We all know that airplane takeoffs can be very noisy. This is because the plane’s engine output and aerodynamic drag are near their maximum.
Optimizing the aerodynamics during takeoff is key to achieving the financial and environmental goals of the aerospace industry.
You might not know it, but for a typical airline, jet fuel is the second biggest expense — next to labor. Since fuel prices are unpredictable we can expect this expense will rise without a cost-cutting intervention.
As previously mentioned, during takeoff engine output and aerodynamic drag are near their maximum. As a result, jet fuel is consumed faster during takeoff than any other time during the flight.
In fact, during short-range flights, a plane can use up 25 percent of the total fuel budgeted for the trip. In Europe, for example, 45 percent of all flights are considered short-range — flying a distance under 500 km (311 miles).
Therefore, reducing drag during takeoff is one of the most effective ways we can minimize an airline’s fuel consumption. To do this, we need computational fluid dynamics (CFD) and high-performance computing (HPC) tools that can help design aircraft to consume less fuel during takeoff.
Optimizing Airplane Takeoff Requires Complex Simulations
In the first 10 or so minutes of a flight, there are a lot of changes to the aircraft’s altitude, configuration and aerodynamic behavior.
For example, the ground effects are initially large but fade as the aircraft climbs.
Additionally, the aircraft leaves the runway with slats, flaps and landing gear extended. However, the gear is retracted when the plane has a positive climbing rate — significantly reducing drag and noise.
If we are to capture all of this complexity in our simulations, we will need to model many CFD phenomena, including:
- Boundary layer transitions.
- Flow separation.
- Flow reattachment.
- Wake-boundary layer interactions.
This complexity means the CFD simulation tool we use needs powerful HPC resources and parallel processing capabilities to offer accurate results in a timely fashion.
How to Simulate Airplane Takeoff in a Few Hours
Properly simulating airplane takeoff aerodynamics takes a mesh with tens of millions, or even hundreds of millions, of cells.
Clearly, we have our work cut out for us, as simulating airplane takeoff with a mesh that large involves a lot of computing power.
We can use ANSYS Fluent on a Cray XC system to simulate airplane takeoff in about two hours. In this case, we should use about 189 million cells on about 2,000 CPU cores to get the results we need in a few hours.
ANSYS recommends these settings because they were tested in an assessment developed by the American Institute of Aeronautics and Astronautics (AIAA) High Lift Prediction Workshop (HiLiftPW). The assessment consisted of three test cases, two high-lift configurations and four mesh refinements.
The results were then validated using experimental data gathered from the Japan Aerospace Exploration Agency’s (JAXA) low-speed wind tunnel. The aerodynamic forces, moments, pressure coefficients and surface oil flow textures in the simulation matched those seen in the wind tunnel data.
For more details on the tests that were performed, read Cray and ANSYS Give Aircraft Takeoffs a Big Lift. To learn more about the deployment of HPC clusters on-premise and in the cloud, read HPC Clusters Made Easy for Engineering Simulation.