Tribological behavior of PLA reinforced with boron nitride nanoparticles using Taguchi and machine learning approaches

Tribological behavior of PLA reinforced with boron nitride nanoparticles using Taguchi and machine learning approaches

This research investigates the tribological characteristics of Polylactic Acid (PLA) reinforced with Boron Nitride Nanoparticles (BNNP). The study employs the Taguchi approach, examining the effects of applied load (20–60 N), speed (100–300 RPM), sliding distance (500–1500 m), and BNNP content (0–0....

Full description

Saved in:
Bibliographic Details
Main Authors: Harishbabu Sundarasetty, Santosh Kumar Sahu
Format: Article
Language:English
Published: Elsevier 2025-06-01
Series:Results in Engineering
Subjects:
Wear
PLA
BNNP
Taguchi
Machine learning
Online Access:http://www.sciencedirect.com/science/article/pii/S2590123025008497
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This research investigates the tribological characteristics of Polylactic Acid (PLA) reinforced with Boron Nitride Nanoparticles (BNNP). The study employs the Taguchi approach, examining the effects of applied load (20–60 N), speed (100–300 RPM), sliding distance (500–1500 m), and BNNP content (0–0.04) on tribological properties. Results indicate that the applied load significantly affects the wear rate, with higher filler concentrations enhancing wear resistance. The observed wear rates range from 0.011×10⁻³ g/m (0.04 PLA/BNNP at 20 N, 300 RPM, 1500 m) to 0.124×10⁻³ g/m (0.02 PLA/BNNP at 60 N, 100 RPM, 1500 m). ANOVA results reveal that applied load is the most influential factor, accounting for 53.01 %. Conversely, speed predominantly affects the coefficient of friction (COF), contributing 78.05 %. The COF values range from 0.069 (0.02 PLA/BNNP at 60 N, 100 RPM, 1500 m) to 0.319 (0.04 PLA/BNNP at 20 N, 300 RPM, 1500 m). A machine learning technique, i.e., Random Forest Regression, is employed to predict the tribological performance. The Relative Root Mean Square Error (RRSME) values decisively confirm that Random Forest Regression (23.86 % wear rate and 18.32 % COF) is more accurate than traditional linear regression approaches. Optical microscopy confirmed enhanced wear resistance for 0.04 PLA/BNNP composites. Integrating the Taguchi approach and machine learning effectively optimized the tribological performance of PLA-BNNP composites for diverse applications.
ISSN:2590-1230

Similar Items