Marble waste is the principal solid waste generated during the processing of marbles in stone processing industries. Managing solid waste create a challenge for society. In the present study, composites are fabricated with different proportions (0, 5, 10, and 15 wt. %) of waste marble powder (WMP) along with epoxy and glass fiber by a simple polymer casting method. Erosion wear characterization of the composites have been conducted by following ASTM G76 standard. The experiment was statistically designed as per Taguchi’s L16 model. The incorporation of WMP had shown an improvement in erosion resistance of composites. A theoretical model has been developed to predict the erosion wear rate of the composites. Subsequent to this, a neural network model is also developed to predict the erosion wear properties of glass-epoxy-marble (GEM) composites based on their experimental results obtained from the erosion test. The input data were trained in such a manner that the neural network model can precisely predict erosion wear properties of these composites. Moreover, the experimental results are compared with the predicted data to validate the neural network model. It is inferred, from the validation that the developed model can be utilized to predict the erosion wear rate of composites up to an accuracy of 95%.
Keywords: glass fiber; waste marble particle; erosion resistance; neural network model