Experimental investigations with machine learning techniques for understanding of erosion wear in advanced aluminum nanocomposites

Abstract

This article addresses the significant challenge of erosion wear in aluminum composites, particularly in industries such as automotive, aerospace and energy, where sustained material performance is crucial. The study focuses on the experimental investigation of erosive wear characteristics exhibited by advanced aluminum nanocomposites, utilizing an air jet erosion wear test apparatus. The erosion wear tests were conducted using diverse parameters, such as angles (30°, 45°, 60° and 90°), air pressures (1–4 bar) and a fixed 600-s duration, maintaining a constant sand particle feed rate of 2.0 (g/min) and utilizing stand-off distances of 10, 15 and 20 mm. Surface assessments were conducted using field emission scanning electron microscopy (FE-SEM) and energy-dispersive X-ray spectroscopy (EDS). The results revealed observable specimen wear, particularly at a 45° angle, with wear rates decreasing at higher impingement angles and reaching a minimum at 90°. Notably, the Al-4 wt.%TiC nanocomposites exhibited a 25% improvement in wear resistance compared to the base alloy. Further analysis of eroded surfaces through (FE-SEM) revealed a mechanism involving micro-cutting, plowing and grain pull-out, attributing wear primarily to plastic deformation and crack formation. Evaluating erosive wear results through six machine learning (ML) models demonstrated that gradient boosting regression emerged as the most accurate, achieving an R2 value of 0.97. This highlights the effectiveness of ML in predicting erosion wear rates and offers insights for improving aluminum composite materials in demanding industrial applications.