Living near an airport could be hazardous to your health. Noise from aircraft taking off and landing can exceed 140 decibels. Even much lower levels of noise pollution can have significant health effects such as hypertension, heart disease, and sleep disturbance.

It’s not just people unfortunate enough to live next-door to an airport that are affected. As aircraft approach landing, FAA rules require that pilots deploy the landing gear and reduce engine thrust. That happens several miles out from the airport, meaning aircraft are flying at a low altitude and producing loud noise right over a much larger swath of the population.

That’s why NASA is working to reduce landing noise and improve the quality of life for communities near major airports.

“As a government agency, our role is to reduce noise pollution for communities on the ground, because those are our stakeholders,” says Mehdi Khorrami of NASA’s Langley Research Center. “We work for the benefit of the public who are adversely impacted by aircraft noise as they come in for landing.”

Finding the source of the sound

In previous generations of aircraft, engine noise used to be the greatest concern. But with the phase out of older models, the airframe is now the dominant source of noise during landing.

As air flows over the complex components of an aircraft’s undercarriage and wing high-lift devices, it generates a lot of turbulence. The different sizes and shapes of those components create a broad range of frequencies. The resulting noise is propagated in all directions, but only the sound radiating down from the plane affects the communities below.

Before noise can be reduced, the location of the loudest sounds must be identified. To do that, scientists use a phased microphone array, which consists of a large number of microphones usually arranged in a spiral pattern. This array works as an acoustic camera.

Just like a visual camera, the array can be directed to “view” specific parts of the plane. A grid is created to map the underside of the aircraft and then the microphones are steered towards each point in the grid. Based on the amount of noise that reaches each microphone and the time it takes for a signal to reach them, researchers can identify, very specifically, where most of the sound waves or pressure fluctuations are coming from.

Once they have that information, researchers turn to computer simulations to start experimenting with potential noise-reduction strategies. In this case, NASA is developing a full-scale Boeing 777 simulation.

No longer a joke

“Even eight or nine years ago, if I told people, ‘Hey, in a few short years we’ll simulate the full-scale aircraft,’ I would have been laughed out of the room,” says Khorrami.

But rapid increases in computational power and significant improvements to algorithms have changed all that. Khorrami points particularly to the computational tools developed for modern supercomputers that enable running a lot of computations quickly and distributing those computations over a large number of processors.

Even initial low-resolution simulations required over 6000 cores and used over 1.4 million processor hours on Pleiades, one of the world’s top 50 supercomputers. “But this is still not our highest resolution,” Khorrami says. “We want to go further.”

He estimates that to capture the high-frequency segment of the noise spectrum, the next finer-resolution simulation will require as many as 7.5 million processor hours, distributed over 10-12,000 Pleiades cores.

Once the simulation results are validated, NASA scientists will work closely with the manufacturer to share their new understanding of airframe noise.

“We are trying to demonstrate that we are at the stage where we can apply some of these tools to our grand challenge,” says Khorrami. “The results we’re getting are extremely promising and show very good correlations with wind tunnel measurements as well as flight testing.”

But ultimately, building the noise reduction hardware and whether it actually flies on a real aircraft is a decision for the manufacturer to make.

Sticking with it

For Khorrami, the multidimensional nature of aircraft noise is the best part of this challenge. Many simulation problems are steady state, meaning that they don’t change over time. But for airframe noise simulations, multiple factors have to come together.

The computer models must accurately depict how the turbulent flow evolves over time, as well as represent fluctuations in hydrodynamic pressure. And then, because the generated sound is only 1/1000^th or less of the magnitude of the pressure, the simulation must be extremely accurate in order to be useful. Says Khorrami, “It’s almost like finding a needle in a haystack.”

By pushing the envelope on noise prediction capability, Khorrami and his team have made advances that will benefit other researchers working in related areas, such as unmanned aerial vehicles and urban air mobility.

“To be honest,” Khorrami says, “a lot of my contemporaries thought that not only would we retire, but we would die before we saw the day that we could do the full aircraft simulation.”