Search

It is often said that turbulence is the last unsolved problem of classical physics \cite{Uriel}. This is caused by its enormous complexity resulting from the mutual interactions of large number of \emph{objects} of different sizes and strengths -- the \emph{vortices} and other \emph{coherent structures}. The classical definitions often mention the chaotic behavior \cite{Pope}. In fact, turbulence is not chaotic: the laminar flow is chaotic, as the random thermal molecular motion averages out and the large-scale flow is governed only by external forces. In the case of turbulence, the molecular motion is organized over large distances (in comparison with molecular scale); the motion of molecules in one part of a vortex is connected with the opposite motion in the second part of this vortex. Therefore the entropy of the turbulent flow is smaller than that of corresponding laminar flow. The rise of turbulence (i.e. system of organized interacting structures) from the laminar flow (uniform motion, random at micro-scale) is one example of the \emph{self-organization} problem \cite{TheRise}. The decreased entropy of the system leads to faster increasing of total entropy as the organized structures are much more effective in mixing and transport of momentum or heat than the molecular diffusion alone. We see the turbulence as an random chaotic process just because we do not understand it. Lets use a parable: imagine You are listening a speech in some language unknown for You. First You hear only a \emph{chaotic} sequence of sounds, which is clearly different from \emph{white noise}. As You study that language, You are able to recognize individual vocals, later on the words and even later the sentences carrying information. The idea of chaos comes never more into Your mind. Contemporary methods of studying turbulent flows are mainly based on measuring some average quantity (velocity, pressure) and its variance or higher statistical moments. More advanced studies use frequency analysis (spectra \cite{DudaUruba2019}), correlation or structure functions \cite{SF} and powerful Eigen-decompositions (e.g. proper orthogonal decomposition POD \cite{POD}). In the light of the above mentioned parable, such trials of understanding are naive. Therefore, we slowly develop an algorithm for identification of individual vortices. The idea is based on the assumption, that vortices are dominant coherent structures in the turbulent flow, thus the induced velocity field produced by the set of vortices might reconstruct the original measured velocity field. The properties of individual vortices (i.e. their fitting parameters: position, radius and circulation) are studied statistically.