Full Download Fluid Structure Sound Interactions and Control Yu Zhou PDF
Full Download Fluid Structure Sound Interactions and Control Yu Zhou PDF
Full Download Fluid Structure Sound Interactions and Control Yu Zhou PDF
OR CLICK LINK
https://textbookfull.com/product/fluid-structure-
sound-interactions-and-control-yu-zhou/
Read with Our Free App Audiobook Free Format PFD EBook, Ebooks dowload PDF
with Andible trial, Real book, online, KINDLE , Download[PDF] and Read and Read
Read book Format PDF Ebook, Dowload online, Read book Format PDF Ebook,
[PDF] and Real ONLINE Dowload [PDF] and Real ONLINE
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
https://textbookfull.com/product/fluid-structure-sound-
interactions-and-control-proceedings-of-the-3rd-symposium-on-
fluid-structure-sound-interactions-and-control-1st-edition-yu-
zhou/
https://textbookfull.com/product/fluid-structure-interactions-
volume-2-second-edition-slender-structures-and-axial-flow-
michael-p-paidoussis/
https://textbookfull.com/product/mechanics-of-flow-induced-sound-
and-vibration-volume-2_-complex-flow-structure-interactions-
william-k-blake/
https://textbookfull.com/product/perrines-literature-structure-
sound-and-sense-13th-edition-greg-johnson/
Biological fluid surface interactions in detection and
medical devices 1st Edition Michael Thompson
https://textbookfull.com/product/biological-fluid-surface-
interactions-in-detection-and-medical-devices-1st-edition-
michael-thompson/
https://textbookfull.com/product/software-adaptation-in-an-open-
environment-a-software-architecture-perspective-1st-edition-yu-
zhou/
https://textbookfull.com/product/combustible-solid-waste-
thermochemical-conversion-a-study-of-interactions-and-influence-
factors-1st-edition-hui-zhou-auth/
https://textbookfull.com/product/computational-fluid-flow-and-
heat-transfer-advances-design-control-and-applications-1st-
edition-mukesh-kumar-awasthi/
https://textbookfull.com/product/variable-structure-control-of-
complex-systems-analysis-and-design-1st-edition-xing-gang-yan/
Lecture Notes in Mechanical Engineering
Yu Zhou
Motoaki Kimura
Guoyi Peng
A. D. Lucey
Lixi Huang Editors
Fluid-Structure-
Sound
Interactions and
Control
Proceedings of the 4th Symposium on
Fluid-Structure-Sound Interactions and
Control
Lecture Notes in Mechanical Engineering
Lecture Notes in Mechanical Engineering (LNME) publishes the latest develop-
ments in Mechanical Engineering—quickly, informally and with high quality.
Original research reported in proceedings and post-proceedings represents the core
of LNME. Also considered for publication are monographs, contributed volumes
and lecture notes of exceptionally high quality and interest. Volumes published in
LNME embrace all aspects, subfields and new challenges of mechanical
engineering. Topics in the series include:
• Engineering Design
• Machinery and Machine Elements
• Mechanical Structures and Stress Analysis
• Automotive Engineering
• Engine Technology
• Aerospace Technology and Astronautics
• Nanotechnology and Microengineering
• Control, Robotics, Mechatronics
• MEMS
• Theoretical and Applied Mechanics
• Dynamical Systems, Control
• Fluid Mechanics
• Engineering Thermodynamics, Heat and Mass Transfer
• Manufacturing
• Precision Engineering, Instrumentation, Measurement
• Materials Engineering
• Tribology and Surface Technology
Lixi Huang
Editors
Fluid-Structure-Sound
Interactions and Control
Proceedings of the 4th Symposium
on Fluid-Structure-Sound Interactions
and Control
123
Editors
Yu Zhou A. D. Lucey
Shenzhen Graduate School Curtin University
Harbin Institute of Technology Perth, WA
Shenzhen Australia
China
Lixi Huang
Motoaki Kimura The University of Hong Kong
Nihon University Hong Kong
Tokyo China
Japan
Guoyi Peng
Nihon University
Tokyo
Japan
This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
part of Springer Nature
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Contents
v
vi Contents
Abstract For the past 50 years or so, Kolmogorov’s (1962) correction (K62) to his
1941 hypotheses (K41) has been embraced by an overwhelming majority of tur-
bulence researchers. Our recent work suggests that there are no valid reasons for
abandoning K41. In particular, analytical considerations, based on the NS equa-
tions, which take into account the finite Reynolds number (FRN) effect, together
with the available experimental laboratory data, seem to confirm a tendency towards
the simple and elegant predictions of K41 as the Reynolds number increases. This
is especially true when the focus is on the length scales which lie in the dissipative
range. Incorrectly accounting for the FRN effect and the inclusion of the atmospheric
surface layer (ASL) data, likely to have been affected by the proximity to the sur-
face, appear to be the major factors which have contributed to a nearly unchallenged
acceptance of K62.
1 Introduction
The theory of small scale turbulence has benefited immensely from the contribu-
tions of Taylor [1], who introduced the concept of isotropy and obtained a simplified
expression for the mean energy dissipation rate 𝜀,
̄ and Karman and Kowarth [2] who
derived a transport equation (hereafter referred to as the KH equation) for the two-
point velocity correlation function in isotropic turbulence. Further
significant progress was made in 1941 when Kolmogorov [3] introduced two impor-
tant hypotheses, specifically the first and second similarity hypotheses dealing with
turbulence scales in the dissipative (DR) and inertial (IR) ranges at very large
Reynolds number. In a separate paper, Kolmogorov [4] also derived, starting with
the KH equation rewritten in terms of 𝛿u, a simple expression for (𝛿u)3 ∕𝜀r,
̄ which is
generally interpreted as the energy flux or mean rate of transfer of energy down the
cascade at a scale r within the IR, viz.
(𝛿u)3 4
− = , (1)
𝜀r
̄ 5
where the asterisk denotes normalization by the Kolmogorov velocity and length
scales, uK = (𝜈𝜀)1∕4 and 𝜂 = (𝜈 3 ∕𝜀)1∕4 respectively. The function fun is a universal.
A consequence of (2) when r → 0 is that the normalized velocity derivative moments
should all be constant, i.e. independent of R𝜆 ,
(𝜕u∕𝜕x)n
Sn = n∕2
= const. (3)
2
(𝜕u∕𝜕x)
Note that S3 and S4 are the skewness and flatness factor (or kurtosis) of 𝜕u∕𝜕x. The
second similarity hypothesis argues that in the IR (𝜂 ≪ r ≪ L, where L is the integral
length scale), the local Reynolds number is sufficiently large to allow the effect of
viscosity to be ignored and yields the well-known result
where Cun are universal constants. The main outcomes of K41, Eqs. (2)–(4), were to a
large extent undermined by the third (revised) hypothesis introduced by Kolmogorov
20 years after K41 [5, 9].
The revision was made to take into account the spatio-temporal intermittency
of fluctuations in 𝜀 and assumed a log-normal model for 𝜀r , the subscript denoting
K41 Versus K62: Recent Developments 5
|Sn | ∼ R𝛼(n)
𝜆
(𝛼 > 0), (5)
where the exponent 𝜁un may depart from n∕3, except when n = 3 (since the exact
result, Eq. (1), needs to be preserved). The log-normal and subsequent probabilis-
tic models indicate that the departure from n∕3 increases as n increases, albeit at a
rate which may differ between models. Predictions from physical models, based on
generally simplistic proposals of how 𝜀 is localized in space, have not always qual-
itatively agreed with (5) (see for example Van Atta and Antonia [10], Sreenivasan
and Antonia [11], Wyngaard [12]).
It is fair to assert that experimental support for both (5) and (6) has been nothing
short of considerable. We do not wish to embark on a detailed discussion here; we
refer the reader to Van Atta and Antonia [10] and the reviews by Sreenivasan and
Antonia [11], Anselmet et al. [13], and Ishihara et al. [14]. Notwithstanding a few dis-
sentions, objections or doubts about K62 (e.g. [15–18]), the experimental evidence
has strongly pointed to K41 having to be abandoned in favour of K62. If we leave
aside the results from the ASL and recall that the most “damning” evidence against
K41, in the context of (6), came from the laboratory investigation of Anselmet et al.
[19], a major criticism that can be levelled at nearly all the laboratory studies is the
failure to properly recognise the influence of the Reynolds number, especially since
R𝜆 has seldom exceeded 1000, on the statistics of velocity derivatives as well as on
the moments of 𝛿u in the IR; the fact that the latter was never satisfactorily estab-
lished should have been cause for concern, if not alarm. The realization that the FRN
effect has to be taken into account came to the fore in 1997 (Qian [20]) and a little
later (Qian [21], Danaila et al. [22], and Lindborg [23]). The essence of the approach
was to revisit the KH equation (or the Karman-Lin equation [24] in the case of Qian
[21] who adopted a spectral approach) which includes the nonstationarity (or 𝜕u∕𝜕t
term) that was ignored by Kolmogorov [4]. The inclusion of the nonstationarity per-
mits an assessment of how the large scale inhomogeneity can affect the small scales
and hence an estimation of how large R𝜆 should be before Eq. (1) is satisfied. Such
an estimation was also carried out by Lundgren [25] and Antonia and Burattini [26].
The latter authors showed that “4/5” is approached more rapidly for forced than for
decaying turbulence; for a substantial IR to exist, the results indicated that R𝜆 may
need to exceed 103 in the former case and 106 in the latter. These results, obtained
via a scale-by-scale energy budget, were confirmed by Tchoufag et al. [27] using the
Eddy-Damped Quasi-Normal Markovian (EDQNM) method. One natural inference
6 R. A. Antonia et al.
from the previous results is that the approach to “4/5” is expected to depend on the
flow since each flow may have different levels of inhomogeneity at large scales.
The above considerations lead to only one conclusion: previous evidence in sup-
port of K62 needs to be critically reexamined in the context of the FRN effect since
this latter effect appears to have been incorrectly mistaken for the intermittency
effect. Note that intermittency was not explicitly taken into account in the previously
described work since xit is intrinsic to the N-S equations.
One objective of this paper is to first point out likely flaws in the approach used
hitherto for examining consequences of K41 and K62. This is done in Sect. 2. A
second is to highlight recent advances in understanding the behaviour of S3 with
R𝜆 , underpinned by the transport equation for (𝛿u)2 or scale-by-scale energy budget.
The limiting form of this budget has provided valuable insight into the Reynolds
number dependence of S as well as that of (𝛿u)3 in the scaling range. Some analytical
considerations are briefly given in Sect. 3. Results associated the DR and IR are
presented in Sects. 4 and 5 resepctively. Conclusions are given in Sect. 6.
Fig. 1, taken from Wyngaard and Tennekes (WT) [28], highlights the impact that
the ASL data has had on bolstering the validity of K62 and provides incisive insight
into how subsequent researchers may have been misled into accepting, if not
2
10
1
F
S
1
10
0.5
0.3
2 3 4 2 3 4
10 10 10 10 10 10
Fig. 1 Dependence of S and F on R𝜆 reproduced from Figs. 5 and 6 of Wyngaard and Tennekes
[28]
K41 Versus K62: Recent Developments 7
L
𝜎 2 = A + 𝜇 ln (7)
r
in the lognormal model; A depends on the large scale (L is the integral scale). WT
concluded that the dependence of S and F on R𝜆 contradicts K41 or the universal
equilibrium theory, whilst recognizing that the good agreement with the prediction
from the lognormal model may need reappraisal due to the choice made for r, the
neglect of A and the uncertainty in the value of 𝜇 (0.5 was used; the present day
consensus value for 𝜇 is about 0.2). With the benefit of hindsight, the methodology
embodied in Fig. 1 can now be criticized on at least three major levels. Firstly, it
ignores the FRN effect (this will be addressed in Sect. 3) which can affect the lab-
oratory data in a significant way and, to a lesser extent, the ASL data. Secondly,
the FRN effect, through the very nature from which it arises (it mainly reflects the
inhomogeneity associated with the large scale motion; the latter has been shown to
depend on the flow), is expected to affect different flows differently. This expecta-
tion is completely ignored in Fig. 1. The compilation by Van Atta and Antonia [10]
of how S and F vary with R𝜆 and its subsequent reappraisal (Sreenivasan and Anto-
nia, [11]) do indeed show that the laboratory values exhibit “apparent” scatter (our
recent work, [29–32], confirms that, at least for S, there is a systematic dependence
on R𝜆 which can be explained analytically). Clearly, one cannot afford to indiscrim-
inately use data obtained from various flows unless the Reynolds number is large
enough to allow the FRN effect in each of theses flows to be dismissed. The need to
achieve a sufficiently high Reynolds number in any given laboratory flow so that Eq.
(1) is approximately satisfied in the IR should, strictly speaking, be considered as
a minimum requirement before examining the consequences of K41 and K62. This
seems to have been largely overlooked. Thirdly, the inclusion of the ASL data when
testing K41 and K62 needs, at the very least, to be discussed or scrutinized more
objectively than in the past. For example, the Kansas data in Fig. 1 were collected at
three heights (z = 5.66, 11.3 and 22.6m) above the ground; with a few exceptions, the
ASL data used to test the dependence of S and F on R𝜆 have generally been obtained
below a height of 30m. Often, no information for the stability conditions is given.
Under “neutral” conditions, the ASL ought to be comparable (at similar values of
z∕𝛿) with the laboratory boundary layer provided the nature of the surface is also
comparable. It is difficult to estimate with confidence the thickness 𝛿 of the neu-
tral ASL. Wyngaard [12] suggests that, under neutral conditions, 𝛿 may scale with
U𝜏 ∕f (U𝜏 is the friction velocity and f is the Coriolis parameter) and estimates, for
8 R. A. Antonia et al.
mid-latitudes, 𝛿 to be within the range 250–2500 m. This implies that the majority of
the neutral ASL data for S and F were taken in the region z∕𝛿 < 0.1. It is not difficult
to conclude that there is significant uncertainty in the estimation of 𝛿, e.g. a value of 𝛿
smaller than 100 m was found by Metzger et al. [33] based on experiments in Utah’s
western desert under almost “exactly” neutral conditions; the authors stressed how-
ever the difficulties associated in making measurements under those conditions. In
any case, it is not clear if any the published ASL data were obtained under “exactly”
neutral conditions. Certainly, it is difficult to believe that measurements of S and F
in region z∕𝛿 < 0.1 have not been affected by the proximity to the surface. Recent
measurements (Djenidi et al. [34]) in a laboratory boundary layer over smooth and
rough walls indicate that both S and F are approximately constant (with respect to
z∕𝛿) in the region 0.3 < z∕𝛿 < 0.6 (R𝜆 is also approximately constant with respect
to z∕𝛿 in this region). However, the magnitudes of S and F can increase, perhaps by
as much as a factor of 3, as the wall is approached. In the region z∕𝛿 > 0.6, S and F
are affected by the intermittency associated with the turbulent/irrotational interface.
The implication of these observations is that it would be unwise to use data outside
the range 0.3 < z∕𝛿 < 0.6 when testing K41 and K62. Naturally, one should also
check that there are no major departures from local isotropy in this range due to the
probable effects of the mean shear. In the region z∕𝛿 < 0.1, a larger departure from
local isotropy is expected due to the combined effect of an increase in mean shear
and the “blockage” caused by the presence of the surface. Notwithstanding the large
values of R𝜆 that can be achieved in the ASL (the ASL results of Sreenivasan and
Dhrava [35] at z ≈ 35m illustrate the difficulty in unambiguously estimating power-
law exponents for (𝛿u)2 and (𝛿u)4 in the scaling range even when R𝜆 is in the range
10000–20000), it would be preferable to err on the side of caution and exclude all
the previously obtained ASL data when examining the R𝜆 dependence of S and F
and examining the behaviour of (𝛿u)n in the scaling range. We do not rule out the
potential viability of data, collected for example in the region 0.3 < z∕𝛿 < 0.6 in a
near-neutral ASL (this may entail making airborne measurements e.g. Sheih et al.
[36]; flying can circumvent, if not minimize, the effect any possible nonstationarity
and facilitate the convergence of statistics) or data collected away from boundaries
as in the tidal channel experiment of Grant et al. [37].
3 Analytical Considerations
For homogeneous isotropic turbulence at very large Reynolds number, the transport
equation for (𝛿u)2 is given by
𝜕 4
−(𝛿u)3 + 6𝜈 (𝛿u)2 = 𝜀r. (8)
𝜕r 5
K41 Versus K62: Recent Developments 9
If we divide all terms in Eq. (8) by 𝜀r, the above relation states, in essence, that
the energy flux −(𝛿u)3 ∕𝜀r in the IR, where the viscous term is negligible, remains
constant (see Eq. (1)). In small to moderate Reynolds number, this equation is usually
not satisfied except at small r since Eq. (8) does not contain any large-scale term. For
example, [22] showed that Eq. (8) is satisfied only for r∕𝜂 ≤ 5 for grid turbulence
at R𝜆 = 66, suggesting that the large-scale term contributes to Eq. (8) for r∕𝜂 > 5.
After including such as term, Iu (r) say, Eq. (8) becomes
𝜕 4
−(𝛿u)3 + 6𝜈 (𝛿u)2 = 𝜀r − Iu (9)
𝜕r 5
where Iu differs from flow to flow, e.g. in decaying HIT [22, 38], along the axis in
the far field of an axisymmetric jet flow [39], and along the centreline of a fully
developed channel flow [40]. This equation is of fundamental importance since it is
an equilibrium relation between second- and third-order moments. Further, Eq. (9)
provides a viable platform for quantifying the FRN effect on (𝛿u)3 since it provides
insight into the budget of energy at each scale. As r → 0, Eq. (9) becomes the trans-
port equation for the mean enstrophy or 𝜀, which can be written in the generic form
[29–32]
G C
S+2 = , (10)
R𝜆 R𝜆
( )
(𝜕 2 u∕𝜕x2 )2
where G = u2 2 is the enstrophy destruction coefficient and C is a constant
(𝜕u∕𝜕x)2
which depends on the flow and may vary across regions of the same flow. Equation
(10) represents, in essence, a constraint on how S varies with R𝜆 in different flows.
Since the ratio G∕R𝜆 approaches a constant relatively rapidly with increasing R𝜆
[29–32] and the terms on the right side of (10) must eventually vanish, Eq. (10)
implies that the magnitude of S should become constant at sufficiently large R𝜆 . This
expectation has been confirmed [29–32] and will be discussed briefly in the next
section.
There has been strong support for K41 in the DR for the second-order statistics. For
example, Kolmogorov-normalized one-dimensional velocity spectra 𝜙∗u (k1∗ ) collapse
very well in the high wavenumber region [41]. In contrast to the stipulations of K41
and K62, viz. R𝜆 must be very large and local isotropy should hold, this collapse
does not require R𝜆 to be large [41], not does it require local isotropy to be satisfied
rigorously; it does however break down [42] when R𝜆 is sufficiently small, typically
when it falls below about 40. Pearson et al. [43] showed that (𝛿u∗ )2 also collapses
reasonably well in the DR over a large range of R𝜆 (40 < R𝜆 < 4250). In flows con-
10 R. A. Antonia et al.
0.5
0.2 1 2 3
10 10 10
Fig. 2 Dependence of S on R𝜆 in different flows. The 4 solid curves (black, red, blue and pink),
inferred from Eq. (10) by assuming that 2G∕R𝜆 = constant (≈ 0.53 for R𝜆 ≥ 70 in all flows except
the far-wake where R𝜆 ≥ 100), correspond to the flow the axis in the far-field of a round jet, decaying
grid turbulence, the centerline of a fully developed channel flow and the axis in the far-wake of a
circular cylinder; corresponding measurements in each flow are shown as symbols with the same
color without identifying the sources [29–32]. The green dashed curve is the prediction by Qian
[16]; the solid green symbol was obtained in the ONERA wind tunnel [45]. The horizontal dashed
line indicates the value of 0.53
0.5
2 3 4 5 6 7 8
12 R. A. Antonia et al.
that the FRN effect, which depends on the flow, needs to be properly accounted for
before determining whether corrections to K41 are required.
6 Conclusions
We, once again, recall that both K41 and K62 are asymptotic R𝜆 → ∞ phenomenolo-
gies. As such, we may never be able to decide if K62 provides a realistic alternative
to K41 when R𝜆 → ∞. What is of paramount importance however is to critically
appraise the manner with which K41 and K62 have been tested. In Sect. 2, we put for-
ward plausible arguments for excluding ASL data obtained in the region z∕𝛿 < 0.1
when testing K41 and K62. Laboratory boundary layer data obtained in a similar
region are not suitable for this purpose, given the likely effect of the strong shear
and the departure from LI, and have not, to our knowledge, been used. If we set
aside the ASL data, the laboratory data for S and F are consistent with K41, at least
when R𝜆 approaches 103 , i.e. S and F approach constant values. Our recent analyt-
ical work [29–32] has allowed a reliable estimation of the way S is affected by the
FRN effect in a variety of flows. This has confirmed that S does indeed approach
a constant along different paths in different flows, thus debunking the practice of
indiscriminately lumping together values of S obtained in various flows before test-
ing the dependence of S on R𝜆 . The analytical approach adopted is based on the
scale-by-scale energy budget, derived from the N-S equations, which allows impor-
tant physical constraints to be satisfied, namely the energy and enstrophy budgets
at large and small r respectively, in each flow. The challenge ahead of us is to now
extend this approach to examine how F varies in each flow; we believe this should be
feasible, see for example [49], with the help of DNS. The laboratory data for F [32,
50] indicate that, like S, F also becomes constant at sufficiently large R𝜆 (this value
is not much larger than 10) thus further supporting K41; of course, confirmation of
this observation, underpinned by the N-S equations, is essential. Further, we have
observed, using only laboratory data, that the scaling range exponents tends towards
the K41 value of n∕3, at least for 2, 3, 4, as R𝜆 increases. The scaling exponent 𝛼n
approaches n∕3 (the K41 value), more slowly as n increases. Naturally, it is highly
desirable that we continue to scrutinize this trend by trying to further increase R𝜆 ,
preferably within the controlled environment of the laboratory.
Acknowledgements SL Tang wishes to acknowledge support given to him from NSFC through
grant 11702074.
K41 Versus K62: Recent Developments 13
References
29. Antonia RA, Tang SL, Djenidi L, Danaila L (2015) Boundedness of the velocity derivative
skewness in various turbulent flows. J Fluid Mech 781:727–744
30. Tang SL, Antonia RA, Djenidi L, Abe H, Zhou T, Danaila L, Zhou Y (2015) Transport equation
for the mean turbulent energy dissipation rate on the centreline of a fully developed channel
flow. J Fluid Mech 777:151–177
31. Tang SL, Antonia RA, Djenidi L, Zhou Y (2015) Transport equation for the isotropic turbulent
energy dissipation rate in the far-wake of a circular cylinder. J Fluid Mech 784:109–129
32. Antonia RA, Djenidi L, Danaila L, Tang SL (2017) Small scale turbulence and the finite
Reynolds number effect. Phys Fluids 29:020715
33. Metzger M, McKeon BJ, Holmes H (2007) The near-neutral atmospheric surface layer: turbu-
lence and non-stationarity. Phil Trans R Soc Lond A 365:859–876
34. Djenidi L, Antonia RA, Talluru MK, Abe H (2017) Skewness and flatness factors of the lon-
gitudinal velocity derivative in wall-bounded flows. Phys Rev Fluids 2:064608
35. Sreenivasan KR, Dhruva B (1998) Is there scaling in high-Reynolds-number turbulence? Prog
Theor Phys Suppl 130:103–120
36. Sheih CM, Tennekes H, Lumley JL (1971) Airborne hot-wire measurements of the small-scale
structure of atmospheric turbulence. Phys Fluids 14:201–215
37. Grant HL, Stewart RW, Moilliet A (1962) Turbulence spectra from a tidal channel. J Fluid
Mech 12:241–268
38. Antonia RA, Zhou T, Danaila L, Anselmet F (2000) Streamwise inhomogeneity of decaying
grid turbulence. Phys Fluids 12:3086
39. Thiesset F, Antonia RA, Djenidi L (2014) Consequences of self-preservation on the axis of a
turbulent round jet. J Fluid Mech 748(R2)
40. Danaila L, Anselmet F, Zhou T, Antonia RA (2001) Turbulent energy scale-budget equations
in a fully developed channel flow. J Fluid Mech 430:87–109
41. Antonia RA, Djenidi L, Danaila L (2014) Collapse of the turbulent dissipation range on Kol-
mogorov scales. Phys Fluids 26:045105
42. Djenidi L, Tardu SF, Antonia RA, Danaila L (2014) Breakdown of Kolmogorov’s first simi-
larity hypothesis in grid turbulence. J Turb 15:596–610
43. Pearson BR, Antonia RA (2001) Reynolds-number dependence of turbulent velocity and pres-
sure increments. J Fluid Mech 444:343–382
44. Pope SB (2000) Turbulent flows. Cambridge University Press
45. Bos WJT, Chevillard L, Scott JF, Rubinstein R (2012) Reynolds number effect on the velocity
increment skewness in isotropic turbulence. Phys Fluids 24:015108
46. Tsuji Y, Ishihara T (2003) Similarity scaling of pressure fluctuation in turbulence. Phys Rev E
68:026309
47. Meldi M, Sagaut P (2013) Pressure statistics in self-similar freely decaying isotropic turbu-
lence. J Fluid Mech 717:R2
48. Tang SL, Antonia RA, Djenidi L, Danaila L, Zhou Y (2017) Finite Reynolds number effect on
the scaling range behavior of turbulent longitudinal velocity structure functions. J Fluid Mech
820:341–369
49. Djenidi L, Antonia RA, Danaila L, Tang SL (2017) A note on the velocity derivative flatness
factor in decaying HIT. Phys Fluids 29:051702
50. Tang SL, Antonia RA, Djenidi L, Danaila L, Zhou Y (2017) Reappraisal of the velocity deriva-
tive flatness factor in various turbulent flows. J Fluid Mech (in revision)
Jets and Waves Generated by an Obstacle
in Stratified or Homogeneous Fluids
Hideshi Hanazaki
Abstract Jets and waves, whose origin is in gravity force, are often observed in flu-
ids. When the fluid has a vertical density stratification, both can be generated due to
the buoyancy force. The jet appears when an obstacle descends vertically in a strat-
ified fluid. The generation is supported by the molecular diffusion of the stratifying
scalar such as salt. If there is no scalar/density diffusion, the density must be con-
served along the fluid motion, and the originally horizontal isopycnal surfaces are
simply deformed as they are dragged down by the obstacle, so that the dragged fluid
will move downward indefinitely and will never go back to its original height. If the
diffusion exists, the fluid can change its original density, and move away from the
isopycnal surface to which the fluid originally belonged. In this study, we demon-
strate the generation process of the jet, along with its interaction with the internal
gravity waves. As another example of waves in fluids, water waves with capillary
effects in a homogeneous fluid are numerically simulated, showing the excitation of
short capillary waves by a long solitary wave. The phenomenon has been predicted
by a weakly nonlinear theory, but has not yet been observed.
Keywords Stratified fluid ⋅ Jet ⋅ Internal gravity wave ⋅ Capillary gravity wave
There have been uncountably many studies on stratified flows, but most of them
are on horizontal flows, since many of the applications of stratified flows are in the
large-scale atmospheric or oceanic flows in which the horizontal velocity is usually
much larger than the vertical velocity. However, importance of the vertical flow gen-
erated by a vertical movement of an obstacle in stratified fluid has been recently
recognized, particularly in smaller-scale flows. For example, accurate control of the
H. Hanazaki (✉)
Department of Mechanical Engineering and Science, Kyoto University,
4 Kyoto daigaku-katsura, Nishikyo-ku, Kyoto 615-8540, Japan
e-mail: [email protected]
vertical movement of floats used for the abyssal observation is necessary to develop a
reliable float, and the prediction of the vertical movement of zoo-planktons is essen-
tial to estimate the carbon circulation in the ocean, which is important in the climate
model to predict the global warming of the earth.
In this study, we consider a sphere moving vertically at constant speed in a linearly
stratified fluid. Both the experiments and numerical simulations are performed, and
various jet/wake patterns are observed and the essential mechanisms of the flow,
including the mechanism of the jet formation, are investigated.
The return journey was for the most part uneventful, but with
empty wagons we could travel more rapidly.
On our reaching the crossing of the Arkansas we found there a
company of dragoons, and the officers informed us that they had
been fighting and chasing the Cheyennes all summer, having just
halted there in following one band of these Indians to the Arkansas
river. They had been forced to abandon their provision wagons some
days before we saw them, and were almost entirely out of food. The
artillery had also been left behind two or three days’ march down the
Arkansas river. These troops, a part of Colonel Sumner’s regiment,
had had several brushes with the Cheyennes, and captured a lot of
horses from the Indians. The soldiers, their horses and equipments,
gave every evidence of having undergone a severe campaign, and
they came around our camp begging for something to eat, tobacco
and whisky, much as the Indians were in the habit of doing. But our
ability to relieve their wants was very limited, having with us only
supplies enough for our own party back to the settlements.
The officers said that it would be hazardous for us to proceed
further, advising our captain to remain until the trains in our rear
could get up, until they had accumulated to at least one hundred
wagons and men, when we would be strong enough to resist any
attack that we were likely to be subjected to.
Acting on this advice, we remained in camp several days, until
five or six trains had arrived and camped in our immediate vicinity.
The journey was then resumed, our train taking the lead, all our
weapons of defense being put in as good order as possible. After the
trains were under way the wagonmasters of those behind us, to the
number of ten or a dozen, mounted on horses and mules, would ride
ahead to join Captain Chiles, Reece and myself, thus forming a lively
and agreeable company of companionable men.
As we were thus riding along down the level bottom of the
Arkansas, some distance in advance of the trains away to our right a
mile or more, out near the bank of the river, where we could see
some scattering cottonwood trees, we observed a smoke rising from
a camp fire. Some one of the party suggested that it was the smoke
of the camp of the artillery company, of which we had been told, so
we rode forward, giving little more attention to the smoke of the
camp fire that went curling upward among the cottonwood. When we
had reached a point about opposite the smoke there suddenly
appeared in our view a company of some fifty horsemen, riding pell-
mell in a fast gallop towards us. They were yet too far off to be
distinctly seen or for us to tell what manner of men they were. In
another moment, Captain Chiles exclaimed:
“Men, they are Indians! Soldiers don’t ride in that disorderly
manner. Form a line and get out your guns. We are in for it!”
Instantly all hands obeyed his command, forming a line, facing
the enemy, each of us drawing a pistol. The lead wagons of our train
were just barely visible, probably two miles from us. When the
approaching horsemen saw that we had formed a line of battle, they
instantly drew rein, slackening their speed to a walk, but kept
steadily drawing nearer us.
“MEN, THEY ARE INDIANS!”
When Reece had got to the top of the mound he saw Captain
Chiles, sitting on a horse, holding by a rope a huge bull elk. The elk
stood in the bottom of a deep, narrow ditch, ten feet deep, with
banks almost perpendicular, so steep that he was unable to get up
them or out of the ditch to assail his captor. Captain Chiles, when he
first caught up with the band of elk, had made an effort to kill one
with the pistols, but for some reason he could only get the pistols to
fire two of the charges, and with these two he only wounded a cow
slightly, not enough to stop her from running. He kept after the band,
all the while trying to get the revolver to fire, trying every chamber,
but with no success. After he had kept up the chase for two or three
miles the large bull elk, being very fat, got too tired to keep up with
the band, but trotted along behind, in fact, so far exhausted that
Chiles could keep up with him with his horse in a trot. The captain
despaired of being able to stop one with the pistols, and, finding a
small lariat I had brought from the Kiowas as we went out, on my
saddle, used for picketing my horse, resolved to try the plan of
lassoing the big fellow.
Being an expert in rope throwing, he had little difficulty in
preparing the noose or getting a fastening around the top prong of
one branch of the elk’s great antlers. As soon as the elk found he
was restrained by the rope he turned about and charged on Captain
Chiles with all the power and fury he could command, and twice or
thrice the captain was forced to cut loose from him in order to
escape his assaults. The rope was long enough to drag on the
ground some distance behind him, so that the captain could recover
hold of it without dismounting, reaching down and picking it up as the
bull trotted away from him. He kept on after him for some distance,
occasionally jerking him back, and worrying him until he could hardly
walk. Coming to the lower end of the ditch, washed out to a depth of
ten feet, at a point a few yards above, he managed to guide the
animal, bewildered as he was by the heat, together with the violent
and prolonged exercise, into it, leading or driving him along up the
ditch until he got him in between the high banks of it to a place
where the animal could not get at him however anxious he was to do
so.
When Reece arrived, as above related, he found Chiles sitting
there on the horse holding the end of the rope, but having nothing
with which to kill the animal, not even a pocket pistol. Reece had
with him a belt revolver, and, under the directions of Chiles, he
carefully crawled to the edge of the ditch to within a few feet of the
elk’s head and killed him with a couple of shots in the forehead.
The bull had not been wounded by Chiles, and no one but a
veritable daredevil as he was would have undertaken the job of
lassoing an elk under such circumstances as he did. But Chiles was
a stranger to fear.
Chiles, Reece and I got to the camp about 2 o’clock, near six
miles from where the elk was killed. After dinner we went out with
pack mules and the necessary hatchets and butcher knives, and two
of the drivers, to butcher the elk. The animal was a splendid
specimen of his kind, supporting a magnificent pair of antlers, fully
hardened and developed, and was fatter than any other animal of
the deer kind I have seen, before or since. We butchered and
brought to camp on the pack mules every part of his carcass,
including the antlers. The latter were brought home to Jackson
county. We feasted on the flesh of the fat elk for several days, and
my recollection is that I never tasted better meat.
The remaining part of the journey was uneventful, the entire
party remaining with the train until we were within eighty miles of the
state line of Missouri. Then, in company with Captain Chiles, I
started, before daylight, to make a forced march to Westport. We
rode forty miles before we halted for breakfast, obtaining it at a
settler’s cabin in the vicinity of Black Jack, arriving in Westport late in
the evening, in the latter part of September, feeling very willing to
rest once more in a comfortable house and bed.
I saw my friend Reece about a year after he had returned to his
home in Missouri still making a fight for life, but during the second
year he struck his flag and made a final surrender.
At Westport the drivers were paid off and disbanded, but I was
not present to witness the separation of the company that had
formed a companionship, offensive and defensive, during this long
and tiresome journey across the plains. Doubtless nearly all of them,
in the vernacular of the Western mountains, have “crossed over the
range.”
Lewis & Clark’s Route
Retraveled.