For people living near large international airports, loud aircraft noise is a never-ending nuisance. The noise pollution has even been linked to increased risk of cardiovascular disease. Some airports are helping pay for the sound-proofing of area homes, but the makers of these engines so necessary to 21st century travel and commerce are approaching the problem from the other end by designing quieter engines.
Engineers are making strides to increase engine efficiency and decrease noise with the help of the country’s most powerful supercomputers. The problem doesn’t just affect jet engines, either, the low-frequency noise is an environmental aggravation that also threatens the wind-energy industry.
One of the companies most associated with jet engines and wind turbines is General Electric, which has been building turbo-machines for over a century. In partnership with Argonne National Laboratory, GE Global Research is using the Argonne Leadership Computing Facility’s (ALCF’s) high-performance computers to study the air as it passes through jet exhaust nozzles and over wind turbine blades.
Using an approach called large eddy simulations (LES), the researchers are working to understand and predict turbulent flow features. By examining variables such as velocity, temperature and pressure, they can characterize some of the key flow physics of this multi-scale turbulent mixing phenomenon. The information revealed by this study will lay the groundwork for more efficient wind turbines and jet engines.
An article on the ALCF website explains that with the demand for lighter, more fuel-efficient structures, understanding the physics of advanced blade design becomes increasingly important – and computation-based insight is the key. Resolving the wall bounded turbulent flow around a wind turbine airfoil takes a lot more computing power than simulations involving jet mixing as in jet engine exhaust. The ALCF’s 10-petaflops Mira supercomputer has proven to be up to the challenge. These complex simulations used to take two to three months, but running on the IBM Blue Gene/Q, they can be computed in just two weeks.
“These are what we call high-fidelity LES, which are very accurate. You are directly predicting the jet engine noise, without modeling the turbulent flow noise sources as in a traditional RANS (Reynolds Averaged Navier Stokes) approach,” says team leader Umesh Paliath, a GE Global Research scientist.
Working together, the researchers from ALCF and GE are using the new data to improve noise prediction models. The project holds the promise for a new breed of quieter, more efficient engines and could also breathe new life into wind power efforts.