Estimation of Mechanical Properties of Copper Powder Filled Linear Low-Density Polyethylene Composites

Purpose: The complex geometry of many composites is in a loose multi-phase and the large difference in the mechanical and electrical properties of the different components makes it difficult to predict the effective properties of the composites. The mechanical properties of copper powder filled linear low-density polyethylene (LLDPE) were predicted using an artificial neural network (ANN) approach. Method: Artificial neural networks have been used to predict the mechanical properties of loose multi-phase material systems. ANN is a network motivated by biological neural networks. ANN is based on Feed Forward Back Propagation (FFBP) using three different training functions (TRAINGDA, TRAINGDM, and TRAINGDX). The ANN approach runs the threshold TANSIG-PURELIN function for 200 epochs with a back propagation algorithm. The input parameters manipulated for the prediction were elongation at break (δ), stress at break (ρ), Young’s modulus (Y), volume fraction of the filler (ϕ) and constants (KE1,KE2,KS1,KS2,KS3,KS4,KY1). Copper powder filled LLDPE has a complex structure which makes it difficult to accurately predict the mechanical properties. This prediction was done using the ANN approach. Results: The theoretical models were compared with the experimental data and there was a good agreement between some models and the data. Conclusion: In line with the experimental data, we found that as we increased the volume fraction of the copper powder, the elongation and stress at break of the composites decreased, while the Young’s modulus increased.