INVESTIGATION OF WEAR BEHAVIOR OF AS FORGED INCONEL 690 SUPER ALLOY USING ARTIFICIAL NEURAL NETWORKS

Abstract

The present study aims to study the wear properties of as forged Inconel 690. The dry sliding wear behavior of as forged Inconel 690 is studied in accordance with ASTM standards G99 i.e. dry sliding on pin on disc wear test apparatus. Three wear parameters namely normal load, sliding distance and sliding velocity were considered in this study. The experiments for wear loss have been conducted as per Taguchi Design of experiments. An L27 Orthogonal array was employed for this purpose. The wear loss obtained for As Forged Inconel 690 is predicted by the Neural Network Toolbox of MATLAB R2015a using the Levenberg-Marquardt (trainlm) algorithm which trains the feed forward neural network having 3-6-1 (three input neurons, six hidden neurons in the single hidden layer and one output neuron). Experimental data sets from obtained from L27 Orthogonal array have been utilized to develop ANN. The results concluded that error for wear loss of As Forged Inconel 690 lies within 10% between experimental data and neural network prediction