Prediction of the mechanical properties of copper powder-filled low-density polyethylene composites. A comparison between the ANN and theoretical models

Abstract

In the present study, the mechanical properties of copper (Cu) powder-filled low-density polyethylene (LDPE) composites are predicted by using artificial neural networks (ANNs) as a function of the filler concentration. An ANN is a form of artificial intelligence, which attempts to mimic the function of the human brain and nervous system. A three-layer feedforward ANN structure was constructed and a backpropagation algorithm was used for training ANNs. The ANN models are based on a feedforward backpropagation (FFBP) network with such training functions as the Levenberg– Marquardt (LM), conjugate gradient backpropagation with Polak–Ribiere updates (CGP), Broyden, Fletcher, Goldfrab and Shanno (BFGS) quasi-Newton (BFG), one-step secant (OSS), and resilient backpropagation (RP). The volume fraction and different mechanical properties of continuous (matrix) and dispersed (filler) phases are input parameters to predict the different mechanical properties such as elongation at break, stress at break, and Young’s modulus. A training algorithm for neurons and hidden layers for different feedforward backpropagation networks runs at the uniform threshold function TANSIG-PURELIN for 1000 epochs. Our ANN approach confirms that the mechanical properties of copper powder-filled LDPE composites are predicted in excellent agreement with experimental results. A comparison with other models is also made and found that the values of mechanical properties predicted by using present model are in good agreement with the reported experimental values.