Prediction of surface roughness and MRR in grinding process on Inconel 800 alloy using neural networks and ANFIS

Cylindrical grinding is a prominent finishing process producing precise tolerances and high quality surface finish. However, in comparison to other techniques, the cylindrical grinding is marginally expensive which necessitates in optimizing of input parameters in order to achieve good output responses that limit the product cost. Surface roughness and Metal Removal Rate (MRR) are the major output responses influencing the quantity and quality in production. These output responses purely depends on numerous input parameters that are broadly classified into following groups like wheel parameters, work piece parameters, process parameters and machine parameters etc. Therefore in the present investigation, a full-factorial design of experiments are carried-out by considering work speed, depth of cut and feed rate as input process parameters while the output responses studied are surface roughness and MRR on Inconel 800 alloy. 27 experimental data sets are used for training the networks and an additional 8 sets of experiments were conducted for validation. Three computational methods such as regression analysis, neural networks and Adaptive-neuro fuzzy inference system (ANFIS) are analysed to predict output responses of work piece for variety of input parameters. A comparison between three computational methods is performed to identify a better method in predicting the output responses. The variation of surface roughness and MRR is also studied with correlation to various process parameters. (c) 2017 Published by Elsevier Ltd.