Search

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]

Based on heat-resistant nickel, superalloys are the basis for creating the critical parts of gas turbines that operate in extreme temperature conditions for the long periods of time. The alloys exhibit remarkable resistance to mechanical and chemical degradation under severe temperature and time loads. One of the major superalloys' mechanical properties is the rupture strength. The time to rupture that is related to the rupture strength is, also, of a particular engineering interest. In order to determine these features, particular tests are carried out. Conditions of the Western and Russian experiment differ, however, they might be united in neural model where test sample features form separately. In this work, we build a Deep Learning artificial neural network, teach it with the Western superalloys results, and predict the rupture strength and time to rupture. The results obtained and the prediction error distribution show the ability to unite different test results in one model and confirm wide prospective for the neural modeling of the superalloys properties. [ABSTRACT FROM AUTHOR]

The nickel-based superalloys are unique materials with complex doping applied to manufacturing the gas turbine engine parts. The alloys show resistance to mechanical and chemical degradation under high pressure, high temperature, and long-term isothermal exposures. One of the main alloys’ service properties is the heat resistance. Numerically, it is expressed in the tensile strength values (MPa). Simulation of the heat resistance behavior is an important engineering task, which would significantly simplify the analysis of existing and designing the new alloys. In this paper, we use results of the heat resistance simulation by an artificial neural network, as well as, experimental data for approximating the changes in the heat resistance vs isothermal exposures expressed in the complex Larson-Miller parameter by a sigmoidal function. [ABSTRACT FROM AUTHOR]

The nickel-based superalloys are unique materials with complex alloying used in the manufacture of gas turbine engines. The alloys exhibit excellent resistance to mechanical and chemical degradation under the high loads and long-term isothermal exposures. The main service property of the alloy is its heat resistance, which is expressed by the tensile strength. Simulation of changes in the heat resistance is an important engineering problem, which would significantly simplify the design of new alloys. In this paper, we apply a deep learning neural network to predict the tensile strength values and to compare the predictive ability of the proposed approach. Also, the results are presented of the feed-forward neural network accounting changes in heat resistance vs isothermal exposures that are expressed in the complex Larson-Miller parameter. [ABSTRACT FROM AUTHOR]