Study on the Fault Diagnosis Method of Compressor Instability by Fusing Multiple Time Domain Feature Parameters

To fulfil the active safety control requirements of high-performance aero-engines, an efficient, accurate, and reliable method has been developed to diagnose compressor aerodynamic instability. my the one-way ANOVA was used to optimize the time-domain parameters selection for detecting the instability precursor state. A model was then trained using a classification algorithm to detect the precursor of compressor instability, and a method for monitoring and warning the axial compressor instability by fusing multiple time-domain feature parameters was proposed. The results showed that the compressor exhibited specific precursor characteristics in the sound signal before instability at both speeds. Through validation with data from other test runs of the compressor, the method proposed in this paper effectively diagnosed and identified the aerodynamic instability precursor.