Analysis of Sliding Wear Characteristics of BFS Filled Composites Using an Experimental Design Approach Integrated with ANN

Abstract

Short fiber-reinforced polymer composites are used in numerous tribological applications. In the present work, an attempt was made to improve the wear resistance of short glass fiber (SGF)-reinforced epoxy composites by incorporation of microsized blast furnace slag (BFS) particles. The effect of various operational variables and material parameters on the sliding wear behavior of these composites was studied systematically. The design of experiments approach using Taguchi's orthogonal arrays was used. This systematic experimentation led to identification of significant variables that predominantly influence the wear rate. The Taguchi approach enabled us to determine optimal parameter settings that led to minimization of the wear rate. The morphology of worn surfaces was then examined by scanning electron microscopy and possible wear mechanisms are discussed. Further, in this article, the potential of using artificial neural networks (ANNs) for the prediction of sliding wear properties of polymer composites is explored using an experimental data set generated from a series of pin-on-disc sliding wear tests on epoxy matrix composites. The ANN prediction profiles for the characteristic tribological properties exhibited very good agreement with the measured results, demonstrating that a well-trained network was created. The simulated results explaining the effect of significant process variables on the wear rate indicated that the trained neural network possessed enough generalization capability to predict wear rate from any input data that are different from the original training data set.