Highly interpretable machine learning framework for prediction of mechanical properties of nickel based superalloys

Superalloys are a special class of heavy-duty materials with excellent strength retention and chemical stability at very high temperatures. Nickel-based superalloys are used commercially in aircraft turbines, power plants, and space launch vehicles. The optimization of mechanical properties of alloys has been traditionally carried out using experimental approaches, which demand massive costs in terms of time and infrastructure for testing. In this paper, we propose a method for mechanical property prediction of Ni-based superalloys by learning from past experimental results using machine learning (ML). Five highly accurate ML models are developed to predict yield strength (YS), ultimate tensile strength (UTS), creep rupture life, fatigue life with stress, and strain values. We have developed an extensive database containing mechanical properties of over 1500 Ni-based superalloys. Basic material parameters such as the composition of the alloy, annealing conditions, and testing conditions are also collected and used as features for developing the ML models. The prediction root mean squared errors for the YS, UTS, creep, and fatigue life models are 0.11, 0.06, 0.19, 0.22, which are minimal, leading to a highly accurate estimation of the target values. These ML models are highly transferable and require a minimum number of input features. In addition, feature analysis performed by SHapley Additive exPlanations (SHAP) for individual properties reveals the relative significance of each descriptor in deciding the target property. We demonstrate that a unified and highly accurate ML framework can be developed using common features for all mechanical properties. The models are developed on experimental data, making them directly applicable for industries.

Figure 1

Schematic representation of the work. (a) A database is established with experimental data from the literature. The data is subjected to rigorous filtering conditions and standardization. (b) Feature engineering is performed for individual data sets to identify the relevant descriptors and study the effect on the target property. (c) ML models are developed for individual properties using the identified crucial features.

Figure 2

The actual versus predicted values of best GPR models for (a) YS, (b) UTS, (c) creep, (d) fatigue with stress as feature (fatigue stress), (e) fatigue with strain as feature (fatigue strain). The data in all the mechanical properties lies near the black line, which implies that the ML models are highly accurate.

Figure 3

Feature importance of composition and experimental conditions (stress and temperature of experiment) for the prediction of YS, UTS, creep rupture life. The correlations are presented as heat map for individual features.

Figure 4

Feature importance for fatigue life predicted through (a) stress range(MPa), (b) strain range (%).

Figure 5

SHAP force plot for some individual data set demonstrating the effects of rhenium addition on the creep rupture life of the material. The red arrows indicate that addition of a feature leads to increase of creep rupture life, while blue arrows represent that addition of the feature leads to decrease of rupture life.

Figure 6

SHAP dependence plot for rhenium addition on creep rupture life of the superalloys.