PARAMETRIC OPTIMIZATION OF MAGNETIC ASSISTED FINISHING OF TITANIUM (GRADE 5) USING ARTIFICIAL NEURAL NETWORKS & GENETIC ALGORITHMS

Abstract

This work is an experimental study of parametric optimization of the Magnetic Abrasive Finishing (MAF) process, during finishing of Titanium Grade 5 using unbonded MAF process. Magnetic Abrasive Finishing (MAF) is a non-traditional finishing process which is initially controlled by magnetic field and by which a high degree of surface finish with very close tolerances can be obtained. The principal idea behind this finishing process bank on using a large number of arbitrarily cutting edges with unfixed alignment and geometry, called the abrasive particles, for potent exclusion of material with chip smaller compared to that come by in machining process using definite edges cutting tools. The abrasive particles don’t have any innate magnetic properties, for this reason magnetic iron ore is mixed, so that they can be glide in air using a magnetic field for brush formation, by which the material surface is grinded. As the particles are loosely held at their positions by the effect of the magnetic field applied, their weight, friction and mechanical boundations caused by the adjoining particles, MAF allows better surface finish, close tolerances, creation of unclear surface features and machining of harder and strenuous to machine materials. In this study, the potential of the process is strived to be capitalize by studying the outcome of input parameters such as coil’s current, abrasive percentage, machining gap, total machining time, and the RPM of spindle, using an L16 orthogonal design of experiments.