Abstract
In Part I, a companion paper of the two-part article, two turbine stages working with the same Reynolds number (8.8 × 105), flow coefficient, loading coefficient and reaction, but the different exit Mach number (Mexit = 0.7 and Mexit = 1.1, respectively) are designed and optimized to provide a direct contrast between the high subsonic and the transonic turbine tip. In Part II, further analysis are carried out primarily aimed at addressing the issues of interest arising from Part I:
a) to understand the driving flow physics for the observed contrasting aerodynamic efficiency sensitivities of the two stages with respect to squealer height.
b) to seek a more suitable heat transfer objective function for the tip aero-thermal design optimization, given the seemingly strong conflicts among the heat transfer objective functions.
The advanced post-processing techniques are adopted to identify and quantify vortical structures in the rotor tip region. Two vortices, the Pressure Side Vortex (PSV) and the Casing-driven Cavity Vortex (CCV), are found to have major influence on the flow pattern within the squealer tip.
On the aerodynamic side, the transonic squealer is shown to be somewhat less effective than the subsonic one (Part I). The OTL mass flow rate in the transonic stage is less sensitive to the squealer height. The corresponding reduced blockage for the transonic squealer may be attributed to two factors: 1) the vena-contracta formed by the PSV as it leaves the squealer cavity induces smaller size of separation in the supersonic flow. Thus, the blockage effect and the mass flux deficit are reduced in the transonic stage; and 2) the choking of the OTL at the rear portion of the transonic tip acts as a limiter for the leakage mass flow. On the heat transfer side, the two vortices also have major signatures on the cavity surfaces, mainly by impingement. For the PSV, the impingement impact is mainly on the cavity floor. For the CCV on the other hand, its impact is mainly on the inner side wall of the suction side rim. The overall linear variations of the heat load with the squealer height primarily correspond to the linear increase of the cavity side wall surface. The vortical structure is found to be influential: the higher slope of increments in the transonic stage is due to the stronger and more persistent CCV with a clear signature on the large portion of the suction side cavity wall at a large squealer height.
In addressing issue b), the coolability weighted non-uniformity parameter is proposed to integrate the local heat transfer and the coolability. The proposed coolability augmented objective function is shown to provide more consistent Pareto fronts for heat transfer non-uniformity, than those adopted in Part I. It thus should serve well as an enabler to help practical applications of blade tip aerothermal design optimizations.
