Spatial Correlation Based Stall Inception Analysis

Abstract

Investigations of stall inception and compressor pre-stall behavior have used a variety of techniques to make inferences about the mechanisms of rotating stall inception. Many of these techniques utilized data from arrays of circumferentially spaced hot-wires or high frequency response pressure transducers. This paper presents results from the application of several typical analysis techniques to the interpretation of unsteady casing pressure measurements recorded during two representative stall event in a high-speed axial compressor stage. Results from visual pressure trace inspection, spatial Fourier decomposition, wavelet filtering, and traveling wave energy techniques are presented and compared. The effects of measurement and analysis parameters are also briefly discussed. A new analysis technique based on windowed two-point spatial correlation between adjacent stall inception sensors is described. The method was found to provide both spatial and temporal information about rotating features in the compressor flow and is insensitive to low pass filtering and parameter selection over a wide range of values. It was also found to be valuable for analysis of both pre-stall and stall inception behavior.