Finite Amplitude Scaling in Transitional Pipe Flows

TL;DR

This study uses direct numerical simulations to classify the scaling exponents of finite amplitude disturbances in pipe flows, linking them to the radial distribution of initial disturbances and clarifying different routes to turbulence.

Contribution

It identifies two distinct scaling exponent ranges and associates them with specific initial disturbance locations, advancing understanding of transition mechanisms in pipe flows.

Findings

Two scaling exponent ranges are classified:  eta  extless  -1.3 and eta  extgreater  -1.

The steep exponent range is linked to boundary disturbances, while the shallow range relates to disturbances elsewhere.

The study clarifies the disturbance characteristics leading to different transition behaviors.

Abstract

Studies on the finite amplitude stability of pipe flows identified a range of different scaling exponents between $\beta\approx -1 $ and $\beta\approx-1.5$, relating $A\sim Re^{\beta}$, where $A$ is the minimum amplitude of disturbance to cause a transition to turbulence and $Re$ is the Reynolds number. The circumstance under which a particular scaling exponent manifests itself is still not clear. Understanding this can shed light on the different routes to turbulence \citep{willis2008experimental} and the mechanisms involved. The exponents observed in previous experiments and simulations were explained based on the spatial localization of initial disturbances. In this paper, through direct numerical simulations (DNS), we classify the exponent, $\beta$ into two ranges; a steeper exponent with $\beta\lessapprox-1.3$ and a shallower exponent with $\beta\gtrapprox-1$. We then determine the nature of the disturbance to produce a specific exponent. Our results clearly show that the two ranges of the scaling exponents are related to the radial distribution of the initial disturbance, where $\beta \lessapprox -1.3$ exists for a disturbance at the boundary, and $ \beta \gtrapprox -1$ exits otherwise. We also compare the previous experiments and simulations on injection-type and push-pull-type initial disturbances. This study clarifies the nature of the initial disturbance that can result in either of the two different scaling exponents observed so far.