Getting a Better Picture of Gas Turbine Combustion

A simulation of combustion within two adjacent gas turbine combustors. GE researchers are incorporating advanced combustion modeling and simulation into product testing after developing a breakthrough methodology on the OLCF's Titan supercomputer. (Image courtesy of GE Global Research.)
As a promising low-emission energy source, natural gas turbines should be developed to maximize their efficiency and reliability. Modeling and simulation of these machines is a very important step in their enhancement.

Until now, the shortcomings of available programs to do so greatly hindered the status of natural gas turbines. However, thanks to the capabilities of a supercomputer and collaborative efforts amongst a highly talented team of engineers, a modeling program fit to handle to the complexities of natural gas turbines has become a reality.

This is in the form of a simulation of a General Electric (GE) gas turbine’s multicombustor dynamic instability.


Fluid Dynamics and Supercomputing

The modeling program is based on Cascade Technologies’ CHARLES, a powerful fluid dynamics tool. CHARLES is a flow solver for large eddy simulation based on a mathematical model utilizing Navier-Stokes flow equations. As a result, CHARLES can take on the complexities of air and fuel mixing during combustion and produce petabytes of simulation data using a supercomputer. 

Titan, OLCF's Crazy XK7 system. (Image courtesy of OLCF.)
That’s why CHARLES was paired up with Titan, Oak Ridge Leadership Computing Facility’s (OLCF) Cray XK7 system. Titan is capable of 27 petaflops, equivalent to 27 quadrillion calculations per second. Using this power couple, and 11.2 million hours on Titan, simulation runs harnessing 8,000 and 16,000 cores at a time were able to speed up code performance to 30 times greater than the original code.

 

Simulating Gas Turbine Combustion

With these advancements in-tow along with a nearly-one billion cell fine-mesh grid, the research team captured snapshots of the air-fuel mixture during turbulent combustion such as particle diffusion, chemical reactions, heat transfer, and energy exchange. Each cell of the grid captured each snapshot at the microsecond scale. The simulation data was then analyzed in high-definition through work with a visualization specialist.

The end result was the most refined model of the GE gas turbine ever. The instability of the air-fuel mix, and its causes, could be adjusted in terms of speed and zoom level. Furthermore, this development enabled the observation of combustion physics at the sub-millisecond level. The results of the simulation to the GE 2015 gas turbine were confirmed upon full-scale testing.


One Percent Makes All the Difference

This development does more than just enhance our understanding of the inner-workings of gas turbines through accurate modeling and simulation. The program may also enable a full percentage gain in efficiency. You might think that a one percent increase may not seem worth the efforts of a supercomputer or microsecond snapshots.

Think again.

A single percent increase in efficiency for a gas turbine could save over $11 billion in fuel for GE’s customers over the next two decades. Furthermore, this efficiency increase would also prevent millions of tons of carbon dioxide from being emitted.

For some perspective, a one percent efficiency increase in a one gigawatt power plant results in a decrease of 17,000 metric tons of carbon dioxide emissions per year. This is equivalent to 3,500 vehicles being taken off the road. If a one percent efficiency gain were to be made across all power plants, around 200 gigawatts, approximately 3.5 million metric tons of carbon dioxide would be saved each year.

The collaborative engineering effort amongst GE, OLCF, and Cascade Technologies is the first of its kind, but it’s also only the beginning. As the simulation results are iteratively applied to the development of gas turbines, we may soon begin to climb the efficiency ladder, one percent at a time.

For further reading on the natural gas industry, check out this article on bioconversion.