Related Experiment Video
Updated: Oct 9, 2025
09:12
Production of Single Tracks of Ti-6Al-4V by Directed Energy Deposition to Determine the Layer Thickness for Multilayer Deposition
Published on: March 13, 2018
9.4K
Selection of effective manufacturing conditions for directed energy deposition process using machine learning methods
Jong-Sup Lim1, Won-Jung Oh2, Choon-Man Lee3
1School of Smart Manufacturing Engineering, Changwon National University, Changwon, 51140, Republic of Korea.
Scientific Reports
|December 18, 2021
Summary
Optimizing directed energy deposition (DED) for titanium alloys involves selecting process parameters. This study found the Random Forest model accurately predicts optimal parameters based on deposited surface color.
Area of Science:
- Materials Science
- Additive Manufacturing
- Machine Learning
Background:
- Directed Energy Deposition (DED) requires extensive empirical testing to determine optimal process parameters for materials like titanium alloys.
- Surface color variations in DED-processed titanium alloys correlate with underlying material properties and process conditions.
Purpose of the Study:
- To investigate the relationship between DED process parameters (laser power, scan speed) and the resulting surface color of titanium alloy single tracks.
- To develop and compare machine learning models for predicting optimal DED parameters based on surface color.
- To identify a surface color indicative of suitable quality for additive manufacturing.
Main Methods:
- Single-track experiments were performed on titanium alloys using varying laser power and scan speed in the DED process.
- Deposited samples were analyzed for surface color, cross-sectional view, hardness, microstructure, and composition.
- Random Forest (RF) and Support Vector Machine (SVM) multi-classification models were trained using surface color data.
Main Results:
- Distinct surface color variations were observed and correlated with specific process parameters.
- Analysis revealed a target color range suitable for additive manufacturing applications.
- The Random Forest model demonstrated higher accuracy in predicting optimal DED parameters compared to the SVM model.
Conclusions:
- Surface color serves as a reliable indicator for optimizing DED process parameters in titanium alloys.
- Machine learning, particularly the Random Forest model, can effectively predict suitable DED parameters based on surface color, reducing empirical testing.
- This approach facilitates the selection of optimal parameters for high-quality additive manufacturing of titanium alloys.
