Indexed by:期刊论文

Date of Publication:2019-04-01

Journal:OPTICS AND LASER TECHNOLOGY

Included Journals:SCIE、Scopus

Volume:111

Page Number:653-663

ISSN No.:0030-3992

Key Words:Underwater laser machining; Channel fabrication; Burr-free; Regression analysis; Machine learning

Abstract:Laser machining in the nanosecond pulse regime utilises the heat induced by laser irradiation to ablate solid material, however the heat-driven ablation is very often accompanied with adverse effects such as oxidation, debris recast and burr formation. An effective technique invented to minimise these effects was underwater laser machining. In this study, we present a multi-scan laser machining process on a copper sample immersed under flowing water with a secured upper surface. The results show that channels with smaller heat affected zone (HAZ), less debris recast and minimal burr can be produced by this process. By applying multi-scan machining, channels with up to 275 mu m depth, 13-22 degrees taper angle and 1.30-6.95 mu m roughness S-a were produced. Moreover, empirical models relating the processing parameters to channels' characteristics were derived using polynomial regression (PR) analysis and a machine learning algorithm, Gaussian process regression (GPR), after which their performances in the prediction of the channels' characteristics were validated.