Machine Vision-Integrated Grey-Taguchi Optimization of Fiber Laser Drilling on Aluminum 6063: A Novel Approach for Minimizing Circularity Error and Heat-Affected Zone Simultaneously

Abstract

In the current study, a novel automated machine vision based technique along with Grey Relational Analysis (GRA) has been presented for multi-objective optimization of parameters in fiber laser drilling process of Aluminum 6063 alloy. In traditional parameter optimization approaches, there is a strong dependence on manual optical measurements while a Digital Image Processing (DIP) technique combining Canny Edge Detection algorithm and Hough Circle Transform was designed in the present investigation for accurate measurement of HAZ area and circularity error. For conducting experiments, L25 orthogonal array design was used while varied combinations of peak laser power, scanning speed, and pulse frequency constituted the controlled variables in the experiment. Material removal rate, automated measured area of HAZ, and entrance/exits circularity errors were considered as response parameters. With the aid of GRA analysis, multi-response optimization problems have been transformed into individual optimal objective function which is referred as Grey Relational Grade (GRG) and optimal parameter set and significance of each independent parameter were found using ANOVA test. The proposed approach showed great improvement in human error associated with traditional parameter optimization methods.