System identification of tool chip interface friction while machining CNT-Mg-Al composites

Abstract

Tools wear directly impacts manufacturing productivity. Tool chip interface friction is a major source of tool wear during composite machining. The current work presents system identification modeling of this important phenomenon based on nanocomposite machining data. Turning experiments were performed on carbon nanotube (CNT) reinforced aluminium-magnesium composites (0.3, 0.5, 0.7% CNT; 1% Mg). The experimental design involved variations of all machining parameters (cutting speed, feed rate and depth of cut) one at a time, maintaining the remaining parameters at their mid-levels. Tool chip interface friction was determined from measured cutting and tangential force components using machining angle relations. Friction results and corresponding machining parameter settings were applied as time step inputs for modeling. Five different model structures were implemented for tool chip friction system identification. Model parameters were optimized to maximize prediction accuracy. The Autoregressive with Exogenous Variables (ARX) model with 16 parameters and 1021 order attained the highest FIT (98.8 %) and the lowest mean squared error (1.477e-05). This result is very encouraging towards effective control of nanocomposite machining.