Abstract Centrifugal compressors are crucial components in several industrial applications, making it fundamental to understand their behavior, particularly under incipient surge conditions, to optimize their operating region. This study examines surge transitions in centrifugal compressors by utilizing a comprehensive approach that combines dynamic mass flow rate and pressure measurements, obtained through anemometric fiber film probes and high-frequency piezoresistive pressure transducers, along with vibroacoustic data collected from micro mono-axial accelerometers and microphones. To better understand the transition from stable to unstable compressor operation, an extensive experimental campaign was conducted on a centrifugal compressor at the test bench for components of propulsion systems of the University of Genoa. These tests involved both steady-state and transient measurements to characterize the compressor’s behavior in the stable region, during incipient surge, and under deep surge condition. The acquired experimental data on the compressor’s fluid dynamics and mechanical response were analyzed using frequency and time-frequency techniques to define the compressor operation as stable or unstable. Time-domain synchronous average was found to be effective in the identification of incipient surge conditions. Additionally, statistical methods such as Skewness and Kurtosis were also employed proving their reliability in surge detection. Finally, coherence functions calculated using multiple sensors in both stable and unstable conditions demonstrated strong potential as a tool for surge detection. The main goal of this study is to provide insights that support the development of advanced control strategies and predictive protocols, helping to mitigate surge-related problems and enhance the overall performance and reliability of systems where centrifugal compressors are installed.
Usai et al. (Mon,) studied this question.