Improving Airfoil Lift Prediction

Abstract

An improved formulation for lift estimation for integral boundary layer (IBL) methods (i.e. RFOIL, XFOIL) for thick airfoils is presented. Lift over-prediction (5 − 10% around (l/d)max) in RFOIL (similarly in XFOIL) is observed for thick airfoils for a wide range of angles of attack. The lift slope is over-predicted resulting in the increasing error in lift with increasing angle of attack. The wake geometry in RFOIL and XFOIL is determined from the inviscid calculations which seems to give rise to the above problem. A scheme has been developed for IBL methods and implemented in RFOIL, to include the effects of viscous flow on the wake geometry which lead to improved lift prediction for thick airfoils. New insights were obtained regarding the discrepancy in predicting the maximum lift. It is observed that lift over-prediction persists in case of thick trailing edge (TTE) airfoils even with the improved method. The cause for this behavior is identified and discussed in order to evaluate the possibilities of improvement. High Reynolds number and increased free-stream turbulence intensity have been observed to inhibit the onset of flow separation and delay stall. These effects are not yet captured accurately in RFOIL indicating a need for further investigation. Nomenclature α Angle of attack c Airfoil chord length c d Sectional drag coefficient c l Sectional lift coefficient h T E Airfoil trailing edge thickness M ∞ Free stream Mach number N crit Critical amplification factor Re Chord length based Reynolds number X, Y Cartesian space coordinates

Airfoil optimisation for vertical‐axis wind turbines with variable pitch

“…In general, the lift gradient is overestimated while the drag is underestimated. The drag effect of thick trailing edges is underestimated as well . XFOIL still misses some accuracy in the stall prediction .…”

“…The drag effect of thick trailing edges is underestimated as well. 42,43 XFOIL still misses some accuracy in the stall prediction. 44 Although small variations on these parameters exist, this has only a small impact on the VAWT simulations.…”

Blade element momentum theory extended to model low Reynolds number propeller performance

“…Similar recommendations of caution were made by Ol et al (13) regarding the interpretation of both experimental and analytical aerodynamic data below Reynolds numbers of 60,000. Additionally, XFOIL is seen to struggle to accurately replicate aerofoil performance in the post-stall regime, and also suffers from accuracy degradation when simulating thick aerofoil performance (41)(42)(43) . These factors highlight the need for high quality low Reynolds number data across a suitable range of angles to validate the aerodynamic tools used to generate aerodynamic polars for small propeller analysis methods.…”

Airfoil Optimization using Design-by-Morphing

“…On the other hand, the viscous/inviscid zonal approach, which combines separate solutions for the inviscid external flow and the viscous shear layer flow iteratively to form a continuous profile, is faster and less expensive. Among the many inviscid/viscous airfoil analysis codes, the XFOIL program [44] has been the most dominant and widely adopted one [45][46][47][48][49][50][51][52]. It couples a vorticity panel method for exterior flow with an integral boundary-layer method for viscous boundary layers and uses an 𝑒 9 -type amplification formulation to determine the transition point [44].…”