Decomposition of pressure fields to hydrodynamic and acoustic part using an EMD–cross-correlation approach

The separation of hydrodynamic (pseudo-sound) and acoustic pressure fluctuations in turbulent flows remains a fundamental challenge in aeroacoustics, particularly in the near field where both components coexist. In this work, we propose a time-domain approach based on Empirical Mode Decomposition combined with cross correlation analysis (EMD-CC) to distinguish between convective and propagative pressure components without relying on empirically defined spectral thresholds. The method decomposes measured pressure signals into intrinsic mode functions, after which each mode is classified according to its dominant propagation velocity, estimated from the time delay between spatially separated microphones. The classification is therefore based on a physically motivated criterion rather than on frequency content alone. The methodology is first validated using synthetic signals constructed to represent simplified hydrodynamic and acoustic pressure fields with distinct propagation characteristics. The method is then demonstrated on a well-documented low-Mach-number round jet, where results show excellent agreement with established wavelet-based and classical EMD techniques in terms of overall sound pressure level and spectral characteristics. The proposed method is further applied to the near field of a low-pressure axial fan. The analysis reveals a spatial separation between hydrodynamic and acoustic regions and provides new insight into the origin of tonal and broadband noise components. In particular, blade-induced pressure fluctuations are observed to be associated with radiated tonal noise at the inlet, while turbulence-driven broadband noise dominates at the outlet. The results demonstrate that the EMD-CC approach offers a robust and physically consistent tool for analysing complex aeroacoustic sources in practical applications.