Abstract
Cold upsetting experiments were carried out on sintered aluminium iron composite preforms in order to analyse the formability. A Neural Network (NN) model has been developed with a radial basis neural network algorithm to find the effect of particle size and iron content on formability. The training data were collected by the experimental setup in the laboratory for the sintered aluminium iron composite preforms with various preform densities, particle size, iron content and aspect ratios. The network is trained to predict the deformation characteristics such as formability index β, instantaneous strain hardening coefficients ki and ni and instantaneous density coefficients Bi and Ci. It is found that formability index increases rapidly at the earlier stages of deformation and followed by a gradual increase with further increase in true axial strain. It is also observed that coarse particle size and higher addition of iron content exhibits improved formability index values. In addition, instantaneous density coefficients Bi and Ci of the aluminium iron composite preforms were mathematically evaluated and simulated. Regression analysis has confirmed a well coincidence between predicted and experimental data with more accuracy. Hence, this approach helps to facilitate a knowledge base in order to generate advice for the designer at the earlier stages of design by forecasting and preventing the surface defects at the computer design stage itself. It also avoids the cost of prototyping or trialing.