Optimum abrasive wear resistance for epoxy composites reinforced with polyethylene (PET) waste using Taguchi design and neural network

Abstract

The current work presents a study of the tribological properties of composite materials designed based on polyethylene terephthalate (PET), which has an important role in the structures of machines, represented by tribological couplings made of composite polymers. The paper examined the effect of two factors, namely recycled waste heating time (HT) and weight percentage (wt. %), on the improvement of the abrasive wear resistance of micro-filler-reinforced epoxy composites. The current research aims to develop epoxy composites by improving abrasive wear resistance while ensuring low cost and weight. Improving wear resistance due to the use of epoxy composites to connect joints that operate under conditions without lubrication in various industrial fields will increase their operational life. The signal-to-noise ratio was analyzed to find out the effect of test parameters HT and wt. % on the wear rate of epoxy composites. Using MINITAB 19 software, regression equations were obtained for each variable to compare it with the Artificial Neural Network (ANN) results. Predictive models based on the regression equation and artificial neural network were developed to predict the wear rate of epoxy composites, and to determine which model is more efficient, their results were compared and the most appropriate model with the low error was determined. The results of the current research showed that the wear resistance of epoxy composites reinforced with RCCF improved by 41 % when increasing wt. % and HT, and also showed that the ANN model is more suitable than the regression model for predicting the wear rate of epoxy composites