The influence of leading-edge geometry on profile and secondary losses in turbine cascades

This thesis presents detailed experimental results of the midspan and secondary flows from two large-scale, low-speed linear turbine cascades. The airfoils for the two cascades differ mainly in their leading-edge geometries. Detailed flow field measurements were made upstream and downstream of the cascades using three- and seven-hole pressure probes and static pressure distributions were measured on the airfoil surfaces. The data were supplemented with extensive surface oil flow visualization. Measurements were made at three different values of incidence: 0, +10 and +20 degrees.; For the off-design profile losses, significant discrepancies were observed between the present measurements and the predictions from the most recent available correlation (Moustapha et al., 1990). Based on data from the present experiment and cases from the open literature, a revised off-design profile loss correlation has been developed that is significantly more successful than the existing correlations. The new correlation includes leading-edge wedge angle as an influential and additional correlating parameter.; The results from the secondary flow measurements at the design incidence suggest that the leading-geometry has only a minor influence on the secondary losses, whereas, the loading on the forward part of the airfoil appears to be of primary importance. At off-design incidence, the most recent off-design secondary loss correlation (Moustapha et al., 1990) includes leading-edge diameter as an influential correlating parameter. The correlation predicts that the secondary losses for the airfoil with the larger leading-edge diameter are lower at off-design incidence; however, the opposite is observed experimentally. The loss results at high positive incidence have also highlighted some serious shortcomings with the conventional method of loss decomposition. An empirical prediction method for secondary losses has been developed to replace the conventional one.; In addition to the experimental study, a computational investigation was conducted using a CFD code that is currently widely used in turbomachinery design. The objective was to determine the degree to which the code is capable of predicting off design profile losses. For cases at moderate values of incidence, the correct loss trend was predicted; however, the code did not predict the trailing-edge separation that occurred at high incidence.