Civil Engineering Reference
In-Depth Information
unstable steady-state solutions for the Couette flow, similar
to those discovered by Nagata [NAG 90]. We will not go into
detail about this aspect in this topic. Interested readers can
consult [WAL 97].
5.9. Triggering of the mechanism
In the previous chapter, we gave a detailed discussion of
the scales governing the frequency of the ejections, and
concluded that f is governed by inner wall scales rather
than outer scales. The reasonable interpretation of f as the
frequency of regeneration of the near-wall structures
implies, a priori , that any triggering mechanism cannot be
linked to the outer flow. The DNS performed by Jimenez and
Pinelli [JIM 99] lend significant support to this hypothesis.
These authors extend their simulations in the minimum
channel (see [JIM 91]) to apply to Reynolds numbers ranging
up to
Re τ =
633
. The spanwise dimension of the channel is
kept at
. However, the spanwise integral scale near
to the center is typically
L + =
100
, according to [KIM 87],
L
=
0.5
h
int
+
which means that at
, the external structures
in the minimum channel are “squashed” at
h e
τ
==
633
L + =
100
, whereas
they should actually extend until
. The
external structure of the flow, therefore, is greatly modified,
the classic cascade mechanism is interrupted, and in spite of
this the profiles of the turbulent velocity- and vorticity
intensities remain essentially unchanged in the inner layer
50
+
+
L
>
0.5
h
=
317
y +
. Thus, a profound structural modification to the
outer layer causes only a slight modification to the internal
structure and the regeneration mechanism. The minimum
transition Reynolds number in a channel flow is
3
<
. Jimenez
and Pinelli [JIM 99] quite rightly stressed that the
mechanism of (bypass) transition does not need an outer
Re
=
hU
/
ν
=
0
[ORS 83], which is close to
h +
=
50
c
Search WWH ::




Custom Search