Tip clearance flow in transonic rotors is known to have a significant effect on compressor performance and stability. The purpose of this study is to employ a novel statistical analysis method as a stall precursor detector and investigate the flow physics underlying stall inception. To allow for natural stall evolution, a full annulus simulation of a transonic axial compressor stage (NASA Stage 35) was utilized. Due to the size of the data set, a statistical analysis was employed to rapidly analyze the entire spatial and temporal simulation domain. The statistical analysis method was performed pointwise on the domain and utilized Entropy in a Grubb's test to reveal statistically anomalous regions, times, and trends which would be of interest for future evaluation with more standard visualization techniques. Through use of the statistical anomaly detecting Grubbs' test, rotating stall, which developed in the stage, was tracked back in time to immediately after mass flow stabilization of the stalled operating condition. The Grubbs' analysis showed a rotating stall cell formed from a rotating disturbance region with a once per revolution mode. Oscillations were also observed in a spiral-type vortex breakdown of the tip clearance vortex. The interaction between the disturbance region and the spiral-type vortex breakdown caused the amplitude of oscillation in the tip vortex to increase in the radial and circumferential directions. The behavior caused local decreases in the mass flow and high loading to the blades in the disturbance region. Eventually, the oscillations led to leading edge spillage and the disturbance region broke down into rotating stall. By tracking anomalies of Entropy in time and space the statistical Grubbs' test showed the efficacy of the method at detecting the earliest signs of rotating stall.
展开▼
