Porosity prediction in aluminium casting using approximate functions generated by evolutionary algorithm

Aluminum alloy is one of the popular materials due to its abundant supply and its good properties. However, its casting process has one undesirable effect - porosity. There are many attempts to predict its formation in order to avoid it. In this research, a way to predict total porosity percent from the initial chemical compositions and cooling rate after casting is proposed. The prediction is in a form of prediction function which utilizes the combination of genetic algorithm and differential evolution to generate the predicting formulae with datasets from other researches. The result functions achieve satisfying accuracy showing that they are capable of the good prediction.