Heat exchangers are very important to the evolution of the aerospace industry, but we need new materials to meet the requirements of production and product performance.
Through the case of the cooperation between Intellegens and GKN Aviation, we will learn how Intellegens' machine learning tool AIchemite can provide Boeing Axal Accelerator with the highest thermal conductivity products without reducing mechanical performance requirements.
Artificial intelligence realizes material design iteration
The article "Artificial Intelligence and Material Technology Achieve Superalloys and Reveal How Fraunhofer’s FutureAM Project Helps Next Generation Aircraft Engines" explains how Fraunhofer IWS experts have adopted "artificial intelligence" (AI) and "Artificial Intelligence" (AI) and "Artificial Intelligence" in the futureAM next-generation additive manufacturing project. The advanced method of "machine learning" to improve the understanding of the processing process was researched by the Fraunhofer IWS Image Processing and Data Management Working Group. Through artificial intelligence, the hidden connections in these data floods can be found.
It can be said that replacing a large number of boring material development processes with artificial intelligence is an established trend.
Challenge
The next key milestone in aviation is the introduction of sustainable fuels, such as batteries or hydrogen energy. Future aircraft design will require internal cooling and heating units, and heat exchangers are a very necessary part.
3D printing-additive manufacturing can realize very complex heat exchangers to improve heat exchange performance. In addition, the heat exchanger is a structural part, and the material must be sufficiently strong.
This combination of high thermal conductivity, high strength and suitability for additive manufacturing makes there are not many materials suitable for manufacturing heat exchangers. Therefore, the industry needs to develop new alloys to meet the 3D printing needs of this heat exchanger and meet the needs of the next generation of fuel aircraft.
Ti-6Al-4V (Ti-64) is an alloy of titanium, aluminum, and vanadium. This material is widely used in 3D printing-additive manufacturing, and has high mechanical properties and corrosion resistance, but its The thermal conductivity is relatively low, so this material has not been adopted as a material for heat exchangers.
However, through artificial intelligence designing new titanium alloys, Intellegens and GKN Aviation have discovered new possibilities.
Alchemite artificial intelligence engine
Solution
Intellegens uses Alchemite artificial intelligence engine to analyze titanium alloy material data with GKN Aviation. In this development process, 256 alloys were considered for analysis by the artificial intelligence engine, and approximately 20 physical properties were estimated.
Optimizing lightweight
The purpose of Alchemite optimization is to obtain materials with higher thermal conductivity and mechanical strength. One of the findings is that there is 3% vanadium, 1.9% molybdenum, 1.5% iron and other elements including 0.31% palladium, 0.41% ruthenium, and no aluminum.
This material is predicted by artificial intelligence to have a thermal conductivity of 18.4W/mK (the error range is plus or minus 1.1), and the ultimate tensile strength is 595MPa (the error range is 50.7).
Alchemite artificial intelligence engine almost finds the best combination in the two directions of maximizing thermal conductivity and tensile strength.
Experts from GKN Aviation confirmed that due to the high content of palladium in the new material combination, this may limit the previous application prospects. How to reduce the content of palladium metal can be adjusted through the Alchemite artificial intelligence engine, without the need for a lot of real material ratio experiments. The true ratio is not only time-consuming, but also expensive.
However, the content of palladium metal is not suitable to be reduced to a minimum level of zero. In this case, the performance of heat exchange will be very bad. The role of palladium metal is to improve the heat exchange performance, which also shows that in other alloy combinations, palladium metal can be added as a key element to improve the heat exchange performance.
The Alchemite artificial intelligence engine can shorten the material development and design process that usually takes 2 years to 3 months. In the next step, this project can be extended to use copper, nickel, and aluminum alloys as research objects and develop heat exchangers that meet performance requirements by adjusting the combination of metal elements. Through the Alchemite artificial intelligence engine, the development process can be visualized and the development progress can be continuously obtained.
Through 3D printing technology can help achieve more cost-effective high-performance materials, artificial intelligence will play a decisive role in development.
Just like the role of artificial intelligence in the pharmaceutical field, it usually takes 10-15 years for a new drug from the beginning of research and development to clinical trials and then to market; with the advent of the digital economy era, the application of big data, artificial intelligence and other technologies will greatly Shorten drug development time, improve efficiency and quality. In the pharmaceutical industry, people are interested in implementing AI-driven solutions to discover new drugs and accelerate their time to market. The Food and Drug Administration has further promoted this interest, promoting innovation in the use of AI-based technologies for drug development. Overall, AI and machine learning aim to change the drug discovery process, thereby reducing financial costs and time to market.
The same thing will happen in material development in the field of 3D printing. Artificial intelligence will play a role in two dimensions: reducing the financial cost and development cycle of material development.
In addition to the development of a single material, as Fraunhofer discovered in the futureAM project, it is usually not very effective to design the entire component of an aircraft engine with a single material, because the components will not receive the same heat at all points. It is best to use expensive high-resistance materials only in places with high temperatures. In other areas, cheaper materials are sufficient. This is exactly what the additive manufacturing system can achieve. Once the artificial intelligence has learned to process the superalloys needed, the next step is to integrate various high-performance materials into one component.
Another new alloy designed by Intellegens’ Alchemite deep learning algorithm. This new alloy is manufactured through a directional energy deposition (DED) metal 3D printing process. The alloy can meet the performance goals required by additive manufacturing for manufacturing Jet engine parts.
There are millions of commercial materials around the world, which are characterized by hundreds of different characteristics. Using traditional techniques to explore the information we know about these materials and propose new substances, substrates and systems is an arduous process that may take months or even years. By understanding the basic correlations in existing material data and estimating missing attributes, artificial intelligence can quickly, efficiently and accurately propose new materials with target attributes-thus speeding up the development process.