Uncertainty Quantification of Aero-Thermal Performance of a Blade Endwall Considering Slot Geometry Deviation and Mainstream Fluctuation

Turbine inlet temperatures (TIT) of modern gas turbines are constantly increased, which imposes higher thermal load on the blade endwall. Besides, the strong secondary flows near the endwall complicate the flow and thermal environment near the endwall. Thus, to ensure the durability of gas turbines, the blade endwall which operates at extreme harsh conditions should be carefully designed. However, the uncertainty factors are always encountered in manufacturing and operation process, which may cause significant performance deviation and degradation. Therefore, it is crucial to investigate the aero-thermal performance of the endwall quantitatively by taking the uncertainty effects into consideration.

The flow field and heat transfer performance of the endwall are significantly affected by the geometry of upstream slot and inflow conditions of mainstream. However, the slot geometry and mainstream conditions inevitably vary and thus deviate from their design points due to manufacturing or assembly errors and operational condition variations. In this paper, the deviation of slot geometry (slot width, endwall misalignment) and the fluctuation of mainstream (turbulence intensity, incidence angle) are taken as uncertainty parameters, and they follow normal and uniform distribution, respectively. Coupled with the three dimensional Reynolds-Averaged Navier-Stokes (RANS) simulations, a kriging-based uncertainty quantification (UQ) method is proposed and employed to investigate the effects of above-mentioned uncertainty parameters on endwall aero-thermal performance. The mean, variance and the probability density distribution of the aero-thermal performance of the endwall are calculated and analyzed. A sensitivity analysis is also conducted to detect and identify the significant parameters. Results show that under the influence of the uncertainty parameters, the practical performance of the endwall has a high probability of deviating from the nominal value, and the overall deviations of endwall thermal load can reach up to 20%. Using sensitivity analysis, the slot width is found to be the most important parameter affecting the endwall aero-thermal performance because it significantly changes the momentum of coolant ejection.