Dynamic Stall: A Catastrophic Loss of Lift

Pilots and aerodynamicists are all familiar with the effects of stall when a wing is at a fairly constant angle of attack: the air flow over the wing separates so that it no longer follows the wing's contours, and the wing can't produce as much lift as expected or desired. To recover from a stalled condition, the pilot must decrease the angle of attack by pushing the stick forward. An increase in power is also helpful to build up airspeed, which is the sine qua non of lift production.

This scenario is called "static stall," since the angle of attack changes very little over the time during which the flow separates from the wing. Typical stall angles of attack range from 10 - 15 degrees.

The flow field development is radically different if the angle of attack continues to increase while the flow separates. At first the flow remains attached, so the lift continues to increase well beyond what is experienced in the static case. Researchers have measured lift up to seven times greater than normal!

This scenario is called "dynamic stall," since the angle of attack is constantly changing. The dynamic stall angle of attack is a function of the pitch rate, as discussed below.

We decided to investigate dynamic stall experimentally using hydrogen bubble flow visualization in our water tunnel. The results gave us a very good qualitative understanding of the dynamic stall process, but they didn't provide any quantitative information beyond the angle of attack and pitch rate.

I modified a fully compressible, time-accurate Navier-Stokes solver to simulate the motions of virtual hydrogen bubbles while computing velocity, pressure, and temperature. This gave us quantitative information which has only been verified to date using surface pressure measurements, and the underlying physical cause that induces leading edge separation has remained elusive.

In the images above you can compare our experimental and computational flow visualizations for angles of attack of 12 deg, 21 deg, and 31 deg. The experiments are on the left, and the computations are on the right. Both capture the physical phenomena of dynamic stall: At the static stall angle of attack (12 deg; the top images), the flow remains attached, as though nothing unusual were happening. At 21 degrees (the middle images), a small "bubble" forms near the leading edge. This bubble grows into a large structure called the dynamic stall vortex, seen dominating the flow over the leading edge at 31 degrees (the lower images). As time progresses, the vortex is swept downstream, where it eventually winds up in the wake of the airfoil. As long as the vortex remains located above the wing, it acts to enhance the lift being produced. After the vortex is shed from the airfoi, lift decreases abruptly, and just when you needed that lift the most, the wing stops flying. (oof!)

The actual angle at which leading edge separation begins is strongly dependent on pitch rate: the faster the angle of attack is increasing, the longer the onset of leading edge separation is delayed.

Dynamic stall is most often encountered in helicopters, where it induces a powerful and abrupt rotational flexing of the rotor blades that can cause catastrophic blade failure within as little as 30 seconds.

As mentioned above, the quantative results of our computations have not been verified experimentally, because it's not been possible to make the appropriate measurements. We think we've solved that problem with our new measurement technique, MTV.

We've already published our flow visualizations and some of the computational results. More quantitative journal publications will be forthcoming next year. In the meantime, you might be interested in perusing these:

This page last updated 17 Sep 97 (cpg).