Open Access

ARTICLE

Prediction of Dendritic Parameters and Macro Hardness Variation in PermanentMould Casting of Al-12%Si Alloys Using Artificial Neural Networks

by

1 MED, NIT Calicut, Kerala, India. Email: joseph@nitc.ac.in

Fluid Dynamics & Materials Processing 2006, 2(3), 211-220. https://doi.org/10.3970/fdmp.2006.002.211

Abstract

Aluminium-Silicon alloys are in high de-mand as an engineering material for automotive,aerospace and other engineering applications. Mechanical properties of Al-Si alloys depend not only on chemical composition but also more importantly on microstructural features such as dendritic alpha-aluminiumphase and eutectic silicon particles. As an additive to Al-Si alloys, sodium improves mechanical properties byforming finer and fewer needles like microstructures.Thus, prediction of the macro and microstructures obtained at the end of the solidification is of great interest for the manufacturer of aluminium alloys. Neuralnetworks are sophisticated nonlinear regression routinesthat, when properly “trained”, allow for the identificationof complex relationships between a series of inputs andone or more outputs. In this paper, an approach using ar-tificial neural networks for predicting alpha aluminiumdendritic parameters(fraction and secondary dendriticarm spacing) and macro hardness variation (Brinell hard-ness number) of permanent mould casting of Al-12%Sialloy is described. This approach has the advantage thatcomplex interactions among cooling rate, solidificationvelocity and chill position on the amount of dendritic al-pha Aluminium phase within a fixed modifier content al-loy can easily be taken into account.

---

Keywords

---