Scientists in the United States are studying shape memory alloys with the help of artificial intelligence (AI) to enable fighter jets to become more efficient and perform better.
The technology could enable the wings of the fighter jets to be folded using electrical heating and cooling, which allows for more efficient movement. The fighter jets (like the US F/A-18) need to be able to fit their wings to be carried on crowded aircraft carriers.
Currently, the system fighter jets use is comprised of heavy mechanical parts. This could change with the use of high-temperature shape memory alloys (HTSMAs).
The HTSMAs would allow the jet to move with less weight and more efficiency, meaning more jets get ready to fly at a faster pace with optimal energy use.
Shape memory alloys for more efficient fighter jets
To date, shape memory alloys have been plagued with one problem: they are usually quite expensive.
The scientists from the Department of Materials Science and Engineering, Texas A&M University, suggest that AI and high-throughput experimentation can be combined to accelerate materials discovery and reduce development costs.
This means the process could be done faster, resulting in more efficient materials at an affordable cost.
Designing new materials requires testing thousands of metal mixtures to find the right one, as even a tiny change can totally alter the way the material behaves.
Finding the right combination for the alloy could therefore be a total hit-and-miss experiment.
The team led by Department Head and Chevron Professor Dr. Ibrahim Karaman and Chevron Professor II Dr. Raymundo Arroyave has developed a data-driven approach to material discovery.
“This work shows that we can design better high-temperature alloys not through expensive trial-and-error but through smart, targeted exploration driven by data and physics,” said Karman.
“This project is exciting as it shows the power of the advanced alloy development frameworks we have been developing in the past years,” Arroyave added.
Designing the alloys
The team has brought in powerful computers and AI to predict how different metal mixtures would interact, so they don’t have to test every single option in their labs. This leads to a huge cut in the number of combinations they have to actually test in the lab.
The team has integrated machine learning and experimental work through a framework known as Batch Bayesian Optimization (BBO). BBO allows the team to refine their alloy prediction based on past experimentations, minimizing waste, and maximizing discovery efficiency.
“This framework not only speeds up discovery,” Karaman says, “but also opens the door to tailoring alloys for specific functions, such as reducing energy loss or improving actuation performance in many applications.”
Their goal is to design materials that change shape in response to heat or electricity, sort of like a muscle for machines.
These special materials – actuators – are used in aerospace, robotics, and medical devices. Once the goal is realized it could make US fighter jets agile and work better.
The findings have been published in the journal Acta Materialia.
Abstract
Chemical composition and thermal processing parameters are used in a first-of-their-kind machine learning (ML) and batch Bayesian optimization (BBO) approach in an iterative fashion in the quaternary NiTiCuHf high-temperature shape memory alloy (HTSMA) composition space to minimize thermal hysteresis in a desired transformation temperature range. The first of three iterations exploited an existing SMA database of lower complexity alloys (binary and ternary), attempting to optimize quaternary NiCuTiHf chemistry and thermal processing for the given constraint and the objective. Alloy synthesis and characterization revealed that the initial ML model displays high error levels between the predicted and experimental values, indicating the need for high-fidelity data in the complex quaternary alloy design space for optimization. The second iteration used this conclusion to explore an expanded design space through tuning Gaussian process (GP) hyperparameters. Utilization of active learning enabled the enlargement of data present in the high-complexity space during the iterative process, improving model accuracy. The third iteration discovered NiTiCuHf HTSMAs with the lowest reported martensitic transformation thermal hysteresis with transformation temperatures between 250 °C and 350 °C to date without precious metals. The effects of optimized secondary heat treatments on the martensitic transformation characteristics were explored and compared to those achieved after the initial homogenization heat treatments to demonstrate the ability of the BBO framework to create optimal alloys with controlled chemistry and thermal processing. In Ni-rich compositions of the designed alloys, the secondary heat treatments suggested by the BBO framework resulted in significant increases in transformation temperatures, suggesting the formation of Ni-rich precipitates.