Much of the aerodynamic design of wind turbines isaccomplished using computational tools such as XFOIL. These codes are not robust enough for predicting performance under the low Reynolds numbers found with small-scale wind turbines. Wind tunnels can experimentally test wind turbine airfoils to determine lift and drag data over typical operating Reynolds numbers. They can also test complete small wind turbine systems to determine overall performance. For small-scale wind turbines, quality experimental airfoil data at the appropriate Reynolds numbers are necessary for accurate design and prediction of power production. A c c e p t e d M a n u s c r i p t N o t C o p y e d i t e dDownloaded From: http://energyresources.asmedigitalcollection.asme.org/ on 05/23/2015 Terms of Use: http://asme.org/terms

A transient method of measuring the local heat transfer under an array of impinging jets has been developed. The use of a temperature sensitive coating consisting of three encapsulated thermochromic liquid crystal materials has allowed the calculation of both the local adiabatic wall temperature and the local heat transfer coefficient over the complete surface of the target plate. The influence of the temperature of the plate through which the impingement gas flows on the target plate heat transfer has been quantified. Results are presented for a single inline array configuration over a range of jet Reynolds numbers.

Wind tunnel testing of wind turbines can provide valuable insights into wind turbine performance and provides a simple process to test and improve existing designs. However, the scale of most wind turbines is significantly larger than most existing wind tunnels, thus, the scaling required for testing in a typical wind tunnel presents multiple challenges. When wind turbines are scaled, often only geometric similarity and tip speed ratio matching are employed. Scaling in this manner can result in impractical rotational velocities. For wind tunnel tests that involve Reynolds numbers less than approximately 500,000, Reynolds number matching is necessary. When including Reynolds number matching in the scaling process, keeping rotational velocities realistic becomes even more challenging and preventing impractical freestream velocities becomes difficult. Turbine models of 0.5, 0.4, and 0.3 m diameter, resulting in wind tunnel blockages up to 52.8%, were tested in order to demonstrate scaling using Reynolds number matching and to validate blockage corrections found in the literature. Reynolds numbers over the blades ranged from 20,000 to 150,000 and the tip speed ratio ranged from 3 to 4 at the maximum power point for each wind speed tested.

With the new generation of gas turbine engines, low Reynolds number flows have become increasingly important. Designers must properly account for transition from laminar to turbulent flow and separation of the flow from the suction surface, which is strongly dependent upon transition. Of interest to industry are Reynolds numbers based upon suction surface length and flow exit velocity below 150,000 and as low as 25,000. In this paper, the extreme low end of this Reynolds number range is documented by way of pressure distributions, loss coefficients, and identification of separation zones. Reynolds numbers of 25,000 and 50,000 and with 1 percent and 8-9 percent turbulence intensity of the approach flow (free-stream turbulence intensity, FSTI) were investigated. At 25,000 Reynolds number and low FSTI, the suction surface displayed a strong and steady separation region. Raising the turbulence intensity resulted in a very unsteady separation region of nearly the same size on the suction surface. Vortex generators were added to the suction surface, but they appeared to do very little at this Reynolds number. At the higher Reynolds number of 50,000, the low-FSTI case was strongly separated on the downstream portion of the suction surface. The separation zone was eliminated when the turbulence level was increased to 8-9 percent. Vortex generators were added to the suction surface of the low-FSTI case. In this instance, the vortices were able to provide the mixing needed to re-establish flow attachment. This paper shows that massive separation at very low Reynolds numbers (25,000) is persistent, in spite of elevated FSTI and added vortices. However, at a higher Reynolds number, there is opportunity for flow reattachment either with elevated free-stream turbulence or with added vortices. This may be the first documentation of flow behavior at such low Reynolds numbers. Although it is undesirable to operate under these conditions, it is important to know what to expect and how performance may be improved if such conditions are unavoidable.

This paper describes an experimental investigation of the use of Gurney flaps to control laminar separation on turbine blades in a linear cascade. Measurements were made at Reynolds numbers (based upon inlet velocity and axial chord) of 28×103,65×103 and 167×103. The freestream turbulence intensity for all three cases was 0.8%. Laminar separation was present on the suction surface of the Langston blade shape for the two lower Reynolds numbers. In an effort to control the laminar separation, Gurney flaps were added to the pressure surface close to the trailing edge. The measurements indicate that the flaps turn and accelerate the flow in the blade passage toward the suction surface of the neighboring blade thereby eliminating the separation bubble. Five different sizes of Gurney flaps, ranging from 0.6 to 2.7% of axial chord, were tested. The laser thermal tuft technique was used to determine the influence of the Gurney flaps on the location and size of the separation bubble. Additionally, measurements of wall static pressure, profile loss, and blade-exit flow angle were made. The blade pressure distribution indicates that the lift generated by the blade is increased. As was expected, the Gurney flap also produced a larger wake. In practice, Gurney flaps might possibly be implemented in a semi-passive manner. They could be deployed for low Reynolds number operation and then retracted at high Reynolds numbers when separation is not present. This work is important because it describes a successful means for eliminating the separation bubble while characterizing both the potential performance improvement and the penalties associated with this semi-passive flow control technique.