament features and the passability of the model's dynamic capture of 3D printing filament features, it is helpful to achieve accurate classification and judgment of interference in production data, reduce ineffective analysis of data sources without obvious interference, and improve the targeting and efficiency of 3D printing filament data processing. If so, after the analysis, the printing filament feature quality decision feedback is performed, which helps to achieve dual verification at the feature level and the quality level, and to correlate the feature analysis results with the actual production quality prediction. If not, the printing filament data source distortion optimization and printing filament data feature optimization are performed. After the optimization is qualified, the loss of 3D printing filament features and the passability of the model's dynamic capture of 3D printing filament features are analyzed, which helps to reduce the distortion of analysis results caused by noise and process parameter fluctuations, and ensure the accuracy of feature analysis and model evaluation.

[0009] 2. By monitoring the quality interference of 3D printing filament data, the noise impact-filament feature analysis value and the filament feature qualification assessment value are obtained. When the noise impact-filament feature analysis value exceeds the preset filament noise impact value, filament data source distortion optimization is performed. When the filament feature qualification assessment value does not meet the filament feature qualification conditions, filament data feature optimization is performed. Compared with the existing technology that performs indiscriminate optimization of noise and process parameter interference, this avoids new data distortion caused by over-adjustment of sensor gain or process sampling strategy. At the same time, it realizes differentiated targeted processing of noise interference and feature deviation, which helps to achieve accurate anti-interference optimization of 3D printing filament production data. This allows feature extraction and model modeling to be based on a higher quality data source, reduces feature loss and model capture failure caused by interference, and thus ensures the stability of the correlation between process parameters and quality features in the 3D printing filament production process, reducing the risk of production quality misjudgment caused by data interference.

[0010] 3. By performing qualification inspection at the 3D printing filament data level, the missing values ​​of printing filament production features and the qualified values ​​of printing filament feature capture are obtained. When the missing values ​​of printing filament production features meet the optimization conditions for printing feature decision, the printing feature loss optimization decision is executed. When the difference between the qualified value of printing filament feature capture and the preset qualified value of printing filament feature capture is not greater than 0, the prediction model is updated. Compared with the limitations of existing technologies with fixed feature extraction strategies and static models, this helps to improve the problem of not optimizing in time after feature loss or continuing to use outdated models. It realizes the quantitative triggering of feature optimization and model iteration, which helps to realize the dynamic adaptation of feature extraction window and the adaptive update of quality prediction model, so that the integrity of feature retention and the effectiveness of model capture always match the changes in production conditions.

[0011] 4. By obtaining 3D printing filament production prediction parameters through feature quality decision feedback, a feature quality feedback prompt is sent when the absolute value of the difference between the 3D printing filament production prediction parameter and the corresponding preset 3D printing filament production prediction parameter is within the predefined acceptable range for printing filament production. Conversely, a feature quality feedback prompt is sent when the absolute value is outside the acceptable range. Compared with the existing technology of manually verifying production quality prediction results offline, this reduces the generation of batch non-conforming products caused by the lag in the discovery of quality problems. It realizes real-time closed-loop verification of production quality prediction, which helps to achieve accurate early warning and rapid intervention for 3D printing filament production quality, thereby improving the controllability of 3D printing filament production quality and product yield, and ensuring the molding accuracy and performance stability of 3D printed products. Attached Figure Description

[0012] Figure 1 This is a flowchart of the IoT-based 3D printing filament production process monitoring data processing method provided in an embodiment of the present invention; Figure 2 This is a general overview diagram of the IoT-based 3D printing filament production process monitoring data processing method provided in this embodiment of the invention; Figure 3 This is a schematic diagram of the filament feature quality decision feedback in the IoT-based 3D printing filament production process monitoring data processing method provided in this embodiment of the invention. Figure 4 This is a schematic diagram illustrating the adaptation of the LSTM cell gating structure provided in this embodiment of the invention to the 3D printing filament production parameter prediction task. Figure 5 This is a schematic diagram of the structure of the IoT-based 3D printing filament production process monitoring data processing system provided in an embodiment of the present invention. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on the present invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] In the following description, references to "some embodiments" refer to a subset of all possible embodiments; however, it is understood that "some embodiments" can be the same or different subsets of all possible embodiments and can be combined with each other without conflict. Embodiments of the present invention provide a method for monitoring and processing data in the 3D printing filament production process based on the Internet of Things. For example... Figure 1 The flowchart shown is for a data processing method for monitoring the 3D printing filament production process based on the Internet of Things. The processing flow of this method may include the following steps: Filament data quality interference monitoring: In a specified 3D printing filament production monitoring scenario, 3D printing filament data quality interference monitoring is performed. Based on the output monitoring results, it is determined whether to analyze the loss of 3D printing filament features and the qualification of the model in dynamically capturing 3D printing filament features. By conducting filament data quality interference monitoring, it is helpful to accurately identify core issues such as noise interference (e.g., mechanical vibration, cooling water flow fluctuation), feature correlation drift, and data source distortion under complex working conditions. It can clarify data quality shortcomings and interference root causes, avoid blind optimization or ineffective analysis, and prevent low-quality data from flowing into subsequent analysis and modeling stages, thus ensuring the accuracy of data processing from the source.

[0015] Feature quality decision feedback: If performed, feature quality decision feedback will be conducted after the analysis to assess the predicted quality of 3D printing filament production. By conducting feature quality decision feedback, it is helpful to objectively assess the quality and risk level of 3D printing filament production based on the dual-dimensional analysis results of feature retention integrity and model dynamic capture effectiveness, avoiding misjudgments caused by relying solely on model prediction or feature analysis.

[0016] Filament data optimization: If not performed, filament data source distortion optimization and filament data feature optimization are performed. After successful optimization, the loss of 3D filament features and the model's ability to dynamically capture 3D filament features are analyzed. Filament data source distortion optimization is used to reduce the adaptability defects of 3D filament feature signals, thereby reducing filament data source distortion. Filament data feature optimization is used to match the variation of process parameters, thereby improving the model's ability to capture dynamic nonlinear relationships. By optimizing filament data, it is helpful to reduce data distortion rate and noise ratio, improve data integrity, continuity and reliability, provide a high-quality data foundation for feature analysis, improve the dynamic quality prediction model's ability to capture nonlinear relationships, alleviate the problem of decreased model adaptability caused by changes in operating conditions, and provide accurate model support for subsequent quality assessment and process parameter adjustment.

[0017] Before designing the IoT-based 3D printing filament production process monitoring data processing method provided in this application, a database storing various set data is established. The database includes, but is not limited to, preset printing filament noise impact values ​​and predefined data source distortion detection pass values. All values ​​are directly set by technical personnel. All data are directly set by technical personnel in combination with 3D printing filament industry standards, IoT sensor calibration data, multi-batch production measurement big data, and full-process process optimization experience. This is used to support the IoT system in real-time verification, feature analysis, anomaly judgment, optimization decision-making, and other automated processing links of production data, ensuring the accuracy, standardization, and efficiency of data processing.

[0018] like Figure 2The diagram shown is a general overview of the IoT-based 3D printing filament production process monitoring data processing method provided in an embodiment of this invention. Figure 2 It can be seen that: by monitoring the quality interference of 3D printing filament data, the noise impact-filament feature analysis value and the filament feature qualification assessment value are obtained; when the noise impact-filament feature analysis value is greater than the preset filament noise impact value, filament noise impact optimization is performed, otherwise, qualification detection at the 3D printing filament data level is performed; when the filament feature qualification assessment value meets the filament feature qualification conditions, qualification detection at the 3D printing filament data level is performed, otherwise, optimization of the impact of process parameter changes is performed; by performing qualification detection at the 3D printing filament data level, the filament feature capture qualification value is obtained, and when the difference between the filament feature capture qualification value and the preset filament feature capture qualification value is greater than 0, filament feature quality decision feedback is performed, otherwise, the prediction model is updated.

[0019] In this embodiment, the integrated approach of monitoring filament data quality interference, feature quality decision feedback, and filament data optimization helps to transform data processing from "static and passive" to "dynamic and proactive." The monitoring stage perceives changes in operating conditions and data in real time, the optimization stage dynamically adapts to changes and adjusts strategies, and the decision-making stage quickly responds to changes and outputs instructions. The three work together to ensure that data processing, feature analysis, and quality assessment are always adapted to the dynamic characteristics of 3D printing filament production (such as raw material batch changes and process parameter fluctuations). This minimizes the quality risks caused by data distortion, feature loss, and insufficient model adaptation, and helps to ensure that key features (such as high-frequency components of diameter perturbation and cooling crystallization temperature gradient) are not missed, model predictions are accurate, and quality decisions are correct during the production process.

[0020] Furthermore, quality interference monitoring of 3D printing filament data is conducted. The specific process is as follows: The impact of production noise on the 3D printing filament characteristic signals is monitored. The monitored noise impact - filament characteristic analysis value is used as feedback conditions for determining the impact of printing filament noise, and to decide whether to optimize the printing filament data source for distortion. The noise impact - filament characteristic analysis value is used to assess the distortion of the 3D printing filament data source caused by production noise. This is achieved by using the average amplitude of the 3D printing filament characteristic signal at each preset frequency point (such as the average frequency of high-frequency tension harmonic characteristic frequency points monitored by a frequency sweep meter) and the pre-set... The difference in amplitude of the corresponding preset 3D printing filament characteristic signal is used to represent the noise impact. The determination of the noise impact is as follows: it is determined whether the noise impact - printing filament characteristic analysis value is greater than 0. If it is greater than 0, the noise impact - printing filament characteristic analysis value is monitored. If it is not greater than 0, a noise impact qualification prompt is sent. It is determined whether the monitored noise impact - printing filament characteristic analysis value is greater than the preset printing filament noise impact value. If it is, printing filament data source distortion optimization is performed. Otherwise, qualification detection at the 3D printing filament data level is performed. The preset printing filament noise impact value is represented by the average value of the noise impact - printing filament characteristic analysis value over a historical time period.

[0021] Specifically, the optimization process for filament data source distortion is as follows: Checking the gain within acceptable range involves determining whether the initial gain of the filament production monitoring sensor is within the pre-set acceptable gain range. If so, a step-by-step gain reduction adjustment is triggered; otherwise, a sensor gain alarm is sent. The gain of the filament production monitoring sensor includes the gain of the temperature sensor, pressure sensor, and tension sensor used to monitor the 3D filament production process. The step-by-step gain reduction adjustment is performed as follows: The combination of the initial filament production monitoring sensor gain monitored by the spectrum analyzer, the noise impact-filament characteristic analysis value, the screw preset point pressure monitored by the pressure sensor, the traction machine preset point tension monitored by the tension sensor, and the average temperature of the 3D filament production process monitored by the temperature sensor is input into the optimal gain prediction table. The output includes the optimal temperature sensor gain, optimal pressure sensor gain, and optimal tension sensor gain. The optimal filament sensor gain combination is selected. Adjustment prompts are sent to designated personnel. Based on the initial gain of the filament production monitoring sensors, adjustments are made according to the corresponding optimal filament sensor gain combination, triggering 3D filament data source distortion detection. Specifically, this detection checks whether the difference between the noise impact and the filament characteristic analysis value in the initial and final states of filament data source distortion optimization exceeds a predefined data source distortion detection pass value. If so, a pass test is performed at the 3D filament data level; otherwise, a maintenance prompt is sent to the 3D filament production equipment (such as a single-screw extruder, high-precision extrusion die, three-meter dual-stage hot and cold water tank, four-meter cooling water tank, four-wheel traction machine, dual-axis diameter gauge, horizontal tension storage and winding rack, single-reel servo winding machine, etc.). The predefined data source distortion detection pass value is represented by the average difference between the noise impact and the filament characteristic analysis value in the initial and final states of filament data source distortion optimization over a historical time period.

[0022] It should be explained that by adjusting the gain of the initial filament production monitoring sensors based on the corresponding optimal combination of filament sensor gains—that is, adjusting to the optimal temperature sensor gain, optimal pressure sensor gain, and optimal tension sensor gain—it helps to improve the recognition and extraction efficiency of key feature signals. For example, the signal amplitudes of core features such as melt viscoelastic pressure fluctuations, cooling crystallization temperature gradients, and high-frequency tension harmonics of diameter perturbations are more stable. This provides higher-quality input data for dynamic feature association and online learning modeling, and promotes further upgrades in model prediction accuracy.

[0023] It should be added that the database constructed in the implementation scheme of this application covers a series of initially created tables or sets with mapping functions. The core advantages of these tables or sets are significant, mainly reflected in their ability to support two types of mapping relationships: on the one hand, they can achieve a precise one-to-one mapping between individual parameters to ensure the accuracy of information transmission; on the other hand, they can achieve a many-to-one mapping from multiple parameters to a single parameter to meet the data processing needs in complex scenarios.

[0024] The pre-defined process involves data collection, focusing on key operational information within a specific historical timeframe. This includes combinations of initial filament production monitoring sensor gains, noise impact-filament feature analysis values, screw preset point pressure, traction machine preset point tension, and average temperature during 3D filament production; combinations of process parameter change analysis data and parameter-filament feature analysis values; combinations of process parameter change analysis data, parameter-filament feature analysis values, and the total number of production time windows; combinations of missing filament production features and the number of 3D filament feature signals; and combinations of filament feature capture pass values ​​and the amount of data from the 3D filament data source. The collected historical key information is then input into a machine learning model (e.g., a decision tree model) that reveals feature importance. This model, through its feature splitting mechanism, deeply analyzes and processes the input information according to predefined mapping rules, extracting results that can be used as weight values ​​or other correlated data. These results include optimal filament sensor gain combinations, filament feature pass assessment values, combinations of production window sampling data and the number of adjustments corresponding to production window sampling operations, feature extraction window overlap area decision values, and preset feature update cycles.

[0025] The collected raw data is correlated and matched with the weights or related data output by the model to ensure that each set of raw data can accurately find the corresponding model output result. This ultimately generates the best gain prediction table, the pre-built parameter change impact comparison table, the pre-built production process window sampling reading table, the preset feature extraction window overlap area reading table, and the update cycle reading table, etc. In the actual production monitoring stage, when real-time monitoring and collection are performed on data such as the initial gain of the printing filament production monitoring sensor, the noise impact-printing filament characteristic analysis value, the screw preset point pressure, the traction machine preset point tension, and the average temperature of the 3D printing filament production stage, the process parameter change analysis data and the parameter-printing filament characteristic analysis value combination, the process parameter change analysis data, the parameter-printing filament characteristic analysis value combination and the total number of production stage time windows, etc., the data is collected. Information such as the combination of missing values ​​of filament production features and the number of 3D printing filament feature signals, and the combination of qualified filament feature capture values ​​and the amount of data from the 3D printing filament data source, can be directly read by inputting such real-time information into pre-built tables such as the optimal gain prediction table, the pre-built parameter change impact comparison table, the pre-built production process window sampling reading table, the preset feature extraction window overlap area reading table, and the update cycle reading table. With the help of the pre-set mapping relationships in the tables, the corresponding optimal filament sensor gain combination, filament feature qualification assessment value, combination of production process window sampling reading data and the number of adjustment times corresponding to the production process window sampling operation, feature extraction window overlap area decision value, and preset feature update cycle can be directly read, providing timely and reliable data support for real-time control and optimization of the production process.

[0026] In this embodiment, the impact of noise during 3D printing filament production on the characteristic signals of the 3D printing filament is monitored to obtain the noise impact-filament feature analysis value. When the monitored noise impact-filament feature analysis value is greater than the preset filament noise impact value, filament data source distortion optimization is performed; otherwise, a pass / fail test at the 3D printing filament data level is performed. This helps to accurately define the critical state of noise interference to the characteristic signals, avoiding the waste of computing power caused by blindly starting optimization when the noise does not exceed the standard, or the subsequent analysis deviation caused by the noise severely overwhelming the effective features but being missed. This improves the targeting of data processing and the efficiency of resource utilization. By linking the monitoring of noise during 3D printing filament production and the filament data source distortion optimization, a closed-loop linkage mechanism from noise interference identification to distortion-targeted repair is constructed. This provides clear "problem targets" for data source distortion optimization, making the optimization operation more targeted and avoiding indiscriminate optimization.

[0027] Furthermore, monitoring the quality interference of 3D printing filament data also includes: monitoring the impact of changes in 3D printing filament production process parameters on the characteristic signals of the 3D printing filament, acquiring process parameter change analysis data, and performing parameter change impact judgment based on the process parameter change analysis data; the process parameter change analysis data includes raw material feed rate change value reflecting changes in raw material feed rate during feed, raw material extrusion rate change value reflecting changes in raw material extrusion rate during extrusion, raw material setting temperature change value reflecting changes in setting temperature during cooling and setting, and raw material traction force change value reflecting changes in traction force during traction and stretching; the raw material feed rate change value is represented by the absolute value of the difference between the average raw material feed rate and the preset raw material feed rate during each preset raw material feed time period monitored by a mass flow meter and a feeder; the raw material extrusion rate change value is represented by the absolute value of the difference between the average raw material extrusion rate and the preset raw material extrusion rate during each preset raw material extrusion time period monitored by a mass flow meter and a feeder; the raw material setting temperature change value is represented by the average raw material temperature gradient and the preset raw material temperature during each cooling and setting time period monitored by a thermocouple. The gradient is represented by its absolute value; the change in raw material traction force is represented by the absolute value of the average traction tension and the preset traction tension during each traction stretching time period monitored by the tension sensor; the specific process for determining the impact of parameter changes is as follows: input the process parameter change analysis data and the parameter-printing filament feature analysis value into the pre-built parameter change impact comparison table, and output the printing filament feature qualification assessment value; the parameter-printing filament feature analysis value represents the average value of the Pearson correlation coefficient between the obtained 3D printing filament data source and the preset target features (such as filament diameter, printing filament grain size, etc.), which is used to reflect the qualification status of the correlation between the 3D printing filament data source and the features; determine whether the printing filament feature qualification assessment value meets the printing filament feature qualification conditions: if yes, perform qualification detection at the 3D printing filament data level; otherwise, perform printing filament data feature optimization; the printing filament feature qualification condition means that the printing filament feature qualification assessment value is greater than the predefined printing filament feature qualification assessment value, where the predefined printing filament feature qualification assessment value is represented by the average value of the printing filament feature qualification assessment values ​​over a historical time period.

[0028] It should be added that the specific process of optimizing the printing filament data features is as follows: First, the process parameter change analysis data, parameter-printing filament feature analysis values, and the total number of production stage time windows monitored by the counter are input into the pre-built production stage window sampling and reading table. The output is the production stage window sampling and reading data and the adjustment number corresponding to the production stage window sampling operation. The production stage window sampling and reading data includes the adjustment values ​​for the number of window sampling points in the raw material feeding stage, the raw material extrusion stage, the cooling and shaping stage, and the traction and stretching stage. The production stage window sampling operations include the step-by-step decrease of the number of window sampling points in the raw material feeding stage, the step-by-step decrease of the number of window sampling points in the raw material extrusion stage, the step-by-step decrease of the number of window sampling points in the cooling and shaping stage, and the traction and stretching stage. The process involves three steps: First, the number of sampling points in each window decreases progressively. Second, the amplitude corresponding to the data read from the production window sampling is used as the adjustment step size for the raw material feeding, raw material extrusion, cooling and shaping, and traction stretching stages, respectively, to perform production window sampling operations. Third, after performing a preset number of production window sampling operations, a pass / fail judgment is triggered: Judgment condition one, after performing a preset number of production window sampling operations, the number of monitored window sampling points is within the predefined pass range set by the personnel. Judgment condition two, after performing a preset number of production window sampling operations, the newly acquired filament feature pass evaluation value meets the filament feature pass conditions. When both judgment conditions one and two are triggered simultaneously, a pass / fail test is performed at the 3D filament data level; otherwise, a 3D filament production equipment maintenance prompt is sent.

[0029] By using the amplitude corresponding to the data sampled at each production stage window as the adjustment step size for the raw material feeding, raw material extrusion, cooling and shaping, and traction stretching stages, the number of window sampling points is gradually reduced in each of the following stages: raw material feeding, raw material extrusion, cooling and shaping, and traction stretching. This helps improve the targeting and efficiency of data sampling at each production stage, ensuring that the sampling frequency matches the intensity of characteristic signal fluctuations (e.g., when the temperature fluctuation amplitude is large in the cooling and shaping stage, the reduction amplitude of sampling points is reduced), and avoiding issues caused by... Overly dense sampling leading to feature redundancy or overly sparse sampling leading to feature loss helps improve the sensitivity of the correlation between feature signals and process parameters at each stage. When a preset number of production stage window sampling operations are simultaneously monitored, and the number of corresponding monitored window sampling points is within the predefined acceptable range, and the re-acquired filament feature qualification evaluation value meets the filament feature qualification conditions after a preset number of production stage window sampling operations, then performing qualification detection at the 3D filament data level helps avoid problems such as reasonable sampling but feature failure or feature compliance but sampling redundancy caused by a single condition trigger, thus controlling the quality from both the sampling structure and feature quality dimensions.

[0030] In this embodiment, by monitoring the impact of changes in 3D printing filament production process parameters on the 3D printing filament feature signals and obtaining filament feature qualification assessment values, when the monitored filament feature qualification assessment values ​​meet the filament feature qualification conditions, a qualification test is performed at the 3D printing filament data level; otherwise, filament data feature optimization is performed. This helps to accurately locate feature signal defects caused by dynamic changes in process parameters (such as feature correlation drift caused by parameter adjustments, weak features being masked by redundant information), ensuring that the feature signals always adapt to the dynamic fluctuations of process parameters. Through the mutual support of 3D printing filament production process parameter change monitoring and filament data feature optimization, it helps to continuously reduce the adaptation deviation between dynamic changes in process parameters and feature signals, improve the stability of feature distribution across operating conditions, and thus enhance the ability of the dynamic quality prediction model to capture nonlinear relationships.

[0031] Furthermore, the qualification detection at the 3D printing filament data level includes analyzing the loss of 3D printing filament features and analyzing the qualification of the model dynamically capturing 3D printing filament features. The specific process for analyzing the loss of 3D printing filament features is as follows: Based on the acquired loss value of printing filament production features, a judgment is initiated to determine whether the loss value meets the printing feature decision optimization conditions. If so, the printing feature loss optimization decision is executed; otherwise, the printing filament feature quality decision feedback is executed. The loss value of printing filament production features is represented by the difference between the preset number of 3D printing filament features (such as the number of pressure features) and the actual number of 3D printing filament features acquired by the counter. The actual number of 3D printing filament features acquired is represented by the sum of the number of pressure features (taking screw head pressure as an example), the number of temperature features (taking printing filament melt temperature as an example), and the number of tension features (taking traction tension harmonic amplitude as an example). The printing feature decision optimization condition is that the difference between the loss value of printing filament production features and the preset loss value of printing filament production features is greater than 0.

[0032] The optimization decision for printing feature loss includes the following: This optimization decision aims to reduce high-frequency feature breakage and loss, balancing feature extraction integrity with data processing efficiency. Measure one involves inputting the loss value of printing filament production features and the number of 3D printing filament feature signals monitored by the counter into a preset feature extraction window overlap area reading table for 3D printing filament features, outputting a feature extraction window overlap area decision value for the 3D printing filament features. Measure two uses the magnitude corresponding to the feature extraction window overlap area decision value as the adjustment step size for increasing the feature extraction window overlap area. This helps improve the temporal continuity and integrity of feature extraction, enhances the adaptability of different feature types, improves the adaptation efficiency between features and the model, and ultimately achieves core feature optimization. The completeness of the extraction; Measure 3: Within the predefined acceptable range of the overlapping area of ​​the feature extraction window, each time the overlapping area of ​​the feature extraction window is increased, the difference between the missing values ​​of the printing filament production features before and after the corresponding operation is re-monitored to see if it is within the pre-set acceptable window overlapping area increase range, and whether the re-acquired missing values ​​of the printing filament production features meet the printing feature decision optimization conditions. If the difference between the missing values ​​of the printing filament production features before and after the operation is within the pre-set acceptable window overlapping area increase range, and the missing values ​​of the printing filament production features meet the printing feature decision optimization conditions, the qualification status of the model's dynamic capture of 3D printing filament features is analyzed; otherwise, a 3D printing filament feature loss alarm is sent.

[0033] In this embodiment, analyzing the loss of 3D printing filament features and the pass / fail status of the model's dynamic capture of 3D printing filament features helps to accurately pinpoint the core shortcomings in the data processing and modeling stages, quantitatively evaluate the matching degree between data quality and model performance, and provide a clear target for subsequent optimization strategies. By first obtaining the loss value of printing filament production features through pass / fail detection at the 3D printing filament data level, and when the loss value of printing filament production features obtained after executing the printing feature loss optimization decision does not meet the printing feature decision optimization conditions, the pass / fail status of the model's dynamic capture of 3D printing filament features is further analyzed. This helps to build a hierarchical problem investigation mechanism of "feature priority and model supplementation", avoid model misjudgment caused by feature defects, and improve the reliability of quality decisions.

[0034] Furthermore, the pass / fail status of the model's dynamic capture of 3D printing filament features is analyzed. The specific process is as follows: Based on the number of successfully captured qualified 3D printing filament features and the total number of 3D printing filament features monitored by the counter, the corresponding ratio is marked as the qualified value of filament feature capture, which reflects the pass / fail status of the preset quality prediction model's capture of 3D printing filament features. A judgment is made based on the difference between the qualified value of filament feature capture and the preset qualified value of filament feature capture: if the difference is greater than 0, a filament feature quality decision feedback is performed, where the preset qualified value of filament feature capture is represented by the average value of filament feature capture qualified values ​​over a historical time period; otherwise, the prediction model is updated, and after the prediction model update is completed, the re-monitored qualified value of filament feature capture is determined. If the difference between the result and the preset qualified value of filament feature capture is greater than 0, a filament feature quality decision feedback is performed; if it is not greater than 0, a prediction model update alarm is sent. The prediction model update is used to dynamically adapt the model capture effect and improve the model's adaptability to dynamic working conditions. Specifically, the qualified value of filament feature capture and the amount of data from the 3D filament data source monitored by the network analyzer are input into the pre-built update cycle reading table of the 3D filament data source. The preset feature update cycle is read, and the corresponding adjustment prompt is sent to the preset personnel. Based on the initial update cycle of the 3D filament data source, the update cycle of the 3D filament data source is adjusted to the preset feature update cycle. This helps to achieve the synchronization of the 3D filament data source and the dynamic changes of features, as well as the timeliness of the original data, thereby helping to improve the analysis efficiency of the dynamic feature association module.

[0035] It should be added that, such as Figure 3The diagram illustrates the filament feature quality decision feedback of the IoT-based 3D printing filament production process monitoring data processing method provided in this invention application. The specific process of the filament feature quality decision feedback is as follows: The acquired 3D printing filament features are input into a preset quality prediction model, which outputs 3D printing filament production prediction parameters. These parameters include the filament extrusion speed and the average screw rotation speed. Based on these parameters, the 3D printing filament production quality is judged as qualified, specifically as follows: If the absolute value of the difference between the filament production prediction parameter and the corresponding preset filament production prediction parameter is within a predefined range of qualified filament production parameters set in advance by a preset person, a qualified feature quality feedback is sent; if the absolute value of the difference is not within the predefined range of qualified filament production parameters, a unqualified feature quality feedback is sent. The preset filament production prediction parameters include the preset filament extrusion speed and the preset average screw rotation speed.

[0036] Specifically, the dataset of 3D printing filament features (pressure features, temperature features, and tension features, etc.) collected over historical time periods and the preset 3D printing filament production prediction parameters are randomly divided into training sets by preset personnel and then input into a machine learning model, such as LSTM (Long Short-Term Memory), and trained based on the Adam optimizer and MSE (Mean Squared Error) loss function to obtain the preset quality prediction model; by inputting the re-collected 3D printing filament features into the preset quality prediction model, the 3D printing filament production prediction parameters are output.

[0037] like Figure 4 The diagram shown illustrates the adaptation of the LSTM cell gating structure provided in this embodiment of the invention to the 3D printing filament production parameter prediction task. Figure 4 It can be seen that: input terminal X t Corresponding to the real-time acquisition of pressure, temperature, and tension characteristics of the 3D printing filament, c t-1 h is the memory of the production characteristics of the preceding time step. t-1 These are the preceding prediction parameters; the gating modules in the figure all use the Sigmoid activation function (labeled σ in the figure), whose output range is 0-1, used to control the proportion of information retained—for example, the σ output of the forget gate determines the degree of discarding invalid noise in historical features, the σ output of the input gate determines the retention weight of the current feature, and the σ output of the output gate controls the proportion of effective information passed to the prediction result; the gating module combines the Tanh activation function and vector operations to complete feature fusion; the output c t Used to store the filtered, valid temporal features, which are then passed to the next time step for subsequent prediction. tThe parameters for 3D printing filament production prediction (preset printing filament extrusion speed and preset screw average speed) are reused as inputs for subsequent time steps. This figure clearly shows the logic of the LSTM unit capturing the long-term dependence of production time sequence features through the σ-gating mechanism, which is the core structural support for the high-precision prediction of this invention.

[0038] It should be explained that by obtaining 3D printing filament production prediction parameters through filament feature quality decision feedback, and sending a feature quality feedback qualified prompt when the absolute value of the difference between the 3D printing filament production prediction parameter and the preset 3D printing filament production prediction parameter is within the predefined qualified range of printing filament production, a feature quality feedback qualified prompt is sent; otherwise, a feature quality feedback unqualified prompt is sent. This helps to accurately locate the quality risk level, strengthen the traceability of the production process, promote the rapid implementation of closed-loop regulation, and improve the transparency and efficiency of production control.

[0039] like Figure 5 The diagram shown is a structural schematic of the IoT-based 3D printing filament production process monitoring data processing system provided in an embodiment of this invention. The IoT-based 3D printing filament production process monitoring data processing system is used to implement an IoT-based 3D printing filament production process monitoring data processing method, including: a filament data quality interference monitoring module, a feature quality decision feedback module, and a filament data optimization module. The filament data quality interference monitoring module is used to monitor the quality interference of 3D printing filament data in a specified 3D printing filament production monitoring scenario, and based on the output monitoring results, to determine whether to address the loss of 3D printing filament features and the passability of the model's dynamic capture of 3D printing filament features. The analysis module includes a feature quality decision feedback module, which, if performed, evaluates the predicted 3D printing filament production quality compliance after the analysis. The filament data optimization module, if not performed, performs filament data source distortion optimization and filament data feature optimization. After successful optimization, it analyzes the loss of 3D printing filament features and the compliance of the model's dynamic capture of 3D printing filament features. Filament data source distortion optimization reduces 3D printing filament feature signal adaptability defects, thereby reducing 3D printing filament data source distortion. Filament data feature optimization matches the variation range of process parameters, thereby improving the model's ability to capture dynamic nonlinear relationships.

[0040] In this embodiment, the pass / fail value of filament feature capture is obtained by analyzing the pass / fail status of the model's dynamic capture of 3D printing filament features. When the difference between the pass / fail value and the preset pass / fail value is greater than 0, filament feature quality decision feedback is performed; otherwise, the prediction model is updated. This helps to accurately define the model's performance boundaries, promptly correct model dynamic adaptation defects, enhance the reliability of quality decisions, and reduce the probability of production misjudgments. By analyzing the interaction between the pass / fail status of the model's dynamic capture of 3D printing filament features, prediction model updates, and filament feature quality decision feedback, a closed-loop linkage mechanism from performance evaluation to model iteration and then to quality decision-making is established. This dynamically balances model stability and adaptability, improves the long-term stability of model prediction accuracy, and thus ensures the continuity and consistency of 3D printing filament production quality, reducing batch quality defects.

[0041] In summary, the embodiments of this invention monitor the quality interference of 3D printing filament data and, based on the output monitoring results, determine whether to analyze the loss of 3D printing filament features and the passability of the model's dynamic capture of 3D printing filament features. This helps to achieve accurate classification and judgment of interference in production data, reduce invalid analysis of data sources without obvious interference, and improve the targeting and efficiency of 3D printing filament data processing. If this is done, after the analysis, a printing filament feature quality decision feedback is performed, which helps to achieve dual verification at the feature level and the quality level, and correlate the feature analysis results with actual production quality prediction. If this is not done, printing filament data source distortion optimization and printing filament data feature optimization are performed. After the optimization is qualified, the loss of 3D printing filament features and the passability of the model's dynamic capture of 3D printing filament features are analyzed, which helps to reduce the distortion of analysis results caused by noise and process parameter fluctuations, and ensure the accuracy of feature analysis and model evaluation.

[0042] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0043] It should be understood that determining B based on A does not mean determining B solely based on A; it also means determining B based on A and / or other information.

[0044] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0045] The above description is only an optional embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A data processing method for monitoring the production process of 3D printing filament based on the Internet of Things, characterized in that, The method includes: Perform quality interference monitoring on 3D printing filament data, and determine whether to analyze the loss of 3D printing filament features and the qualification of the model's dynamic capture of 3D printing filament features based on the output monitoring results; If performed, after the analysis is completed, a filament feature quality decision feedback will be conducted to assess the predicted 3D printing filament production quality compliance. If not, perform filament data source distortion optimization and filament data feature optimization, and after the optimization is qualified, analyze the loss of 3D filament features and the qualification of the model to dynamically capture 3D filament features.

2. The IoT-based 3D printing filament production process monitoring data processing method as described in claim 1, characterized in that, The specific process for monitoring quality interference in 3D printing filament data is as follows: The impact of noise during 3D printing filament production on the characteristic signals of 3D printing filament is monitored. The noise impact minus the filament characteristic analysis value is used as the feedback condition for determining the impact of filament noise and whether to optimize the data source distortion of filament. The noise impact-printing filament feature analysis value is represented by the average amplitude of the 3D printing filament feature signal at each preset frequency point and the difference between the preset amplitude of the 3D printing filament feature signal. The determination of the impact of printing filament noise is as follows: Determine if the noise impact - filament characteristic analysis value is greater than 0. If it is greater than 0, continue monitoring the noise impact - filament characteristic analysis value. If it is not greater than 0, send a noise impact qualified prompt. Determine whether the monitored noise impact - filament feature analysis value is greater than the preset filament noise impact value. If so, perform filament data source distortion optimization; otherwise, perform 3D filament data level pass inspection.

3. The IoT-based 3D printing filament production process monitoring data processing method as described in claim 2, characterized in that, The specific process for optimizing the distortion of the printing filament data source is as follows: Determine whether the gain of the initial filament production monitoring sensor is within the set acceptable gain range. If so, trigger the gain to be gradually reduced and adjusted; otherwise, send a sensor gain alarm. The gain of the filament production monitoring sensor includes the gain of the temperature sensor, the gain of the pressure sensor, and the gain of the tension sensor. The gain is adjusted by decreasing it step by step, and the specific method is as follows: The combination of the initial filament production monitoring sensor gain, noise impact-filament characteristic analysis value, screw preset point pressure, traction machine preset point tension, and average temperature of the 3D filament production process is input into the optimal gain prediction table, and the output includes the optimal filament sensor gain combination including the optimal temperature sensor gain, the optimal pressure sensor gain, and the optimal tension sensor gain. An adjustment prompt is sent to a preset personnel. Based on the initial gain of the filament production monitoring sensor, the adjustment is made according to the corresponding optimal combination of filament sensor gains, and a 3D filament data source distortion detection is triggered. Specifically, the 3D filament data source distortion detection involves determining whether the difference between the noise impact and the filament feature analysis value in the initial and final states of filament data source distortion optimization is greater than a predefined data source distortion detection pass value. If so, a pass detection is performed at the 3D filament data level; otherwise, a 3D filament production equipment maintenance prompt is sent.

4. The IoT-based 3D printing filament production process monitoring data processing method as described in claim 2, characterized in that, The process of monitoring quality interference in 3D printing filament data also includes: Monitoring the impact of changes in 3D printing filament manufacturing process parameters on the characteristic signals of 3D printing filaments. Acquire process parameter change analysis data, and based on the process parameter change analysis data, perform parameter change impact determination; The process parameter change analysis data includes changes in raw material input, raw material extrusion, raw material setting temperature, and raw material traction force. The specific process for determining the impact of parameter changes is as follows: Input the process parameter change analysis data and parameter-printing filament characteristic analysis values ​​into the pre-constructed parameter change impact comparison table, and output the printing filament characteristic qualification assessment value; The parameter-printing filament feature analysis value represents the average value of the Pearson correlation coefficient between the obtained 3D printing filament data source and the preset target features; Determine whether the filament characteristic acceptance assessment values ​​meet the filament characteristic acceptance criteria: If yes, perform a pass / fail test on the 3D printing filament data; otherwise, perform filament data feature optimization. The filament feature qualification condition means that the filament feature qualification assessment value is greater than the predefined filament feature qualification assessment value.

5. The IoT-based 3D printing filament production process monitoring data processing method as described in claim 4, characterized in that, The specific process for optimizing the filament data features is as follows: The first step is to input the process parameter change analysis data, parameter-printing filament characteristic analysis value, and the total number of production time windows into the pre-built production window sampling and reading table, and output the production window sampling and reading data and the number of adjustments corresponding to the production window sampling operation. The production process window sampling data includes the adjustment values ​​for the number of window sampling points in the raw material feeding stage, the raw material extrusion stage, the cooling and shaping stage, and the traction and stretching stage. The production process window sampling operation includes a step-by-step decrease in the number of window sampling points in the raw material feeding stage, a step-by-step decrease in the number of window sampling points in the raw material extrusion stage, a step-by-step decrease in the number of window sampling points in the cooling and shaping stage, and a step-by-step decrease in the number of window sampling points in the traction and stretching stage. The second step is to use the amplitude corresponding to the data sampled and read from the production process window as the adjustment step size for the raw material feeding, raw material extrusion, cooling and shaping, and traction stretching processes, respectively, and to perform production process window sampling operations. The third step involves performing a preset number of production process window sampling operations, followed by triggering a pass / fail judgment for the window sampling operation. Judgment condition one: After performing a preset number of production process window sampling operations, the number of corresponding monitored window sampling points is within the predefined qualified range; Judgment condition two: After performing a preset number of production process window sampling operations, the newly acquired filament feature qualification assessment value meets the filament feature qualification condition; When both discrimination condition one and discrimination condition two are detected simultaneously, a pass / fail test is performed on the 3D printing filament data; otherwise, a maintenance reminder for the 3D printing filament production equipment is sent.

6. The IoT-based 3D printing filament production process monitoring data processing method as described in claim 5, characterized in that, The pass / fail detection at the 3D printing filament data level includes analyzing the loss of 3D printing filament features and analyzing the pass / fail status of the model dynamically capturing 3D printing filament features. The analysis of feature loss in 3D printing filaments is performed as follows: Based on the obtained missing values ​​of printing filament production features, determine whether the missing values ​​of printing filament production features meet the conditions for printing feature decision optimization. If so, execute the printing feature loss optimization decision; otherwise, execute the printing filament feature quality decision feedback. The loss value of the filament production feature is represented by the difference between the preset number of 3D filament features and the actual number of 3D filament features acquired. The printing feature decision optimization condition is that the difference between the missing value of the printing filament production feature and the preset missing value of the printing filament production feature is greater than 0.

7. The IoT-based 3D printing filament production process monitoring data processing method as described in claim 6, characterized in that, The print feature loss optimization decision includes the following: Input the missing value of the filament production feature and the number of 3D filament feature signals into the preset feature extraction window overlap area reading table of the 3D filament feature, and output the feature extraction window overlap area decision value of the 3D filament feature; The magnitude corresponding to the decision value of the overlapping area of ​​the feature extraction window is used as the adjustment step size for performing the operation of increasing the overlapping area of ​​the feature extraction window; Within the predefined acceptable range of the overlapping area of ​​the feature extraction window, each time the overlapping area of ​​the feature extraction window is increased, the difference between the missing values ​​of the printing filament production features before and after the corresponding operation is re-monitored to see if it is within the preset acceptable window overlapping area increase range, and whether the re-acquired missing values ​​of the printing filament production features meet the printing feature decision optimization conditions. If the difference between the missing values ​​of the printing filament production features before and after the operation is within the preset acceptable window overlapping area increase range, and the missing values ​​of the printing filament production features meet the printing feature decision optimization conditions, the acceptable status of the model's dynamic capture of 3D printing filament features is analyzed; otherwise, a 3D printing filament feature loss alarm is sent.

8. The IoT-based 3D printing filament production process monitoring data processing method as described in claim 7, characterized in that, The analysis of the passability of the model's dynamic capture of 3D printing filament features is as follows: Based on the number of successfully captured qualified 3D printing filament features and the total number of 3D printing filament features, the corresponding ratio is marked as the qualified value for filament feature capture. The determination is performed based on the difference between the qualified value of the filament feature capture and the preset qualified value of the filament feature capture: If the difference between the qualified value of the filament feature capture and the preset qualified value of the filament feature capture is greater than 0, the filament feature quality decision feedback is performed. Conversely, if the prediction model is not updated, the difference between the newly monitored qualified value of the filament feature capture and the preset qualified value of the filament feature capture is greater than 0. If it is greater than 0, the filament feature quality decision feedback is performed. If it is not greater than 0, the prediction model update alarm is sent. The prediction model update is specifically as follows: inputting the qualified value of the filament feature capture and the amount of data from the 3D filament data source into the pre-built update cycle reading table of the 3D filament data source, reading the preset feature update cycle, and prompting the preset personnel to adjust the update cycle of the 3D filament data source to the preset feature update cycle based on the initial update cycle of the 3D filament data source.

9. The IoT-based 3D printing filament production process monitoring data processing method as described in claim 8, characterized in that, The specific process of the printing filament feature quality decision feedback is as follows: The acquired 3D printing filament features are input into a preset quality prediction model, and the 3D printing filament production prediction parameters are output. The 3D printing filament production prediction parameters include the printing filament extrusion speed and the average screw rotation speed; 3D printing filament production quality assessment based on 3D printing filament production prediction parameters: If the absolute value of the difference between the 3D printing filament production prediction parameter and the corresponding preset 3D printing filament production prediction parameter is within the predefined qualified range of printing filament production, a qualified feature quality feedback prompt will be sent. If the absolute value of the difference between the 3D printing filament production prediction parameter and the corresponding preset 3D printing filament production prediction parameter is not within the predefined acceptable range for printing filament production, a feature quality feedback failure message will be sent. The preset 3D printing filament production prediction parameters include the preset printing filament extrusion speed and the preset average screw rotation speed.

10. A data processing system for monitoring the production process of 3D printing filament based on the Internet of Things (IoT), used to implement the data processing method for monitoring the production process of 3D printing filament based on the IoT as described in any one of claims 1-9, characterized in that, The system includes: a filament data quality interference monitoring module, a feature quality decision feedback module, and a filament data optimization module; The filament data quality interference monitoring module is used to monitor the quality interference of 3D printing filament data, and based on the output monitoring results, it determines whether to analyze the loss of 3D printing filament features and the qualification of the model's dynamic capture of 3D printing filament features. The feature quality decision feedback module is used to perform feature quality decision feedback on the printing filament after the analysis is completed, if performed, to evaluate the predicted 3D printing filament production quality qualification status. The filament data optimization module is used to perform filament data source distortion optimization and filament data feature optimization if not performed, and to analyze the loss of 3D filament features and the pass rate of model dynamic capture of 3D filament features after optimization is qualified.

Images