Prediction of Friction Coefficient for Ductile Cast Iron Using Artificial Neural Network Methodology Based on Experimental Investigation

The key objective of the present study is to analyze friction coefficient and wear rate for ductile cast iron. Three different microstructures were chosen upon which perform experimental tests under sliding time, load, speed conditions. These specimens perlite + ferrite, bainitic. Moreover, an artificial neural network (ANN) model was developed in order predict using a set data collected during experiments. ANN structure made up four input parameters (namely number, nodule diameter) one output parameter (friction coefficient). Levenberg–Marquardt back-propagation algorithm applied train feed-forward back propagation (FFBP). results experiments revealed that reduced as increased constant load. Additionally, it exhibits same pattern action when test run with heavy load speed. increased, dropped. also show bainite harder wears less quickly than ferrite structure. pertaining showed single hidden layer more accurate double model. highest performance validation stage, however, observed at epochs 8 20, respectively, 0.012346 epoch 20.