AI-driven solution to high entropy alloy design for engine application

THE CHALLENGE

High entropy alloy (HEA) application ranges from state-of-the-art race cars, spacecrafts, submarines, nuclear reactors, and jet aircrafts among others. HEAs characterize the cutting edge of high-performance materials with complex compositions of multiple elements and striking attributes in contrast to conventional alloys. HEAs have considerably better strength-to-weight ratios with higher degrees of fracture resistance, tensile strength, and corrosion and oxidation resistance than conventional alloys. These attributes allow suitable alloying elements to increase the properties’ structural materials suitable for engine applications. Conventional alloy design approaches involve selection of the primary element (e.g. Fe, Cu, or Al), followed by inclusion of small amounts of additional elements to improve properties. For rapid advancement in HEA development, accelerating optimal HEA composition design for target properties is necessary. materialsIN, with its proprietary AI-powered solution, can enable rapid screening of optimized HEA composition from a seemingly infinite compositional space. materialsIN’s AI-powered product aims to expedite alloy design to address HEA design challenge.

THE MATERIALSIN SOLUTION

materialsIN employs its AI-driven solution for HEA design. materialsIN combines published HEA compositional and property data with proprietary data to create models to predict hardness (GPa), solidus temperature (°C), and Ductility ranking (0-5). Leveraging the learned knowledge of these models, materialsIN’s product, materialsINsight, builds an HEA atlas, which inherently embeds crucial chemistry-property association information. With advanced methodologies, materialsINsight can screen for optimal HEA chemistry for target properties utilizing chemical selection pathways. materialsINsight accelerates identification of new compounds from an infinite HEA compositional space with desired properties for aircraft engine application.

THE PRODUCT

materialsINsight delivers on multiple fronts, from developing robust predictive models to leveraging embedded chemistry-property information to identify new compounds via chemical selection pathway. A vital aspect of materialsINsight is to empower its user to make informed decisions regardings chemical(s) selection.

Property Prediction

The parity plots show the model performances by linking composition to solidus, hardness, and ductility. materialsIN methodologies achieve better model accuracy and predictive capability compared to other approaches. Additionally, based on these models, materialsINsight allows customers to categorize the elemental importance in determining the HEA properties. For solidus temperature prediction, the importance of elements in the HEA composition decreases as Ta > Hf > Re > Ru > W… indicating the presence of Ta in HEA having a significant impact on the solidus temperature. Similarly, for hardness the importance of elements influencing the hardness of the HEA decreasing with Ru > Re > Nb > Ta > Hf… For ductility, the impact of elements is categorized for overall ductility and different ductility ranks. materialsINsight reveals crucial insights on the importance of elemental composition for better HEA design.

Chemical Selection Pathways

materialsINsight provides a unique, simultaneous exploration of chemistry-property association framework to identify optimal HEA via chemical selection pathways over a HEA atlas. The HEA atlas embeds crucial chemistry-property association, which can be traversed for HEA design. The chemical selection pathway described below initiates at 45Mo28Re28W HEA with the objective to find HEA with solidus temperature < 2100°C and compositional constraint refraining inclusion of Hf. The tabulated data details changes in ductility, solidus, and hardness properties along with HEA chemistry as shown in the pathway figure through the pathway. The chemical selection pathway achieves the objective by identifying 20Mo20Re20Nb20Ru20W as an alternative HEA with lower solidus temperature than the initial 45Mo28Re28W HEA.

materialsINsight can provide exceptional predictive models and properties for hypothetical HEA. materialsINsight can recommend elemental variation required to achieve optimized property, and search for the best property HEA and its alternative with optimized composition for target property. materialsIN empowers customers with AI-driven informed, efficient, and cost-effective solution to screen optimized HEA design.