Bjorn’s Corner: Aircraft drag reduction, Part 10

By Bjorn Fehrm

December 22, 2017, ©. Leeham Co: In the last Corner, we described how the Wright Brothers flew a manned aircraft for the first time, propelled by its own power.

Now we will disscuss what was known about what stopped so many projects from achieving the flight distances the Wright’s could do, the aircraft’s drag.

Figure 1. The Wright Flyer flies at Kitty Hawk, NC, at 17th December 1903. Source: Wright-Brothers.org.

Aircraft drag

When the Wrights and others experimented with gliders working in kite mode (tethered, Figure 2), they could measure the lift pulling in the tether lines. They also felt what the Wrights called “Drift.”

Figure 2. The Wright’s flying their glider in kite mode. Source: Wright-Brothers.org

Drift was the horizontal part of the force in the lines, with lift the vertical. It was measured in pounds with coil spring-based grocery scales.

The brothers knew they must have lift high enough to counter the weight of the Flyer, including its pilot. By measuring the “Drift” and lift components from their experiments, they could work out the Lift-to-Drag ratio, L/D.

But they had vague understating of the cause of the Drift. It was clear it was related to lift. The more angle of attack they had on their gliders, the stronger the lift and drag forces.

They also knew that putting out a blunt object against the wind on a train or in the wind tunnel cause a drag force. Different forms created different drag forces.

Drag component knowledge

It was intellectually understood during the 19th century that air passing over a surface would cause friction against the surface. But the speeds were low and so friction drag was low. It was thought it was a negligible drag component at any speed.

Today, we know air friction drag is the dominant drag component, once one flies away from the back of the drag curve (the left slope in Figure 3).

Figure 3. The drag curve as a function of speed. Source: LNC.

We will cover the two dominant drag components (induced drag and friction drag) of this curve in following Corners. Now, we will dwell on a drag that was high at the time and was subsequently mastered to a non-dominant component. It was Form drag.

Form drag

The theoretical understanding of fluid dynamics was rather advanced at the time of the Wright Flyer. Euler had set up the Euler equations for inviscid (non-sticky or non-frictional) fluid flow in 1757. But no one could use the equations as the math was too complicated.

By 1822, Navier had further developed those equations to include friction, today called the Navier-Stokes equations. This knowledge was of no use to anyone, as the math was even more complicated (three-dimensional differential equations).

The equations couldn’t be solved until computers could hack the problem into small pieces and step through it in an iterative way. This is what Computational Fluid Dynamic (CFD) programs do. They divide the total flow problem into either a piecemeal Euler problem (simpler, as no friction is involved) or Navier-Stokes problem (more complex and therefore very compute intensive). At the time, these equations were theoretical work with no practical use.

There was some theoretical work which had some use. D’Alembert had in 1744 theoretically showed that if there was no flow separation around an object moving in non-friction (nonsticky) fluid, there could be no Form drag, Figure 2.

Figure 2. Flow around a symmetrical body, without and with flow separation. Source: Google images.

As the flow curves perfectly around an object, the air’s pressure will have a force equal on all sides of the object, therefore no drag force.

As people could feel the force on the hand when putting it out the window on the train, it was called D’Alambert’s paradox.

In practice, there are flow separations as soon as there are air friction against the surface and higher speeds. This causes separation and the resulting wake causes energy loss and by it Form drag.

The Wright Flyer had quite a bit of Form drag. Its many wires and braces with non-aerodynamic forms (a drop form would be best) caused separations and wakes and therefore drag.

As aircraft designers sought higher speeds, they could see from practical experiments and wind tunnel tests a cleaner aircraft without wires and braces would have lower drag. It could fly faster with the same propulsive power. This was the beginning of streamlining of aircraft.

Streamlining took a step forward with the air races in the early 1900s. The most famous was the Schneider Trophy for flying boats. From Wikipedia:

“Announced in 1912 by Jacques Schneider, a French financier, balloonist and aircraft enthusiast, the competition offered a prize of approximately £1,000. The race was held twelve times between 1913 and 1931. It was intended to encourage technical advances in civil aviation but became a contest for pure speed with laps over a (usually) triangular course (initially 280 km, later 350 km). The contests were staged as time trials, with aircraft setting off individually at pre-agreed times, usually 15 minutes apart. The contests were very popular and some attracted crowds of over 200,000 spectators.”

Reginald Mitchell, the legendary Spitfire designer, learned his streamlining trade through designing aircraft for the Schneider Trophy for the Supermarine company, Figure 5.

Figure 5. The Supermarine S6B of 1931, the ultimate flying boat racer. Source: Google images.

Through these races, the value of streamlined shapes and lower thickness wings and stabilizing surfaces was learned. The racers had a minimum of separated flow; they were streamlined. Their form drag was low and has stayed low on aircraft since.

As the speed was over 300mph, friction drag had increased. The mechanisms around friction drag had been researched and explained by the 1930s. It’s the subject for the next Corner.

2 Comments on “Bjorn’s Corner: Aircraft drag reduction, Part 10”

1. The leap from The Wright Flyer to the Schneider Cup machines is a large one.

2. One normally introduces artificial viscosity into the Euler equations when doing CFD calculations and the mesh size also gives some artificical effects. Even numerically solving the N-S equations in a computational mesh for engineering size problems introduces artificicial viscosity thru the selected turbulence models and mesh. So doing engine/aircraft CFD analysis and getting accurate results repeatedly is still not straight forwad but requires skilled hands often by trained PhD’s like Prof. Jameson.