Application of probability density function analysis in the monitoring of engine aerodynamic instability precursor

Experiments were performed to investigate surge problem of a single shaft axial flow turbojet engine during throttling. If the discharge bleed valve of the engine is closed manually, the working line of the engine will traverse the stalling region in a mass flow range, thus throttling from full speed may surely transit the engine into unstable operation. It is observed that the casing wall static pressure of the first compressor stage stator is most sensitive to throttling operation, and that probability density functions of the pulsating components of the static pressure vary accordingly. An eigenvalue, θ, is then introduced to quantify the variation of probability density functions. Further data analysis demonstrates that θ is capable of reflecting some aspects of flow separation in the tip region of the first compressor stage, and that θ may serve as an indicator of imminent aerodynamic instability of the engine, thus long stall warning time of the examined turbojet engine is achievable than that to directly detect modes or spikes.