The Art of Science: Supercomputers Help Scientists See What Microscopes and Cameras Can’t Capture

Scientists at GE Global Research have been using the world’s most powerful supercomputers to simulate everything from fuel flowing through jet engine nozzles to water drops turning into ice. The results can be rewarding beyond solving research riddles. “Many times our work generates images that are visually breathtaking,” says Rick Arthur, who leads the Advanced Computing Lab at GRC.

Supercomputers are helping GE engineers speed up innovation, crack previously intractable problems, and shorten the business cycle. Take a look at our slideshow featuring a hypnotizing turbine flow, density gradients and other arresting images generated by GRC scientists.

[slides image_align="left"]
[image src="http://files.gereports.com/wp-content/uploads/2013/08/ArtInScience1.jpg"]
The “blue blobs” shown in this picture represent particles in an advanced nickel alloy used to manufacture high-pressure turbine rotors and cooling systems for jet engines. The model is a simulation of what happens to the size and distribution of the particles when the alloy rapidly cools at a rate of 200 degrees Fahrenheit per minute.
[/image]
[image src="http://files.gereports.com/wp-content/uploads/2013/08/injector36in.jpg"]
This picture shows a simulation of a liquid spray from a jet engine fuel injector. Fuel injectors have an intricate design and must handle punishing heat and pressure. They are notoriously difficult to test and build. “High-fidelity computer simulations can significantly reduce the number of trials and can provide insights into why a fuel injector behaves the way it does,” says Madhu Pai, computational combustion engineer at GRC. This image was generated on the Sierra supercomputer at Lawrence Livermore National Laboratory.
[/image]
[image src="http://files.gereports.com/wp-content/uploads/2013/08/ArtInScience2.jpg"]
Arthur calls this computer-generated image of a low-pressure turbine “chromatic ring.” His team used the Jaguar supercomputer based at Oak Ridge National Laboratory to model fluid dynamics inside the turbine. They were looking for tiny variations that could help them improve turbine efficiency.
[/image]
[image src="http://files.gereports.com/wp-content/uploads/2013/08/ArtInScience3.jpg"]
Here is a close-up of the previous image.
[/image]
[image src="http://files.gereports.com/wp-content/uploads/2013/08/ArtInScience4.jpg"]
This image does not represent a serving of Rice Krispies treats but three nickel-alloy computer models generated by a GRC server cluster. They help scientists understand the microstructure of the alloy molecules and gain insight into the properties of the metal.
[/image]
[image src="http://files.gereports.com/wp-content/uploads/2013/08/ArtInScience5.jpg"]
This picture shows the density of a jet engine exhaust flow. GE engineers are using it to increase jet engine performance and reduce noise. The image was created on the Intrepid computer network at Argonne National Lab.
[/image]
[image src="http://files.gereports.com/wp-content/uploads/2013/08/ArtInScience6.jpg"]
This picture is a two-dimensional cut-away from the previous image.
[/image]
[image src="http://files.gereports.com/wp-content/uploads/2013/08/ArtInScience.gif"]
This hypnotizing image shows an unsteady flow inside the low-pressure turbine of a jet engine. GRC scientists are using an in-house code to visualize the “unsteadiness” and get a better understanding of the aerodynamic losses inside the turbine. This helps them design more efficient engines.
[/image]
[image src="http://files.gereports.com/wp-content/uploads/2013/08/Nucleation.gif"]
This simulation shows ice spreading through a water droplet. The model shown above was developed on Titan at Lawrence Berkeley National Lab, currently the top ranked supercomputer in the world. GRC scientists are using the research to develop icephobic surfaces that prevents ice creation and build up on turbine blades, oil and gas rigs and elsewhere. Video credits: Mike Matheson, Oak Ridge National Lab.
[/image]
[image src="http://files.gereports.com/wp-content/uploads/2013/08/CheckFlow2.jpg"]
This image shows noise generation due to turbulent flow over the trailing edge of a wind-turbine blade. High fidelity computer simulations provide engineers with better insights into noise sources and noise generation mechanisms, and help them design low-noise blades. These images were generated by the Red Mesa, one of the world’s fastest supercomputers based at Sandia National Laboratory in New Mexico. Image courtesy of Prof. Sanjiva Lele, Stanford University
[/image]
[/slides]