Stall and surge represent two prevalent forms of aerodynamic instability in engine. Stall is characterized by the circumferential propagation of rotating stall cells， while surge manifests as large-amplitude axial oscillations in pressure and mass flow. Failure to promptly mitigate such instabilities once initiated can lead to a rapid degradation in engine performance and， in severe cases， catastrophic mechanical damage. Current stall/surge detection or prevention systems deployed on component test facilities typically rely on feature vectors extracted from time- and frequency-domain signal analyses for instability detection. However， these conventional approaches often exhibit limited responsiveness and insufficient sensitivity to incipient stall events. In light of recent advances in active engine control methodologies， there is a growing need for high-fidelity， real-time monitoring of the initial perturbation waves associated with aerodynamic instability. This demand necessitates stall detection algorithms with enhanced speed and accuracy in identifying the onset of instability. To address this challenge， this paper presents a stall identification algorithm based on cross-correlation of power spectral densities. By computing the correlation coefficient between signals acquired from two spatially separated sensor channels， the proposed method enables timely detection of stall or surge inception and triggers an early warning. The algorithm effectively advances the lead time for stall prediction， thereby establishing a critical technical foundation for the implementation of active instability control strategies in engine.