Sensitivity-driven deep learning model for tribological prediction in Al7075/B₄C nanocomposites

The development of lightweight and durable nanocomposites for industrial use is constrained by traditional tribological evaluation methods that are costly, time-consuming, and inadequate for capturing nonlinear interactions between material and operational parameters. This study proposes an integrated framework combining Global Sensitivity Analysis (GSA) and Machine Learning (ML) to predict the Coefficient of Friction (COF) and wear loss for Al7075/boron carbide (B₄C) nanocomposites . Four GSA techniques - Sobol indices, delta index, PAWN index, and mutual information - were employed to rank the significance of input parameters, including applied load, B₄C reinforcement percentage, time, sliding velocity, and sliding distance. Using 10,800 experimental records from pin-on-disc tests, a Deep Residual Regression Network (DRRN) was developed to model tribological behavior. The nanocomposites were fabricated with Al7075 as the matrix and B₄C particles at weight fractions of 0%, 4%, 8%, and 12%. The results show that increasing B₄C content significantly improves wear resistance, with the 12 wt% B₄C composites exhibiting a reduction of 79% in wear loss and 19% in COF under severe operating conditions. The proposed deep learning (DL) framework achieved high predictive performance, with coefficients of determination ( R ² ) of 0.93 for COF and 0.99 for wear loss. The GSA results show that reinforcement percentage ( Wt%) is the most influential parameter, followed by sliding velocity ( V ) and sliding distance ( D ).