摘要
A modified Miedema model using four atomic parameters and pattern recognition or artificial neural network has been used to study the factors that affect the entropy of mixing of liquid binary alloy systems. It has been found that the systems with larger electronegativity difference (△Φ) usuallg have negative △Sxs of mixing, while the systems with larger valence electron density difference(denoted by △n) and small △Φ usually have positive △Sxs of mixing. The artificial neural network-atomic parameter method can be used to predict the △Sxs of binary alloy systems consisting of non-transition elements.
A modified Miedema model using four atomic parameters and pattern recognition or artificial neural network has been used to study the factors that affect the entropy of mixing of liquid binary alloy systems. It has been found that the systems with larger electronegativity difference (△Φ) usuallg have negative △Sxs of mixing, while the systems with larger valence electron density difference(denoted by △n) and small △Φ usually have positive △Sxs of mixing. The artificial neural network-atomic parameter method can be used to predict the △Sxs of binary alloy systems consisting of non-transition elements.
