Predicting The Hardness and Porosity of a Smart Alloy (Cu-Al-Ni) with Nanoparticles Added, Using Smart Neural Networks.

Abstract

This research paper is dealing with determined the best ratio between the experimental and prediction results and comparison between them by using the Artificial Neural Network for the mechanical properties of Hardness and porosity of smart (83% Cu-13% Al-4% Ni) alloys by adding aluminum nanoparticles (0%, 5%, 10%, 15%). Those mechanical properties have major technological and commercial interest like industrial and aerospace applications, and also in high damping composites, sensors, actuators, and filters.

To ensure the phase responsible for the creation reliability of this type of alloy, this is by employing scanning electron microscopy (SEM) and X-ray diffraction (XRD).

The prediction process utilizing the ANN tool in Matlab R2020a software is separated into two stages: the first is to select the best network to predict the best outcomes for the experiment's inputs. In order to decrease the expense, effort, and time necessary to carry out numerous further trials in order to attain these findings, the second stage entails using this best network for comparison between the predicted and experimental results. The forward back propagation algorithm was used in all networks of ANN.

Show the results that increasing the percentage of nanoparticle addition leads to an increase in hardness where its value was (136) for the sample without addition, while it reached its maximum value (190.7) when 15% of the nanoparticles were added. The porosity test showed a reverse behavior from the hardness test, where the porosity increased when no nanoparticles were added, and its value was (21.54), while its value (3.245) when added was 15%.