STEADY AND UNSTEADY FLOW IN WAVE-INDUCED BOUNDARY LAYERS

D. Mouazé, Y. Ourmières and J. R. Chaplin,

Department of Civil and Environmental Engineering, University of Southampton, Southampton SO17 1BJ, UK
E-mail: [email protected]; [email protected]; [email protected]

SUMMARY

This paper describes measurements of velocities inside the seabed boundary layer in conditions where the boundary is
slightly rough. Agreement with linear and non-linear theory is found to deteriorate as the higher harmonic components
of the waves become more important. In all cases the steady streaming at the edge of the boundary layer is considerably
weaker than predicted.

Previous measurements of the flow inside wave-induced

1. INTRODUCTION
boundary layers include those by Beech [6], Sleath [7],
Hwung & Lin [8] and Liu, Davis & Downing [9]. These
Wave-induced boundary layers on the seabed and on the
are small scale experiments over smooth beds, and
surface of floating bodies have some remarkable
properties that are fully understood only over smooth generally reveal good agreement with the oscillatory
boundaries and in waves of moderate amplitude at small flow predicted by equation (1), and moderate agreement
scale. Owing to its progressive and oscillatory with the steady flow predicted by equation (2). The
characteristics, the boundary layer at the seabed importance of scale effects, transition to turbulence, and
generates steady streaming in the direction of wave wave non-linearities are however not known. In this
propagation that may be responsible for material paper we present measurements of velocities inside the
transport over large distances [1]. Secondary flow seabed boundary layer in conditions where the boundary
similarly generated around floating and submerged is slightly rough. Agreement with equations (1) and (2),
bodies can have surprising effects on their loading and and with the non-linear theory, is found to deteriorate as
response [2]. the higher harmonic components of the waves become
more important.
For the simple harmonic case in which the wave-induced
flow just outside the boundary layer is the real part of 2. EXPERIMENTAL ARRANGEMENTS
u = u1 exp[i (kx − ωt )] , a solution to second order in the
expansion parameter (u1k2ν/βω), where β-1 is the These velocity measurements were made in advance of
boundary layer length scale, β = (ω / 2ν)1/2 and ν is the separate investigations of the flow around a horizontal
kinematic viscosity, was obtained by Longuet-Higgins cylinder at two different scales [10]. They were
[3]. The first order part consists of the oscillatory flow in performed by Laser Doppler Velocimetry in two wave
the boundary layer flumes. The larger one, at the University of Hanover, is
120m long, 2.2m wide, and has a still water depth of
u (1) = u1 exp[i (kx − ωt )] {1 − exp[−(1 + i )βy ]} (1) 1.0m. Other tests were carried out in a smaller flume at
the University of Caen, 22m long, 0.8m wide and with a
still water depth of 0.5m. Both wavemakers have active
while the second order component has a steady Eulerian absorption control; reflections from the beach were
velocity around 10% in the Hanover flume, but much smaller in
the Caen flume. The fibre-optic based LDV system
u12 k
u (2s) = [3 − 2(βy + 2) exp(−β y ) cos(β y ) − provided a spatial resolution of about 0.15mm, and was
4ω (2) positioned in turn with the centre of its measurement
2(βy − 1) exp(−βy ) sin(βy ) + exp(−2βy ) volume at elevations between 0.1mm and 300mm above
the floor of the tank. The roughness of the floor was
where y is measured from the surface. As y tends to estimated at 0.8mm and 0.15mm for the Hanover and
infinity, u(2s) approaches 3u12 k /4ω , which is the steady Caen flumes respectively. Elevations y given below are
measured from the top of the roughness elements. A
streaming velocity that might be expected at the outer
wave gauge was placed directly over the LDV system.
edge of the boundary layer. Higher order solutions were
developed by Sleath [4] for the bed boundary layer and
Results given below were obtained from velocity records
by Riley [5] for the related case of the viscous flow that had been phase-averaged over a sequence of similar
around a cylinder driven around a circular path without waves. The wave conditions for the 8 test cases
rotation. Comparisons are made here with the former, to discussed below are set out in Tables 1 and 2, in which
second order, in conditions where the amplitude of u2/u1 is the ratio of the amplitudes of the first two
second harmonic component of the wave-induced flow is harmonic components of the wave-induced flow just
as much as 30% of that of the fundamental. outside the boundary layer.
 Case C1 C2 C3 C4 flow inside, it seems unlikely that a suitable boundary
T (s) 1.33 1.50 1.80 2.20 condition can be formulated explicitly. In numerical
H (m) 0.08 0.13 0.16 0.16 solutions to be presented later, we follow the approach
L (m) 2.39 2.82 3.57 4.53 adopted by Riley [11]. The computation is repeated with
β (m-1) 1439 1355 1237 1119 different steady outer velocities superimposed on the
u2/u1 0.021 0.049 0.199 0.269 wave-induce oscillatory flow, until a solution is found in
which the vorticity tends uniformly to zero at large
Table 1. Wave conditions for the Caen tests distances from the boundary layer. This has been found
to give results that are in close agreement with second
Case H1 H2 H3 H4 order analytical solutions in appropriate conditions.
T (s) 2.00 3.00 3.50 5.00
5. CONCLUSIONS
H (m) 0.15 0.15 0.22 0.18
L (m) 5.21 8.69 10.36 15.24
Velocity measurements in the seabed boundary layer
β (m-1) 1174 958 887 742 beneath waves are compared with non-linear analytical
u2/u1 0.035 0.064 0.197 0.300 solutions. For the oscillatory flow, agreement is
generally very good, but it deteriorates as the harmonic
Table 2. Wave conditions for the Hanover tests components of the wave-induced flow become more
important. In all cases the mean flow at the outer edge of
3. DISCUSSION OF THE MEASUREMENTS the boundary layer is found to be considerably weaker
than the predictions. This is consistent with earlier
Figures 1 and 2 show the oscillatory flow through the observations [2] that the steady streaming around a
thickness of the boundary layer at eight phases equally circular cylinder beneath waves is weaker than might be
separated over one wave period. For the mildest cases expected from boundary layer theory.
(C1, C2, H1, H2) the data generally agree very well with
the second order theory [4], but in more non-linear 6. REFERENCES
conditions there are significant differences. These occur
when the amplitude of the second harmonic component [1] Batchelor, G.K., An Introduction to Fluid Dynamics
of the wave-induced flow becomes an appreciable Cambridge University Press, 1967.
proportion of that of the fundamental. [2] Chaplin, J.R., ‘Non–linear forces on a horizontal
cylinder beneath waves’ J. Fluid Mech. 147, 449–
The mean Eulerian velocities through the boundary layer 464, 1984.
are plotted in figure 3, and compared with the results of [3] Longuet-Higgins, M.S. ‘Mass transport in water
the same analytical solution. Here the agreement is waves’ Phil. Trans. Roy. Soc. Lond. A245, 535-581.
much less satisfactory, and in each case the steady [4] Sleath, J.F.A., ‘A second approximation to mass
streaming at the outer edge of the boundary layer is transport by water waves’ J. Marine Res. 30, 295-
considerably weaker than predicted. Some of the 304, 1972.
differences can be attributed to the very sensitive nature [5] Riley, N. ‘Stirring of a viscous fluid’ ZAMP 22,
of the measurements, but similar observations were made 645-653, 1971.
by Liu, Davis & Downing [9]. They put forward as [6] Beech, N.W., ‘Laser Doppler measurements in the
possible explanations the effects of wave reflections, oscillatory boundary layer beneath water waves’
secondary currents in the tank, and higher order non- DISA Information, No. 23, 6-10, 1978.
linearities. In the case of the present series C tests, wave [7] Sleath, J.F.A., ‘Measurements of mass transport
reflections were very small. We propose to explore over a rough bed’ 19th Conf. Coastal Engrg, 1149-
further the importance of non-linear contributions 1160, 1984.
through a full application of Sleath’s solution, and by a [8] Hwung, H.H & Lin, C., ‘The mass transport of wave
fully non-linear numerical model outlined below. propagating on a sloping bottom’ 22nd Conf.
Coastal Engrg, 544-556, 1990.
4. NUMERICAL COMPUTATIONS [9] Liu, P.L.-F., ‘Wave-induced boundary layer flows
above and in a permeable bed’ J. Fluid Mech. 325,
The chief difficulty in formulating a Navier Stokes code 195-218, 1996.
for this flow is in specifying the boundary conditions. If [10] Mouazé, D.P.M., ‘Experimental study of the
the computational domain covers only a thin region close boundary layer around a cylinder subject to waves’,
to the seabed, a method has to be found for setting the PhD thesis, University of Caen, 2001.
appropriate steady streaming velocity at its upper [11] Riley, N. Circular oscillations of a cylinder in a
boundary. Since the steady streaming is created by the viscous fluid ZAMP 29, 439-449, 1978.
 4 4

3 3
y (mm)

y (mm)
2 2

1 1
(a) (b)
0 0
-0.1 0 0.1 0.2 -0.1 0 0.1
4 4

3 3

y (mm)
y (mm)

2 2

1 1
(c) (d)
0 0
-0.2 0 0.2 -0.2 0 0.2
4 4

3 3
y (mm)

y (mm)

2 2

1 1
(e) (f )
0 0
-0.4 -0.2 0 0.2 0.4 -0.4 -0.2 0 0.2 0.4
4 4

3 3
y (mm)

y (mm)

2 2

1 1
(g) (h)
0 0
-0.2 0 0.2 0.4 -0.2 0 0.2 0.4
u (m/s) u (m/s)

Figure 1. The oscillatory flow in cases C1 (a,b), C2 (c,d), C3 (e,f) and C4 (g,h). The right-most line in each left hand
plot corresponds to the phase of the wave crest, and the other profiles (of which the last four are shown on the adjacent
plot on the right) are at equal phase intervals through one wave period. Measurements are shown as points, and the lines
represent second order theory [4].

4 4

3 3
y (mm)
y (mm)

2 2

1 1
(a) (b)
0 0
-0.2 -0.1 0 0.1 0.2 -0.2 -0.1 0 0.1 0.2

Figure 2(a) and (b). See caption below.

 4 4

3 3
y (mm)

y (mm)
2 2

1 1
(c) (d)
0 0
-0.2 0 0.2 -0.2 0 0.2
4 4

3 3
y (mm)

y (mm)
2 2

1 1
(e) (f)
0 0
-0.2 0 0.2 0.4 -0.2 0 0.2 0.4
4 4

3 3
y (mm)

y (mm)

2 2

1 1
(g) (h)
0 0
-0.2 0 0.2 0.4 -0.2 0 0.2 0.4
u (m/s) u (m/s)

Figure 2. As for figure 1, cases H1 to H4.

6
y (mm)

2 (d)
(a) (b) (c)
0
0 0.003 0.006 0 0.008 0.016 0 0.02 0.04 0 0.02 0.04
8

6
y (mm)

2 (g)
(e) (f) (h)
0
0 0.01 0.02 0.03 0 0.008 0.016 0 0.01 0.02 0 0.01 0.02
u(s) (m/s) u(s) (m/s) u(s) (m/s) u(s) (m/s)

Figure 3. The Eulerian mean flow for cases C1 (a) to C4 (d), and H1 (e) to H4 (h). Measurements are shown as points,
and the lines represent theory [1,4].