The most recent correlations for turbine profile losses at off-design incidence include the leading-edge diameter as the only aspect of the leading-edge geometry which influences the losses. Cascade measurements are presented for two turbine blades which differ primarily in their leading-edge geometries. The incidence was varied over a range of ±20 degrees and the results show significant discrepancies between the observed profile losses and those predicted by the available correlations. Using data from the present experiments, as well as cases from the literature for which sufficient geometric data are given, a revised correlation has been developed. The new correlation is a function of both the leading-edge diameter and the wedge angle, and it is significantly more successful than the existing correlations. It is argued that the off-design loss behaviour of the blade is influenced by the magnitude of the discontinuity in curvature at the points where the leading-edge circle meets the rest of the blade profile. The wedge angle appears to be an approximate and convenient measure of the discontinuity in curvature at these blend points.

The off-design performance of axial turbines is usually predicted by calculating the incidence losses using empirical correlations. Periodic review and improvement to these prediction methods, to reflect recent turbine designs and test results, are essential for the accurate assessment of losses in turbine airfoils. The purpose of the present work is to evaluate existing turbine incidence loss correlations, and present an improved prediction method for profile and secondary losses at off-design conditions which correlates better with the available experimental results. The incidence losses are shown to be a function of leading edge diameter, pitch, aspect ratio and channel convergence.

Despite its wide use in meanline analyses, the conventional loss breakdown scheme is based on a number of assumptions that are known to be physically unsatisfactory. One of these assumptions states that the loss generated in the airfoil surface boundary layers is uniform across the span. The loss results at high positive incidence presented in a previous paper (Benner et al. [1]) indicate that this assumption causes the conventional scheme to produce erroneous values of the secondary loss component. A new empirical prediction method for secondary losses in turbines has been developed, and it is based on a new loss breakdown scheme. In the first part of this two-part paper, the new loss breakdown scheme is presented. Using data from the current authors’ off-design cascade loss measurements (Benner et al. [1]), it is shown that the secondary losses obtained with the new scheme produce a trend with incidence that is physically more reasonable. Unlike the conventional loss breakdown scheme, the new scheme requires a correlation for the spanwise penetration depth of the passage vortex separation line at the trailing edge. One such correlation exists (Sharma and Butler [2]); however, it was based on a small database. An improved correlation for penetration distance has been developed from a considerably larger database, and it is detailed in this paper.

The off-design performance of axial turbines is usually predicted by calculating the incidence losses using empirical correlations. Periodic review and improvement to these prediction methods, to reflect recent turbine designs and test results, are essential for the accurate assessment of losses in turbine airfoils. The purpose of the present work is to evaluate existing turbine incidence loss correlations, and present an improved prediction method for profile and secondary losses at off-design conditions which correlates better with the available experimental results. The incidence losses are shown to be a function of leading edge diameter, pitch, aspect ratio and channel convergence.

A new empirical prediction method for design and off-design secondary losses in turbines has been developed. The empirical prediction method is based on a new loss breakdown scheme, and as discussed in Part I, the secondary loss definition in this new scheme differs from that in the conventional one. Therefore, a new secondary loss correlation for design and off-design incidence values has been developed. It is based on a database of linear cascade measurements from the present authors’ experiments (Benner [1]) as well as cases available in the open literature. The new correlation is based on correlating parameters that are similar to those used in existing correlations. This paper also focusses on providing physical insights into the relationship between these parameters and the loss generation mechanisms in the endwall region. To demonstrate the improvements achieved with the new prediction method, the measured cascade data are compared to predictions from the most recent design and off-design secondary loss correlations (Kacker and Okapuu [2], Moustapha et al. [3] using the conventional loss breakdown. The Kacker & Okapuu correlation is based on rotating-rig and engine data, and a scaling factor is needed to make their correlation predictions apply to the linear cascade environment. This suggests that there are additional and significant losses in the engine that are not present in the linear cascade environment.