Active stall control with practicable stall prediction system using auto-correlation coefficient

Here we report that a practical prediction system of impending compressor stall was established, which can be incorporated into a real time control system. The system uses a static pressure sensor located on the compressor casing to detect stall precursor. The leading edge of the rotor blades was selected as the axial location of the sensor after analyzing the test data of the blade passage flow implying the tip vortex interaction on the leading edge at stall point. The impending stall detection algorithm has been developed, which computes the auto-correlation coefficients of the past and present static pressure profiles and declares an impending stall if the auto-correlation coefficient decreases below a pre-set threshold. Post experiment analysis proved the effectiveness of the algorithm. The auto-correlation coefficients remained nearly 1.0 at stable operation conditions but showed an abrupt decreasing only when the flow coefficient became next to stall. The algorithm consistently detect the stall inception about 100 rotor revolutions prior to the fully developed stall, which is promising and encouraging for practical active stall control. The effectiveness of the developed auto-correlation algorithm was proved by the successful closed loop demonstration with the dynamically inserted plate in front of the compressor.