Quantifying interaction among POD modes in turbulent flows using transfer entropy
Stoļarovs K, Toosi S, Yang L, Massaro D, Garcia E, Yao J, Mihajlović M, Hussain F, Rezaeiravesh S, Schlatter P (2026)
Publication Type: Conference contribution
Publication year: 2026
Journal
Publisher: Institute of Physics
Book Volume: 3230
Conference Proceedings Title: Journal of Physics: Conference Series
Event location: ESP
DOI: 10.1088/1742-6596/3230/1/012028
Abstract
Proper Orthogonal Decomposition (POD) can effectively identify large-scale coherent structures in turbulent flows. To quantify causal relationships between such structures, applying non-intrusive (data-driven) information theoretic measures, particularly transfer entropy (TE), has been receiving growing attention. In this study, the TE is computed for the POD temporal coefficients of two turbulent flows, focusing on the impact of the hyperparameters of the TE estimator and, more importantly, the arbitrary phase shift of the time series. The POD is applied to the velocity snapshots obtained from direct numerical simulations of a canonical turbulent pipe flow and a stepped cylinder flow. We demonstrate that the Kraskov-Stögbauer-Grassberger (KSG) estimator provides stable and reproducible estimates for TE when investigating the impact of its main hyperparameter. The inherent phase shifts in the POD coefficients have been mainly overlooked in the TE-POD literature. We found that they notably affect the TE magnitude in convection-domainated flows (stepped cylinder wake), but have smaller impact in flows with continuous symmetry in the axial direction (canonical pipe flow). The same analyses can be extended to other turbulent flows to quantify causal relations among various turbulent structures.
Authors with CRIS profile
Siavash Toosi
Lehrstuhl für Strömungsmechanik (LSTM)
Philipp Schlatter
Lehrstuhl für Strömungsmechanik (LSTM)
Involved external institutions
Beijing Institute of Technology (BIT) / 北京理工大学
China (CN)
Texas Tech University (TTU)
United States (USA) (US)
University of Manchester
United Kingdom (GB)
Royal Institute of Technology / Kungliga Tekniska Högskolan (KTH)
Sweden (SE)
China (CN)
Texas Tech University (TTU)
United States (USA) (US)
University of Manchester
United Kingdom (GB)
Royal Institute of Technology / Kungliga Tekniska Högskolan (KTH)
Sweden (SE)
How to cite
APA:
Stoļarovs, K., Toosi, S., Yang, L., Massaro, D., Garcia, E., Yao, J.,... Schlatter, P. (2026). Quantifying interaction among POD modes in turbulent flows using transfer entropy. In Journal of Physics: Conference Series. ESP: Institute of Physics.
MLA:
Stoļarovs, K., et al. "Quantifying interaction among POD modes in turbulent flows using transfer entropy." Proceedings of the 6th Madrid Turbulence Workshop, MTW 2025, ESP Institute of Physics, 2026.
BibTeX: Download