Using the group method for the synthesis of copper/ZrO2 nanocomposites to achieve high wear resistance by ball milling and spark plasma sintering

Abstract

The use of ZrO2 nanoparticles is common in the Cu matrix to manufacture Cu/ZrO2 nanocomposites. Many approaches and parameters produce Cu/ZrO2 nanocomposites with an appropriate wear rate and other properties. Hence, proposing an accurate and reliable model was imperative. The main goal of this paper is to find out the best models among the ant colony optimization (ACO), gene expression programming (GEP), artificial bee colony (ABC), and gray wolf optimization algorithm (GWOA) to predict the wear rate of/ZrO2 nanocomposites. To the best of our knowledge, this is the first research that predicts Cu/ZrO2 nanocomposites wear rate using group modeling. To develop models, 105 data were collected from the Cu/ZrO2 manufacturing by ball milling and spark plasma sintering in different conditions. The input parameters are Cu and ZrO2 concentration, milling time, and sintering temperature. The proposed rules for the prediction of wear rate based on the collected dataset are assessed based on several metrics such as (R square), mean absolute error percentage (MAPE), root mean square error (RRSE), mean square error (MSE), relative error and average absolute relative. According to the results, the ACO enables to predict wear rate by R square = 0.9923, MAPE = 5.2, RRSE = 0.0143, and MSE = 0.0043. Also, the sensitivity analysis results revealed that the milling time and ZrO2 content are the most effective parameters on the wear rate of Cu/ZrO2 nanocomposites. Moreover, experiments were done at the best model to prove the validity of the ACO. According to the results, it can be concluded that the ACO model and real analysis are fully functional to accurately estimate the optimal conditions for ball milling conditions in the production of Cu/ZrO2 nanocomposites.