Hypersonic Flight Race: NASA, ANL Develops Aircraft Engine Simulation Using AI and Machine Learning

US fighter jets like the F-15 vehicles have a maximum speed that could surpass Mach 2, or twice the speed of sound. At this level of velocity, the jets undergo a supersonic phase. Compared to Mach 2, hypersonic flights are much faster. Mach 5 is the perfect example, as an aircraft moving at this rate can reach almost 5,000 kilometers per hour.

Supersonic and Hypersonic Flight Simulations

Hypersonic speeds are undoubtedly among the fastest displacement measurements in the modern-day. The type of speed was first introduced in the 1950s as part of the development of propulsions for aircraft and rockets.

Today, the search for the perfect hypersonic speed is still ongoing. The main interest of propulsion specialists behind the advancements is to develop a more affordable and accessible design than conventional rocket launches.

Hypersonic speed in a vehicle can also help many industries that use propulsion techniques, such as space missions, commercial flights, and even national defense.

A new study carried out by experts from NASA and the Argonne National Laboratory (ANL) incorporated machine learning systems to gain data and reduce the required memory and cost in computational fluid dynamics (CFD) simulations.

Computational fluid dynamics are essential tests that simulate flights with respect to fuel combustion rates and an aircraft's supersonic and hypersonic capacity.

The neural network-based supersonic combustor research was discussed last January at the prestigious American Institute of Aeronautics and Astronautics SciTech Forum.

Before any aircraft's major assembly and testing, aviation engineers utilized CFD simulations to have specific data about the forces that can affect and interact with an aircraft during flight.

Aircrafts at supersonic speed produce shockwaves when they exceed the speed of sound and pass through what we know as a sound barrier. During this phase, the vehicle creates a disturbance around itself, making the air denser, reaching high temperatures, and a violent pressure, according to The EurAsian Times.

Hypersonic speed, compared with supersonics, can reach measurements on a whole new level. In addition, the friction during sound barrier bypass during a hypersonic speed can melt some parts of a conventional vehicle, especially those in the commercial industry.

Hypersonic Flight Race

In the study, it was shown that air-breathing engines could utilize oxygen more effectively while on flights. CFD simulations could help how this sustainability can be perfected and predict other factors that could impact the aircraft, not just outside the vehicles but also during the exchange in the plane's engines and the oxygen's activity with the fuel.

Argonne National Laboratory's Advanced Propulsion and Power Research interim director Sibendu Som, who also co-authored the study, explained that aircraft engines are delicate with regard to the chemical and turbulence interaction.

With that said, their team developed advanced simulation models that show combustion rates and CFD codes, which could lead to advanced information on the combustion physics of supersonic flights.

NASA previously developed a hypersonic CFD code called VULCAN-CFD. This system has flamelet tables, where each flamelet serves as the small unit of flame in an entire combustion model. It can also store snapshots accumulated during the fuel-burning phase. All of these processes take up a large amount of computer memory.

Together with ANL, the space agency developed a machine learning AI that could simplify the CFD computations and simulations while reducing the computational costs as well;l as the memory requirements.

Their AI-based software was tested and worked perfectly along the flamelet table produced by NASA's VULCAN-CFD. Experts believe that future studies on aviation and propulsion will be accompanied by the assistance of the ever-growing technologies of artificial intelligence and machine learning.

Many countries are now joining the race to achieve the peak hypersonic speed of air flights. China's Space Transporation announced last February plans to create a hypersonic plane that can fly over 11,000 kilometers per hour.

An exhibition from the firm is expected by 2050; wherein a plane will fly from Beijing to New York in just an hour or less.

The study was published in the AIAA ©SCITECH 2022 Forum, titled "Deep neural network based unsteady flamelet progress variable approach in a supersonic combustor."

Check out more news and information on Artificial Intelligence in Science Times.