Ocean Engineering 259 (2022) 111911

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

Wave forces on crown wall of rubble mound breakwater under swell waves
Xinyu Han a, Yunpeng Jiang b, Sheng Dong a, *
a
College of Engineering, Ocean University of China, Qingdao, 266100, China
b
Tianjin Research Institute for Water Transport Engineering, Tianjin, 300456, China

A R T I C L E I N F O A B S T R A C T

Keywords: Crown walls are often used to reduce costs and increase the top of the rubble mound breakwater. Swell period
Rubble mound breakwater waves often occur along the coast of South Asia, Africa, and South America and have attracted increasing
Crown wall attention. An experiment consisting of 744 tests was conducted to study the wave force on the crown wall under
Swell waves
swell waves. Pressure transducers of 100 and 1000 Hz were arranged on the crown wall. Through comparison,
Wave force
Sampling frequency
the acquisition frequency of 1000 Hz can capture greater impact pressure than 100 Hz. The effects of wave
Improved method height, water depth, wave period, slope, and berm width on the wave force were studied. Subsequently,
experimental data were compared with the Pedersen, Martin, Nørgaard, and Code of China Design methods, and
the applicability of these methods was analysed. A linear modified method was proposed to improve the original
Pedersen, Martin, and Nørgaard methods. The results indicated that the improved methods have significantly
smaller errors than those of the original methods. The improved Martin method was found to be the most
effective in estimating horizontal force. The improved Nørgaard method was the most accurate for determining
the uplifting force.

1. Introduction conservation design. Günback and Göcke (1984) proposed a method to

assess the wave pressure on a crown wall based on physical model tests
The breakwater is a common coastal defence structure. According to of the run-up process and its interaction with the structure. They
the section, it can be divided into vertical breakwater and mound considered that the maximum force was caused by the impact pressure
breakwater (Yuksel et al., 2020). According to the shape of axis line, it (Pm) and hydrostatic pressure (Ph), and the height of the run-up was used
can be divided into straight type (Liao et al., 2018) and arc type (Han to calculate Ph.
et al., 2021). A crown wall is often used on top of a rubble mound The wave force on a crown wall structure under the action of
breakwater to increase the height of the breakwater and enhance the irregular waves is stochastic and thus extremely complicated. Jensen
protection of the harbour. The use of a crown wall can reduce con­ (1984) studied the influence of variations in wave height (Hs), wave
sumables. In addition, infrastructure and equipment can be installed on period (Tp), and sea level on the maximum wave force per metre wall for
this part of the structure. Under extreme wave conditions, the crown 1000 waves (Fh0.1%). Jensen (1984) concluded that with an increase in
wall experiences a horizontal force and an uplifting force, which may the wave period, the wave force increases, and the influence of sea level
cause the crown wall to slip and turn. To design a crown wall, it is variation can be expressed through the crest freeboard Hs/Ac (Ac refers
necessary to accurately predict the wave load on the crown wall. to Table 1). This correlation between the crown wall crest freeboard and
Several studies have been performed to predict the wave forces on pressure was later used by more recent methods (Molines, 2016; Molines
crown walls on rubble mound breakwaters. Irivarren and Nogales et al., 2018). Bradbury et al. (1988) investigated the effect of the slope
(1954) first defined the forces exerted on a crown wall on top of a rubble on the loads over a superstructure by performing experiments with five
mound breakwater. However, their method is only applicable to the different rock slope configurations under irregular wave attacks. They
specific geometry recommended by them and it is based on the also fitted their coefficients to Jensen’s (1984) equations, obtaining
assumption that the waves are broken before arriving at the super­ good results. However, their approach is only useful for geometries
structure (Valdecantos et al., 2013). Additionally, this method does not similar to those tested. For cases that do not match the indicated
consider the influence of the berm width and wave height, resulting in a cross-sections, new coefficients need to be obtained (Valdecantos et al.,

* Corresponding author.
E-mail address: [email protected] (S. Dong).

https://doi.org/10.1016/j.oceaneng.2022.111911
Received 8 March 2022; Received in revised form 15 June 2022; Accepted 3 July 2022
Available online 12 July 2022
0029-8018/© 2022 Elsevier Ltd. All rights reserved.
X. Han et al. Ocean Engineering 259 (2022) 111911

Table 1 the applicable ranges of the methods proposed by Irivarren and Nogales
Ranges of cross-section parameters and wave conditions. (1954), Günback and Göcke (1984), Martin et al. (1999), Bradbury
Parameter Symbol Value/Range et al., (1988), Pedersen and Burcharth (1992), and Berenguer and
Baonza (2006). Bekker et al. (2018) studied the influence of the free­
Slope angle cot α 1.5, 2
Width of crown wall (m) F 0.25 board on the uplifting force of the crown wall and found that the method
Crown wall freeboard (m) Rc 0.28, 0.34, 0.40, 0.46 of Pedersen (1996) and Nørgaard et al. (2013) is conservative. They
Bottom of crown wall to still water d1 − 0.09, − 0.03, 0.03, 0.09 proposed a reduction coefficient γv to correct the prediction formula.
level (SWL) (m) However, the range of applications for γ v is small (Sop = 0.01; 0.04).
SWL to the crest level of the armour Ac 0, 0.06, 0.12, 0.18
layer (m)
On the west coast of South America, and the east and west coasts of
Berm width (m) B 0.07, 0.14, 0.21, 0.28 Africa, East Asia, South Asia, waves with a period of 10–20 s are com­
Ratio crest level of crest wall and Rc/Ac ≥1.5556 mon, which belongs to swell waves. In recent years, with the develop­
armour (− ) ment of coastal constructions, swell waves have received increasing
Non-dimensional crest level (− ) Rc/H 1.4–11.5
attention. Jiang et al. (2021) studied the wave force on a crown wall for
Non-dimensional protruding part of (Rc-Ac)/ 1.4–7
crest wall (− ) H swell waves using an actual engineering case. The wave force was found
Non-dimensional level of base plate, d1/H − 2.25–2.25 to be considerably greater than the results presented by the Coastal
w.r.t. SWL (− ) Engineering Manual (U.S. Army Corps of Engineers, 2002) and Code of
Thickness of the armour layer (m) da 0.07 China Design (CCCC First Harbour Consultants CO., LTD, 2015). How­
Water depth (m) d 0.63, 0.57, 0.51, 0.45
Wave period (s) T 1.10, 1.46, 1.83, 2.19, 2.56,
ever, few studies on the hydrodynamics of crown wall under swell waves
2.92, 3.29 exist. In the coastal area of the world, the wave period is generally less
Incident wave height at toe (m) H 0.04, 0.08, 0.12, 0.16, 0.20 than 20 s (Amrutha et al., 2017; George and Kumar, 2020; Ahn, 2021).
The ratio of wave height and water H/d 0.0635–0.4444 Therefore, the wave forces on crown wall under swell waves with 10–20
depth (− )
s is studied in this work.
Wave steepness (s = H/L) s 0.0051–0.1153
Ratio of wave length and depth d/L 0.0671–0.3452 In this study, the wave force on the crown wall of a rubble mound
Surf-similarity parameter (ξ = tanα/ ξ 1.5110–13.6083 breakwater under swell waves was investigated using a series of
s0.5) experimental tests. The remainder of this paper is organised as follows.
In Section 2, the physical model test in a wave flume is presented. In
Section 3, the influence of the sampling frequency on the wave force is
2013).
discussed. In Section 4, the influence of the wave condition parameters
Pedersen (1996) performed 373 physical tests and developed a
and structural parameters on the wave force is studied. In Section 5, the
semi-empirical method based on a theoretical approach similar to that of
methods proposed by Pedersen (1996), Martin et al. (1999), Nørgaard
Günback and Göcke (1984), in which the wave force on a crown wall is
et al. (2013), and Code of China Design are introduced. In Section 6, a
linked to the fictitious run-up height, combined with a parametric
comparison of the experimental data and the results of the calculation
analysis of the results of small-scale physical model tests under the ac­
methods is presented. In Section 7, improved methods are proposed and
tion of non-breaking irregular waves. The van der Meer (1988) run-up
compared with the original methods. Finally, conclusions are drawn.
formula was used in this study. The experimental tests were limited to
deep and intermediate water wave conditions (0.5 < Hm0/Ac < 1.5). The
2. Physical model tests
wavelength in Pedersen’s formula represents the deep-water wave­
2
length associated with Tm (L0m = gTm /2π) (Molines, 2016). Nørgaard
Physical model tests were performed in a wave flume at the Tianjin
et al. (2013) indicated that Philips P13-OEM pressure transducers used
Research Institute for Water Transport Engineering of the Ministry of
by Pedersen (1996) can be influenced by dynamic amplification;
Transport of China. The flume was 68 m long, 1 m wide, and 1.3 m high,
amplification predominates only when the wave impacts the unpro­
as shown in Fig. 1. The left side of the flume was equipped with a piston-
tected part of the crown wall. Their tests used Drück PMP UNIK pressure
type wavemaker, which could generate regular and irregular waves.
transducers, which are uninfluenced by dynamic amplification, and they
Sponges were set at the right end of the flume to absorb waves and
proposed an update formula with a wider range of applications.
prevent wave reflection. The toe of the breakwater was 38 m from the
Nørgaard et al. (2013) calculated the run-up using the method proposed
wave-maker piston. A physical model of the breakwater is depicted in
by Kobayashi et al. (2008), and modified the moment empirical formula.
Fig. 2. Fig. 3 shows the tested cross-section with the definition of the
Martin et al. (1999) proposed a method to calculate the horizontal and
most important parameters. The involved parameters in the physical
uplifting forces on the crown wall under regular waves. This method was
model tests are listed in Table 1. The crown wall was made of plastic,
compared with the results from laboratory tests conducted by Burcharth
thus avoiding the problem of water swelling in the case of wooden
et al. (1995) and Jensen (1984), revealing a fairly good agreement.
materials. The thickness of the plastic plate was 2 cm, and there was no
Similar to that of Günback and Göcke (1984), Martin’s formula assumes
deformation under the action of waves. To measure the wave force on
an impact pressure and a hydrostatic pressure. The methods proposed by
the crown wall, pressure transducers were arranged on the crown wall,
Losada and Gimenez-Curto (1981) and Losada (1992) were used to
as shown in Fig. 4. Fifteen pressure transducers were placed on the
calculate the run-up. Berenguer and Baonza (2006) introduced a for­
vertical surface to measure the horizontal force. Seven pressure trans­
mula to calculate the maximum forces on the crown wall for
ducers were placed on the bottom surface to measure the uplifting force.
non-breaking waves based on laboratory tests. This formula considers
The height and width of the prototype crown wall are 10.5 m and 7.5 m.
the influence of the foundation level on the wave loads on the crown
The experimental parameters were based on Froude’s law of similarity
wall.
using prototype parameters. The scale of space was 1:30, and the scale of
Camus and Flores (2004) evaluated the formulas presented by
time was 1:5.48. The wave conditions are regular waves. Each case is
Günback and Göcke (1984), Jensen (1984), Bradbury et al., (1988),
only run 20 wave cycles to guarantee that the measurements are not
Pedersen (1996), and Martin et al. (1999). They concluded that Peder­
affected by the secondary wave reflection from the wavemaker. The
sen method is the most accurate for the maximum horizontal force.
periods of the prototype waves corresponding to the test periods were 6,
Braña and Guillé (2004) used the Alicante, Ferrol, and Coruña break­
8, 10, 12, 14, 16, and 18 s. Therefore, T = 6 s and 8 s are short-period
waters as cross-sections and compared the empirical formulas and test
waves, and T = 10 s, 12 s, 14 s, 16 s, and 18 s are swell waves. Before
data. Pedersen’s method was found to be the most reliable to estimate
the tests, the wave condition before the breakwater should be deter­
the wave force on the crown wall. Valdecantos et al. (2013) summarised
mined. Three wave gauges were arranged at the location of the

2
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 1. Layout of the model test in 2D wave flume.

Fig. 2. Physical model.

Fig. 3. Scheme of the main parameters of the rubble mound breakwater with crown wall.

breakwater to measure the wave conditions to ensure that the waves in of the pulsation pressure. For example, the maximum impact pressure at
front of the breakwater meet the design requirements. P4 in Fig. 6 was approximately eight times the pulsating pressure. Under
the action of a regular wave, each wave produces impact pressure;
3. Comparison of the pressure measured by different sampling however, the peak value of the impact pressure of each wave is not
frequencies necessarily equal. Both sampling frequencies of 100 and 1000 Hz can
capture the peak value of the impact pressure. At positions P1–P6, the
The impact wave pressure may be generated when the waves attack impact pressure measured by the 1000 Hz sampling frequency is slightly
the crown wall. To measure the impact force, two pressure transducers higher than that measured by the 100 Hz sampling frequency. At posi­
with different acquisition frequencies were used: 100 and 1000 Hz. In tions P7–P9, the impact pressure measured by the 100 Hz sampling rate
the physical test, two rows of pressure measurement points were ar­ is slightly higher than that measured by the 1000 Hz sampling rate. This
ranged as shown in Fig. 5, where the sampling frequency of the one on difference may be attributed to the difference in the cross-sections of the
the left was 1000 Hz and that of the one on the right was 100 Hz. rows of the 100 and 1000 Hz pressure sensors. In general, the impact
Figs. 6–12 depict the comparison of the time series of the 100 and pressure measured at a sampling frequency of 100 Hz is basically the
1000 Hz pressure sensors under d = 0.63 m, H = 0.16 m, T = 1.10, 1.46, same as that measured at a sampling frequency of 1000 Hz.
1.83, 2.19, 2.56, 2.92, and 3.29 s, respectively. In the case of a short When the period increases from 1.46 to 1.83 s, it can be seen from
period (Figs. 6 and 7), the impact pressure appears on the crown wall. Fig. 8 that the wave impact pressure decreases significantly, indicating
The peak value of the impact pressure was significantly higher than that that the phenomenon of wave impact on the crown wall become less

3
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 4. Arrangement of the pressure transducers on the crown wall (unit: cm). Fig. 5. 100 and 1000 Hz pressure transducer arrangement.

significant. This is because as the period increases, the wavelength also high-frequency signal at a sampling frequency of 1000 Hz is clearer than
increases. When the waves propagate to the breakwater, the wave that at a sampling frequency of 100 Hz. Fig. 15 depicts the magnitude
deformation is not as large as in the shorter period. Therefore, the wave scalograms of P1 when d = 0.63 m, H = 0.16 m, and T = 3.29 s. The
does not violently attack the crown wall. With an increase in the wave magnitude is smaller at high frequencies, indicating that the impact
period (Figs. 9–12), the wave impact phenomenon disappears. The pressure is zero.
pressure–time series of each measurement point represents the pulsating According to the pressure of each measurement point, the horizontal
pressure. The results measured at a sampling frequency of 1000 Hz are and uplifting wave forces on the crown wall was calculated as follows:
consistent with those measured at a sampling frequency of 100 Hz. ∑
14
Then, the pressure peaks measured by the pressure sensor at 100 Hz FH = 0.04p1 + 0.02 pi + 0.05p15 (1)
sampling frequency and 1000 Hz sampling frequency under the same i=2

regular wave are compared, as shown in Fig. 13. The water depth is 0.63
m and wave height is 0.16 m. It can be seen that the peak pressure of ∑
21
FU = 0.45p16 + 0.03pi + 0.55p22 (2)
1000 Hz is higher than that of 100 Hz at T = 1.10 s and 1.46 s. Because of i=17
shorter period, the waves produce impact pressure on the crown wall,
By selecting a stable wave force time series, the maximum wave force
and 1000 Hz sampling frequency can better capture the peak value of
of each regular wave is calculated, and then the average value is taken as
impact pressure. With the increase of the period, it can be seen that the
the wave force under each case.
pressure peaks measured at 1000 Hz and 100 Hz are similar.
Fig. 16 depicts a comparison of the horizontal wave force on the
Wavelet transform can provide a ‘time-frequency’ window that
crown wall using the pressure data collected by the 100 Hz and 1000 Hz
changes with frequency. This is an ideal tool for signal time-frequency
sensors. The wave force of the 100 Hz sampling frequency is the same as
analysis. This can fully highlight the characteristics of the signal
that of the 1000 Hz sampling frequency under the same wave conditions
through a transformation. To further analyse the influence of the 100
and structure parameters. This shows that a sampling frequency of 100
and 1000 Hz sampling frequencies on the pressure signal, these were
Hz is sufficient to satisfy the measurement requirements of the wave
analysed using the Morlet wavelet transform. Fig. 14 depicts the
force on the crown wall.
magnitude scalograms of P1 when d = 0.63 m, H = 0.16 m, and T = 1.46
s. The magnitude scalograms of the 100 Hz sampling frequency are
4. Influence of structure and wave parameters on wave force
similar to those of the 1000 Hz sampling frequency, while the charac­
teristics of the pulsating and impact pressures differ. The pulsating
4.1. Influence of slope
pressure component is of low frequency and high magnitude, whereas
the impact pressure component has a high frequency and low magni­
In the experimental tests, two slopes of 1:1.5 and 1:2 were used.
tude. With an increase in frequency, the band of the magnitude becomes
Fig. 17 shows a comparison of the maximum horizontal forces measured
smaller. For a more detailed comparison, The same area is selected in the
at slopes of 1:1.5 and 1:2. All graphs in Fig. 17 show that the wave force
spectrum of 100 Hz and 1000 Hz. The characteristic description of the

4
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 6. Time series of P1–P9 of 100 and 1000 Hz sampling rates under d = 0.63 m, H = 0.16 m, and T = 1.10 s.

Fig. 7. Time series of P1–P9 of 100 and 1000 Hz sampling rates under d = 0.63 m, H = 0.16 m, and T = 1.46 s.

of slope 1:1.5 is similar to that of slope 1:2 under different berm widths. 4.2. Influence of wave height and water depth on horizontal wave force
Fig. 18 depicts a comparison of the maximum uplifting wave force
measured for slopes 1:1.5 and 1:2. The uplifting force of slope 1:1.5 is Fig. 19 depicts the change in the non-dimensional horizontal wave
larger than that of slope 1:2 under the berm widths 0.07 and 0.14 m. FH
force ρgH H FH
2 with the non-dimensional wave height R − A . ρgH2 decreases
c c
However, the uplifting force of slope 1:1.5 is slightly smaller than that of rapidly with an increase in Rc −HAc under d1 = +0.09 m. At d1 = +0.03 m,
slope 1:2 under the berm widths 0.21 and 0.28 m. This indicates that FH
under different berm widths, the variation in the uplifting force with ρgH2 decreases slightly with an increase in Rc −HAc . When d1 = − 0.03 m and
H FH
slope differs. − 0.09 m, with an increase in Rc − Ac , ρgH2 demonstrates an increasing
trend. As can be observed in Fig. 19, the variation trend of the non-

5
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 8. Time series of P1–P9 of the 100 and 1000 Hz sampling rates under d = 0.63 m, H = 0.16 m, and T = 1.83 s.

Fig. 9. Time series of P1–P9 of the 100 and 1000 Hz sampling rates under d = 0.63 m, H = 0.16 m, and T = 2.19 s.

dimensional wave force with non-dimensional wave height at different 4.3. Influence of wave period on horizontal wave force
water depths is consistent under different wave periods and berm
FH H FH
widths. When d1 is positive, ρgH 2 decreases with an increase in R − A .
c c
Fig. 20 shows that the non-dimensional horizontal wave force ρgH 2

FH H
√̅̅̅̅̅̅̅̅̅
When d1 is negative, ρgH 2 increases with an increase in R − A . In addition,
c c
changes with the non-dimensional wave period T g/H. When the water
through the comparison of wave force under different d1, the wave force depth and wave height are small, the wave force on the crown wall is
decreases with a decrease in water depth. zero because the waves do not impact the crown wall. Fig. 20 depicts the
conditions under which waves of different periods can impact the crown
wall under the same wave height and water depth. Under the same wave
height and water depth, the wave force increased as the wave period

6
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 10. Time series of P1–P9 of the 100 and 1000 Hz sampling rates under d = 0.63 m, H = 0.16 m, and T = 2.56 s.

Fig. 11. Time series of P1–P9 of the 100 and 1000 Hz sampling rates under d = 0.63 m, H = 0.16 m, and T = 2.92 s.

increased. When d = 0.63 m, in the case of a short period (T = 1.10 and period wave force is 2–3 times the shorter-period wave force. When the
1.46 s), a large wave force may be generated owing to the wave impact static water level is lower than the bottom of the crown wall, the longer-
on the crown wall. Under d = 0.57, 0.51, and 0.45 m, the impact force period wave force is 3–7 times the shorter-period wave force. Overall, in
disappears under short wave periods. With an increase in the wave the range of swell waves, the wave force increases with an increase in
period, the wave force gradually increases; however, the rate of increase the wave period. When the static water level submerges the armour
gradually decreases. The wave period has a significant influence on the blocks, the short-period waves impact the crown wall and produce a
wave force on the crown wall. When d = 0.63 m, B = 0.21 m, and 0.28 large wave force.
m, the wave force of waves with longer periods is approximately 1–3
times that of waves with shorter periods. When d = 0.53 m, the longer-

7
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 12. Time series of P1–P9 of the 100 and 1000 Hz sampling rates under d = 0.63 m, H = 0.16 m, and T = 3.29 s.

Fig. 13. Pressure peaks comparison of 100 Hz and 1000 Hz under different wave periods.

4.4. Influence of berm on horizontal wave force areas wherein swell waves often occur, the berm width can be reduced
without increasing the wave force on the crown wall, and the con­
Figs. 21 and 22 depict the changes in the non-dimensional horizontal struction cost can be reduced simultaneously.
FH
wave force ρgH 2 with the non-dimensional berm width B/H under H =

0.16 and 0.20 m, respectively. In the case of a short period (T = 1.10 and 5. Methods for determining the wave force on the crown wall
1.46 s), the horizontal wave force decreases significantly with an in­
crease in the berm width. In the scope of swell waves (T = 1.83, 2.19, 5.1. Pedersen method
2.56, 2.92, and 3.29 s), the berm width has little effect on the horizontal
wave force. When the waves propagate to the berm, the volume of water Pedersen (1996) assumed a vertical pressure distribution, as shown
above the slope increases with an increase in the wave period. The ar­ in Fig. 23 (right). The definition of the relevant parameters during the
mour units on the berm have a blocking effect on the waves, with a calculation is shown in Fig. 23 (left). Pedersen’s (1996) crown wall wave
significant effect on short-period waves. However, the blocking effect of force calculation method is as follows:
the berm armour blocks on swell waves is insignificant. Therefore, in sea

8
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 14. Comparison of the magnitude scalograms of P1 under d = 0.63 m, H = 0.16 m, and T = 1.46 s.

Fig. 15. Comparison of the magnitude scalograms of P1 under d = 0.63 m, H = 0.16 m, and T = 3.29 s.

Fig. 16. Horizontal wave force comparison at a sampling frequency of 100 and 1000 Hz.

Fig. 17. Horizontal wave force comparison of slopes 1:1.5 and 1:2.

9
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 18. Uplifting wave force comparison of slopes 1:1.5 and 1:2.

{
1.12H1/3 ξm0 ξm0 ≤ 1.5 tan α tan α
Ru,0.1% = ξm0 = √̅̅̅̅̅̅̅̅̅̅̅̅ (3) ξ0 = √̅̅̅̅̅̅̅̅̅ (13)
1.34H1/3 ξ0.55
m0 ξm0 > 1.5 2π H1/3 2π H
g T 2m g T 2m

[ ]
tan(α − 15∘ ) Ru ( )
y = (Ru − Ac ) 1 − (4) = Au 1 − e B u ξ 0 (14)
tan α H
[y ] ⎧ ( )
yeff = min ; fc (5) ⎨ H 1 − Ac
⎪
Ac < Ru
2 S0 = Ru (15)
⎪
⎩
√̅̅̅̅̅̅̅ 0 Ac ≥ Ru
Lm0
FHu,0.1% = a pm yeff b (6)
B ( )2
Ru
√̅̅̅̅̅̅̅ ca = 2.9 cos α (16)
1 Lm0 H
FHl,0.1% = a pm Vhprot (7)
2 B
PS0 = ca ρw gS0 (17)
( )
pm = gρw Ru,0.1% − Ac (8) ( )
B
⎧ λ = 0.8 exp − 10.9 (18)
L
⎨ V2 V2 < V1
V = V1 (9) {
⎩ PS0 z > Ac
1 V2 ≥ V1 Pd (z) = (19)
λPS0 w f < z < Ac
√̅̅̅̅̅̅̅( )
Lm0 1 ( )
FH,0.1% = FHu,0.1% + FHl,0.1% = a pm yeff b + pm Vhprot (10) m = e1 exp e3 (H/L − e2 )2 (20)
B 2
( ) (21)
MH,0.1% = c hprot + yeff FH,0.1% (11) Pr (z) = mρw g(S0 + Ac − z) wf < z < Ac + S0
( )
Pb,0.1% = eVpm (12) pre = λpS0 + pr wf (22)

The definition of relevant parameters in Eqs. 3–12 is shown in pra = 0 + pra,r (23)
Fig. 23. Lm0 = gT2 /2π in Eqs. (6) and (7). The parameters a and b in Eq.

⎧ ( )
⎪
⎪ ( ) Ac + S0 − wc wc − wf
⎪
⎨ λ⋅PS0 ⋅ Ac − wf + PS0 (wc − Ac ) + Pr ⋅ 1 + A − w + S ⋅ 2 Ac − wf + S0 > wc − wf
(24)
c f 0
FH =
⎪
⎪ ( ) A − wf + S0
⎪
⎩ λ⋅PS0 ⋅ Ac − wf + PS0 ⋅S0 + Pr ⋅ c Ac − wf + S0 ≤ wc − wf
2

(6) are 0.21 and 1.6, respectively. c = 0.55 in Eq. (11) and e = 1 in Eq.
(12). Variables Au and Bu in Eq. (14) are determined by the porosity of the
armour layer. Martin et al. (1999) provided reference values for Au and
5.2. Martin method Bu for some armour units (cubic block, quadrapod, and tetrapod). a, b
and c in Eq. (20) depends on the armour unit number (B/le) on the berm,
The method proposed by Martin et al. (1999), referred to as the as presented in Table 2.
Martin method, was performed on a series of regular wave tests based on
the Gijόn breakwater section. Martin et al. (1999) divided the wave
force into two parts, namely dynamic and reflecting pressure, as shown 5.3. Nørgaard method
in Fig. 24 (right). The definition of the relevant parameters during the
calculation is shown in Fig. 24 (left). Nørgaard et al. (2013) improved the Pedersen (1996) method to
cover loads in both deep and shallow water using 162 new physical
model tests. The calculation process is shown in Eqs. 25–29.

10
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 19. Variation of non-dimensional horizontal force respect to non-dimensional wave height.

11
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 20. Variation in non-dimensional horizontal force with non-dimensional wave period.

Fig. 21. Change in non-dimensional horizontal wave force with non-dimensional berm width when H = 0.16 m.

{ √̅̅̅̅̅̅̅
0.603H0.1% ξm0 ξm0 ≤ 1.5
Ru,0.1% = (25) FHl,0.1% = 0.21⋅
1 Lm0
⋅pm ⋅V⋅hport (27)
0.722H0.1% ξ0.55
m ξm0 > 1.5 2 B
√̅̅̅̅̅̅̅
Lm0 FH,0.1%,mod = FHu,0.1% + FHl,0.1% (28)
FHu,0.1% = 0.21 ⋅pm ⋅yeff (26)
B

12
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 22. Change in non-dimensional horizontal wave force with non-dimensional berm width when H = 0.20 m.

Fig. 23. Calculation diagram of Pedersen (1996) and vertical distribution on the crown wall.

Fig. 24. Calculation diagram of Martin et al. (1999) and vertical distribution on the crown wall.

5.4. Code of China design method

Table 2
Coefficients of e1, e2, and e3.
The Code of Hydrology for Harbour and Waterway (JTS 145–2015)
armour units on the berm e1 e2 e3 (CCCC First Harbour Consultants CO. LTD, 2015) stipulates a method to
1 0.446 0.068 259.0 calculate the wave force on the crown wall under the condition of no
2 0.362 0.069 357.1 shielding. This method assumes that the horizontal pressure distribution
3 0.296 0.073 383.1 is uniform and that the uplifting pressure distribution is triangular, as
shown in Fig. 25.
(
1
)
1 The average pressure p on the crown wall during the wave crest is
MH,0.1%,mod = hprot + ⋅ yeff ⋅ f1 FHu,0.1% + ⋅ hprot ⋅ FHl,0.1% ⋅f2 (29)
2 2 p = 0.24γHKp (30)

where f1 = 0.95 and f2 = 0.40. The parameters of Eqs. 25–29 refer to where γ is the specific gravity of water. H is the wave height. Kp is the
Fig. 23. mean pressure coefficient related to ξ and L/H.

13
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 25. Scheme of the China design code method for determining the wave force on the crown wall.

( )( )2πH/L
d1 d (34)
ξ= (31) P = p(d1 + z)
d H
The uplifting wave force is
where d1 is the water depth in the front of the crown wall. If the still
Bp
water level is lower than the bottom of the crown wall, d1 is negative; Pu = μ (35)
2
otherwise, d1 is positive. The maximum limit of ξ is ξb:
( ) where μ is the reduction factor, which is recommended to be 0.7. The
H
ξb = 3.29 + 0.043 (32) coefficients of Kp and Kz are selected according to Fig. 26.
L
The height of horizontal pressure distribution d1+z is 6. Comparison of experimental data and wave force methods
( )
2πd
d1 + z = H tanh Kz (33) 6.1. Comparison of experimental data with Martin method
L
In the Martin method, the influence of the slope is considered to be
where z is the vertical distance between the top of the pressure distri­
the Irivarren number. Therefore, the Martin method was used to
bution and still water level. Kz is the action height coefficient related to ξ
calculate the wave force on the crown wall with slopes of 1:1.5 and 1:2.
and L/H. Therefore, the horizontal wave force per unit length is
A comparison of the experimental data and calculation results is shown
in Figs. 27 and 28. In Eq. (20), Martin et al. (1999) only provided the
values of a, b, and c for one, two, and three rows of armour units on the
berm. Therefore, according to the interpolation method, the values of a
= 0.229, b = 0.088, c = 283.1 in the four rows on the berm can be
obtained. Hence, we compared the data and results under berm widths
of 0.07, 0.14, and 0.21 m. As seen in Fig. 27, the calculation results
obtained using the Martin method are significantly higher than the test
results when the bottom of the crown wall is submerged. Under different
berm widths, the results obtained using the Martin method are roughly
multiples of the test results. This multiple relationship is discussed in the
next section. Under the bottom of the non-submerged crown wall, the
calculation results are close to the experimental data. When d1 = − 0.03
m, the horizontal wave force calculated using the Martin method is
slightly larger than the test results. When d1 = − 0.09 m, the results
obtained using the Martin method are in good agreement with the test
results. Martin et al. (1999) indicated that ‘the application of the new
method requires wave breaking on the armour layer; that is, only broken
waves will reach the crown wall’. From the experimental phenomenon, the
waves break before reaching the crown wall at lower water depths, as
shown in Fig. 28. Therefore, the Martin method can better predict the
wave force when d1 = − 0.09 m.
Fig. 29 depicts the comparison of the test results of the uplifting force
and the results obtained using the Martin method. Regardless of whether
the bottom of the crown wall is submerged, the results obtained using
the Martin method are greater than the experimental results. When d1 is
positive and B = 0.07 m, the results obtained using the Martin method
are multiples of the test results. This multiple differs under different
berm widths and submerged depths. As the berm width increases, this
multiple relationship becomes less significant. When d1 is negative, the
Martin method cannot be applied to estimate the uplifting force.

Fig. 26. Kz and Kp curves under different values of L/H.

14
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 27. Comparison of the horizontal force of the experimental data and the results of the Martin method.

Fig. 28. Wave breaking on the slope.

Fig. 29. Comparison of the uplifting force of the experimental data and the results obtained using the Martin method.

6.2. Comparison of the experimental data with the CCD method wave force of the crown wall is small. This is consistent with the com­
parison results of the wave force for slopes of 1:1.5 and 1:2. Therefore,
According to the calculation method of CCD, it is believed that the slope is not considered in Eqs. 30–35. The CCD results were
within the slope of practice engineering, the impact of the slope on the compared with the test results at a slope of 1:1.5, as shown in Figs. 30

15
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 30. Comparison of the horizontal wave force of the experimental data with the CCD method.

Fig. 31. Comparison of the uplifting wave force of the experimental data with the CCD method.

and 31. Fig. 30 depicts the comparison results of the horizontal wave calculating the irregular wave force is used to calculate the wave force
force. When the bottom of the crown wall is submerged, the test results under regular waves. The regular wave height is considered as H0.1% of
are consistent with the CCD results. When the bottom of the crown wall the irregular waves, and the period of regular waves is considered as the
is non-submerged, the test wave forces are greater than the calculation mean period of the irregular waves. Because H1/3 is used in Eq. (3), Eq.
results in certain cases. Fig. 31 shows the comparison results of the (36) is used to calculate the mean wave height H according to H0.1%.
uplifting force. When the bottom of the crown wall is submerged, the H13% is calculated using H. For irregular waves, H13% is generally
CCD results are consistent with the experimental data when B = 0.07 m, considered to be equal to H1/3. Eq. (3) was proposed by van der Meer
and the CCD results are slightly larger than the experimental data when and Stam, 1992 and is only valid for relatively deep water and Ru,0.1%,
B = 0.14, 0.21, and 0.28 m. When the bottom of the crown wall is non- max = 2.58H1/3.
submerged, the uplifting wave force calculated using the CCD is invalid.
This indicates that the CCD method is suitable for cases wherein the [ ( ) ]1− H/d
4 1 H 2

crown wall is submerged; however, it is not applicable in cases wherein HF = H − 1 + √̅̅̅̅̅ ln F (36)
π 2π d
the crown wall is non-submerged.
Fig. 32 depicts the comparison results of the horizontal wave force.
6.3. Comparison of the experimental data with the Pedersen method When the bottom of the crown wall is submerged and B = 0.07 and 0.14
m, the results obtained using the Pedersen method are greater than the
Because the Pedersen and Nørgaard methods are used to calculate experimental data. When B = 0.21 and 0.28 m, the results obtained
the F0.1% of irregular waves, an equivalent relation between regular and using the Pedersen method are in good agreement with the experimental
irregular wave forces is established. Therefore, the method of results. When the bottom of the crown wall is non-submerged, the

16
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 32. Comparison of the horizontal wave force of the experimental data with the Pedersen method.

Fig. 33. Comparison of the uplifting wave force of the experimental data with the Pedersen method.

calculation results when B = 0.21 and 0.28 m are more accurate than 6.4. Comparison of the experimental data with the Nørgaard method
that when B = 0.07 and 0.14 m. The Pedersen method overestimates the
horizontal force when the berm consists of one or two rows of armour Nørgaard et al. (2013) used the run-up calculated by H0.1% (Eq. (25))
blocks, and the Pedersen method is more accurate when the berm con­ instead of the effective wave height. The run-up formula was proposed
sists of three or four rows of armour blocks. by Kobayashi et al. (2008). The regular wave height was taken as H0.1%,
Fig. 33 depicts the comparison results of uplifting wave forces. The and the regular wave period was taken as the mean period Tm. Fig. 34
uplifting forces calculated using the Pedersen method are greater than depicts the comparison results of the horizontal force. Regardless of the
the test results. The calculation results when the bottom of the crown condition of the bottom of the crown wall, whether submerged or
wall is submerged are better than those when the bottom of the crown non-submerged, as the berm width increases, the calculation results
wall is non-submerged. The uplifting force of the Pedersen method is obtained using the Nørgaard method approach the test results. However,
equal under certain conditions. Because the van der Meer and Stam compared with that of the results obtained using the Pedersen method,
(1992) run-up formula is used, and the maximum value of the run-up is the dispersion degree of the results obtained using the Nørgaard method
Ru,0.1% = 2.58H1/3, when Ru,0.1% is greater than 2.58H1/3, Ru,0.1% = is greater. This results from the wave run-up were calculated by Eq. (25).
2.58H1/3. Here, the increase in period has no effect on the uplifting wave Fig. 35 shows the uplifting force comparison between the experi­
force. In the tests, it is found that as the wave force increases, the mental data and the results obtained using the Nørgaard method. The
uplifting force increases. The Pedersen method fails to calculate the calculation results obtained using the Nørgaard method are generally
uplifting force when d1 = − 0.09 m. higher than the test results. For the cases wherein the bottom of the
crown wall is submerged, the calculation results have a similar linear
relationship with the test results. This shows that Kobayashi et al. (2008)
run-up formula is more effective than that presented by van der Meer

17
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 34. Comparison of the horizontal wave force of the experimental data with the Nørgaard method.

Fig. 35. Comparison of the uplifting wave force of the experimental data with the Nørgaard method.

and Stam (1992) in calculating the uplifting force. When the bottom of 7. Improved method to predict the wave force on the crown wall
the crown wall is non-submerged, the results of the Nørgaard method are
scattered, and the results when d1 = − 0.09 m are invalid. 7.1. Improved Martin method

From Figs. 27 and 29, a certain multiple relationship exists between

Table 3
Fitting coefficient of the improved Martin method.
tanα 1:1.5 1:2

coefficient CH,1 CH,2 CU,1 CU,2 CH,1 CH,2 CU,1 CU,2

a0 0.4597 − 3.1000 0.2706 − 8.5100 0.4890 − 10.1510 0.2629 − 46.8500
a1 0.1121 − 5.2900 0.0990 − 2.9040 − 0.0191 20.4000 − 0.0869 7.9820
b2 − 0.0562 8.4850 − 0.0997 10.0500 0.2456 − 3.6080 − 0.1145 − 15.1700
b1 0.1560 3.6130 − 0.0224 10.8500 − 0.1086 − 1.4110 − 0.0236 − 22.3900
a2 0.0270 6.4010 − 0.0683 2.8470 0.0085 − 0.9276 − 0.0593 − 5.6830
ω1 40.3800 26.9700 46.6200 45.7900 46.7800 7.2530 42.6100 52.4200
c1 0.0051 − 3.4260 0.0156 − 8.2300 − 0.1942 − 14.4000 0.0980 − 61.4500
d2 − 0.1013 1.2260 − 0.0122 − 5.3500 − 0.1520 0.6602 − 0.0118 73.3700
d1 0.0351 − 7.2420 − 0.0768 0.0037 − 0.1003 − 7.2880 − 0.0334 33.5500
c2 0.0394 − 4.0300 − 0.0313 − 4.9690 − 0.0774 − 2.1490 0.0029 43.5000
ω2 61.7900 36.8000 61.1500 9.1800 73.7800 72.3100 13.5800 3.4340

18
X. Han et al. Ocean Engineering 259 (2022) 111911

the wave forces obtained using the Martin method and the test results. process, it was discovered that the large calculation results of the orig­
Therefore, the Martin method was revised to improve the accuracy of inal Pedersen method could be mainly attributed to the fact that the
the calculation. The improved Martin method is defined as follows: calculated deep-water wavelength was significantly large, which caused
the wave force calculated using Eqs. (6) and (7) to be extremely large.
FH,improved = CH,1 ⋅FH,original + CH,2 (37)
Therefore, the deep-water wavelength in the Pedersen method is
replaced by an intermediate-depth wavelength. The intermediate-depth
FU,improved = CU,1 ⋅FU,original + CU,2 (38)
wavelength is calculated using the dispersion relation. The Irivarren
By linearly fitting the results obtained using the Martin method with number is used as the intermediate-depth wavelength instead of the
the test results, the parameters CH,1, CH,2, CU,1, and CU,2 can be obtained. deep-wave wavelength. Eqs. 37 and 38 are used for modification. Eq.
Then, the Fourier series is used to fitting CH,1, CH,2, CU,1, CU,2 with B/d. (39) is used to fitting the coefficient of CH,1, CH,2, CU,1, CU,2. The fitting
The fitting expression is as follows: results are shown in Table 4 and Fig. 39. It can be seen that the fitting
( ) ( ) ( ) ( ) results are in good agreement with the experimental results. The results
B B B B
CH(U),1(2) =a0 +a1 cos ω1 +b1 sin ω1 +a2 cos 2ω1 +b2 sin 2ω1 of the improved Pedersen method are shown in Figs. 40 and 41. Fig. 40
d d d d
( ) ( ) ( ) ( ) shows the comparison results of the horizontal wave force. When B =
B B B
+c1 cos ω2 +d1 sin ω2 +c2 cos 2ω2 +d2 sin 2ω2
B 0.07 and 0.14 m, the accuracy of the improved Pedersen method is
d d d d significantly higher than that of the original Pedersen method. When B
(39) = 0.21 and 0.28 m, the results of the two methods are similar. From the
statistical histogram of the error, the improved Pedersen method results
Table 3 is the fitting coefficient of the improved Martin method.
are closer to the experimental data.
Fig. 36 shows the comparison of CH,1, CH,2, CU,1, and CU,2 with fitting
Fig. 41 depicts the comparison results of the uplifting wave force.
results. It can be seen that the fitting results are in good agreement with
Although the improved Pedersen method reduces the dispersion of the
the correction coefficients. Under different slopes, the trends of CH,1 and
data, it still does not solve the problem of the same wave force in
CU,1 are similar, indicating that the influence of the slope on the hori­
different wave periods. This is because the run-up formula is not suitable
zontal force is small. In practice, by determining B/d, CH,1, CH,2, CU,1,
for swell waves. Nevertheless, the improved Pedersen method still
and CU,2 can be obtained. Thus, the horizontal force on the crown wall
significantly improves the accuracy of the original Pedersen formula to
can be calculated using the improved Martin formula.
calculate the uplifting force.
Fig. 37 depicts the comparison results of the horizontal force. For
different berm widths, the accuracy of the improved Martin method is
7.3. Improved Nørgaard method
significantly higher than that of the original Martin method. The results
of the improved Martin method are very close to the test results. Sub­
Similar to the way in which the Pedersen method was improved, the
sequently, a statistical analysis of the error between the calculation and
Nørgaard method was improved. First, the deep-water wavelength used
test results was performed. The results show that the error of the
in the calculation was replaced with the intermediate-depth wavelength,
improved Martin method is mainly concentrated between − 50 and 50
and the fitting was performed according to Eqs. (37) and (38). Table 5 is
N/m, accounting for 80–95%. In contrast, the error of the original
given the fitting coefficient of Eq. (39) of CH,1, CH,2, CU,1, CU,2. Fig. 42
Martin method ranging from − 50 N/m to 50 N/m accounts for 20–40%.
shows the fitting results are in good agreement with experimental re­
Fig. 38 depicts the comparison results of the uplifting wave force.
sults. Fig. 43 shows the comparison results of the horizontal wave force.
The uplifting forces calculated using the improved Martin method are in
Under different berm widths, compared with that of the original
good agreement with the test results. For the uplifting force, the
Nørgaard method, the calculation accuracy of the improved Nørgaard
improved Martin method also has a significantly higher accuracy than
method is significantly improved, particularly when B = 0.07 and 0.14
that of the original Martin method. Statistical results show that under
m. Fig. 44 shows the comparison results of the uplifting wave force.
different berm widths, the statistical frequency of the error ranging from
Under different berm widths, the improved Nørgaard method increases
− 50 N/m to 50 N/m is above 80%.
the statistical frequency of the original Nørgaard method in the error
interval [− 50 N/m 50 N/m] from 10%–35% to 80–90%.
7.2. Improved Pedersen method

The calculation results obtained using the original Pedersen method

are generally greater than the test results. Thus, a method was used to
improve the accuracy of the Pedersen method. During the calculation

Fig. 36. Coefficient comparison of the experimental and fitting results.

19
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 37. Comparison of the horizontal force and statistical frequency of the Martin method and improved Martin method.

Fig. 38. Comparison of the uplifting force and statistical frequency of the Martin method and improved Martin method.

Table 4
Fitting coefficient of the improved Pedersen method.
tanα 1:1.5 1:2

coefficient CH,1 CH,2 CU,1 CU,2 CH,1 CH,2 CU,1 CU,2

a0 0.6694 9.3060 0.4657 − 10.5630 1.0796 6.5283 0.4236 − 6.8150
a1 − 0.4216 26.5600 0.0071 13.6500 0.2263 3.5560 0.2193 − 25.4600
b2 0.0481 10.7400 0.0098 − 9.7420 − 0.4551 5.1350 − 0.0339 − 16.8800
b1 − 0.0522 3.5670 − 0.0577 14.6200 0.2889 − 25.4600 − 0.0807 − 10.1500
a2 0.0890 − 10.1200 − 0.2173 − 2.3350 − 0.1633 11.9300 − 0.0861 − 13.2200
ω1 6.4360 55.3700 21.6700 49.2700 8.4470 39.8200 14.1600 74.8200
c1 0.0269 4.6110 0.0233 − 16.5300 0.0214 3.3100 − 0.0058 − 16.3200
d2 0.0395 − 1.7470 0.0671 8.0890 0.1363 − 5.7450 0.0430 2.9710
d1 − 0.0522 6.4740 − 0.1486 − 12.3600 − 0.1619 − 15.1100 − 0.1177 7.4550
c2 − 0.0014 11.5900 − 0.0408 11.3600 0.0555 − 4.2940 − 0.1426 − 13.4800
ω2 50.5700 32.5800 40.5800 8.1520 46.7400 45.1100 21.4000 42.7200

7.4. Comparison of the methods for predicting the wave force on the wave forces on the crown wall. The mean square error (MSE), root mean
crown wall square error (RMSE), mean absolute error (MAE), and mean relative
error (MRE) were used to estimate the accuracy of the methods, as
Seven methods were used to calculate the horizontal and uplifting shown in Eqs. (40)–(43). When the wave force is small and relative error

20
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 39. Coefficient comparison of the experimental and fitting results.

Fig. 40. Comparison of the horizontal force and statistical frequency of the Pedersen method and improved Pedersen method.

Fig. 41. Comparison of the uplifting force and statistical frequency of the Pedersen method and improved Pedersen method.

21
X. Han et al. Ocean Engineering 259 (2022) 111911

Table 5
Fitting coefficient of the improved Nørgaard method.
tanα 1:1.5 1:2

coefficient CH,1 CH,2 CU,1 CU,2 CH,1 CH,2 CU,1 CU,2

a0 − 8.3126 9.1492 0.0697 − 35.6940 − 4.9819 20.5403 0.1076 − 43.4944
a1 7.2140 − 8.4140 − 0.0243 − 14.0800 2.3960 41.8900 − 0.0095 − 1.9830
b2 10.5800 4.9960 − 0.0045 − 38.0600 8.7320 − 15.0700 − 0.0139 − 22.6800
b1 0.8601 16.0600 − 0.0257 18.9700 1.6790 − 0.4263 0.0322 − 4.8790
a2 − 2.9710 − 6.2040 − 0.0135 − 17.5600 − 1.2820 − 6.8890 − 0.0056 − 24.9400
ω1 2.4230 26.7900 40.8700 76.0100 3.5890 6.4690 52.1400 38.0200
c1 − 0.0750 4.0930 − 0.0147 − 0.2791 − 0.1946 − 6.1620 − 0.0234 − 4.9790
d2 0.0572 − 1.5800 0.0050 − 7.5940 0.3309 − 19.4200 0.0273 26.5800
d1 − 0.0420 10.6200 − 0.0061 − 9.0850 0.2089 − 10.7200 − 0.0322 11.2000
c2 − 0.0225 4.8380 0.0198 − 10.5300 − 0.0420 − 10.1800 0.0189 4.7530
ω2 55.9000 58.9600 30.7700 17.1600 51.2200 50.3700 39.6800 53.3100

Fig. 42. Coefficient comparison of the experimental and fitting results.

Fig. 43. Comparison of the horizontal force and statistical frequency of Nørgaard method and improved Nørgaard method.

is large, the MRE is significantly large. Therefore, MRE was calculated √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
√
for all working conditions where the experimental wave force exceeded √1 ∑ N
( )2
20 N/m. RMSE = √ Fi,Exp − Fi,Cal (41)
N i=1

1 ∑N
( )2
MSE = Fi,Exp − Fi,Cal (40)
N i=1 1 ∑N ⃒ ⃒
MAE = ⃒Fi,Exp − Fi,Cal ⃒ (42)
N i=1

22
X. Han et al. Ocean Engineering 259 (2022) 111911

Fig. 44. Comparison of the uplifting force and statistical frequency of Nørgaard method and improved Nørgaard method.

of the CCD method is smaller than that of the Martin, Pedersen, and
Table 6
Nørgaard methods. The improved Martin, Pedersen, and Nørgaard
MSE, RMSE, and MAE of the horizontal force when the bottom of the crown wall
methods have significantly smaller errors than those of the original
is submerged.
methods. Among the seven methods, the improved Pedersen method
method MSE RMSE MAE MRE
exhibits the smallest error and most accurate calculation.
Martin method 256595.79 506.55 402.25 137% Table 7 presents the error results for the horizontal force when the
Pedersen method 67007.10 258.86 140.23 51% bottom of the crown wall is not submerged. The errors of the Martin and
Nørgaard method 47281.73 217.44 129.53 51%
Pedersen methods are similar. Although the CCD method is superior to
CCD method 6218.16 78.86 58.33 34%
improved Martin method 5670.43 75.30 45.56 26% the Martin, Pedersen, and Nørgaard methods, CCD still has a large error
improved Pedersen method 2977.56 54.57 38.14 24% in calculating the horizontal wave force under the non-submerged berm.
improved Nørgaard method 4216.78 64.94 43.58 33% The improved Martin, Pedersen, and Nørgaard methods have signifi­
cantly smaller errors than those of the other four methods. The improved
Martin method has the smallest calculation error.
Table 7 Table 8 presents the horizontal force error results for all the cases.
MSE, RMSE, and MAE of the horizontal force when the bottom of the crown wall For the horizontal wave force on the crown wall under regular waves,
is non-submerged. the Martin method is invalid. The Nørgaard method is slightly better
method MSE RMSE MAE MRE than the Pedersen method. The MRE of the CCD method are 39% smaller
than those of the Pedersen and Nørgaard methods, respectively. The
Martin method 10893.21 104.37 66.96 77%
Pedersen method 12928.93 113.71 66.63 78% accuracy of the improved Martin, Pedersen, and Nørgaard methods is
Nørgaard method 15860.12 125.94 72.38 69% significantly better than those of the other four methods. The improved
CCD method 1959.69 44.27 33.09 48% Martin method reduces the MAE of the original Martin method from
improved Martin method 742.22 27.24 19.73 31%
110% to 28%. The improved Pedersen method reduces the MRE of the
improved Pedersen method 912.16 30.20 22.32 39%
improved Nørgaard method 816.44 28.57 19.88 36%
original Pedersen method from 60% to 29%. The improved Nørgaard
method reduces the MRE of the original Nørgaard method from 57% to
⃒ ⃒ 34%. The improved Martin method is the best to estimate the horizontal
1 ⃒Fi, ​ Exp − Fi, ​ Cal ⃒ force.
MRE = (43)
N Fi,Exp Table 9 presents the error results of the uplifting force when the
bottom of the crown wall is not submerged. The MREs of the Martin,
Table 6 presents the error results for the horizontal force when the Pedersen, Nørgaard, and CCD methods are all greater than 50%. The
bottom of the crown wall is submerged. By comparing the four in­ improved Martin, Pedersen, and Nørgaard methods have significantly
dicators, the Martin method is found to have the largest error. The errors minor errors. The error of the improved Nørgaard method is the
of the Pedersen and Nørgaard methods are similar. The calculation error

Table 9
Table 8 MSE, RMSE, and MAE of the uplifting force under the bottom of the submerged
MSE, RMSE, and MAE of the horizontal force of each method. crown wall.
method MSE RMSE MAE MRE method MSE RMSE MAE MRE

Martin method 156201.18 395.22 265.25 110% Martin method 104974.20 324.00 289.96 137%
Pedersen method 44910.65 211.92 110.16 60% Pedersen method 18472.61 135.91 102.41 72%
Nørgaard method 34442.80 185.59 106.18 57% Nørgaard method 14194.65 119.14 103.59 83%
CCD method 4478.14 66.92 48.02 39% CCD method 5501.11 74.17 65.11 60%
improved Martin method 3656.75 60.47 35.01 28% improved Martin method 2256.72 47.50 37.69 42%
improved Pedersen method 2133.63 46.19 31.67 29% improved Pedersen method 1778.08 42.17 33.99 33%
improved Nørgaard method 2827.40 53.17 33.89 34% improved Nørgaard method 1222.73 34.97 26.09 28%

23
X. Han et al. Ocean Engineering 259 (2022) 111911

Table 10 period increased, the wave force increased, and the rate of in­
MSE, RMSE, and MAE of the uplifting force under the bottom of the non- crease gradually decreased. When the bottom of the crown wall
submerged crown wall. was submerged, the wave force of the swell waves was approxi­
method MSE RMSE MAE MRE mately 1–3 times that of the short-period waves. When the still
Martin method 25923.14 161.01 141.44 126%
water level was lower than the bottom of the crown wall, the
Pedersen method 10184.90 100.92 81.63 91% longer-period wave force was 3–7 times the shorter-period wave
Nørgaard method 12273.86 110.79 89.49 106% force. The berm width had a significant influence on the short-
CCD method 11166.07 105.67 98.85 101% period wave force. As the berm width increased, the wave force
improved Martin method 599.37 24.48 16.62 47%
decreased significantly. However, the berm width had little effect
improved Pedersen method 583.46 24.15 17.98 50%
improved Nørgaard method 474.58 21.78 14.51 37% on swell waves.
(3) The comparison of the experimental data and the method of
Martin et al. (1999), CCD (CCCC First Harbor Consultants CO.
LTD, 2015), Pedersen (1996), and Nørgaard et al. (2013) is
Table 11
analysed. (i) Martin method obviously overestimates the hori­
MSE, RMSE, and MAE of the uplifting force for each method.
zontal force and uplifting force on crown wall under the bottom
method MSE RMSE MAE MRE
of crown wall submerged. In this experimental scope, Martin
Martin method 72673.77 269.58 229.27 134% method is suitable for horizontal force of d1 = − 0.09 m. (ii) CCD
Pedersen method 15086.23 122.83 93.92 76% horizontal force results has a good agreement with experimental
Nørgaard method 13409.81 115.80 97.83 87%
data under the bottom of crown wall submerged. While under the
CCD method 7815.82 88.41 78.90 66%
improved Martin method 1579.52 39.74 29.08 43% bottom of crown wall is non-submerged, the CCD underestimates
improved Pedersen method 1289.95 35.92 27.45 37% the horizontal force when longer period waves, and it is over­
improved Nørgaard method 917.03 30.28 21.36 30% estimate the uplifting wave force. (iii) Because Pedersen method
and Nørgaard method are used to calculate F0.1%. To calculate the
wave force under regular waves, the height and period of regular
smallest, and its MAE is 28%. Table 10 presents the uplifting error re­
waves is taken as H0.1% and Tm. Under the condition that the berm
sults when the bottom of the crown wall is non-submerged. Because the
is 3 rows and 4 rows armour blocks, the calculated horizontal
water level is lower than the bottom of the crown, wave breaking occurs
force results of Pedersen method are in great agreement with the
on the slope and the waves may not reach the crown wall. The Martin,
test results. When the berm is 1 row and 2 rows blocks, the
Pedersen, Nørgaard, and CCD methods have large MREs of over 90%.
Pedersen horizontal results are significantly higher than the test
These four methods are invalid in calculating the uplifting force under
results. Pedersen method overestimates the uplifting force and
the non-submerged bottom crown wall. All the improved methods have
cannot consider the influence of wave period on uplifting force.
smaller errors than those of the original methods. The improved
(iv) Nørgaard method has greater dispersion of horizontal wave
Nørgaard method has the smallest error. Table 11 shows the uplifting
force than Pedersen method. Under the bottom of crown wall
force error results for all the cases. The MSE, RMSE, MAE, and MRE of
submerged, Nørgaard’s results are more and more consistent with
the improved Nørgaard method are the smallest. This indicates that the
the test results as the berm width increases. For the uplifting
improved Nørgaard method is the most accurate in determining the
force, the Nørgaard results under the bottom of crown wall sub­
uplifting force.
merged is more accurate than the bottom of crown wall
non-submerged.
8. Conclusions
(4) An linear modified method is proposed to improve the original
method of Martin et al. (1999), Pedersen (1996), and Nørgaard
A total of 744 physical tests were performed to study the wave force
et al. (2013). The fitting formulas of linear modified coefficients
on the crown wall under swell waves. The wave period was emphatically
under different B/d are given. The application scope is
considered in the test design to study the influence of swell waves on the
Rc /Ac ≥ 1.5556, 1.4 ≤ Rc /H ≤ 11.5, 1.4 ≤ (Rc − Ac ) /H ≤ 7 −
wave load. The wave periods of model tests were 1.10, 1.46, 1.83, 2.19,
2.25 ≤ d1 /H ≤ 2.25, 1.51 ≤ ξ ≤ 13.61, 1.5 ≤ cot α ≤ 2. The
2.56, 2.92, and 3.29 s, which corresponded to the prototype wave pe­
comparison of the improved Martin method, improved Pedersen
riods of 6, 8, 10, 12, 14, 16, and 18 s (6–8 s are short period; 10–18 s are
method, and improved Nørgaard method with the original
swell wave period). The following conclusions were drawn.
method is studied. The improved method is more accurate than
the original methods. The improved Martin method is the best
(1) By adopting pressure sensors of 100 and 1000 Hz to collect the
method to estimate the horizontal force. The improved Nørgaard
pressure, the influence of the sampling rate on the wave force was
method is the most accurate for uplifting force.
studied. Under short-period waves, both could capture the impact
pressure of each pressure point. Under swell waves, the pressur­
e–time series of each measurement point represented the pul­ CRediT authorship contribution statement
sating pressure. The pressure–time series of the 100 and 1000 Hz
sensors were the same. The wavelet energy spectra of the 100 and Xinyu Han: Methodology, Software, Validation, Investigation, Data
1000 Hz time series at the same measurement point were curation, Writing – original draft, preparation, Writing – review &
consistent. Although the comparison of the wave forces was editing, Visualization. Yunpeng Jiang: Validation, Investigation.
performed using the 100 Hz and 1000 Hz pressure data, the re­ Sheng Dong: Conceptualization, Methodology, Formal analysis, Re­
sults were found to be consistent, indicating that a 100 Hz sam­ sources, Writing – original draft, preparation, Writing – review & edit­
pling rate was sufficient to collect the pressure on the crown wall. ing, Supervision, Project administration, Funding acquisition.
(2) The influence of the wave conditions and structural parameters
on the wave load of the crown wall was analysed. The slope had
little effect on the horizontal and uplifting forces. Under the same Declaration of competing interest
water depth, the changing trend of the non-dimensional wave
force with non-dimensional wave height was basically observed The authors declare that they have no known competing financial
when the berm and wave period were changed. As the wave interests or personal relationships that could have appeared to influence
the work reported in this paper.

24
X. Han et al. Ocean Engineering 259 (2022) 111911

Acknowledgements Irivarren, R., Nogales, C., 1954. Obras Marítimas: Oleaje Y Diques. Editorial Dossat
Madrid, Spain.
Jensen, O.J., 1984. A Monograph on Rubble Mound Breakwaters. Danish Hydraulic
The study was partially supported by the National Natural Science Institute, Hørsholm, Denmark.
Foundation of China (52171284). Jiang, Y., Chen, H., Ji, Z., Zhang, H., 2021. Experimental study on wave-induced load on
crest wall in North Indian Ocean. In: IOP Conference Series: Earth and
Environmental Science, vol. 621, 012159.
References Kobayashi, N., de los Santos, F.J., Kearney, P.G., 2008. Time-averaged probabilistic
model for irregular wave runup on permeable slopes. J. Waterw. Port, Coast. Ocean
Ahn, S., 2021. Modeling mean relation between peak period and energy period of ocean Eng. 134 (2), 88–96.
surface wave systems. Ocean Eng. 228, 108937. Liao, C., Tong, D., Jeng, D., Zhao, H., 2018. Numerical study for wave-induced
Amrutha, M.M., Kumar, V.S., George, J., 2017. Observations of long-period waves in the oscillatory pore pressure and liquefaction around impermeable slope breakwater
nearshore waters of central west coast of India during the fall inter-monsoon period. heads. Ocean Eng. 157, 364–375.
Ocean Eng. 131, 244–262. Losada, M.A., 1992. Recent developments in the design of mound breakwaters. In:
Bekker, J., Hofland, B., Smith, G., 2018. Experiments of the effect of freeboard on the Herbich, J.B. (Ed.), Handbook of Coastal and Ocean Engineering, pp. 939–1050.
stability of a breakwater crown wall. In: Proceedings of the International Conference Losada, M.A., Gimenez-Curto, L.A., 1981. Flow characteristics on rough, permeable
on Coastal Engineering 2018. ASCE. slopes under wave action. Coast. Eng. 4, 187–206.
Berenguer, J.M., Baonza, A., 2006. Rubble mound breakwater crown wall design. In: Martin, F.L., Losada, M.A., Medina, R., 1999. Wave loads on rubble mound breakwater
Proceedings of National Conference of the Port and Coastal Technical Association. crown walls. Coast. Eng. 37 (2), 149–174.
Algeciras, Spain. (in Spanish). Molines, J.L., 2016. Wave overtopping and crown wall stability of cube and Cubipod-
Bradbury, A.P., Allsop, N.W.H., Stephens, R.V., 1988. Hydraulic Performance of armored mound breakwaters, PhD Thesis. Universitat Politectina De Valencia.
Breakwater Crown Walls. Report No SR 146. HR Wallingford. HR Wallingford. Molines, J.L., Herrera, M.P., Medina, J.R., 2018. Estimations of wave forces on crown
Braña, P.C., Guillé, J.F., 2004. Wave forces on crown walls: evaluation of existing walls based on wave overtopping rates. Coast. Eng. 132, 50–62.
empirical formulae. In: Proceedings 29th International Conferences on Coastal Nørgaard, J.Q.H., Andersen, T.L., Burcharth, H.F., 2013. Wave loads on rubble mound
Engineering. World Scientific, pp. 4087–4099. breakwater crown walls in deep and shallow water wave conditions. Coast. Eng. 80,
Burcharth, H.F., Frigaard, P., Berenguer, J.M., Gonzalez, B., Uzcanga, J., Villanueva, J., 137–147.
1995. Design of the ciervana breakwater, bilbao. In: Proceedings of 4th Coastal Pedersen, J., 1996. Wave Forces and Overtopping on Crown Walls of Rubble Mound
Structures and Breakwaters. Institution of Civil Engineers. Breakwaters. Ph.D. thesis, Series paper 12. Hydraulics & Coastal Engineering Lab.,
Camus, P.B., Flores, G.J., 2004. Wave forces on crown walls: evaluation of existing Dept. of Civil Engineering, Aalborg University, Denmark, 0909-4296.
empirical formulations. Coast. Eng. 4, 4087–4099, 2004. Pedersen, J., Burcharth, H.F., 1992. Wave forces on crown walls. In: Coastal Engineering
CCCC First Harbor Consultants CO., LTD, 2015. Code of Hydrology of Harbour and 1992. ASCE, New York, pp. 1489–1502.
Waterway (JTS 145-2015). Ministry of Transport of People’s Republic of China, U.S. Army Corps of Engineers, 2002. Coastal Engineering Manual. Engineer Manual
Beijing. 1110-2-1100. U.S. Army Corps of Engineers, Washington, D. C (in 6 volumes).
George, J., Kumar, V.S., 2020. Climatology of wave period in the Arabian Sea and its Valdecantos, V.N., Gutiérrez, J.S.L., Flors, J.I.P., 2013. Comparative study of breakwater
variability during the recent 40 years. Ocean Eng. 216, 108014. crown wall - calculation methods. Mari. Eng. 166 (MA1), 25–41.
Günback, A.R., Göcke, T., 1984. Wave screen stability of rubble mound breakwaters. In: van der Meer, J.W., 1988. Rock Slopes and Gravel Beaches under Wave Attack. Doctoral
Proceedings of the International Symposium on Maritime in the Mediterranean Sea. Thesis. Delft Technical University.
Athens, Greece, 2999-2112. van der Meer, J.W., Stam, C.J.M., 1992. Wave run-up on smooth and rock slopes. J. WPC
Han, X., Dong, S., Wang, Y., 2021. Interaction between oblique waves and arc-shaped OE 188 (5), 534–550.
breakwater: wave action on the breakwater and wave transformation behind it. Yuksel, Y., Cevik, E., van Gent, M.R.A., Sahin, C., Altunsu, A., Yuksel, Z.T., 2020.
Ocean Eng. 234, 109252. Stability of berm type breakwater with cube blocks in the lower slope and berm.
Ocean Eng. 217, 107985.

25