Optimising machine-learning-based fault prediction in foundry production

Abstract

Microshrinkages are known as probably the most difficult defects to avoid in high-precision foundry. The presence of this failure renders the casting invalid, with the subsequent cost increment. Modelling the foundry process as an expert knowledge cloud allows properly-trained machine learning algorithms to foresee the value of a certain variable, in this case the probability that a microshrinkage appears within a casting. Extending previous research that presented outstanding results with a Bayesian-network-based approach, we have adapted and tested an artificial neural network and the K-nearest neighbour algorithm for the same objective. Finally, we compare the obtained results and show that Bayesian networks are more suitable than the rest of the counterparts for the prediction of microshrinkages. © 2009 Springer Berlin Heidelberg.