CO/k AL

EPlGllqEEiUPlG
Coastal Engineering, 22 (1994) 3-29

Review and analysis of vertical breakwater failures

lessons learned
H. Oumeraci
Franzius-lnstitut, SFB 205, University of Hannover, Germany
Received 15 March 1992; accepted after revision 15 May 1993)

Abstract

The failures experienced by vertical and composite breakwaters are briefly reviewed and the main
reasons for failures are examined.
The objective of this study is threefold: (a) identification of the weak components of the structure-
foundation system, (b) illustration of the weak points of past and present design procedures and (c)
development of a research strategy for the improvement of the technical basis for the design of vertical
structures.
First, three main categories of the reasons for failures are suggested: (a) reasons inherent to the
structure itself, (b) reasons inherent to the prevailing hydraulic and loads conditions, and (c) reasons
inherent to the foundation and seabed morphology. These reasons are then sytematically discussed
and lessons are drawn which are related to each of the aspects considered. The results suggest that the
present design approaches cannot explain most of the failure modes reported, and that the stability of
vertical breakwaters is an integrated and complex problem which can satisfactorily be solved only by
dynamic analysis and probabilistic design approaches.

I. Introduction

At the end of the seventies and at the beginning of the eighties, catastrophic failures were
experienced by a series of large rubble mound breakwaters. This shock to the profession
was comparable to the first shock half a century earlier, as several vertical breakwaters
collapsed. The irony of the technical development which followed these two events was
that in both cases the profession returned to old solutions and concepts that appeared to
have been almost abandoned. The failed vertical breakwaters were rebuilt as conventional
rubble mound structures, and for the rehabilitation of the damaged rubble mound break-
waters with concrete armour units the old b e r m breakwater concept was rediscovered.
The principal difference between the two cases certainly lies in the attitudes adopted by

0378-3839/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved

SSDI0378-3839 (93)E0042-H
4 H. Oumeraci / Coastal Engineering 22 (1994) 3-29

the profession with regard to the further application of such structures. After the first shock
in the thirties, the vertical breakwater was almost abandoned - - except in some countries
like Japan, Taiwan, China, Korea and Italy - - in favour of the rubble mound type. On the
other hand, the second shock in the eighties gave rise to extensive research activities towards
improving the design and construction of rubble mound breakwaters. The latter certainly
represents the better attitude, although no definitive solutions to the most urgent problems
(structural strength of armour units, geotechnical stability, crown-wall stability etc.) have
yet been achieved.
Meanwhile, the need for breakwaters at greater depths to suit the increasing draught of
large vessels in the last decades has made the costs of such structures more prohibitive
(construction costs in the range of US$100.000 per linear metre structure). In this respect,
monolithic structures may represent a better alternative, in terms of performance, total costs,
standardisation, quality control, environmental aspects, construction time and maintenance.
In addition, the situation is quite different from that in the thirties, since a number of
important developments which might promote the revival of vertical breakwaters have taken
place in the last decades. These are for instance:
- availability of more reliable wave observations, wave records, meteorological data and
sophisticated wave hindcast and refraction models;
-considerable knowledge which has been accumulated since the thirties with respect to
wave breaking and impacts on structures;
- remarkable development of hydraulic modelling by using irregular waves and further
sophisticated techniques for the measurement of transient loading and response of struc-
tures;
-availability of large-scale testing facilities (super wave tanks) in which the dynamic,
hydraulic and geotechnical aspects can simultaneously be investigated;
- developments in the offshore oil industry, especially with regard to the numerical mod-
elling of wave-structure-soil interaction, as well as to the technology of caisson structures
and marine foundation work. In this respect, a large experience in the technology related
to breakwaters has also been accumulated in Japan (Tanimoto et al., 1987; Tanimoto and
Goda, 1991; Takahashi et al., 1992; Tanimoto and Takahashi, 1994).
The aformentioned, relatively new and ever-increasing need for protective structures in
deeper water, induced by the rapid increase of ship sizes in the last decades, together with
the relatively recent developments enumerated above, constitute a good opportunity to
recover the chance missed in the thirties by learning more from the failures and, based on
the lessons learned, to establish an integrated research programme which will allow vertical
breakwaters to get at least as competitive as traditional rubble mound structures.
It is believed that this action towards the revival of vertical breakwaters should necessarily
start with a comprehensive review of past failures, since one can learn more from these
difficult experiences, where the forces of the sea have prevailed over the defenses of men,
than from successful experiences. It is in this spirit that this study has been undertaken.
The main objective of this study is to contribute to the revival of vertical breakwaters, in
the sense that it might help the profession:
- to see better how unfortunate it was to abandon almost fully this type of structure;
- to identify the weak components of the structure-foundation system and to point out the
weak points of past and present design procedures;
 H. Oumeraci / Coastal Engineering 22 (1994) 3-29 5

- to develop an efficient research strategy for promoting this type of structure which may
not only be technically and economically better than the rubble mound type, but may also
exhibit potential advantages with respect to environmental issues.
On the other hand, it is also hoped that this study will be considered by contractors,
consulting engineers, harbour authorities, owners and further decision makers in this field
as an appeal to overcome their "cover-up attitude" towards failure, and to report and
publish on such events objectively. This, however, will only be possible if it is really
understood that failures are an integrated part of the technical progress and that coastal and
harbour engineering, like science in general, advances by both successful experience and
mishaps.
The general approach adopted for this study is based on the consideration of three
questions, since the engineer generally wants to know what happened, why it happened and
how to avoid similar failures in the future. Therefore, the facts and design conditions are
described first, and then the major causes are pinpointed. Finally, the lessons are drawn
which may help not only the designer to avoid similar future problems, but also the researcher
to define the most urgent research tasks.
There is, of course, no pretention to perform a comprehensive analysis of all reasons for
failures, since detailed information is generally lacking, for instance with regard to design
waves, water levels and further design conditions like seabed morphology and soil condi-
tions, circumstances of the storm (peak, increasing and decreasing phases), sequences and
manner in which the different components of the structure-foundation system were dam-
aged. Nevertheless, numerous sources of information as well as conflicting opinions and
data are used to reach the best possible conclusions.

2. Review of analysed failure cases

The failure cases which have been reviewed and analysed in this study are related to
vertical breakwaters and to "composite breakwaters" of the Japanese type. The latter are
called in the following "annoured vertical breakwaters".
For vertical breakwaters, 17 failure cases have been considered which are summarized
in Table 1. The type of structure, the design wave conditions, the wave conditions respon-
sible for the damage, the water depth conditions, the various characteristics of the structure
and its foundation, the nature of the seabed and the major reasons for failure are given as
far as the information are available. The notations used in Table 1 are defined in Fig. 1. A
unique failure of a caisson breakwater (Naples habour) in Europe has recently been reported
by Franco and Passoni (1992).
For "armoured vertical breakwaters", five failure cases have been considerd which are
summarized in Table 2. The type of blocks for the protective armour, the design wave
conditions, the wave conditions responsible for the damage, the water depth conditions, the
characteristics of the structure and the major reasons of failure are given. Most of these
failures have been reported by Hattori et al. (1984). The failure of Ventotene breakwater
has been described by Franco ( 1991 ). The notations used in Table 2 are defined in Fig. 2.
 O~

Table 1
Review of vertical breakwater failures (for notation see Fig. 1 )

Breakwater Type H/T Nature of B d hfo hfo di Major reasons for failure
(country, year) seabed hc dw ha bj l/m (major mode of failure)
Design Actual

Madras BCY 7.3 22 7.2 14.6 Breaking waves

(India, 9.1 7.2 7.2 1/1 Overtopping ?~
1881)
Scour and erosion ---.
(seaward tilt)
Bizerta CAI 8 17 10 10 Breaking waves
(Tunesia, 13 8 5 4/5 Overtopping
1915)
Erosion of RMF 0~
Valencia UCY 7/14 fine 12 12 5 6.7 4.0 Exceedance design wave ~.
( Spain, sand, 14.4 9.5 2.7 10 1/3 Breaking waves
1926) mud Overtopping ~,,
Seabed scour
',O
(seaward tilt)
Antofa-Gasta UCY 6/8 9/15 10 30 7.5 7.5 12 Exceedance design wave I
(Chile, 8/47 16.9 9.4 3.5 3.0 4/3 Non-monolithicity
1928-29)
Breaking waves
Catania UCY 6/7 7/9 dense 12 17.5 7.5 Exceedance design wave
(Italy, 7.5/12 sand 20 12.5 4.0 Non-monolithicity
1930-33 )
Breaking waves
Overtopping
Differential settlement
(sliding and seaward tilt)
 7/12 fine 12 17.5 7.4 6.0 8/7 Exceedance design wave
Genoa WE 5.5/7
sand 17.9 10.5 3.0 12 1/1 Non-monolithicity
(Italy, UCY
Breaking waves
1955) CEL
Wave overtopping
(sliding)

5/7.4 6.5/11 silty 11 20 6.5 7.3 7.7 Exceedance design wave .~

Algiers BCY
9/14 sand, 216 13 3.0 3.7 1/2 Breaking and overtopping
(Algeria,
mud Differential settlement
1930,1934)
Seabed scour
Erosion RMF "--
(seaward tilt)

7/13 7/13.5 silty 15 17 4.5 9.0 40 Breaking waves

Niigata CAI
sand 10 4.5 7.0 1/3 Overtopping o~
(Japan,
Differential settlement ~"
1976)
(seaward tilt) 5.

9.5 15.5 4.5 5 5 Exceedance design wave

Bari WEL
15 9.0 3.0 2 1/2 Wave breaking
(Italy,
Erosion RMF
1974)
5.5/7 6.1/11.3 9 36 9.0 t0 36 Exceedance design wave
Palermo UCY
2.6 l0 5.0 5 1/15 Wave breaking
( Italy,
Erosion RMF
1973)
13 19 7.5 10 - Wave breaking
Naples WEL
18 125 2.0 3.5 1 / 1.5 ( sliding and overturning)
(Italy, CEL
1987) CAI
(Continued on p. 8)
 Table 1 (continued)

Breakwater Type H/T Nature of B d ht~ hfo di Major reasons for failure
( country, year) seabed h~ dw hfi bj l/m (major mode of failure)
Design Actual

Mashike CAI 5.5/10 6.6/12 14.5 7 1.0 6.9 2.8 Exceedance design wave
(Japan) 6.0 3.9 1.0 6 1/3 Wave breaking
Overtopping ( non-
completed at head)
( sliding = 2.9 m)
Fukaura CAI 7.6 / I 1 6.3 / 13 20.5 15 2.5 17 4.3 Wave breaking
(Japan) 12.5 11 2.5 12 1/3 Overtopping (non- 2.
completed at head)
Erosion RMF
( sliding = 3.0 m)
Sakata CAI 3.3/8.8 5.5/13 6.5 7.0 5.5 10 2.5 Exceedance design wave
(Japan) 10 2.9 2.5 4.5 1/1.5 Wave breaking
Differential settlement ~.
(sliding and shoreward i-,a
tilt) ~
Onahama CAI 6.1/14 6.1/14 15 16 5.0 7.5 9 Wave breaking .~
(Japan) 13.5 4.5 5.0 5.0 1/3 Erosion RMF
(sliding)
Niigata- CAI 7.5/13.5 7.5/11 18 18 4.5 12 6.5 Wave breaking
West Jetty 17.5 11 4.5 11 1/3 Overtopping
(Japan)
Erosion of RMF
( sliding = 26 m)
Niigata- CAI 7/13 7.1/13.5 15 15 4.5 12 4 Wave breaking
West Breakw. 18 11 4.5 8 1/2 Overtopping
(Japan)
Differential settlement
 H. Oumeraei / Coastal Engineering 22 (1994) 3-29 9

I, - B- ,I
~ m

• 1---1
hfo

i
C

bi

Notations in Tab. 1

T = Wave Period UCY = Cyclopean Block ~ype (unbounded) 1)

RMF = Rubble Mound Foundation CEL = CellularBiock.13,p¢ 1)
CA = Caisson Type WEL = Well Block Type 1)
BCY = Cyclopeun Block Type (bounded) 1)

1 ) For definitions of block types see CAVANEL (1955)

Fig. 1. Vertical breakwaters - - definition sketch for Table 1. W a v e periods are given in seconds; all length
d i m e n s i o n s are in metres.

I< b > :< B ~1

Notations used in Tab.2: G = Weight o f A r m o u r Unit R = Rock

TT = Tetrapods T = Wave Period
U = Unknown

Fig. 2. A r m o u r e d vertical b r e a k w a t e r (Japanese type) - - definition sketch for Table 2. W a v e periods are given
in seconds; all length dimensions are in metres.
Table 2
Review of damaged armoured vertical breakwaters (for notation see Fig. 2)

Breakwater Type of H/T Nature of B hf bl 1/m Major reasons for failure

(country, year) Block Seabed hc d hR ( major mode of failure)
Design Actual

Ventotene R 8 4.5 2 12 I/1.5 Wave breaking

(Italy, 10 5 Erosion RMF
1966)
(sliding) e5
Rumoi TT 8.9/13 6.8/10 15 4.5 6.3 45 4/3 Wave concentration ?
Japan) 16 18 4.5 at bound
Erosion RMF
Ishikari- TI" 5.6/11 3.8/75 10 2.5 4.6 20 4/3 Seabed scouring
New Port 8.5 8 3.9 Erosion RMF 9e~
(Japan)
Oshidomari 5/11 5.8/9.5 11 3.2 6 16 1/1.5 Settlement
(Japan) 13.2 14.6 3.2 Erosion RMF (S-shape)
Wave breaking 4~
(sliding = 4.1 m)
I
Miyako TI" 4.4/10 4.6/11 t~a
12.5 4.0 4.0 12.5 4/3 Wave breaking
( J ap an ) 11 11.5 4.0 Overtopping
Erosion RMP
(sliding= 1.5 m)
 H. Oumeraci / Coastal Engineering 22 (1994) 3-29 11

3. Reasons for failure

3.1. Classification of reasons for failures

In this study, account is only made for the search of technical causes, but one should
always keep in mind that pinpointing the technical causes of failures is only one part of the
problem, since there are also economic, political, social and human factors that must be
considered.
In addition, it should be emphasized that almost none of the failures reported to date have
occurred without prior warning due to previously experienced less violent storms or to
failure case histories of similar structures. Although this problem of difficulties of perception
and communication constitutes one of the main reason of the failures which have particularly
occurred after the fifties, it is not further discussed here.
A further reason which is not considered in this paper is that some structures that have
failed have done so because they represented attempts to design and build boldly beyond
experience.
For the evaluation of the reasons of failures, it ist important that all relevant modes of

Inadequacy of the "concept

of reflective structures"

Non.monoitthicity
Reasons Inherent to of the structure
the structure itself
I I
Too low crest of the structure I

Too high too borm

Exceedcnce of wave
design conditions

Wave concentration
Reasons inherent to along the structure
hydraulic conditions
and loads Wave breaking &
Impact load

Wave overtopping

Unfevoureble
seabed topography

Reasons inherent to Scour end erosion

foundation and
seabed morphology Sefflement

Slip failure (shear)

Fig. 3. Reasons for the failures of vertical structures.

12 H. Oumeraci / Coastal Engineering 22 (1994) 3-29

failure are considered and that any cause which might have contributed to the resulting
failures observed after the storms should be accounted for.
The reasons which have led to the failures summarized in Tables 1 and 2 may be
subdivided in the following three categories which are shown in Fig. 3:
(i) reasons inherent to the structure itself;
(ii) reasons inherent to the hydraulic conditions and loads;
(iii) reasons inherent to the foundation and seabed morphology.

3.2. Reasons inherent to the structure itself

Among the reasons for failures which are inherent to the concrete structure and its rubble
mound foundations, one may distinguish the inadequacy of the concept of reflective struc-
tures (vertical structures have to reflect all the incoming wave energy), the crest level of
the rubble mound foundation which is generally too high, and the crest level of the concrete
structure which is too low. In addition, there are further reasons related to the non-monol-
ithicity of the structure and the weakness of the concrete material. The latter reasons are
however less important, since the modern caisson used nowadays for vertical structures are
necessarily monolithic structures made of good concrete.

Inadequacy of the ' 'concept of reflective breakwaters"

Under "vertical breakwaters" one primarily understood a structure composed of a ver-
tical (or almost vertical) structure founded on a rubble mound base extending to the sea
bottom which withstands the destructive action of the waves by reflecting their energy
towards the open sea. In other words, it was believed, that the basic characteristic of this
breakwater type was to reflect waves, i.e. without the occurrence of any shock (PIANC,
1976).
However, during the exceptionally violent storms which caused the failures it was always
observed that:
- T h e storm waves were highly irregular, short-crested and often broke (Renaud, 1935,
1936; Lira, 1935);
- Waves sometimes broke at water depths where any breaking would have been excluded
if the vertical structure was not present. For instance, waves of 7 m height often broke in
depths of more than 20 m before reaching the breakwater, and 4 m high waves broke at a
vertical breakwater founded at - 4 6 m (Lira, 1935). These and further similar observa-
tions in Algiers (Benezit and Renaud, 1935; Renaud, 1935, 1936), Genoa (Grimaldi,
1955), Palermo etc. have clearly shown, that breaking is not solely caused by shoaling.
In fact, the recent developments in water wave research have now shown that breaking
might also be attributed to reasons inherent to the irregularity of the waves (frequency
modulated wave groups, storm wave-long wave interaction), to wave interferences
between incident and reflected waves, as well as to the effect of wind and current on
waves (Oumeraci et al., 1992).
- T h e failures were almost always induced by breaking waves, even in the cases where
depth-induced breaking was not expected. In fact, the results of extensive hydraulic model
tests performed at MIT on deep-water breaking wave forces on vertical structures have
 H. Oumeraci/ Coastal Engineering 22 (I 994) 3-29 13

hown that impact pressures are comparable to those obtained in shallow water impact
studies, even though the probability of impact is much less in deep water (Chan, 1986).
Therefore, it is surprising that this "concept of reflective breakwaters" is still implicitely
admitted in recent textbooks, recommendations and codes of practice, despite the reported
failure cases and the new accumulated knowledge on waves and wave breaking. For instance,
the rules of PIANC implicitely restrict the application of vertical breakwaters to deeper
water (PIANC, 1976) whereas similar statements in well-established reference books
suggest that vertical breakwaters should be abandoned where wave breaking is expected
(see also Section "Breaking waves and wave impact loads" below).

Non-monolithicity o f the structure and weakness o f the concrete

The non-monolithicity of the structure has generally led to two principal modes of failures:
- Shoreward sliding of the superstructure and the underlying upper two or three layers of
unbonded blocks, essentially caused by the impact of breaking waves in this area. Depend-
ing on the prevailing depth and wave conditions, the last one or two layers on the rubble
mound foundation generally remained in place. This mode of failure often resulted in a
very wide breach in the breakwater;
- Collapse of the superstructure (generally seaward tilt) and development of a small breach
which is then widened by wave action. This generally occurred where differential settle-
ments were observed. A gap between the settled single blocks and the longer monolithic
superstructure could develop and, as a result the latter was undermined.
This mode of failure will not be considered further below, since the type of structure
using unbonded blocks has long been abondoned. However, care must be taken when
designing the superstructure of caisson breakwaters in order to avoid the shearing of the
superstructure which may be induced by the impact of waves or by armour units. The latter
case has often been observed on armoured vertical breakwaters (Hattori et al., 1984).
The failure mechanisms due the degradation of the concrete by repetitive shocks consist
in a gradual breaking up of the end faces in the wave impact zone, so that at certain locations
the superstructure was undermined and the underlying blocks broken up. As a result, small
breaches developed. This mode of failure will not be further discussed here, since these
problems have now been overcome by the recent caisson and concrete technology.

Low structure crest and high toe berm

According to the PIANC recommendations (PIANC, 1976) the free board should be 1.3
to 1.5H u (Hu ---design wave height related to the limit state of use = Hz~/1o) whereas the toe
berm should be at least at - 1.72Hr (Hr = design wave height related to the limit-state of
rupture=Hmax). According to the PIANC notations Hzl/lO is the average heigth of the
highest one thenth of all wave heights Hz, as determined by the zero-up crossing method,
and Hmax is the maximum height that might be observed only once during a time interval
equal to about 1.5 to twice the return period considered. These requirements are valid for
"vertical breakwaters", according to the terminology used by PIANC.
By examining the structures which failed and which were built as "vertical breakwaters",
it can easily be seen that most of them had too high a toe berm; i.e. they are not "composite
breakwaters" which, according to the PIANC definition, are basically characterised by the
14 H. Oumeraci / Coastal Engineering 22 (1994) 3-29

15.0 m
sea side harbour side
+4.5m
-+O.Om ew,

a) Niigata Breakwater ( H S = 7 m, Tp = 13.5 s )

sea side + 6.5~ 11.0'. rn ~,

,~..... harbour side

-- 0 . 0 n"

b) A l g i e r s " b l u s t a p h = B r e a k w a t e r ( 1Is = 9 = , T !, = 13.5 s )

Fig. 4. Seawardtilt of vertical breakwaters.

fact that "waves are reflected by the wall when the water level is high, but break against
the wall or the rock fill slope when the water level is low" (PIANC, 1976).
On the other hand, most of the damaged structures had a low crest, and were hence
heavily overtopped. In this respect, a number of failures also occurred during construction
while the superstructure was not completed. As a result, heavy wave overtopping and
breaking on the structure took place which generally led to differential settlements, thus
resulting in the seaward tilt of the breakwater, irrespective of the type of structure (break-
waters of Madras, Valencia, Catania, Algiers and Niigata). Although these failure mecha-
nisms has often been attributed to seabed scour (Renaud, 1935, 1936) and to liquefaction
(Zen and Umehara, 1985), the actual reasons for this "abnormal" behaviour and the
"abnormal" forces which prevailed are still not understood (see also Sub-sections "Wave
overtopping" and "Failure of purely geotechnical nature" below). Two examples for the
seaward tilt of low crest vertical breakwaters are shown in Fig. 4. The significant wave
height H~ and the peak period Tp of the corresponding storm are also given.
For very wide caissons (e.g. head caissons), however, sliding appears to be the predom-
inant failure, when a superstructure is lacking or the crest elevation is too low.
 H. Oumeraci/ Coastal Engineering22 (1994) 3-29 15

3.3. Reasons inherent to hydraulic conditions and loads

Among the reasons due to hydraulic influencing factors and loads there are the exceedance
of design wave conditions, the focusing of wave action at certain zones along the breakwater,
wave breaking, subsequent impact loads and wave overtopping.

Exceedance of design wave conditions

The vertical breakwaters built before World War II were generally designed for maximum
waves with heights and periods in the range of H = 5-6 m and T = 7 s, respectively. The
storms which caused the disasters, however, were characterised by H = 7-9 m and T = 11-
14 s at their peak phase. This represents an exceedance of the design wave conditions by
about 40 to 80% in terms of the wave heigth, 40 to 100% in terms of the wave period and
200 to 600% in terms of the wave energy.
Even when considering the safety factor of 1.3-1.5 commonly used at that time, it is
obvious that in most cases, the design wave conditions were largely exceeded.
On the other hand, this was not likely the case for most of the caisson breakwaters built
after the fifties. In fact, the storms which injured the breakwaters were generally in the same
range or even slightly less severe than the design storm. This was particularly the case in
Japan where the design principles were based on tests with regular waves, using the concept
of maximum wave height.
This means that for the "old breakwaters" the design wave conditions were exceeded,
due to the lack of wave data and reliable methods of extreme value and long term statistics.
On the other hand, the modem caisson structures failed, rather due to the exceedance of
certain "critical loads" without any apparent exceedance of the design wave; i.e. the latter
failures are merely due to the inadequacy of the concept of "design wave height" which is
still being in use.
Actually, the large experience now available by using irregular waves for model testing
suggests that it is not the highest wave that necessarily causes the extreme impact load, but
rather a rare combination of a succession of certain wave heights and periods. Indeed, the
latter greatly affects the shape of the breaking wave just at the very moment before hitting
the wall which in turn influences the magnitude and duration of the impact load.

Concentration o f wave action at certain zones along the breakwater

The following observations were made on many of the old and more recent breakwaters:
-During severe storms, wave action was generally focussed at certain zones along the
breakwater where the wave height exhibits an increase in the range up to 20% as compared
to the other sections;
- sometimes, a number of small breaches developed at regularly distributed distances from
eachother, the first breach starting generally near the head. Very often, these breaches
correspond to the locations where wave concentrations were observed;
- in other cases, the concentration of wave action was characterised by the occurrence of a

huge breaking and overtopping wave acting over a large front width and suddenly opening
a breach of about the same width in the structure (for instance 150 m in the case of
Mustapha Breakwater) which is then widened by the following waves;
- undulated deformations of the structure in plan developed, the first maximum deformation
 16 H. Oumeraci / Coastal Engineering 22 (1994) 3-29

generally located at about L/10 from the head and the next maxima at about L/3 from
each other (L = wave length). Two examples of undulated deformations of vertical break-
waters are shown in Fig. 5. In a number of other cases, however, these maxima were not
so regurlarly distributed along the breakwater;
- generally, the areas of maximum erosion of the rubble mound foundation and maximum
seabed scour were observed to approximately follow the same distribution as the maximum
deformations of the structure along the breakwater (Renaud, 1935, 1936; Hattori et al.,
1984).
A diffraction theory was advanced by Larras (1942) which appears to satisfactorily
explain the undulated deformations observed for instance at the Mustapha Breakwater, as
well as the location of the breaches developed along the Genoa breakwater. However, this
theory fails in a number of other cases, especially in those where no regularity occurred in
the distribution of deformations and breaches along the structure. In these cases, the reasons

Mean wave
direction

Initialjetty
alignment I Breakwater
0 / 50 100 ~o ~ I ~/head
, z I I , , z l i I

(9oOE
_¢
I ,,
Centme of displaced
caissons 7x7.5xl 3m
Q-
._~ -20~, after storm
e~
shoreward

a) S a k a t a jetty ( after Hat'taft et al, 1984 )

( wave length L -- 100 - 110 m )

Mean wave
direction

E
O
seaward
Alignment of ixeakwater
J
C
20- Initial alignment __
(9 I I of breakwater r"l II Breakwater
E 0 J J 50 / 100 150 ! "-, 200 a/ head
(9
O 0.0
o.
-20-

-40"

shoreward

b) Algiers" M u s t a p h a B r e a k w a t e r ( after Larras, 1942 )

( w a v e length L ,~ 90 - 100 )
Fig. 5. U n d u l a t e d d e f o r m a t i o n s o f vertical b r e a k w a t e r s .
 H. Oumeraci / Coastal Engineering 22 (1994) 3-29 17

were probably due to the unfavourable configuration of the bottom contour lines offshore
of the breakwater.
A further concentration of wave action was also observed at singular points of the
breakwater like heads, bounds and junctions between two different types of structure. For
instance, heads which were not reinforced as compared to the current cross-section generally
failed by sliding of the caisson or/and by toe erosion and seabed scour. Most of the failures
experienced by heads, however, occurred during construction, and the most predominant
modes of failure were sliding and seabed scour. Most of the failures experienced by bounds
also occurred in composite structures of the Japanese type (erosion of the rubble mound
protection).

Breaking waves and wave impact loads

Wave breaking and breaking clapotis represent the most frequent damage source of the
disasters experienced by vertical breakwaters. The recognition of this fact had led to the
development of extensive experimental research activities on impact loading of structures
subject to breaking waves. These activities are still going on, but they have not yet yield
complete and satisfactory results required for the development of methods to evaluate design
wave loads to be used in dynamic analysis.
In fact, wave breaking was often observed during the most severe storms which preceded
the disasters, even in water depths where breaking (due to shoaling) was not expected at
all. Indeed, vertical breakwaters were designed at that time following the "concept of
reflective breakwaters" already discussed above. It is a little surprising that this concept
has been further adopted by the PIANC Commission, althoug it was supplemented by the
requirement that the water depth at the vertical structure should be at least about twice the
expected wave height (PIANC, 1976). As a matter of fact, the latter requirement may help
to considerably reduce the frequency of occurrence of plunging breakers at the structure,
but not to completely eliminate them. Owing to the high irregularity of the sea during
extreme storms and to the aforementioned influencing factors, plunging breakers cannot be
definitely excluded. It is probably useful on this occasion to quote d'Arrigo (1955) who
summed up the experience of past failures with the reminder: "The free reflection from
vertical face is a delicate experience in nature not devot of peril".
The common modes of failure caused by the horizontal wave load due to plunging
breakers are sliding, shear failure of the foundation, but rarely overturning (Goda, 1973).

Wave overtopping
The observations reported so far show that all the structures were heavily overtopped by
the wave(s) which provoked the collapse. A number of suggestions have been advanced
to explain the mechanisms which had led to the collapse of monolithic structures under
overtopping conditions (Miche, 1933; Lira, 1935, 1936), but they rather appear of specu-
lative nature. Actually, the failure mechanisms which are induced under such conditions
are still not fully understood.
In fact, the relatively low crested structures allows a large amount of wave energy to be
transmitted by overtopping, thus enhancing the stability against sliding and overturning
towards the shore side. On the other hand, the observed failures suggest that this overtopping
would have produced "abnormal forces" which are prejeducial to the stability against
 18 H. Oumeraci / Coastal Engineering 22 (1994) 3-29

seaward tilt. Although the nature of these forces and the mechanisms leading to this mode
of collapse have not yet been investigated, it is suggested that at least the following forces
would have contributed to the failure of low crest breakwaters:
- Suction drag on the monolithic structure as well as in the foundation during wave rush
down;
- wave impacts on the deck of the breakwater which result in a high vertical eccentric force.
This effect increases after the initiation of the tilting process.
These aspects are believed to be of considerable importance, especially during construc-
tion. In fact, the period between placing the caissons and casting on site the superstructure
is generally long enough to allow a relatively severe storm to occur before the superstructure
is completed. On the other hand, the design criteria yet available (PIANC, 1976) to assess
the crest level might be totally revised by accounting for these aspects.

3.4. Reasons inherent to the foundation and seabed morphology

Among the factors which have contributed to the failures reported to date the following
may be mentioned:
- Unfavourable configuration of the contour lines of the seabed in front and offshore of the
breakwater,
- seabed scour and erosion of the rubble mound foundation,
- settlement of the structure and shear failure of the foundation (failure of geotechnical
nature).

Configuration of bottom contour lines

Due to unfavourable configurations of the bathymetry in front and offshore of the break-
water, the longest and highest waves in the spectrum were observed to focus on certain
zones along the breakwater, resulting in an increased wave breaking and overtopping, and
in the initiation of failure at these points. In this respect, Goda (1985) suggested "to build
the breakwaters cutting obliquely across the bottom contour lines in order to considerably
reduce the probability of being exposed to dangerous impulsive breaking wave load". This,
however, did not always prevent from failure some of the Japanese caisson breakwaters
which were built accordingly (Hattori et al., 1984).

Seabed scour and erosion of toe berm

Under the worst conditions ( soft soil, breaking waves), the scour depth in front of vertical
breakwaters may reach values up to 0.7 times the initial water depth (PIANC, 1976). The
failure of Mustapha Breakwater may well illustrate seabed scour as one of the reasons which
have essentially contributed to the failure of the breakwaters which tilted seaward. In fact,
this failure case is representative for further structures built on soft soils without any proper
scour protection. One of the major reasons why the latter was practically neglected is that
the period of the design wave - - and hence the bottom velocity - - was largely underesti-
mated (Oumeraci and Tabet, 1987).
Bottom scour has not only contributed to the failure of vertical breakwaters, but also to
those of amoured vertical breakwaters (Hattori et al., 1984).
Unfortunately, seabed scour has been considered in the past as the sole reason for seaward
 H. Oumeraci / Coastal Engineering 22 (1994) 3-29 19

tilting. As a result, the reasons which actually initiated the damage (exceedance of bearing
capacity, stepwise failure etc.) were overlooked.
Scour also took place at a number of Japanese breakwaters, generally resulting in the
dislocation of the slopes of the rubble mound foundation. However, repairs were generally
made before the stability of the monolithic structure was threatened. This was made possible
by the practice of providing quite broad berms and gentle slopes of the rubble mound
foundation. This practice, however, did not prove successful during severe storms, since
the more frequent wave breaking induced herewith diminishes any benefit conferred by
these countermeasures. According to Goda (1985), no effective method has yet been found
to stop seabed scour.
The design of the rubble mound foundation and toe berm has always been a difficult task,
since the dimensions (height and width) and further characteristics (block weight, slope
steepness) of the berm are generally obtained from an optimisation procedure which
accounts for a number of conflicting factors and effects (hydraulic stability, scour protection,
generation of wave breaking, construction aspects). Therefore, it is not surprising that the
instability of the rubble mound foundation has on many occasions contributed to failure.
The failure mechanisms are generally characterised by:
(i) the erosion of the upper layers of the toe berm and its slope,
(ii) the sliding of the berm slope into a scour hole developed at its toe, and
(iii) the shear failure which was initiated at the seaward or/and shoreward edge of the
monolithic structure.

Failures o f purely geotechnical nature

It is very unfortunate that not enough attention has been devoted in the past to geotechnical
design aspects, as this was the case for instance for the hydraulic and structural problems
inherent to the monolithic structure itself. Even the PIANC Commission did "not feel
competent to examine the soil mechanics problems involved in breakwaters" (PIANC,
1976). In fact, the stability of the beating soil has usually been analysed as a static problem,
neglecting the effects of the transient and cyclic loads. In order to better understand the
modes of failure which have been observed to date, one should bear in mind that:
- the impact and cyclic wave loads are transferred to the rubble mound foundation and the
seabed through the rocking and swaying motions of the monolithic structure;
- t h e cyclic loads are directly transferred by the wave motion into the rubble mound
foundation and the subsoil;
- the only resistance (granular soils) opposed to these loads is provided by friction, inertia
and damping forces,
- under the impact and cyclic loads of the waves, the saturated soils may experience large
unacceptable permanent deformations in the following two possible ways:
(i) the soil loses progressively most of its strength (pore pressure build-up after each
loading cycle), resulting in large permanent deformations (liquefaction) ;
(i) during the short time in which the induced stress peak exceeds some admissible
strength, small permanent deformations occur. The repetitive stress peaks during a
storm will then lead to large irreversible deformations without any pore pressure build-
up between successive loadings.
However, due the low frequency of the cyclic loads and the drainage condition of a
20 H. Oumeraci / Coastal Engineering 22 (I 994) 3-29

normal sand bed, it is generally unlikely that a significant pore pressure build-up of the pore
pressure will take place. In this respect, results of model tests were described by Oumeraci
(1989), showing simultaneously measured wave impact loads on a concrete slope, the pore
pressure and soil deformations induced in 10 cm depth (fine sand) underneath the concrete
slab revetment. It was shown that despite unfavourable drainage conditions there was a
summing-up of the soil deformation but no build-up of pore pressure between successive
impacts.
The general settlements which have been observed to date at vertical breakwaters also
took place essentially during the storms. This is illustrated by the principle sketch in Fig. 6.
On the other hand, the foundation very often settles differently on the shore and sea side,
mainly due to the intermittent action of the waves hitting the structure front and receding
from the wall. This generally results in slip failures characterised by the fact that the
geomaterial is expelled out from underneath of the base of the foundation two-sidedly over
curvilinear rupture surfaces. This is for instance illustrated by the failure of the Mustapha
Breakwater (Fig. 7).
Both breakwater toes were pushed out and up, clearly indicating that a shear failure of
the foundation should have developed after some settlements of the structure had occurred.
The mechanisms responsible for this failure may be explained as follows.
Under the action of standing and breaking waves, the breakwater exhibits forced and free
oscillations, respectively. The amplitudes of the forced oscillations of the Mustapha Break-
water evaluated by means of acceleration measurements are in the range of some centimetres

Storm 1 Storm2 Storm3 Time

~ - Settlement without any Storm

i=
oE
S1' S2' S3 : Settlementdue to Storm1,2,and3,respectively

Fig. 6. Settlement of vertical breakwaters.

SWL

~- slip surfaces ~ /
Fig. 7. Shear failure of the foundation of Algiers' Mustapha Breakwater.
 H. Oumeraci / Coastal Engineering 22 (1994) 3-29 21

for incident waves of 3-4 m height and 7-9 s period (Renaud, 1935, 1936; Kulmatch,
1958). Accelerations of caisson breakwater oscillations due to breaking wave impacts in
the range of some dm/s 2 have also been measured in Japan (Muraki, 1966), the natural
frequency of the caisson breakwater being in the order of 5 Hz. Common vertical monolithic
breakwaters have a natural frequency of about 2 to 20 Hz.
It is obvious that the soil beneath the monolithic structure will settle under the action of
such rocking motions. Since the highest amplitudes of these motions are located at the
seaward and shoreward edges of the structure, the latter necessarily transfer this repeated
percussive loads to the rubble mound foundation.
The load transmitted to the rubble mound foundation by these cyclic percussions was
evaluated to about 100 t/m 2 just beneath the edges of the structure (Minikin, 1963;
McCornick, 1962) which is excessive as compared to the generally allowed limits of 40-
50 t/m 2 (PIANC, 1976; Goda, 1985; Kobayashi et al., 1987). This, together with the
suction action of the receding waves which eroded the rubble mound toe on the seaside,
has brought the structure to tilt seaward rather than shoreward.
Depending on the magnitude of the horizontal wave loads and vertical forces, on the
caisson width and depth of the rubble mound foundation as well as on the strength of the
subsoil, the rupture surface may be located:
(i) directly at the base of the monolithic structures,

..._JI li I1~ & I,

I
I:
I i rubblemounO
ll',,il"
I |~ I ~i ,
I ~ ~ foundation , // ; / / ~. rubble mound

subsoil shear surfaces

a) Sliding b) Shallow Slip Failure

it,"i'" .I',. ',/I

~ - - -- " ' - - ' ~ " ~ - - ~ " ~ ' ~ foundation ~
1 -'>~ ~' / _ / 1'~ rubble mound

. X . . . . . . . . . . . . . . . . . . . . . ~ " - '...' . .' . '.....'..... . . . . . . . . . . . . . . . . ".'.'.'.'.'-'.'.'-"-~'-~

... ~:~"s'u'~'l~:.
subsoil .'.~¢j'~,~.""J" shear surfaces ~ ..,;,~r,-¢ra

c) Deep Slip Failure d) Overall Slip Failure

Fig. 8. Sliding and slip failure modes.
22 H, Oumeraci / Coastal Engineering 22 (1994) 3-29

(ii) within the mound foundation and at the boundary between the mound and the subsoil
with a higher strength,
(iii) within the mound which is so deep that the subsoil has not any influence on the shear
failure patterns, and
(iv) within the mound and the subsoil.
These modes of failure are shown in Fig. 8.
In view of the complexity of the phenomena leading to failures of geotechnical nature, it
is obvious that conventional bearing capacity calculations are of limited use to explain the
failure mechanisms observed. Indeed, more reliable models are needed to simulate the
dynamic behaviour of the structure-foundation-water system for complex loads, boundary
conditions and soil characteristics.

4. Lessons learned

The failure cases reported so far not only well illustrate when, how and why things may
go wrong, but also what can be done to reduce such occurrences. The principal lessons
which can be learned from such mishaps are given below.

4.1. Lessons related to design wave conditions, breaking waves and wave loads

(a) The criteria of "design wave height" which are still in use (PIANC, 1976) must be
abondoned and replaced by design wave loads criteria. Thereby, account must be made for
the single load which, when exceeded, will cause a sudden failure ( "first excursion failure" )
as well as for the repetitive loads ("fatigue failure") which gradually weaken the foun-
dation, thus leading to a reduction of the threshold value of the single load assessed. The
design loads should be specified for two limit-states: (i) a limit state of use which corre-
sponds to more or less repairable damage, and (ii) a limit-state of rupture which corresponds
to the failure either of the whole structure or of one of its essential components. However,
one should bear in mind that the most realistic design criteria can only be provided by a
"design load history" resulting from a representative sea storm history.
(b) Care should be taken by using extreme-value and long-term statistics for assessing
the design conditions. Extrapolation of a few years' data to a period of life time of the
structure or, as often seen in the literature, to periods of 500-1000 years, should carefully
be evaluated before application. In this respect, engineering intuition and professional
judgement are still more required. Generally, l 0 years' data cannot be extrapolated to more
than 30-50 years. Therefore, when building a vertical breakwater in areas where wave
heights are not depth limited, it is indispensable to use all possible data (ship observations,
wave measurements, wheather charts etc) and methods of analysis (mathematical models
for wave hindcast, wave refraction etc) available in order to obtain the most realistic extreme
wave conditions. Gumbel' s reminder' 'The improbable is bound to happen one day !" should
always be kept in mind when designing vertical breakwaters in deep water, since the latter
are very sensitive to the exceedance of design conditions.
(c) Because of the high irregularity and short-crestedness of the sea during very severe
storms and a number of further influencing parameters, it appears difficult:
 H. Oumeraci/ Coastal Engineering 22 (1994) 3-29 23

- t o avoid totally well-developed plunging breakers ("mass breaking"). The results of

recent investigations have shown, and experience in offshore activities has confirmed that
this is true, even in deep-water;
- to determine the less dangerous location of a vertical breakwater in shallow water, since
the location of incipient breaking as related to the position of the wall continuously varies
according to the changes of the wave characteristics; i.e. smaller waves can be more
detrimental to the structure than higher design waves.
(d) Breaking waves should be recognized as being the most significant damage source,
clearly calling for:
- emphasis on the development of prediction methods which are appropriate for this type
of dynamic loading and the inferred response of the structure-foundation system. A better
description of the loads by specifying the temporal and spatial distribution of the impact
and uplift pressure is needed for this purpose;
- constructional measures to reduce the wave loads and their effects on the structure and its
foundation.
(e) The nature of the failures reported in general, and the geotechnical failures in
particular, suggest that the concept of "equivalent static load" can only be used as a starting
basis for preliminary design evaluation and decision. For final design purposes, a more
realistic evaluation of the dynamic loading and response of the structure and its foundation
is needed.
(f) Care should be taken in the design and construction of the link between the super-
structure and the placed caissons, since this link is directly located in the impact zone.
(g) The "loading rate" of the structure and its foundation during the storm appears to
be an important parameter to be considered in the design. Most of the severe storms have a
short violent ascending phase and a longer descending phase, and the major failures were
observed to be initiated during the ascending phase and at the peak of the storm.

4.2. Lessons related to breakwater layout

(a) A careful planning of the layout of the breakwater, as related to the configuration of
the bottom contour lines and wave directions, is as important as the design of the cross-
section for the stability of the structure. In this respect, numerical modelling (far and near
field) as well as 3D-hydraulic stability model testing should be performed to determine the
optimal layout.
(b) Singular points (heads, bounds, junctions, toes) are weak points and should always
be carefully designed and constructed.
(c) An experimental verification of the reasons for the undulated deformations of the
breakwater in plan, and for the occurrence of small breaches at regularly distributed locations
along the structure should be undertaken in order to define the countermeasures required
for the reinforcement of the most critical sections.

4.3. Lessons related to overtopping effects

(a) For excessively overtopped structures low crest structure, the most common mode
of failure is seaward tilt. In this respect, care should particularly be taken during construction,
24 H. Oumeraci / Coastal Engineering 22 (1994) 3-29

while the superstructure is still not completed. The different phases of the work should be
carefully planned, and the uncompleted work should be protected accordingly. This aspect
is very important because the construction generally lasts more than a single season.
(b) A further aspect which in certain cases may be important is the action of the upward
projected water mass due to wave breaking on a vertical front. In fact, depending on the
water depth prevailing at the structure when the waves recede, the falling water mass may
strike the toe protection and cause considerable erosion. Eventual facilities and installations
on the superstructure may also be damaged (Gaillard, 1904).

4.4. Lessons related to seabed scour and geotechnical aspects

(a) Seabed scour may represent an important damage source which should properly be
accounted for in the design stage.
(b) Since a number of failures initiated by seabed scour occurred during construction,
scour protection should keep pace with the progress of work, despite the extra cost which
would result. The protection of the uncompleted work should be carefully planned, espe-
cially at the head.
(c) The effect of seabed scour is twofold since it may lead to the gradual dislocation of
the rubble mound foundation and modify the waves and currents at the breakwater toe.
Because of the progressive nature of the failure, a follow up procedure is needed to monitor
seabed scour, and to define alarm and threshold values for the extent of scour which does
not threaten the stability of the structure.
(d) In some circumstances, an underestimate in design wave periods may be even more
detrimental than that of design wave heights.
(e) The rough guidance generally found in codes of practice and textbooks that scour
protection should consist of a blanket of 0.2-0.5 t stones covering the sea bottom to a
distance of 1/4 to 3/4 of the expected wave length in front of the vertical wall does not
appear to be the most economically feasible solution and should be revised according to the
more recent knowledge (Xie, 1981 ).
(f) It is not likely that the large permanent deformations of the subsoil observed to date
underneath vertical breakwaters are due to liquefaction. Indeed, it is an accumulation of
small irreversible strains at repetitive peak stresses rather than a pore pressure build up after
each load cycle which occurs; i.e. the importance of liquefaction due to wave loading has
been overemphasized in the relatively recent publications.
(g) During the storms which provoked the collapse of the vertical breakwaters, the
admissible bearing capacity of the rubble mound foundation (40-50 t / m 2) was generally
exceeded at the edges of the monolithic structures. As a result, shear failure generally
developed at both shore and seaside.
(h) The consideration of the geotechnical aspects in the design of vertical breakwaters
should not be limited to the conventional static analysis of the bearing capacity of the
foundation. The impact and cyclic wave loads should also be accounted for. For this purpose,
the simulation of a realistic (representative) storm history rather than a series of large waves
successively striking the structure is needed. This will, in fact, allow the structure-foun-
dation system to perform relatively small motions before the onset of the largest waves.
 H. Oumeraci / Coastal Engineering 22 (1994) 3-29 25

4.5. Lessons related to monitoring, inspection and maintenance

(a) An adequate monitoring, inspection and maintenance programme may be an impor-

tant measure to possibly prevente catastrophic failures. Breakwaters in deep water generally
represent very large investments comparable to those of large dams, but built with a much
higher risk. It is therefore surprising why the surveillance and monitoring of vertical break-
waters have been almost ignored, even if the consequences are much smaller (generally no
loss of life). In fact, before the actual collapse occurs, it is generally preceded by small
disturbances or reparable damages (settlements, seabed scour, toe berm erosion, slight
tilting) induced by previous less violent storms. In some cases, the collapse would certainly
have been prevented if monitoring, inspection and subsequent maintenance had been under-
taken. This is confirmed by a number of failure cases where the damage was only restricted
to certain sections weakened by previous storms. An example for the reinforcement of
certain vertical breakwater sections which slightly tilted seaward before a severe storm will
occur is shown in Fig. 9.
This may clearly stress the importance of monitoring, inspection and repair as a possible
measure to prevent catastrophic failures. Beside the verification of the integrity of the
structure, the prediction of incipient failures and the determination of the repair specifica-
tions, a well-planned monitoring and inspection procedure is also expected to provide a
better understanding of the behaviour of the structure-foundation system through various
damaging loading conditions. In addition, this may also help to develop a tool for the
evaluation of the reliability of existing structure (feedback into further design and optimi-
zation procedures).
(b) Since it is technically and economically much better to adapt or reinforce an undis-
turbed vertical breakwater than to rebuild a collapsed one, and owing to the knowledge and
experience accumulated to date, the question arises whether the old existing breakwaters
(which did not yet experience the severest possible storm) should be adapted, or would
this be disproportionate compared to the risk that would be taken by leaving things as they
are. An adequate procedure should be developed to properly answer this question.

+6.5m
~1~ ~ 3.0 m Harbourside
± 0.0 ~,SWL Seaside

Fig. 9. Reinforcementof the verticalbreakwaterMustaphain Algiers.

 OVERALL FAILURE MODES

::::::::::::::::::::::::::::::::::::::::::::::: ::::::::::::::::::::::::::::::::::::::::::::::: ::::::::::::::::::::::::::::::::::::::::::::::: :::::::::::::::::::::::::::::::::::::::::::::::

a) Sliding b) Overturning c) Settlement followed by silp d) Settlement followed by slip e~

failure & seaward tilt failure & shoreward tilt
2.

LOCAL FAILURE MODES e~

2.

4~

P
:;::..::::;:::..;::;::;:;::.,::..;::;;:;:..::::::::. :;;::;':::::::;::;::::::::::;::..:;.,;:.::::::;:. t,,a

e) Erosion beneath seaward & f) Punching failure at seaward & g) Seabed scour & toe erosion
ahomward e d g ~ ahoreward e d g ~

Fig. 10. Modes of failure of monolithic vertical breakwaters.

 H. Oumeraci/ Coastal Engineering 22 (1994) 3-29 27

4.6. Lessons related to the design approach

(a) It is necessary to up-date the recommendations of the PIANC Commission (PIANC,

1976). In particular, the concepts of "vertical breakwaters" and "composite breakwaters"
should be revised and adapted to the recent developments in waves, wave breaking and
wave forces.
(b) Owing to:
- the very random nature of the waves during severe storms,
- the stochastic and transient nature of the dynamic loads involved,
- t h e large variation of the dynamic characteristics of the subsoil and of the structure-
foundation-water system,
- the stochastic nature of the dynamic response of the structure and its foundation, and
- the large number of possible failure mechanisms (Fig. 1O) and their interaction,
it is obvious that a probabilistic design approach based on the dynamic analysis of the
structure and its foundation will represent the most reliable solution to the stability problem
of vertical breakwaters.

5. Concluding remarks and research needs

It is questionable whether only one of the afore mentioned reasons of failure can by itself
explain the respective types of collapse observed after the storm. The nature and sequences
of these collapses not only illustrate why a number of reasons have together led to the
catastrophic failures, but also reveal how truly complex the various failure mechanisms and
their interaction are. Breakwater instability is indeed an integrated problem which cannot
simply be reduced to sliding and overturning stability calculations by using static equivalent
loads.
Among other things, methods should be developed to account for the small progressive
failures (settlement, seabed scour, toe berm erosion, slight tilting) which are generally the
precursors to the ultimate collapse.
In developing design tools for vertical breakwaters, one should bear in mind that most of
uncertainties originate from the difficulties to predict the design wave load conditions, as
well as the dynamic soil characteristics which are also the most predominant factors affecting
the stability of the structure. Subsequently, the development of improved methods to eval-
uate these factors should become the highest research priority.
Further urgent research needs are the development of reliable breaking criteria for waves
in front of vertical structures, the evaluation of the "abnormal" forces acting on low crest
structures, the elaboration of prediction tools for the extent of seabed scour and its effect
on the stability of the structure.
In the light of the modes of failures reported to date and their interaction, dynamic analysis
and probabilistic design approaches seems to be the only feasible solution to the integrated
stability problem of vertical breakwaters.
28 H. Oumeraci/ Coastal Engineering 22 (1994) 3 29

Acknowledgements

T h i s s t u d y is p a r t o f a r e s e a r c h p r o g r a m m e o n b r e a k w a t e r s w i t h i n t h e C o a s t a l E n g i n e e r i n g
R e s e a r c h U n i t " S F B 2 0 5 " at t h e U n i v e r s i t y o f H a n n o v e r w h i c h is s u p p o r t e d b y t h e G e r m a n
Research Council (DFG), Bonn. Additional support by the European Community within
the Research Programme "MAST G6-0032" is a l s o g r a t e f u l l y a c k n o w l e d g e d .

References

Benezit, V. and Renaud, P.J.M., 1935. Bauweise senkrechter Dfimme. In: XVlth PIANC Congress, Brussels, S. 2
Q. 2, Paper No. 76.
Cavanel, R., 1961. Port de G~nes: DEgats caus6s ~. la digue verticale de protection du port par la tempfite des 19-
20 Fevrier 1955. CR Soci6t6 Hydrotechn. de France. Q Vii, pp. 630-638.
Chan, E.S., 1986. Deep water breaking wave tbrces on structures. Sc.D. Thesis, Massachussetts Institute of
Technology.
D'Arrigo, A., 1955. The recent damage to the Genoa b r e a k w a t e r - the effect of sea conditions on vertical walls.
Dock Harbour Auth.: 54-56 ( see also discussion by Dobbie, pp. 225-226).
Franco, L., 1991. Vertical breakwaters: the Italian experience and lessons from failures. In: Proc. First Proiect
Workshop - - MASTG6-S, Project 2, Hannover.
Franco, L. and Passoni, G., 1992. Further analysis on the failure of Naples breakwater. In: Proc. Third Project
Workshop - - MASTG6-S. Project 2, Hannover,
Gaillard, D.D., 1904. Wave action in relation to engineering structures. Professional papers No. 31, US Army
Corps of Eng.
Goda, Y., 1973. A new method of wave pressure calaculation for the design of composite breakwater. Rep. Port
Harbour Res. Inst., 12( 3 ): 31-70 ( in Japanese).
Goda, Y., 1985. Random Seas and Design of Maritime Structures. University of Tokyo Press, Tokyo.
Grimaldi, M., 1955. The disastrous cyclone damage at Genoa - - Causes of the breakwater failures examined.
Dock Harbour Auth.: 117-120 (see also discussion by Larras, p. 193, Minikin, pp. 164-165 and Dobbie, pp.
225-226).
Hattori, C., Shibata, K. and Ohori, K., 1984. Disasters of breakwaters by wave action ( Part 3 ). Techn. Note, No.
485, Port and Harbour Research Institute ( PHR1 ), pp. 1-28 ( in Japanese).
Kobayashi, M., Terashi. M., Takahashi, K., Nakashima, K. and Odani, H., 1987. Bearing capacity of a rubble
mound supporting a gravity structure. PHRI Report No. 5, Vol. 26. pp. 215-252.
Kulmatch, P.P., 1958. On the stiffness of soil foundation of massive structures. Gidrotechn. Stroitelstvo No. 4,
pp. 35-38 (in Russian).
Larras, J., 1942. La deformation ondulatoire des digues verticales. Travaux, No. 108.
Lira, J.. 1935. Bauweise senkrechter Dfimme. In: XVlth PIANC Congress, Brussels, S. 2 Q. 2, Paper No. 67.
Lundgren, H., Sorensen, T. and Agerschou, H.. 1983. Breakwater design. In: H. Agerschou et al. tEditors),
Planning and Design of Ports and Marine Terminals. Wiley, Chichester.
McCornick, H.F., 1962. Dock and Harbour Engineering. Charles Griffin and Co Ltd,. London, Vol. 2, Ch. I I.
Miche, M.R., 1933, Les digues maritimes du type vertical. Sci. Ind., (January): 23-27.
Minikin, R.R., 1963. Wind, Waves and Maritime Structures. 2nd ed. Griffin, London.
Muraki, Y., 1966. Field investigations on the oscillations of breakwaters caused by wave action. Coastal Eng.
Jpn,, 9: 97-96.
Oumeraci, H., 1989. Beanspruchung yon Betonplattendeckwerken und ihre Beriicksichtigung bei der Bemessung.
Mitteilungsbl. BAW, Karlsruhe, No. 66.
Oumeraci, H. and Tabet, E.H.. 1987. Extension of the port of Algiers: Caisson vs. rubble mound solution for the
main breakwater. In: Proc. 2nd COPEDEC, Beijing, Vol. I, pp. 662-676.
Oumeraci, H., Partenscky, H.W., Tautenhain, E. and Nickels, H., 1991. Large-scale model investigations: a
contribution to the revival of vertical breakwaters. In: Proc. Conf. on Coastal Structures and Breakwaters, Inst.
Civ. Eng., London, pp. 907-220.
 H. Oumeraci / Coastal Engineering 22 (1994) 3-29 29

Oumeraci, H., Partenscky, H.W. and Tautenhain, E., 1992. Breaking wave loads on gravity structures. In: 2nd Int.
Conf. ISOPE, San Francisco, CA, Vol. III, pp. 532-539.
PIANC, 1976. Final report of the International Commission for the study of waves. Annex to Bull. PIANC No.
25, Vol. !II.
Renaud, P.J.M., 1935, 1936. La jetde Mustapha au Port d'Algier. Ann. Ponts ChaussEes, 553-586 ( 1935 ), 751-
828 (1936).
Takahashi, S., Tanimoto, K. and Shimosako, K., 1992. Experimental study of impulsive pressure on composite
breakwaters. Report of the Port and Habour Research Institute (PHRI) Vol. 31, No. 5 ( Dec. ), Tokyo.
Tanimoto, K. and Takahashi, S., 1994. Design and construction of caisson breakwaters - - The Japanese experience.
Coastal Eng., 22: 57-77.
Tanimoto, K., Takahashi, S. and Kimura, K., 1987. Structure and hydraulic characteristics of breakwaters - - the
state of the art of breakwater design in Japan. Report PHRI Vol. 26, pp. 11-55.
Tanimoto, K. and Goda, Y., 1991. Historical development of breakwater structures in the world. In: Proc. Conf.
on Coastal Structures and Breakwaters, Inst. Civ. Eng., London, pp. 193-206.
Xie, S.L., 1981. Scouring patterns in front of vertical breakwaters and their influences on the stability of the
foundations of the breakwaters. Report, Dept. Civil Eng., Delft Univ. of Technol., pp. 1-61.
Zen, K. and Umehara, Y., 1986. Mishap of a breakwater due to wave-induced liquefaction - - Techn. Note, Port
and Harbour Research Institute, Ministry of Transport, Japan (in Japanese).