Large-size light-cured forming feed defect detection and repair method

Abstract

The application discloses a large-size light-cured forming feed defect detection and repair method, and belongs to the field of light-cured 3D printing. In view of the problems of traditional process detection lag, poor targeted repair, insufficient interlayer bonding force and material waste, a sealed pressure storage tank and a pressure regulating valve are used to build a feed hardware, and an industrial camera and an infrared range finder are used to form a multi-modal layer-by-layer detection module. Through CNN fusion of image and three-dimensional topography data, defects such as material shortage, material overflow, bubbles and scratches are accurately identified according to the height deviation threshold. Different defects are targeted for repair: material shortage is pre-filled and compensated, material overflow is scraped and recycled, and the bubble area is de-bubbled; after repair, secondary detection is performed, and micron-level deviation is qualified, forming a double closed-loop control of detection-identification-repair-verification. The application can detect and repair feed defects online in real time, improve the large-size printing forming precision, interlayer bonding strength and printing success rate, reduce material loss, and is suitable for large-format light-cured 3D printing industrial application.

Description

Technical Field

[0001] This invention relates to the field of additive manufacturing photopolymerization 3D printing technology, specifically to a method for online detection and precise repair of defects in the feeding system of large-sized formed workpieces in SLA and DLP processes. Background Technology

[0002] Photopolymer 3D printing technology, which uses laser or projection light sources to irradiate photosensitive resin and cure it layer by layer, has been widely used in industrial parts, mold manufacturing, and other fields. As the forming size expands, the stability of the feeding system becomes increasingly important for the final precision of the formed product. In large-size applications, the resin storage tank needs to hold more resin, making it prone to defects such as overflow, uneven feeding, localized material shortages, and residual air bubbles due to differences in resin flowability, fluctuations in feeding pressure, and aging of seals. Furthermore, the greater layer thickness of large-size workpieces means that traditional offline inspection can easily lead to significant material waste, and insufficient interlayer bonding after repair can result in reduced workpiece strength.

[0003] Currently, the detection of defects and quality in the photopolymer 3D printing process is somewhat lagging. Most defect detection relies on manual inspection after printing, which cannot be intervened in real time. Furthermore, the repair methods are not targeted effectively; some defect detection solutions only target individual defects and lack compensation measures after defect detection. The "material feeding-detection-repair" process is disconnected, lacking coordinated control of dynamic parameters of the material feeding system and defect types. Therefore, an integrated method for "real-time detection-precise identification-targeted repair" adapted to large-size photopolymerization is needed. Summary of the Invention

[0004] The purpose of this invention is to solve the problem of real-time detection and accurate classification of defects (overflow, insufficient material, uneven material supply, and residual air bubbles) in the large-size photopolymer 3D printing process; at the same time, it classifies different types of defects and designs differentiated repair schemes for different defect types to solve the quality problems of molded parts caused by defects during the printing process, such as surface roughness, dimensional deviation, and poor interlayer bonding; and finally achieves coordinated control of the material supply system, detection module, and repair module to realize a closed-loop repair scheme of "detection-repair-material supply adjustment".

[0005] The objective of this invention is achieved through the following technical solution:

[0006] A method for detecting and repairing defects in the material supply process of large-size photopolymerization molding is proposed. This method detects and classifies defects generated during the material supply process in the printing process, and repairs the identified defects in a targeted manner by adjusting the temperature, coordinating the scraper, and calling the vacuum pump. This achieves a temperature-visual dual closed loop for identifying and repairing defect types, effectively improving the forming quality and printing success rate of photopolymerization 3D printed parts.

[0007] The method specifically includes the following steps:

[0008] Step S1: Build the core hardware system for large-scale molding, including supporting system hardware such as sealed pressure storage tank and pressure regulating valve;

[0009] Step S2: Complete the installation of the detection components, path planning and accuracy calibration. The multi-module defect detection device includes a visual inspection device and an infrared ranging device to assist in the identification of different defect types.

[0010] Step S3: Start the device printing and simultaneously trigger the detection module to collect and process data from various areas of the large-format printer;

[0011] Step S4: The control unit performs real-time analysis of the collected data to complete defect identification and classification;

[0012] Step S5: For different defect types identified, activate the corresponding repair module and simultaneously optimize parameters such as material supply and vacuum level to repair the identified defects;

[0013] Step S6: After the repair is completed, a second inspection is performed to ensure that the defects are eliminated and to optimize the subsequent process, so as to achieve closed-loop verification and model optimization;

[0014] In step S1, the large-capacity sealed pressure storage tank is used to store the resin in a sealed environment to prevent the resin from coming into contact with air and affecting the resin quality. At the same time, the large-capacity resin can continuously supply resin to the large-format and large-size printing process, avoiding the phenomenon of interruption of the material supply and termination of the printing process. The pressure regulating valve is used to regulate the pressure inside the tank to ensure that the resin output inside the tank is stable and controllable, prevent uneven material supply due to pressure fluctuations, and ensure that its viscosity and flowability are always at the optimal forming window.

[0015] In step S2, the visual inspection device uses a high-resolution industrial camera to accurately capture bubbles, scratches, and uneven thickness on the surface of the material layer; the infrared ranging module acquires the material layer height data in real time through non-contact scanning, and performs multi-source fusion analysis with the visual image to more accurately distinguish different defect types.

[0016] In step S3, the printing layer thickness, conventional exposure power, and exposure time are set as basic parameters, and the multimodal detection module is synchronously triggered to scan layer by layer according to the preset path. After each material feeding and leveling is completed, the system automatically pauses once and performs data acquisition to capture defects before and after printing and before and after resin curing. State differences are captured to ensure that minor deformations and uncured abnormalities are captured in time. Visual image data and infrared ranging height data are synchronously transmitted to the control unit.

[0017] In step S4, the control unit fuses the liquid surface morphology features of the acquired visual image with infrared height data and inputs it into the preloaded CNN defect recognition module for defect recognition. It directly identifies easily detectable phenomena such as printing pits and bubbles. For defects that are difficult to determine, a defect discrimination threshold is set. If the resin height deviation exceeds the set discrimination range, it is determined that there is a shortage / overflow of material on the liquid surface. According to the defect characteristics, the defects are classified into several categories such as shortage defects, overflow defects, bubble defects, and scratch defects, and the corresponding repair strategy is triggered based on the classification results.

[0018] In step S5, based on the defect classification results, the system automatically retrieves the corresponding repair strategy library: if the defect is determined to be a material shortage, the feeding scraper is called to replenish resin to the platform; if the defect is an overflow, the feeding scraper is called a second time to level the liquid surface, adjust the resin liquid level in the overflow area to the average height, recover the overflowed resin and send it back to the feeding system in real time to calibrate the resin output of the next layer; when a bubble defect is identified, the vacuum pump of the vacuum scraper is called to perform regional defoaming operation, and at the same time, the control device is adjusted appropriately to reduce the filling resin temperature, increase the resin viscosity, reduce the resin flow rate, and avoid the regeneration of bubbles.

[0019] In step S6, the secondary repair verification closed-loop effect refers to the system immediately starting a secondary multimodal scan after the repair action is performed, comparing the deviation of the liquid surface morphology and height data before and after repair; if the deviation converges to the specified value and the defect features disappear in the visual image, the repair is determined to be successful, and the next layer of printing continues; if the detection result shows that the defect has not been eliminated, return to S5 to perform targeted repair again until the accuracy requirements are met; record the defect type, location, size and corresponding repair parameters in each printing process to optimize the identification speed and repair parameter matching degree of subsequent defects of the same type.

[0020] The beneficial effects of this invention are as follows:

[0021] The detection module moves with the optical engine, enabling it to detect and identify quality issues in each printing area before and after printing in real time and provide real-time feedback, thus solving the lag problem of traditional offline detection.

[0022] It can repair printing defects in real time, achieve precise material control by replenishing missing material and recovering overflow material, avoid resin waste during the printing process, avoid printing failures caused by defects, and save printing costs.

[0023] The system utilizes a dual closed-loop mechanism of material feeding and vision to repair printing defects, making it compatible with all large-format printing and various photosensitive resins.

[0024] Combined with infrared ranging and visual imaging, dual data recognition can accurately identify different defect types. With differentiated repair solutions, it can significantly improve the repair success rate and printing accuracy. Attached Figure Description

[0025] The present invention will now be described in further detail with reference to the accompanying drawings and specific implementation methods.

[0026] Figure 1 This is a flowchart of the defect identification and repair process;

[0027] Figure 2 This is a schematic diagram of the system structure of the present invention;

[0028] In the diagram: Y-axis displacement platform 1; forming and printing platform 2; X-axis displacement platform 3; DLP / SLA light source 4; high-precision industrial camera 5; infrared rangefinder 6; vacuum scraper 7; printed part 8. Detailed Implementation

[0029] The present invention will now be described in further detail with reference to the accompanying drawings.

[0030] like Figures 1 to 2 As shown below, the steps and functions of a method for detecting and repairing defects in the feeding of large-size photopolymer molding materials will be explained in detail.

[0031] Step S1: A large-capacity, sealed pressure storage tank is used to store resin in a sealed environment to prevent air contact from affecting quality and to ensure a continuous supply. The supply process involves... Figure 2 The vacuum feeding scraper 7 shown feeds material to the printing platform to avoid printing interruption; the pressure regulating valve regulates the pressure inside the tank to ensure stable resin output, prevent uneven feeding, and ensure optimal viscosity and flowability.

[0032] Step S2: In such Figure 2 The system structure diagram shown illustrates that the high-resolution industrial camera 5 of the vision inspection device, together with the high-precision infrared rangefinder 6, enables precise capture of micron-level bubbles, scratches, and uneven thickness on the surface of the material layer. The infrared ranging module 6 acquires the three-dimensional morphology data of the material layer in real time through non-contact scanning and performs multi-source fusion analysis with the visual image to more accurately distinguish different defect types.

[0033] Step S3: Set the printing layer thickness, conventional exposure power, and exposure time as basic parameters, and simultaneously trigger the multimodal detection module to scan layer by layer according to the preset path; the DLP projection optical engine 4 can achieve X / Y displacement through the Y-axis displacement platform 1 and the X-axis displacement platform 3 to achieve multiple projections to cover the entire large-format printing platform 2. After each projection is completed, the system automatically pauses once, and performs millisecond-level data acquisition through the high-resolution industrial camera 5 and the high-precision infrared rangefinder 6 to capture defects before and after printing and before and after resin curing; analyze the state differences to ensure that minute deformations and uncured anomalies are captured in time; and synchronously transmit the visual image data captured by the industrial camera 5 and the infrared ranging height data collected by the infrared rangefinder 6 to the control unit.

[0034] Step S4: Based on the printing platform defect data collected in step S3, such as Figure 1 As shown in the defect identification and process repair diagram, the control unit fuses the liquid surface morphology features of the acquired visual image with infrared height data and inputs it into the pre-loaded CNN defect identification module for defect identification. It determines whether there are printing material supply defects as shown in the system flowchart. At the same time, it directly identifies easily detectable phenomena such as printing pits and bubbles. For defects that are difficult to determine, a defect discrimination threshold is set. If the resin height deviation exceeds the set range, it is determined that there is a material shortage / overflow phenomenon on the liquid surface. According to the defect characteristics, the defects are classified into several categories such as material shortage defects / overflow defects / bubble defects / scratches, and the corresponding repair strategy is triggered based on the classification results.

[0035] Step S5: Based on the defect classification results of step S4, the system automatically retrieves the corresponding repair strategy library: if it is determined to be a material shortage defect, the feeding scraper is called to replenish material for the printing platform; if it is a material overflow defect, a second call is made. Figure 2 The feed scraper 7 shown scrapes the liquid surface, adjusting the resin level in the overflow area to the average height, recovering the overflowed resin and transmitting the data back to the feed system in real time to calibrate the resin output for the next layer; when a bubble defect is detected, it calls... Figure 2 The vacuum pump of the vacuum scraper 7 shown is adjusted to adjust the vacuum level to perform regional defoaming operation, and the scraper speed is adjusted appropriately to avoid the regeneration of bubbles.

[0036] Step S6: Based on the printing status after defect identification and repair in Step S5, perform secondary repair verification on the printed parts; the closed-loop effect of secondary repair verification means that after the repair action is executed, the system immediately starts a secondary multimodal scan and re-calls the following... Figure 2 The industrial camera 5 and infrared rangefinder 6 are used to collect and compare data before and after repair, comparing the deviations in liquid surface morphology and height data before and after repair. If the deviation converges to within the set range and the defect features disappear in the visual image, the repair is considered successful, and the next layer of printing continues. If the detection result shows that the defect has not been eliminated, the process returns to S5 to perform targeted repair again until the accuracy requirements are met, and the final product is as shown. Figure 2 The printed part 8 is shown. Multiple printing experiments were conducted, and the defect type, location, size, and corresponding repair parameters were recorded during each printing process to optimize the identification speed and repair parameter matching degree of subsequent similar defects.

Claims

1. A method for detecting and repairing defects in the feeding of large-size photopolymer molding materials, characterized in that: Suitable for large-format SLA and DLP photopolymer 3D printing processes, this system ensures stable resin delivery in a pollution-free environment through a sealed feeding system. Combined with multimodal detection technology, it monitors the material layer status in real time, uses intelligent recognition algorithms to accurately locate defects, performs targeted repair operations, and ensures repair effectiveness through a closed-loop verification mechanism. This achieves an integrated process of online detection and precise repair, ultimately forming a dual closed-loop control system with feeding control and visual feedback at its core, improving the forming quality and reliability of large-size printing.

2. The method according to claim 1, characterized in that, Specifically, the following steps are included: S1: Construct a hardware system consisting of a sealed pressure storage tank and a pressure regulating valve to ensure the stability and controllability of the material supply process; S2: Install visual inspection devices and infrared ranging devices to complete the inspection path planning and accuracy calibration, laying the foundation for subsequent data acquisition; S3: Start printing and collect large-format fabric layer data layer by layer to obtain the distribution and morphology information of the resin in each layer in real time; S4: Integrates multi-source data to complete defect identification and classification, and combines image and height data for comprehensive analysis; S5: Perform differentiated repairs based on defect type and adjust material supply and vacuum parameters to achieve targeted treatment; S6: After repair, secondary testing and verification are performed to form a closed loop and optimize the parameter model to continuously improve system performance.

3. The method according to claim 2, characterized in that: In step S1, the sealed pressure storage tank achieves sealed storage and continuous supply of resin through its efficient sealing structure, effectively preventing the intrusion of external impurities such as dust and moisture, and ensuring the purity of raw materials; the pressure regulating valve stabilizes the output pressure through a real-time monitoring and feedback mechanism to maintain a constant flow rate of resin delivery, avoids process instability caused by pressure fluctuations, and improves the uniformity and reliability of the final product.

4. The method according to claim 2, characterized in that: In step S2, the visual inspection device uses a high-resolution industrial camera and provides uniform illumination to capture clear images. The infrared ranging device obtains material layer height information through non-contact scanning and measures height changes. The data fusion of the two enables accurate differentiation of defect types and improves inspection accuracy.

5. The method according to claim 4, characterized in that: Employing a layered detection mechanism, the system immediately identifies defects during the deposition or curing of each layer of material, enabling real-time monitoring of the entire process from two-dimensional to three-dimensional. By combining high-speed image sensors with intelligent analysis algorithms, the system can instantly capture subtle abnormal changes and feed the detection results back to the main control unit in real time through a data interface, providing immediate basis for online evaluation of printing quality and process adjustment.

6. The method according to claim 2, characterized in that: In step S4, the visual image and height data are input into the defect recognition module. The defects are classified into four categories: missing material, overflow, bubbles, and scratches, using the resin height deviation as the judgment threshold. The recognition accuracy is improved through a deep learning model.

7. The method according to claim 6, characterized in that: Once a defect is accurately identified, the system will immediately and automatically match it with the built-in repair strategy library to establish a precise one-to-one correspondence between defect types and repair solutions. This process ensures that each specific defect can be associated with the most suitable repair solution. Then, based on the identified defect category, the system intelligently calls preset optimization parameters to achieve rapid response and efficient processing, thereby improving the overall repair efficiency and accuracy.

8. The method according to claim 2, characterized in that: Step S5 involves adjusting the temperature to regulate resin fluidity, controlling the scraper movement to smooth the surface, and adjusting the vacuum pump to remove air bubbles, thereby completing material shortage compensation, overflow scraping, and air bubble removal to ensure repair effectiveness.

9. The method according to claim 2, characterized in that: In step S6, the disappearance of defect features is taken as the standard for completion of repair. If it is not qualified, return to step S5 to repeat the targeted repair. At the same time, record data to optimize the identification and repair parameters, forming an adaptive learning loop.

Images