International Symposium on Water Management and Hydraulic Engineering

Ohrid/Macedonia, 1-5 September 2009

Paper: A66

Modelling of Wave Interaction with Submerged

Breakwater Using MIKE 21 BW
Dalibor Carevi, Marko Pri, Eva Ocvirk
University of Zagreb, Faculty of Civil Engineering, Water Research Department,
Zagreb, Croatia, e-mail: [email protected]

Abstract. Submerged breakwater are simple constructions which attenuate

wave energy in area beyound breakwaters. Appliance of such constructions can
achieve multiple benefits like coast erosion reduction, cheapest coast
constructions, overtopping reduction, force reduction etc. Also, the great
contribution from cultural point is preservation of old historical town sights and
landscapes. Submerged breakwater design requires determining of transmission
energy amount, which defines the protection level of aquatory or constructions
behind breakwater.
Irregular waves transmission over submerged breakwater was analyzed in
numerical model, MIKE 21 BW 1D. Results from this model were compared
with laboratory measurements in physical model tests (Johnson, 2005), also
with empirical models (DArgenmond, 1996, Seabrook, 1998, Buccino, 2007).
In model MIKE 21 BW 1D, wave breaking was calculated with modified
breaking angle parameters of surface roller concept. Transmission coefficients
obtained from numerical model have shown good agreement with measured
data and empirical model data. Also, transmission of wave period (mean and
significant) was analized with numerical model for different wave parameters
and geometrical parameters of brekawater.
Keywords: Submerged breakwater, Boussinesq model, surface roller, transmission
coefficient, wave period, wave height

1 Introduction
With development of wave numerical models, modelling of wave transmission over
submerged breakwaters becomes a challenge and the criterion of the quality of the
numerical model. Wave transformation at arival into shallow water has been so far
rather well described by various numeric models, while the process of wave

170
deformation from shallow into deep water, so-called deshoaling is still the subject of
scientific efforts in numerical modelling. Deshoaling effect includes complex nonlinear interactions, i.e. transition of wave energy into lower and higher harmonics.
Numerical model (MIKE21-BW, [1]) has been verified in the works [2], and [3]. For
verification, the results of laboratory tests conducted in [4] were used. Laboratory
tests were carried out on a submerged breakwater with mild slopes (ofshore slope
1:20, shore slope 1:10). Verification was done by comparison of measured and
calculated wave profiles, for a narrow band of wave parameters.
This paper will demonstrate the application of MIKE21-BW-1D for calculation of
the transmissmited wave parameters over the submerged breakwater. Possibility of
application would be presented in two steps:
Transmission coefficients of wave heights. a) Transmission coefficients of wave
heights calculated by MIKE21 model will be compared with the published transmitted
coefficients (Table 1) from laboratory tests, published in the paper [5] for irregular
waves (in Chapter 2.1). b) Also, comparison of transmission coefficients calculated
with MIKE21 and with empirical models will be shown, (DArgenmond, 1996, [6],
Seabrook, 1998, [7], Buccino, 2007, [8]). (in Chapter 6.2), for wave parameters from
Table 2.
Transmission coefficients of wave period. Transmission coefficients of wave
periods calculated by MIKE21 model will be presented (in Chapter 6.3).
Characteristic statistical and spectral wave periods are compared (in Chapter 6.4).
1.1 Laboratory Experiment
Laboratory experiments were done in the basin 9.7x12 m in the laboratory of Aalborg
University, Denmark, [1]. The investigation measured the wave setup, transmission
coefficients and flow around the breakwater. Out of all investigations, this paper will
use the results of measuring of transmission coefficients of wave heights.
Fig. 1 shows the bathymetry used for laboratory experiments. The breakwater has a
2 m opening in the middle. Two berm widths were used, wide 0.6 m, and narrow 0.2
m. Points from 1 to 21 mark the measuring probes.
Breakwaters were made with the core and armour layer of nominal particle size
(Dn50=45mm), and side slope 1:2. Behind the breakwater, the beach was made for
dissipation of waves, of quarry rock (Dn50=15mm), slope 1:5.
Experiments were carried out for the submerged breakwater, emerged breakwater
and zero freeboard breakwater. This paper will deal only with the data referring to the
submerged breakwater. Water depth on the breakwater berm is 7 cm, depth at toe is
27 cm, and depth at the wave generator is 43 cm. Tests were carried out for regular
and irregular waves. Irregular waves were generated as the JONSWAP spectrum with
direction spreading of 22.7. This paper uses only the results of transmission
coefficients for irregular waves.

171

Fig. 1. Cross-section and floor plan of basin with wide breakwate crest 0.6 m

Transmission coefficients (Table 1), are determined as the relation of transmitted

significant wave height, Hs0t, and incident significant wave height, Hs0i. Hs0t is
determined as the average of measured significant wave heights in gauges 19, 20 and
21, (Fig. 1). Hs0i is determined as the average of measured significant wave heights
in gauges 9, 10 and 11.
Table 1. Program of laboratory experiments and measured coefficients of transmission Kt. Hs0incident significant wave height, Tp-peak period, d-water depth on the toe, [1]
Test

Hs0

Tp

d/Lp

33
17
34
18
21
35
22
36

0.12
0.12
0.12
0.12
0.05
0.05
0.05
0.05

1.97
1.97
1.40
1.40
1.32
1.32
0.93
0.93

0.04
0.04
0.09
0.09
0.10
0.10
0.20
0.20

wave type
J
J
J
J
J
J
J
J

3D
3D
3D
3D
3D
3D
3D
3D

berm

Lp/Hs0

Kt-meas

wide
narrow
wide
narrow
narrow
wide
narrow
wide

50.03
50.03
25.17
25.17
50.34
50.34
24.99
24.99

0.52
0.62
0.58
0.69
0.81
0.73
0.76
0.74

172
1.2 Numerical Model MIKE21-BW
The numerical model is set up with identical conditions as the longitudinal cross
section of laboratory model (Fig. 2). Numerical model was enlarged by 20 times,
because calibration of parameters MIKE21-BW was made for realistic wave
conditions. In other words, bathymetry, wave heights and wave lengths were enlarged
by 20 times. Grid spacing used was Dx=1m, time step Dt=0.02s.The waves were
generated as the Jonswap spectrum, and the period of calculation is 10 min, for
achieving of stationary conditions.

Fig. 2. Bathymetry with narrow breakwater berm used for numeric model.
Table 2. Wave parameters for comparison 2).

Test
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

Hs [m]
3,0
3,0
2,4
2,4
2,4
2,4
2,4
2,4
2,1
2,1
2,1
2,1
2,1
2,1
1,5
1,5
1,5
1,5
1,5
1,5
1,1
1,1
1,1
1,1
1,1
1,1
0,8
0,8
0,8
0,8
0,8
0,8

Tp [s]
4,9
4,9
8,8
8,8
6,3
6,3
4,8
4,8
4,5
4,5
5,8
5,8
8,1
8,1
3,8
3,8
4,9
4,9
6,9
6,9
3,2
3,2
5,9
5,9
4,2
4,2
2,8
2,8
3,6
3,6
5,1
5,1

berm [m]
W IDE
NARRO
W IDE
NARRO
W IDE
NARRO
W IDE
NARRO
W IDE
NARRO
W IDE
NARRO
W IDE
NARRO
W IDE
NARRO
W IDE
NARRO
W IDE
W IDE
NARRO
W IDE
NARRO
W IDE
NARRO
W IDE
NARRO
W IDE
NARRO
W IDE
NARRO
W IDE

173
Bed friction is defined according to the theory presented in [9]. Bed roughness
parameter, kN, for calculation of the wave friction factor is calculated as kN=2.5Dn50.
The wave friction factor is limited to the maximum value fwmax =0.8, and the mean
velocity and the velocity at bottom were calculated according to the linear wave
theory. The Manning friction coefficient is determined according to the mentioned
theory, and with the given parameters, for water depth equal to mean height of the
breakwater, and applied as constant along the entire length of the breakwater.
Wave breaking in the MIKE21-BW model is calculated according to roller
concept. Standard set parameters of wave breaking were used, initial breaking angle
Fb=20; final breaking angle F0=10, roller form factor, 1.5, roller cellerity factor, 1.3,
half time, Tp/5). Also for comparison, the calculation was made with
recommendations for submerged breakwaters according to [2], Fb=14 i F0=7, others
remaining the same.
The sponge layer was placed o the left side of bathymetry, and on the right end.
For comparison 1b), described in introduction chapter, calculation was made in
MIKE21 according to wave parameters in Table 2. Wave parameters in Table 2 are
limited by stability of the model; i.e. at large wave heights and lengths in relation to
depth of water at the breakwater, the model becomes unstable. The results of the
numeric model according to Table 2 are compared with empirically obtained
transmission coefficients of wave height in Chapter 6.2.
1.3 Empirical Models
The empiric equations were obtained on the basis of laboratory tests of hydraulic
behaviour of the breakwater, and are used to calculate average transmission
coefficients of wave heights, Kt, for given wave parameters and geometric paramaters
of the breakwater.
Seabrook and Hall, 1998:
BF

+ 0.047

L p D 50a

H
F
1.09 si
K t = 1 exp 0.65
H
B
si

0.067 H si F
BD

50 a

(1)

where: F- freeboard [m], Hsi - significant wave height [m], B- berm [m], Lp- peak
wave lenght [m], Dn50 - nominal diameter of armour layer [m].
DAngremond, 1996:
K t = 0.4

B
F
+ 0.64
H si
H si

B
F
+ 0.51
H si
H si

K t = 0.35

0.31

0.65

1 e 0.5 ,

1 e 0.41 ,

B / Hi < 8

B / H i > 12

(2)

(3)

174
where: = tg /(

H si 0.5
) -Irribaren number. For values 8 B / H s i 12 , the values
Lp

of transmission coefficient are interpolated linearily.

Buccino, 2007:
Kt =

1
H
1.18 si
F

0.12

H
+ 0.33 si
F

for

1.5

F
2
H si

0.83

(4)

H si L p

B
K t = min(0.74; 0.62 0.17 ) 0.25 min 2.2;

Hsi L p

for F = 0
H

si

(5)

For values 0.83 F / H si 0 , the values of transmission coefficient of wave heights

are linearily interpolated.

2 Results
To enable comparing of results, the following statistic parameter was used:
Mean square error (MSE). It represents dispersion of data around the line of
absolute agreement.

( y y )

MSE =

(6)

where: y is actual value (Kt measured in laboratory or empirical model), y is

estimated value of y (Kt calculated by numeric model), and n is number of
comparated transmission coefficients.
2.1 Transmission Coefficients of Wave Heights-Comparison of MIKE21 and
Laboratory Measuraments
This paper compares transmission coefficients obtained by laboratory experiments for
Jonswap waves with directional dispersion (22.7), and transmission coefficients
obtained by 1D numeric model without directional dispersion. As wave breaking is
the dominant process influencing dissipation and transmission of wave energy, and
does not depend essentially on the incident wave angle, it is assumed that 1D numeric
model will describe the transmission of wave energy well enough. Also, wave
diffraction through the opening in the breakwater is neglected, as the gauging probes
19, 20 1nd 21 are deep in the shadow of the breakwater (Fig. 1).

175
The comparison will be shown between transmission coefficients obtained by
numeric model and by measuring in the laboratory (Fig. 3). Transmission coefficients
on the numeric model (Kt MIKE) are calcualted in the identical points as those in
laboratory experiment (Kt measur).
1.0
0.9

Kt MIKE

0.8
0.7
0.6
i=20, 0=10
0.5

i=14, 0=7

0.4
0.4

0.5

0.6

0.7
Kt measur

0.8

0.9

1.0

Fig. 3. Comparison of transmission coefficients obtained by numeric model Kt MIKE, and by

laboratory tests Kt measur, for different initial (Fi) and final (F0) wave breaking angles.
MSE(Fb=14;F0=7)=0.05; MSE(Fb=20;F0=10)=0.09

The results match better for applied breaking angles Fb=14 and F0=7, as
recommended in the paper [2]. Calculated mean square error is MSE(Fb=14;F0=7)=0.05.
For standard breaking angles (Fb=20, F0=10) numeric model overestimates the
transmission coefficients, and mean square deviation is MSE(Fb=20;F0=10)=0.09. When
smaller initial and final breaking angles are used, the waves start to break earlier than
in the case of standard parameters. This increases dissipation of wave energy for all
tests, and the largest influence is exerted on waves with lower wave heights. (Test 21,
35, 22 i 36). Matching of transmission coefficients is satisfactory for applied breaking
angles Fb=14 and F0=7.
2.2 Transmission Coefficients of Wave Heights-Comparison of MIKE21 and
Empirical Models
Numeric model was set up as described in Chapter 4. Due to reflection from the
breakwater, incident significant wave height, Hsi, is defined at sufficient distance
from the breakwater to avoid influence of reflection. The averaging zone from point
100 to 120, Fig. 4 (right), was defined where there is no influence of reflection even at
longest waves. In this section, the values of relevant wave heights were averaged. Hst
was obtained by averaging on the section from point 190 to 200.
Fig. 4 (left), presents the comparison between transmission coefficients obtained
by numerical model and by empirical equations for wave parameters given in Table 2.

176
Transmission coefficients were calculated like: Kt=Hst/Hsi. The best agreement of the
numerical model is that for DAngremond empirical equation, while according to the
other two equations, Seabrook and Buccino, the numeric model slightly
underestimates the results.
1.0
0.9

Kt empiric

0.8
0.7
0.6
0.5

Seabrook

0.4

D'Angremond
Buccini

0.3
0.2
0.2

0.3

0.4

0.5 0.6
Kt MIKE

0.7

0.8

0.9

1.0

Fig. 4. Left: Comparison of transmission coefficients obtained by empirical equations and by

numeric model. MSEd'Angrem=0.06, MSESeabrook=0.08, MSEBuccino.=0.08; MSE was calculated by
introducing into equation (6): y-result of empirical model (Kt empiric.), y -result of numeric
model (Kt MIKE). Right: Evolution of characteristical statistic, (Hs, Hmean), and spectral, (Hs0,
Hrms0), wave heights through numerical wave chanel, for wave conditions in Test17IRR, (Table
2).

2.3 Transmission Coefficients of Wave Periods - MIKE21 Results

It is usuall for calculation of run up and overtopping over the coastal structures and
reflection from perforated constructions to use periods of incoming waves. If any of
mentioned coastal structures are defended by submerged structures, it is important to
calculate transmitted period. Period of single wave which past over submerged
breakwater remain unchanged, because of continuity conservation, but it wave height
becomes smaller. When waves passing over submerged breakwater, wave breaking
occures. At greater wave heights, from wave time series, dissipation of wave energy
is greater then at smaller waves. In that situation smaller waves falls into the one
third of greater waves and change significant wave period. Because of that significant
period becomes smaller. Well known fenomen, triad interactions, cause transition of
wave energy from primar harmonics to heigher and lower harmonics. Because of
greater transition on heigher harmonics, wave field, influenced by those components
with lower periods, has more shorter periods than wave field before submerged
breakwater. In that situation wave periods Tmean, T2 and also Ts, becomes lower after
breakwater.
In Fig. 6 (left), it is shown the influence of submerged breakwater freeboard F in
ratio with wave length on period transmission coefficient. Transmission coefficients
are calculated with spectral (T1, T2) and statistical (Ts, Tmean) wave parameters like:

177

K t Ts =

Tsi
T
T
T
; K t Tmean = meani ; K t T1 = 1i ; K t T2 = 2i ;
Tst
Tmeant
T1t
T2 t

(7)

where: Ts is statistical significant wave period, Tmean is statistical mean wave period ,
T1 is spectral mean centroiod wave period, T1 is spectral mean zero-crossing wave
period.
Spectral periods were calculated with:
m
T1 = 0 ; T2 =
m1

m0
m2

(8)

where: m0, m1, and m2, are zero, first and second moment of spectral energy density
function.

Fig. 5. Left: Comparison of period transmission coefficients (Kt Ts, Kt Tmean, Kt T1, Kt T2)
obtained by MIKE21. Transmission coefficients are calculated with different spectral (T1, T2)
and statistical (Ts, Tmean) wave parameters. Wave lengths, (L1, L2, Ls, Lmean) are calculated in
corresponding to different periods. Right: Evolution of characteristical statistic and spectral
wave periods through numerical wave channel, for wave conditions in Test17IRR (Table 2).

At Fig. 6 (right), it is presented evolution of characteristical statistic and spectral

wave periods through numerical wave chanel, for wave conditions in Test17IRR
(Table 2).

178

Fig. 6. Left: Characteristic ratios of statistical and spectral wave periods in numeric wave
channel. Right: Characteristic ratios of statistical and spectral wave heights in numeric wave
channel.

2.4 Comparison of Statistical and Spectral Wave Periods - MIKE21 Results

MIKE21BW solves Boussinesq equation in time domain, where result of calculation
is surface elevation timeseries in each point of modeled space. Analisys of timeseries
could be made by spectral and statistical approach.
Firstly, for every Test from Table 2, should be calculated evolution of
characteristic wave parameters, like those presented in Fig. 4 (right), and Fig. 5
(right).
For example, Hs(x). Secondly, for every point, and every Test (Tbl. 2.), should be
calculated ratios of characteristic wave parameters. For example, Hs(x)/ Hs0(x).
Thirdly, for every point, should be calculated mean ratio for values from different
Tests (Table 2.). For example [Hs(x)/ Hs0(x)]mean. Result of those calculatins are
presented in Fig.6. Ratio, [Tmean(x)/T2(x)]mean, should be 1, becouse this parameters
present same period.
Variations occure near breaking and deshoaling zone on submerged breakwater.
Ratio, [Ts(x)/Tmean(x)]mean, should be ~1.1, [10], which is nearly satisfied. Ratio,
[Hmean(x)/Hs(x)]mean, sholud be ~0.63, [10], which is also satisfied. Ratio,
[Hs(x)/Hs0(x)]mean, sholud be ~1, where we can observe little underestimation of Hs(x)
in corresponding to Hs0(x) Ratio, [Hmean(x)/Hrms0(x)]mean, sholud be ~0.886, [10],
which is also satisfied.
All ratios of wave heights have deviation in breaking and deshoaling zone at
submerged breakwater.

179

3 Conclusions
In this paper, calibration of numerical model MIKE21-1D was conducted using
laboratory results.
Calibration was conducted by fitting of breaker angles (initial-Fi, final-F0).
Numerical results have shown good agreement with laboratory results for breaking
angles Fb=14 and F0=7.
- Validation of numerical model was conducted with empirical formulas.
- Results indicate that the greater part of numerical transmission coefficients falls
inside limits defined by +/-0.1 of average values of transmission coefficients
according to different authors.
Also, possibility of wave period transmssion calculation was presented and it
dependance of ratio freeboard/wave length.
Comparison of statistical and spectral wave parameters showed a good agreement
and consistency.

References
1. Madsen, P.A., Srensen, O.R.: a new form of the Boussinesq EQUATIONS WITH
IMPROVED LINEAR DISPERSION CHARACTERISTICS. PART 2. A SLOWLYVARYING BATHYMETRY; Coastal Engineering 18 (1992), 183-204.
2. Madsen, P.A., Srensen, O.R., Schffer, H.A.: SURF ZONE DYNAMICS SIMULATED
BY A BOUSSINESQ TYPE MODEL. PART I. MODEL DESCRIPTION AND CROSSSHORE MOTION OF REGULAR WAVES; Coastal Engineering 32 (1997), 255-287.
3. Bredmose, H., Schffer, H.A., Madsen, P.A.: Boussinesq evolution equations: numerical
efficiency, breaking and amplitude dispersion; Coastal Engineering 51 (2004), 1117-1142.
4. Beji, S., Battjes, J.A.: EXPERIMENTAL INVESTIGATION OF WAVE PROPAGATION
OVER A BAR; Coastal Engineering 19 (1993), 151-162.
5. Johnson, H. K., Karambas, T. V., Avgeris, I., Zanuttigh, B., Gozalez-Marco, D., Caceres, I.:
'MODELLING OF WAVES and currents around submerged breakwaters ''; Ocean
Engineering 52 (2005), 949-969.
6. DAngremond, K., van der Meer, de Jong, R.J.:WAVE TRANSSMISION AT LOWCRESTED STRUCTURES; ASCE, Proc. ICCE, Orlando,1996, Florida, 3305-3318.
7. Seabrook, S. R., Hall, K. R.:WAVE TRANSSMISION AT SUBMERGED RUBBLE
MOUND BREAKWATERS; Proc. 26TH Int. Conf. on Coast. Engineering, 1998, ASCE,
2000-2013.
8. Buccino, M., Calabrese, M.: CONCEPTUAL APPROACH FOR PREDICTION OF
WAVE TRANSMISSION AT LOW-CRESTED BREAKWATERS; Journal of waterway,
port, coastal, and ocean engineering , ASCE, May/June 2007.
9. Svendsen, I. A., Jonsson, I. G., HYDRODYNAMICS OF COASTAL REGIONS, (1980),
Technical University of Denmark, 220-229.
10.Goda, Y. (1988) Statistical variability of sea state parameters as a function of a wave
spectrum. Coastal Eng. in Jpn., JSCE 31, no. 1, 3952.

180
11.Madsen, P.A., Srensen, O.R.: a new form of the Boussinesq EQUATIONS WITH
IMPROVED LINEAR DISPERSION CHARACTERISTICS. PART 2. A SLOWLYVARYING BATHYMETRY; Coastal Engineering 18 (1992), 183-204.