Determination of thermal conductivity of eutectic Al–Cu compounds utilizing experiments, molecular dynamics simulations and machine learning

Abstract In this study, the thermal conductivity ( κ ) of Al–Cu eutectics were investigated by experimental and computational methods to shed light on role these compounds in properties connections compound casting. Specifically, nonequilibrium molecular dynamics (MD) method was utilized simulate lattice ${\kappa _{\text{l}}}$?> l six compositions with 5–30 at.% Cu. To extend results MD simulations bulk materials, instead using conventional linear extrapolation methods, a machine learning approach developed for dataset acquired from simulations. The bootstrapping find most suitable among support vector (SVM) polynomial radial basis function (RBF) kernels random forest method. showed that SVM model RBF kernel performed best, thus used predict . Subsequently, chosen produced induction casting their electrical conductivities measured via eddy current calculating electronic contribution Wiedemann–Franz law. Finally, actual alloys xenon flash compared values. It shown is capable successfully simulating system. addition, demonstrated decreases almost linearly formation Al 2 Cu phase due increase content. Overall, findings can be enhance cooling devices made