Machine Learning-Enabled Quantitative Analysis of Optically Obscure Scratches on Nickel-Plated Additively...

Area of Science:

• Materials Science
• Surface Engineering
• Additive Manufacturing

Background:

• Additively manufactured metal components often exhibit surface roughness, requiring post-processing.
• Surface hardness is crucial for component properties, particularly wear resistance.
• Electroless nickel plating is a common surface treatment for enhancing component characteristics.

Purpose of the Study:

• To investigate the relationship between relative surface hardness and scratch width in electroless nickel plating on additively manufactured composite components.
• To explore the effectiveness of K-means unsupervised machine learning for surface characterization.
• To assess the impact of different processing parameters on surface quality.

Main Methods:

• Utilized Taguchi Design of Experiment (TDOE) L9 orthogonal array for parameter optimization.
• Employed scanning electron microscopy (SEM) for high-resolution surface imaging and 3D height mapping.
• Applied K-means clustering algorithm with Python for scratch region identification and width quantification.

Main Results:

• Achieved distinct electroless nickel plating hardness levels across nine experimental samples.
• Observed a non-linear correlation between increased scratch force and scratch width.
• Successfully quantified scratch widths using SEM and K-means clustering, overcoming limitations of optical microscopy.

Conclusions:

• Electroless nickel coatings significantly enhance the surface quality of additively manufactured components.
• The developed K-means machine learning approach provides precise surface characterization.
• Improved surface properties have substantial implications for industrial applications and future surface engineering.