Wind Turbine Aerodynamics Part A: Basic Principles - Wind Turbine Technology: Fundamental Concepts of Wind Turbine Engineering - page 338

Figure 5-39. Experimental dimensionless power curves for the Sandia/DOE 17-m
VAWT compared with streamtube theories. Test data are the same as in Figure 5-37.
Theoretical predictions from [Strickland 1975, Parashiviou 1981, Berg 1983].
Dynamic-Stall Factors
Relatively early in the development of VAWTs, differences between performance test
behavior and theory were attributed to dynamic stall [Strickland et al . 1980], which has a
measurable effect on HAWTs but appears to be even more important to VAWT aerodynam-
ics. Simply stated, dynamic stall on a blade can produce lift and pitching-moment values that
are much larger than static values. These additional forces are brought on by a strong vortex
that forms at the leading edge and is quickly convected over the blade and into the wake. Its
formation and movement strongly depend on the following factors:
-- type of airfoil, leading edge radius, and thickness distribution;
-- initial angle of attack;
-- rate of increase of angle of attack;
-- excursion of angle of attack past the static stall angle;
The effect of this vortex on airfoil properties begins with a rapid increase in forces and
ends with full flow separation and a large drop in lift. It is clear that the dynamic stall process
depends both on the amplitude and the history of the angles of attack on the airfoil. There is
also a hysteresis effect , involving a delay in re-attachment of the flow and recovery of blade
forces to their static levels after the event. Stall time delays related to rapid increases in angle
of attack have been used to reproduce analytically the flattening of VAWT power curves at
high advance ratios [Massé 1984].