Statistics of turbulent-to-laminar transition in plane Couette flow
Abstract
A study of turbulence decay in plane Couette flow at large aspect ratio suggests an alternative spatiotemporal interpretation to the current chaotic transient paradigm It is supported by numerical simulations of a reduced semi-realistic two-dimensional partial-differential model. Evidence comes from the distribution of the sizes of laminar domains nucleating in the turbulent state. Statistical analysis shows that, above some well-defined Reynolds number R(low), its variance stays finite while it diverges below. Accordingly, at large aspect ratio, turbulence has vanishingly small probability to decay when R >= R(low), while decay is unavoidable for R < R(low)