Background
Additive Manufacturing (AM) technologies are transforming rapid prototyping and production of high-performance components across aerospace, medical, and automotive industries. However, a primary impediment to their widespread adoption is the difficulty in process control and the resulting variability in quality. For complex AM processes like in-situ alloying, where multiple metal powders are mixed to form an alloy during fabrication, the interaction of numerous parameters makes defect prediction and management extremely challenging. These subtle variations can induce defects such as porosity, lack of fusion, and microstructural inhomogeneity, severely impacting final part performance and reliability. Traditional physics-based models and empirical rules often fall short in addressing these complex multivariate relationships. Machine learning, with its ability to extract patterns and build predictive models, offers a powerful complement or alternative, marking a crucial step in the ‘smartification’ of advanced manufacturing aiming for more robust and reliable AM parts.
Key Findings
A recent study published in the Journal of Manufacturing and Materials Processing demonstrates a significant advancement in additive manufacturing quality control: a decision tree LSBoost ensemble machine learning method effectively optimizes defect management in in-situ alloyed additively manufactured parts, substantially enhancing manufacturing reproducibility and quality. This research addresses critical quality challenges inherent in AM, particularly with in-situ alloying, by leveraging LSBoost—an algorithm designed for robust predictive accuracy even with limited experimental data.
The developed model proved capable of accurately predicting key quality indicators, including relative density, tensile strength, and defect size distribution, based on critical input manufacturing parameters (e.g., laser power, scan speed, powder composition, layer thickness). Beyond prediction, the model precisely identified the most influential parameters contributing to defect formation and successfully mapped optimal ‘processing windows.’ This capability allows for the efficient discovery of manufacturing conditions tailored to achieve specific quality targets, offering significant time and resource savings compared to traditional trial-and-error methods.
This machine learning-guided defect optimization approach represents a crucial step towards robust quality assurance and process control in AM. It is expected to accelerate the adoption of AM technology in high-stakes applications requiring stringent quality and reliability, such as jet engine components or custom medical implants. Looking forward, such AI-based quality control systems are envisioned for direct integration into manufacturing lines, enabling ‘closed-loop manufacturing’ with real-time defect detection and automated process adjustments. This innovation is poised to become an indispensable component for enhancing the global competitiveness of advanced manufacturing, promising reduced production costs, shorter lead times, and maximized product performance.
Source: https://www.mdpi.com/2504-4494/10/7/254
Get our weekly technology intelligence — free
Receive an infographic that lets you judge at a glance whether each field’s analysis report is worth reading.
Subscribe Free — Weekly Tech Intelligence
By subscribing, you’ll receive Troy-Technical’s weekly technology intelligence newsletter.
- Your email and selected fields are used only to deliver the newsletter.
- We never share your information with third parties.
- You can unsubscribe anytime via the link in each email.
See our Privacy Policy for details.
Takes about a minute · Unsubscribe anytime
Comments