Characterization of Surfaces Generated in Milling and Abrasive Water Jet of CFRP Using Wavelet Packet Transform

Abstract

Polymer matrix composites find extensive application in aerospace and automotive industry. Although composites are molded to a near net shape, secondary machining such as hole-drilling and edge trimming is unavoidable. Conventional milling and Abrasive water jet are widely employed processes to cater the machining needs. Because of the nature of material-tool interaction, both the processes have their limitations resulting in machining induced damages such as matrix cracking, exit-delamination, fiber pull-outs, spring-back, interply delamination, striations, abrasive embedment, and kerf taper. Surface roughness plays a pivotal role in determining the damage induced during the machining processes. Traditionally roughness parameters - average roughness (Ra) and ten-point average (Rz) are used to characterize the surfaces. Often, the use of cut-off wavelengths, governed by standards, result in poor distinguishability for machined surfaces of anisotropic materials such as polymer matrix composites (PMC). The insensitivity and poor distinction in certain cases leads to unsound judgment in characterizing the machining induced damage, and also in the selection of process parameters. This necessitates other methods and indicators that better identify the surface signature and correlate with the process parameters. In this study, a novel approach is proposed in which wavelet packet transform (WPT) is applied to the machined surface profiles. An indicator (ratio of wavelet packet energy and entropy) is defined to characterize the surface and better predict the surface quality as a function of machine tool, material and process variables. Regardless of the type of CFRP and machining process (AWJ and conventional milling), WPT indicator proved to be a better predictor of the machining process (R2 > 80%) when compared with Rz (R2 < 67%).