Performance-driven Computational Design of Multi-terminal Compositionally Graded Alloy Structures using Graphs

TL;DR

This paper presents a unified, performance-driven computational approach for designing complex multi-terminal compositionally graded alloys using graph-based modeling, addressing challenges in high-dimensional alloy spaces and structural integration.

Contribution

It integrates recent graph-based design automation methods to enable the first performance-driven multi-terminal CGA design applicable to complex engineering structures.

Findings

Successfully designed a CGA for a gas turbine blade

Demonstrated applicability to other material systems

Addressed high-dimensional alloy design challenges

Abstract

The spatial control of material placement afforded by metal additive manufacturing (AM) has enabled significant progress in the development and implementation of compositionally graded alloys (CGAs) for spatial property variation in monolithic structures. However, cracking and brittle phase formation have hindered CGA development, with limited research extending beyond materials design to structural design. Notably, the high-dimensional alloy design space (systems with more than three active elements) remains poorly understood, specifically for CGAs. As a result, many prior efforts take a trial-and-error approach. Additionally, current structural design methods are inadequate for joining dissimilar alloys. In light of these challenges, recent work in graph information modeling and design automation has enabled topological partitioning and analysis of the alloy design space, automated design of multi-terminal CGAs, and automated conformal mapping of CGAs onto corresponding structural geometries. In comparison, prior gradient design approaches are limited to two-terminal CGAs. Here, we integrate these recent advancements, demonstrating a unified performance-driven CGA design approach on a gas turbine blade with broader application to other material systems and engineering structures.