A defect high-precision real-time online monitoring method and system based on molten pool characteristics
By identifying the highly sensitive region of molten pool defect response and combining photoelectric signals with convolutional neural networks, the accuracy problem of real-time online monitoring of molten pool defects in additive manufacturing was solved, achieving high-precision defect detection.
CN117517341BActive Publication Date: 2026-07-14BEIJING INST OF TECH
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIJING INST OF TECH
- Filing Date
- 2023-11-07
- Publication Date
- 2026-07-14
AI Technical Summary
Technical Problem
The lack of high-precision real-time online monitoring methods for molten pool defects in existing additive manufacturing technologies limits the application of manufactured parts in fields such as aerospace and marine engineering.
Method used
By identifying the highly sensitive region of the molten pool defect response, and utilizing the molten pool temperature field and photoelectric signals, combined with a convolutional neural network model, high-precision real-time online monitoring of molten pool defects can be achieved.
Benefits of technology
It improves the accuracy and efficiency of molten pool defect detection, and realizes high-precision real-time online monitoring of molten pool defects in additive manufacturing process using only photoelectric signals.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN117517341B_ABST
Abstract
The application discloses a kind of based on molten pool feature's defect high-precision real-time online monitoring method and system, method includes calibration stage and measurement stage;System includes photoelectric signal acquisition module, temperature signal acquisition module and computing module;Calibration stage acquires the temperature signal of different regions of molten pool, and the degree of association between the amplitude-frequency information of different regions of molten pool temperature signal and defect determines the high sensitivity area of molten pool defect response;Acquire the photoelectric signal of the radiation light of high sensitivity area of molten pool defect response, convert its amplitude-frequency information into two-dimensional image, and the correlation model between the two-dimensional image of photoelectric signal and defect is established by convolutional neural network;Measurement stage real-time on-line acquisition photoelectric signal of the radiation light of high sensitivity area of molten pool defect response;Two-dimensional image converted by its amplitude-frequency information is input into correlation model, and molten pool defect is detected in real time whether it is generated.
Need to check novelty before this filing date? Find Prior Art