Experimental Methods Experimental Facilities

The current study was conducted on a flat plate test rig with a moving bar system, as shown in Figure 1. The test section was fitted to the exit of an open – return wind tunnel. The test rig consists of an aluminum plate that is 738 mm long by 458 mm wide and 12.8 mm thick. The fht plate has a 6:1 elliptical leading edge to avoid the separation of ft>w at the leading edge. In the present study, only the front 500 mm of the plate was used to achieve the test Reynolds number range and a reasonable measurement resolution.

A pair of contoured walls was used to impose the pressure distribution on the flat plate. The shape of the contoured walls was designed by ITP, a partner in the project, to match the same pressure distribution as that on the suction side of T106C LP turbine blade. Bleed slots and two overlapping perforated sheets at the outlet were used to avoid the flow separation from the contoured walls due to the large diffusion after the throat. A bi-planar cylindrical rod type turbulence grid can be placed upstream to increase the free stream turbulence intensity up to 3.5% to create a more realistic test environment.

Figure 2. Wavy type roughness element geometry

To simulate the rotor-stator interaction, moving bars were used to generate the incoming wakes. Carbon fibre bars of 7.8 mm diameter were attached to a pair of reinforced nylon belts. The axial distance between the bar moving plane and the plate leading edge is 465 mm. As the bars pass across the inlet of the test section, they shed wakes, which convect over the fkt plate and sim­ulate the wake passing conditions in a turbomachine. In the fkt plate test, the ft»w coefficient (ф) was defined as the ratio of the exit velocity (V2is) to the bar speed (Ub), i. e. ф = V2is/Ub (Stieger, 2000) and ф = 0.83 was chosen to be representative of a repeating stage of the T106C profile. The reduced fre­quency, defined as fr = fbar • C/V2is, was set at 0.84. The Reynolds numbers based on the plate surface length (So) and exit velocity (V2) covered a range between 134000 and 285000.

The roughness elements used in the current study include straight wires, straight steps, wavy steps and wavy wires. The roughness elements were mounted on the plate surface using double-sided tape, which has a thickness of 0.1 mm. Table 1 summaries the parameters of the roughness elements. The relative roughness height k/S* is calculated from the displacement thickness at 50%So with Re=174000 and Tu = 0.5% on smooth surface in steady ft»w condition. Rek is based on the height of the roughness element and the local freestream fbw velocity at Re=174000. The wavy roughness elements were proposed by MTU, a partner in the present program, having previously been described in a lapsed patent. The geometry is shown in Figure 2. The thickness or diameter of the element was changed but the other dimensions were kept the same as those shown in the figures.