Comparative analysis for the prediction of WEDM responses for machining spark plasma sintered boron carbide ceramic sample by RSM and ANFIS

Abstract

Unconventional wire electrical discharge machining (WEDM) process is successfully used to cut different metals, alloys, composites and recent addition is engineered ceramics which possess sufficient electrical conductivity. Boron carbide is one of the hardest ceramic materials that unable to be processed with conventional machine tools and can be machined by WEDM compulsorily with proper selection of machine parameters. This study is based on boron carbide samples which were prepared using spark plasma sintering (SPS) furnace and machined with WEDM. Five machining parameters were analyzed such as pulse on time, pulse off time, peak current, water pressure and servo feed rate. Surface roughness (Ra) and machining speed were considered as output parameters and design of experiment was derived using central composite design (CCD) of response surface method (RSM) with 32 numbers of different test runs. Adaptive neuro-fuzzy inference system (ANFIS) was used with a new set of 16 numbers of experiments to predict results and seen to be more reliable than predicted results of response surface method.