Advanced fault prediction in high-precision foundry production

Microshrinkages are known as probably the most difficult defects to avoid in high-precision foundry due to the large number of factors involved in their apparition. The presence of this failure renders the casting invalid, with the subsequent cost increment. Bayesian networks allow to model the foundry process as a probabilistic constellation of interrelated variables. In this way, after a suitable learning process, the Bayesian network is able to infer causal relationships; in other words, it may guess the value of a variable (for instance, the presence or not of a defect). Against this background, we present here the first microshrinkage prediction system that, upon the basis of a Bayesian network, is able to foresee the apparition of this defect in order to avoid it. Further, we have tested this system in two real foundries and present here the obtained results.