AI Print-Head Control From Layer Image Feedback in 3D Printing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current additive manufacturing systems lack precise feedback and corrective action capabilities during the printing process, leading to suboptimal mechanical, optical, and electrical properties in the final product, as they typically provide feedback only after the entire object is printed, which can result in deviations from the intended design.
Innovation Solution
The implementation of artificial intelligence process control (AIPC) using reinforcement learning, hidden Markov models, and other AI algorithms to analyze images of printed layers in real-time, identify anomalies, and adjust printing parameters dynamically to optimize the printing process and achieve desired properties.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If feedback is provided only after the entire object is printed, then the printing process is simple and continuous, but the manufacturing precision and quality control deteriorate
Solution Approach 1:
The patent implements real-time feedback mechanisms during the additive manufacturing process by capturing images of each printed layer, analyzing them using AI algorithms to detect anomalies, and providing corrective feedback to adjust printing parameters before subsequent layers are deposited. This continuous feedback loop maintains high manufacturing precision without requiring complete object printing before intervention.
Solution Approach 2:
The patent replaces traditional mechanical measurement and inspection systems with optical imaging and AI-based analysis systems. Instead of using physical probes or post-printing mechanical measurement devices, the system uses cameras to capture layer images and employs machine learning algorithms to analyze printed quality, thereby improving precision while managing system complexity through software-based solutions.
2Manufacturing precision
If real-time feedback and corrective action are implemented during printing, then manufacturing precision improves, but device complexity increases
Solution Approach 1:
The patent implements self-service capabilities where the additive manufacturing system automatically monitors its own printing process, detects anomalies in real-time, and performs self-correction by adjusting printing parameters without external intervention. The system uses AI algorithms to analyze printed layers and automatically modifies subsequent printing operations, enabling the system to maintain high precision while managing complexity through automation.
Solution Approach 2:
The patent introduces dynamic adaptability to the printing system by enabling real-time modification of printing parameters based on detected anomalies. Instead of static, pre-programmed printing processes, the system dynamically adjusts extrusion rates, temperatures, and other parameters during printing based on AI analysis of each layer, thereby improving precision while the modular dynamic control architecture manages system complexity.
3Manufacturing precision
If AI algorithms are used to analyze each printed layer in real-time, then manufacturing precision and quality improve, but use of energy and computational resources increases
Solution Approach 1:
The patent applies partial analysis strategies where AI algorithms focus on detecting specific critical anomalies rather than performing exhaustive analysis of every printed layer. The system identifies and prioritizes key defect types and applies targeted analysis only where needed, maintaining high manufacturing precision while reducing unnecessary computational energy consumption through selective rather than comprehensive layer-by-layer inspection.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Systems, methods, and media for additive manufacturing are provided. In some embodiments, an additive manufacturing system comprises: a hardware processor that is configured to: receive a captured image; apply a trained failure classifier to a low-resolution version of the captured image; determine that a non-recoverable failure is not present in the printed layer of the object; generate a cropped version of the low-resolution version of the captured image; apply a trained binary error classifier to the cropped version of the low- resolution version of the captured image; determine that an error is present in the printed layer of the object; apply a trained extrusion classifier to the captured image, wherein the trained extrusion classifier generates an extrusion quality score; and adjust a value of a parameter of the print head based on the extrusion quality score to print a subsequent layer of the printed object.