Experimental Investigation of Stall Cells on NACA0015 Airfoils

A particular type of 3-D separation, known as a stall cell, was investigated experimentally on two NACA0015 airfoils with aspect ratios of AR = 4 and 2.67. A parametric map of the angles of attack and Reynolds number conditions under which stall cells form was created using oil flow visualization. It was observed that stalls cells form naturally under specific conditions when the Reynolds number exceeds a critical Reynolds number, Re c ≥ Recrit. Based on the work of Weihs & Katz, the formation of a stall cell requires sufficient 3-dimensionality in the flow field. Next, full and partial span trips (composed of either zig-zag tape or an artificial step) were added to the airfoil and it was found that the introduction of additional 3-dimensional disturbances reduced the value of Recrit. For full-span step trips, where no additional 3-dimensionalities were introduced to the flow field, a stall cell was not formed at conditions where one was otherwise not present. However, a partial step trip did cause the formation of a stall cell (under specific conditions) through the introduction of three dimensionalities associated with the trip's ends. These results confirm that three dimensionalities need to be present in order for a stall cell to form. Flow field data were used to explore stall cell characteristics with and without external trips. Under conditions where a stall cell was present, two recirculation regions (i.e., stall cell foci) were observed, outboard of which flow abruptly reattached due to entrainment by the foci. Within the stall cell, flow was funneled away from the middle of the stall cell and into the associated focus point. In addition, at mid-span, the separated flow rotated about the spanwise direction. Outboard, the structure also began to rotate about the chord-normal direction; near the foci, all rotation occurred about the chord-normal direction. The fluctuating flow field was also considered, and elevated levels of chordwise (u'u'/Uinfinity 2) and spanwise (w¯'w¯'/Uinfinity 2) components of the normal stress were observed when stall cells were present, concentrated near the foci. Finally, a partial-span dynamic oscillating step trip was incorporated into the NACA0015 model with AR = 2.67. Initially, the actuator was driven by a square wave and the transitory behavior of flow field was explored as the trip moved from the extended to the flush position. It was shown that during this motion the flow was temporarily attached before settling into a state where a small cell was present. The intermediate reattachment was due to the natural oscillations of the actuator at its resonant frequency (ƒres = 100 Hz). This result suggested that actuating the trip at a frequency that is associated with the separated shear layer, which also coincided with the resonance frequency of the actuator, might enable mitigation of the stall cell. Therefore, the trip was driven using a sine wave with ƒ = 100 Hz (corresponding to a dimensionless frequency St = 0.35) when the airfoil was set at alpha = 13.4° and U infinity = 55 m/s, and it caused nearly complete reattachment of a 3-D separated region. At alpha = 16°, the size of the stall cell was very large and extended throughout most of the span when the trip was in the flush position; thus, the dynamic motion of the trip only affected the separated flow directly downstream of the actuator, which was reduced in size and magnitude. Phase-averaged data were also acquired, and it was shown that, during the periodic motion of the trip, coherent vortices were formed and advected downstream as they grew in size. This resulted, in a time average sense, in tilting of the flow towards the surface. However, the reattachment was unsteady.