Simulation of the nickel superalloys solvus temperature by the deep learning artificial neural network with differential layer.

Simulating the properties of complex alloys is an extremely challenging scientific task. The model should take into account a large number of uncorrelated factors, for many of which information may be absent or vague. The individual contribution of one or another chemical element out of a dozen possible ligants cannot be determined by traditional methods, and there are no general analytical models describing the effect of elements on the characteristics of alloys. Artificial neural networks are one of the few statistical simulation tools that may account many implicit correlations and establish correspondences that cannot be identified by other, more familiar mathematical methods. However, networks require complex tuning to achieve high performance. Data engineering and data preprocessing also makes a great contribution. This paper focuses on combining deep network configuration selection based on physics and input engineering to simulate the solvus temperature of nickel superalloys. [ABSTRACT FROM AUTHOR]