Virtual defects preventing real failures
Aerospace materials undergo the harshest environments. Not only is there nearly no margin for error or failure, but also the materials and components cannot be designed in an overly conservative manner, as that can lead to inefficient energy use and more fuel and exhaust gas byproducts. We’re using computational tools to mathematically simulate and track material microstructures and defects to address their probability of failure during operation.
Structural aerospace components are made of materials like high-temperature alloys, lightweight alloys, fiber-reinforced composites, and ultra-high-temperature ceramic-based composites. In what’s called the “hot section path” of a gas turbine engine, a subset of these materials, such as superalloys designed for high temperatures, are subjected to extreme temperatures of around 2,500°F, as well as extreme loads. At full speed, the turbines rotate at far more than 3,000 revolutions per minute — a speed at which the energy stored in the disk is equivalent to the energy required to launch a small car over a 7-story building.
The polycrystalline materials in aerospace components have a large number of grains, each with a lattice-like arrangement of atoms. The materials deform through the generation and movement of defects within the grains (each deformation step typically is sub-nanometer), and the accumulation of these defect movements leads to deformation that initiates a crack. The problem is complex — the materials have a wide distribution of microstructural features at various length scales, so the micromechanical fields are non-uniform and constantly evolving.
While the deformation and failure of aerospace materials have been studied since the conception of aerospace, a universal and consensus understanding is not available. Previous predictive tools have relied on empirical relationships.
We’re using physics-based computational simulation tools to create “virtual instantiations” of microstructures and defects and view them through a mathematical framework to track how the variables interact and evolve. Based on localization of strain and concentration of stress, we monitor the local energy in the system; when this local energy reaches a critical point with respect to service life, we define it as failure. Through the creation of many virtual instantiations, we can calculate the probability of failure for this material in service. These simulations are coupled with companion experiments that track the evolution of thousands of data points, providing unprecedented information for model comparison and validation.
The advantage of this approach is that we can predict service life based on physics, as opposed to observations and experience. This is important because it obviates running hundreds of thousands of tests to qualify usage of a material, thus saving time and money. Moreover, our method enables us to optimize the design, manufacturing and performance of a component based on its fatigue assessment. Traditional approaches lock in the material and then conduct tests to determine its reliability, meaning any changes require the tests to be repeated.
We have worked with the U.S. Department of Defense, National Science Foundation, and NASA, as well as gas turbine industry partners. While we have answered many questions about fatigue behavior, we still need to develop additional trust in this approach — “mature its reliability” — in order to obtain widespread acceptance and update the tools and techniques for determining the lifetime of aerospace components.
There have been tremendous advances in the processing of aerospace materials, as well as in the digital engineering tools that designers, manufacturers, aerodynamicists and structural analysts use. But for fatigue assessment, approaches developed in the 1940s to 1970s still are being used. We see enormous potential to modernize these approaches. The advances we’re pursuing will allow us to interface with others in the digital engineering supply chain and begin to optimize for fatigue-resistant designs — leading to higher performance and more energy-efficient aerospace designs.
Michael D. Sangid, PhD
Elmer F. Bruhn Associate Professor of Aeronautics and Astronautics
Associate Professor of Materials Engineering (by courtesy)
College of Engineering, Purdue University
Related Links
Advanced Computational Materials and Experimental Evaluation (ACME) Laboratory
Professor Michael Sangid’s team receives 2020 Manufacturing Leadership Award
Professor Sangid tapped for Realizing Next-Generation Smart Manufacturing team
