Turbulence and Scale Relativity

TL;DR

This paper introduces a novel formalism based on scale relativity to model turbulence, transforming the Navier-Stokes equations into a Schrödinger-like equation, and validates predictions with experimental data showing new acceleration components and intermittent behaviors.

Contribution

It develops a new turbulence modeling approach using scale relativity, predicting a divergence in acceleration at minima of velocity PDFs and validating these predictions with experimental data.

Findings

Detection of empty zones in velocity PDFs.

Identification of a new acceleration component in turbulence data.

Theoretical prediction and experimental confirmation of acceleration PDF shape.

Abstract

We develop a new formalism for the study of turbulence using the scale relativity framework (applied in $v$-space according to de Montera's proposal). We first review some of the various ingredients which are at the heart of the scale relativity approach (scale dependence and fractality, chaotic paths, irreversibility) and recall that they indeed characterize fully developped turbulent flows. Then we show that, in this framework, the time derivative of the Navier-Stokes equation can be transformed into a macroscopic Schr\"odinger-like equation. The local velocity PDF is given by the squared modulus of a solution of this equation. This implies the presence of null minima $P_v(v_i)\approx 0$ in this PDF. We also predict a new acceleration component in Lagrangian representation, $A_q=\pm D_v \: d \ln P_v/dv$, which is therefore expected to diverge in these minima. Then we check these theoretical predictions by data analysis of available turbulence experiments: (1) Empty zones are in effect detected in observed Lagrangian velocity PDFs. (2) We give a direct proof of the existence of the new acceleration component by directly identifying it in the data of a laboratory turbulence experiment. (3) It precisely accounts for the bursts and calm periods of the intermittent acceleration observed in experiments. (4) Moreover, the shape of the acceleration PDF can be analytically predicted from $A_q$, and this theoretical PDF precisely fits the experimental data, including the large tails. (5) Finally, numerical simulations of this new process allow us to recover the observed autocorrelation functions of acceleration magnitude and the exponents of structure functions.