Integrated experimental design and machine learning framework for predicting UV influenced mechanical properties in polyurethane nanodiamond nanocomposites

Abstract

This study investigates the influence of ultraviolet (UV) irradiation on the mechanical performance of nanodiamond (ND) reinforced polyurethane (PU) nanocomposites. The Taguchi method was employed to systematically design the experiments, while analysis of variance (ANOVA) was used to evaluate the statistical significance and percentage contribution of each factor. An L27 orthogonal array was adopted to examine the effects of composition (pure PU, 0.2 wt% PU/ND, and 0.5 wt% PU/ND), UV exposure duration (0, 200, and 400 h), UV irradiation intensity (1.0, 1.20, and 1.40 W/m²), and UV temperature (40, 50, and 60 °C) on tensile strength, Young’s modulus, and hardness. The results indicate that 200 h of UV exposure enhances tensile strength and Young’s modulus for all samples, with the most pronounced improvement observed in the 0.5 wt% PU/ND nanocomposite, whereas hardness decreases with increasing exposure due to UV-induced surface degradation. ANOVA results indicate that composition and UV exposure duration are the most influential parameters, contributing 48.76% and 17.58% to tensile strength and 40.21% and 19.13% to Young’s modulus, respectively. For hardness, UV exposure duration is the dominant factor, contributing 49.6%. Machine learning models, including linear regression, artificial neural networks, and Gaussian process regression, were developed for prediction. Among them, the Gaussian process model showed the highest accuracy with R² values of 0.99, 0.95, and 0.98 for tensile strength, Young’s modulus, and hardness, respectively. These findings highlight the potential of PU/ND nanocomposites for applications in automotive components, robotic parts, and aerospace structures.