Tribo-informatics analysis of in-situ TiC reinforced ZA27 alloy: Microstructural insights and wear performance modeling by machine learning

Abstract

The study examines the performance of ZA27 alloy reinforced with in-situ TiC under high-stress abrasive wear conditions. The ZA27 + 10 wt% TiC composite demonstrates superior wear resistance among all the materials tested. Response surface analysis revealed that increasing the applied load reduces the wear resistance of the specimens. Analysis of worn-out surfaces revealed mild grooves at low loads and deeper grooves at higher loads. The removal of material is mainly caused by ploughing and cutting actions from abrasive surfaces. Machine Learning models were employed to forecast wear rate, coefficient of friction, and temperature. The MLP model excelled, achieving an R² value of 0.88 during testing. Furthermore, feature importance analysis revealed that reinforcement wt% has the maximum influence in predicting the wear rate.