The performance of various artificial neurons interconnections in the modelling and experimental manufacturing of the composites
Predstavitev različnih umetnih nevronskih povezav pri modeliranju in eksperimentalni izdelavi kompozitov

english

2012

volume 46, issue 2

composite    hardness    mechanical properties 

Inštitut za kovinske materiale in tehnologije
(Obvezni izvod spletne publikacije)

This study reports the performance of different artificial neural network (ANN) training algorithms in the prediction of mechanical properties. First, an experimental investigation was carried out on the mechanical behavior of an A356 composite reinforced with B4C particulates and then an ANN modeling was implemented in order to predict the mechanical properties, including the yield stress, UTS, hardness and elongation percentage. After the preparation of the training set, the neural network was trained using different training algorithms, hidden layers and the number of neurons in hidden layers. The test set was used to check the system accuracy for each training algorithm at the end of the learning. The results show that the Levenberg-Marquardt learning algorithm gave the best prediction for the yield stress, UTS, hardness and elongation percentage of the A356 composite reinforced with B4C particulates.

10.07.2013