An artificial neural network-based system is proposed to predict mechanical
properties in Spheroidal cast iron. Several castings of various compositio
ns and modules were produced, starting from different inoculation temperatu
res and with different cooling times. The mechanical properties were then e
valuated by means of tension tests. Process parameters and mechanical prope
rties were then used as a training set for an artificial neural network. Di
fferent neural structures were tested, from the simple perceptron up to the
multilayer perceptron with two hidden layers, and evaluated by means of a
validation set. The results have shown excellent predictive capability of t
he neural networks as regards maximum tensile strength, when the variation
range of strength does not exceed 100 MPa. (C) 2000 Elsevier Science S.A. A
ll rights reserved.