Key Findings
A new machine learning framework can simultaneously predict four mechanical properties of aluminum WAAM parts: tensile strength, elongation, flexural strength, and hardness. In a confirmation build using optimized settings, the model predicted a tensile strength of 295.20 MPa and 40.10% elongation; the actual measured values were 288.90 MPa and 38.25%, respectively. While the predicted strength gain was not fully realized, the elongation result was 2.53 percentage points higher than the best previous run.
Technical Details
The study utilized a Fronius iWave 500i AC/DC welding unit with an ABB IRB 1520ID robotic arm operating in cold metal transfer (CMT) mode. Researchers used a Taguchi L27 design plus three center-point runs to create 30 builds on a fixed circular path, varying four parameters:
- Wire feed speed: 7.5 to 9.5 m/min
- Travel speed: 6 to 10 mm/s
- Interpass temperature: 80 to 160°C
- Shielding gas flow: 12 to 18 L/min
The team employed a stacked ML model called PG-MOSR. In five-fold cross-validation, R² values were 0.7799 for tensile strength, 0.8299 for elongation, 0.8142 for flexural strength, and 0.8599 for hardness. A version trained on raw settings performed even better, with R² values between 0.8779 and 0.8885.
Background & Context
Most previous WAAM research focused on single responses, such as bead geometry or surface roughness. This study differs by predicting multiple mechanical properties for circular-path ER5356 deposits. For comparison, other frameworks have targeted Ti-6Al-4V via laser powder bed fusion (LPBF) or Al-Si-Mg parts using microstructural and void data to reduce physical testing.
Significance & Outlook
The framework allows for the screening of thousands of parameter combinations—50,000 in this study—to identify optimal settings. However, the researchers note that fracture surfaces showed clustered pores, which may explain why some strengths fell short of predictions. Future work is expected to incorporate CT-based porosity measurement, in-situ thermal monitoring, and larger datasets to improve reproducibility and validation.
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
