Aeroengine Stall Warning by Multicorrelation Analysis
Dakun Sun, Ruize Xu, Xu Dong, Jia Li, Xiaofeng Sun
Abstract
The present work aims to develop a stall warning method that is reliable even when the compressor encounters inlet distortion. For this purpose, multicorrelation analysis (MCA) is proposed to provide an effective indicator for the stall warning method. The relationship between MCA and correlation measures was discussed. It was demonstrated that MCA can provide an overall evaluation of both temporal and spatial correlations of multiple signals collected by different sensors. An experimental study of MCA was carried out on a low-speed axial compressor. Sensors were mounted on the casing to sample the dynamic pressure signals used for stall warning. The correlation of the signals was found to be influenced by the steady blade loading and the disturbance energy. Autocorrelation and cross-correlation were adequate when the inlet flow was uniform, but the accuracy of these methods could be affected by distortion. In contrast, MCA was robust under uniform and distorted inlet conditions.