Data analytics approach for melt-pool geometries in metal additive manufacturing

Area of Science:

• Materials Science and Engineering
• Additive Manufacturing
• Data Science

Background:

• Melt-pool dynamics are crucial for controlling the quality of parts fabricated using powder bed fusion (PBF).
• Understanding the physics-based melt-pool formation is essential for process optimization in PBF of nickel alloys.

Purpose of the Study:

• To employ modern data analytics for understanding and predicting physics-based melt-pool formation.
• To develop machine learning models for reliable prediction of melt-pool geometries in Ni alloy single tracks.

Main Methods:

• Fabrication of Ni alloy single tracks using powder bed fusion.
• Creation of an extensive database of melt-pool geometries with processing parameters and material characteristics.
• Correlation analysis to identify relationships between process parameters and melt-pools.
• Development of machine learning models using highly correlated features.

Main Results:

• Established correlations between process parameters and melt-pool characteristics.
• Successfully developed machine learning models for predicting melt-pool geometries.
• Demonstrated that data analytics enhances understanding of melt-pool physics.

Conclusions:

• Data analytics provides a powerful approach to understand and predict melt-pool formation in PBF.
• The developed models facilitate reliable prediction of melt-pool geometries.
• This methodology can serve as a foundation for melt-pool control and process optimization in additive manufacturing.