Effects of Incidence on Three-Dimensional Flows in a Linear Turbine Cascade

Abstract

The paper describes effects of the incidence on cascade three-dimensional flows and on the associated loss mechanisms occurring in a low-speed linear turbine rotor cascade. For each of five nominal incidences ranging from 7.2 to −53.3 deg, the cascade flow was surveyed at 15 or 16 planes located axially throughout the cascade. Blade-to-blade flows at the cascade midspan and near the endwalls, meridional flows along the blade surfaces, and static and total pressures within the cascade were analyzed in detail. The results were represented by two- or three-dimensional tomograms, on the surfaces of which the cascade flows were drawn by vectors, scalar contours, and streaklines, and from which one can easily understand the extraordinarily complicated cascade flows and the loss generation mechanisms. The present study will not only give new insight into the incidence effects, but will also contribute many solid experimental facts of a quantitative nature to our current knowledge of turbine cascade flows.

Numerical simulation of turbulent cascade flows involving high turning angles

“…They simulated the cascade¯ow with and without tip clearance treating accurately the geometry of the blade tip. Their computed results showed good agreements with a series of experimental results of Yamamoto (1988), Yamamoto and Nouse (1988), and Yamamoto (1989), which were obtained by utilizing a low-speed linear turbine rotor cascade with a high turning angle, thick leading and trailing edges, low aspect ratio and low solidity.…”

“…For comparison with the experimental data of Langston et al (1977), who did not consider any tip clearance effect, 100 Â 35 Â 19 grids are used. For comparison with the experimental data of Yamamoto and Nouse (1988), in which the tip clearance effect was not considered, 90 Â 43 Â 40 grids are used, while for comparison with Yamamoto (1988Yamamoto ( , 1989 in which the tip clearance effect was considered, 90 Â 43 Â 47 grids are used. In the latter case, the calculations using the second order upwind scheme requires 48 seconds of CPU time per iteration with vectorized program on the Cray-YMP Supercomputer and the typical iteration number for convergence is 1500±2000.…”

“…To prescribe the boundary conditions accurately along the complex geometry of the turbine rotor cascade and to facilitate the arrangement of the computational grids in accordance with the¯ow phenomena, a boundary-®tted coordinate system has been adopted, where all the boundary surfaces of the computational region coincide with coordinate surfaces. At the inlet boundary, the experimental data measured by Langston et al (1977) are given, while the pitch-averaged velocity pro®les and the yaw¯ow angles measured by Yamamoto and Nouse (1988) and Yamamoto (1988) are given with the free stream turbulence intensity of 0.5% being assumed, as in Yamamoto (1987). Inside the inlet boundary layer, empirical values and formula obtained from the¯at plate boundary layer experiment (Cebeci and Smith 1974) are used.…”

“…Therefore, the validity of the present method has been con®rmed as far as the blade passage¯ows are concerned. Next, experimental results obtained by Yamamoto and Nouse (1988) for the case of linear turbine rotor cascade without tip clearance are used to further verify the validity of the present calculation method. The setting angle of the inlet guide plates was 7.2 degrees, which was close to the design point.…”

“…In addition, we apply the second order upwind scheme to treat the convective terms. Computed results will then be compared with the experimental results of Langston et al (1977), and those of Yamamoto and Nouse (1988), and Yamamoto (1988and Yamamoto ( , 1989.…”