Modeling and detection of rotating stall in axial flow compressors, II: Experimental study for a low-speed compressor in Beihang University

Early detection of rotating stall and surge in axial flow compressors is of great importance for improving the working efficiency and stability of the compressor. Based on deterministic learning (DL) theory and dynamical pattern recognition, this paper presents experimental research for approximately accurate modeling and rapid detection of stall precursors, and then employs a low-speed axial flow compressor test rig of Beihang University for online experimental verification. Firstly, by installing high response dynamic pressure transducers arranged circumferentially around the casing of the axial compressor, the dynamic pressure data are collected. Based on deterministic learning theory, the system dynamics underlying prestall and stall inception patterns are identified. Secondly, based on modeling results, rapid detection of small oscillation faults is used to perform the detection of stall precursors. Sufficient online experiments are conducted to investigate the efficiency of the approach. Results show that, in different working speeds, this approach successfully detects inception signal of aerodynamic instability of the compressor 0:3 s~1 s in advance to the start of rotating stalls.