You are on page 1of 15

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/275640204

RIPRAP STABILITY STUDIES ON DAMS AND DYKES OF THE LA GRANDE


COMPLEX (Phase I)

Conference Paper · September 1993

CITATIONS READS

0 379

5 authors, including:

Jean-Pierre Tournier
Hydro-Québec
25 PUBLICATIONS 137 CITATIONS

SEE PROFILE

All content following this page was uploaded by Jean-Pierre Tournier on 29 April 2015.

The user has requested enhancement of the downloaded file.


1993 CDSA/CANCOLD Conference
Sept. 19-23, St.John's, Nfld

RIPRAP STABILITY STUDIES ON DAMS AND DYKES


OF THE LA GRANDE COMPLEX (Phase I)

Jerry Levay Jean-Pierre Tournier


Société d'énergie de la Baie James, 500 boul. René-Lévesque Ouest
Montréal, Québec H2Z 1Z9

Raymond Arès Denis LeBoeuf


Rousseau, Sauvé, Warren Inc., 500 boul. René-Lévesque Ouest
Montréal, Québec H2Z 1W7

ABSTRACT

Since impounding of the La Grande-2 reservoir in 1978 and the filling of the Opinaca, La Grande-
3, La Grande-4 and Caniapiscau reservoirs, the riprap of several of the dams and dykes has suffered
some damage during the 8 to 12 year period which followed. A mandate was given to the Société
d'énergie de la Baie James (SEBJ) in January 1992 to review the riprap design, taking into account
the actual conditions of dams and dykes and to evaluate the repairs to be done using the different
techniques presently available. This paper describes the geotechnical aspects of the study carried
out by SEBJ. The hydraulic aspects are treated in a separate companion paper.

RÉSUMÉ

Depuis le remplissage du réservoir La Grande-2 en 1978 et celui des réservoirs Opinaca, La Grande-3, La
Grande-4 et Caniapiscau dans le nord du Québec, le riprap de plusieurs digues et barrages a subi des
dommages au cours des 8 à 12 années qui ont suivi. La Société d'énergie de la Baie James (SEBJ) a
reçu en janvier 1992 le mandat de revoir la conception du riprap, en tenant compte de l'état actuel
des digues et barrages, et d'évaluer les réfections à entreprendre en utilisant les diverses
techniques présentement disponibles. La communication présente les aspects géotechniques de l'étude
effectuée par la SEBJ. Les aspects hydrauliques de ces études sont présentés dans un article séparé à
cette conférence.

Dam Safety 1993 137


INTRODUCTION

The construction of the various structures of the La Grande Complex (Phase I), in northern Québec
(Figure 1) was done over a period of twelve (12) years, between May 1973 (the start of the diversion
tunnel excavation at La Grande-2) and December 1985 (the commissioning of the last group at La
Grande-4). This work was preceded by preliminary exploration and the construction of an access road
that began two (2) years before. The total cost, including interest during construction was 13,7
billion dollars (Canadian); 10,6 billion for the support facilities and hydroelectric structures and 3,1
billion for the transmission system.

The Complex required the building of 215 dams and dykes, 149 of which have waterheads over 1 m in
height. Since the reservoir filling at La Grande-2 in 1978, followed by the Opinaca Reservoir in
1979, La Grande-3 and Caniapiscau in 1981, La Grande-4 and the Fontanges Basin in 1983, the riprap
protection of some of the structures has been damaged and several have required repairs. In all, 19
structures required various repairs over the years involving the replacement of riprap with
quantities varying from a few hundred cubic metres to several thousand cubic metres in a few cases.

In January 1992, Hydro-Québec gave the Société d'énergie de la Baie James a mandate to review
the criteria for determining the size of riprap and to evaluate different methods of repair from the
techniques available.

An intense field campaign was undertaken in 1992 which was crucial to the study. Analysis of the
historical data and the studies which followed the field investigation allowed the condition

138 Dam Safety 1993


of the structures to be evaluated, the nature of the hydrological solicitations to be determined and the
amount and methods of repair to be evaluated.

This paper presents the geotechnical aspects of the study done by SERI which allowed criteria to be
established for the structures to be repaired and the determination of stone size for the repairs. Also,
criteria are proposed for the riprap of future structures. The hydraulic aspects of the study are
presented in an accompanying paper entitled "The Hydraulics of Riprap Design Applied to the Reparation
of Dams and Dykes of the La Grande Complex (Phase I)".

SLOPE PROTECTION ASSESSMENT

Study Methods

The assessment of the riprap protection was done mainly in two ways: by an analysis of
existing construction and surveillance documents and by a field investigation done in the summer
of 1992.

The object of the file study was to verify the design methods and the condition of the protection after
construction, where possible, and to obtain information on the damage reported over the years. The
field studies were carried out in the summer of 1992 to obtain the actual condition of the riprap. This
was done in the following ways:

• by inspection of the damaged zones. This allowed the failure mechanism of the riprap to be
determined and to evaluate the amount of repairs to be done;
• by taking above water profiles of 37 structures and underwater profiles of 14 structures. These
surveys were done to determine the actual profile of the structures;
• by taking measurements of the D50 size of 79 structures (both damaged and undamaged; in all,
over 250 D 50 samples have been taken to date). This work was done to determine the D50 size that
was representative of the different upstream zones as they exist presently;
• by doing photointerpretation of the upstream faces of 67 structures. This analysis allowed
the upstream faces of the studied structures to be classed in zones according to the D50 size;
• by doing acoustic imagery of the submerged slopes of the important structures at La Grande-2, 3
and 4. The investigation was done using side scan radar to evaluate the level of damage below the
water surface.

Characteristics of Slope Protection

In general, the retaining structures on the Complex used two different methods to protect the upstream
faces, one for homogeneous and sand and gravel dykes and another for rockfill dykes.

In the case of homogeneous or sand and gravel dykes, characterized by flat upstream slopes varying
between 2,1 and 2,5 H to 1,0 V, a layer of selected stone was placed on a bedding layer to form the
riprap protection. The rockfill dams and dykes were characterized by steep upstream slopes of 1,7 H to 1,0
V, which in some cases went to 1,5 H to 1,0 V because of the

Dam Safety 1993 139


camber. For the rockfill dykes at La Grande-2 and several dykes on the Caniapiscau reservoir, the
upstream protection was formed with standard dumped riprap, with maximum stone size, to protect
the underlying transition layers, from the crest to 3 m below the minimum reservoir level. For some
dams and dykes, such as the main dam at La Grande-3 and La Grande-4, random rockfill was
specified with the larger sizes to be pushed to the outside of the fill to form the protective layer.

Wind and Wave Solicitation

A measurement of the waves and wind was done to evaluate the wave regime on the reservoirs of the
Complex. Two reservoirs were studied in 1992. The La Grande-3 reservoir represented a regularly
contoured reservoir and the Caniapiscau reservoir represented an irregular contour with many islands.
The results and conclusions of the study are presented in a separate paper and indicate that the
factor used to increase the velocity of the wind over water was underestimated in the original
design for wave heights.

SUMMARY OF THE STUDY RESULTS

Original Design and Specifications

The upstream protection of the retaining structures of the La Grande Complex was, in general, achieved
by selecting stone size which could resist the design waves. The thickness of this protective layer
was of the order of 1,5 times the average stone size diameter. The riprap was placed on a cushion of tr
ansition material made up of random or screened rockfill which served as a filter for the underlying fill.

The riprap was designed by computing the significant wave height (He) using the SMB wave prediction
method as presented in the 1966, 1975 and 1977 versions of the Shore Protection Manuel (SPM) and
the Bertram-Taylor method to determine the stone size. The effective fetch, as proposed by T. Saville, is
used to compute the significant wave height. For the stone size, t h e d e si g n wa v e (H d ) w as
consider ed t o be e qua l t o t he sig n i fi c a n t wa v e h e i g h t H, (Hd = 1,0 Hg), where H, was
calculated using a 1/1000 year wind storm.

The riprap was specified as a maximum weight (W.), average weight (W 50 ) and minimum weight
(Wm;,,), where the W50 corresponded to the calculated value, W. equivalent to 4W50 and Wmin , to 0,
25W 50 . The stone for the riprap was to be strong, weather resistant and free of visible
weaknesses (such as joints, fractures, etc.). In certain cases, so as not to vary the specifications,
a stone size for the most solicitated structure was used for all the structures in a particular
construction package; also, although not always followed, an arbitrary weight of 1000 pounds was
adopted as a minimum W50 to take into account the effect of ice on the riprap.

The specifications for the riprap were applied uniformly, with only a few exceptions, from the crest to
approximately 3 m below the minimum reservoir level.

140 Dam Safety 1993


The riprap specifications were sometimes relaxed, for economical reasons, since because the structures
were of the rockfill type, it was felt that by specifying random rockfill, the larger stones could be
pushed to the outside to form a good protective layer.

Riprap Performance

Evaluation of the Riprap Performance

Due to the nature of the protection used for the slope protection (specified stone size), it is normal
to have a dynamic evolution of the riprap and to have some local failures. More important
localized damage can eventually occur due to zones of weakness developed during placement, caused
principally by poor interlocking and the presence of voids. This type of disorder shows up during
the first filling of the reservoir and when important wave action occurs and should be corrected by
normal riprap maintenance. If the disorder becomes worse or becomes generalized, it is due to poor
placement, the presence of fines, the concentration of undersized stone or possibly undersized riprap.
This leads to more or less important local or general repairs.

A study of the file showed that up to the end of the summer of 1992, some 94 000 m 3 of rockfill
was used to replace eroded rockfill. This quantity was placed essentially on 12 structures.
Aside from La Grande-3, the average quantity of material placed was equivalent to 3 m3 per metre of
repair. In the case of La Grande-3, 9 m3 per metre was used on the north dam and 23 m 3 per metre on
dyke TA-13 which needed important repairs over a length of 175 m in 1988.

Performance of the Structures as a Function of Damage

As a preliminary analysis, it is possible to appreciate the performance of the riprap according to the
amount of damage. From a review of information obtained from the Hydro-Québec files and the
inspection of the structures in 1992, the global performance of the riprap can be represented
according to 3 simple classification levels. The first level (class 1) represents an excellent performance
and is used for structures with no damage or very minor local damage. The second level (class 2)
represents structures where the upstream slope suffered intermediate to important, but local, damage an
d the performance is considered acceptable. The third level (class 3) corresponds to structures where the
upstream face sustained generalized or systematic damage.

This system of classification was applied to all of the structures on the Complex. The structures in class
1 are in good condition and require only normal maintenance. Class 2 structures require local
repair. The structures grouped in class 3 are structures that have systematic damage and may
reflect undersized riprap or riprap that contained a large amount of fine material. Structures which
were poorly repaired because the lower limit of the protection was not stabilized are also included in
this class. Ten (10) structures on the Complex fall into this last category.

Dam Safety 1993 141


The study of the riprap performance, according to the levels of damage, showed that 92 % of the
structures (excluding the freeboard dykes) had an excellent to acceptable performance, that is 98% of
the homogeneous and sand and gravel dykes (flat slopes) and 82% of the rockfill dams and dykes (
steep slopes). Also, 81% of the rockfill used for the riprap repairs by Hydro-Quebec, up until 1992, was
used on rockfill dams and dykes.

Performance of the Riprap as a Function of Wave Solicitation

The maximum wave sustained (H,) was calculated for each structure for which the D50 size was known.
For structures that had systematic damage, the H, was calculated for the period before the repair and
another H, for the period after the repair.

A correlation was made between the Hs and the D50 of all the structures for which the D50 was available.
This information was plotted on a graph for rockfill structures (figure 2) and on another for
homogeneous and sand and gravel structures (figure 3). Because the D 5 0 measurements were
taken after the maximum H, occurred, each of the points of figures 2 and 3 represents a D 50 size that
has resisted the calculated wave height that occurred to date. Therefore, the lower envelope of
points represents the best available information of the minimal D 50 size that can resist a given wave
solicitation. These results show that the Bertram-Taylor formula (for deep water) is a good
approximation for determining the D50 minimal size.

Figure 4 compares the maximum H, sustained with the H, of the revised design and allows the severity of
the solicitation to be judged. For the majority of cases, the waves sustained by the structures are situated
between 50 and 80% of the revised design waves. For comparison, figure 5 compares the waves
sustained by the structures to the original design waves. For several structures, it can be noted that
the wave sustained by the structure is equal to and even surpasses the wave calculated in the original
design.

It can be noted on figure 6 that conventional criteria for the minimum and maximum values for weight (
W m i n = 0,125W 5 0 and W m a x = 4W 5 0 ) of the stone making up the riprap lead to a material
with a very broad grain size and include a significant amount of material with a diameter smaller
than required for the W50. A uniform riprap is preferable.

Model Test

Model testing was done at the National Research Council (NRC) in Ottawa on two typical
structures (dyke TA-13 on a scale of 1 to 10 and dam KA-3 on a scale of 1 to 15) which
represent structures with flat (2,1 H and 2,5 H to 1,0 V) and steep slopes (1,5 H and 1,7 H to 1,0 V).
Irregular waves were produced to simulate natural conditions in terms of significant wave height and
period. The tests allowed the validation and optimization of the configurations proposed for the repairs.

The pertinent data gained from the tests are the following:

• the use of the Hudson formula to calculate the stone size was corroborated by the tests;

142 Dam Safety 1993


Dam Safety 1993 143
144 Dam Safety 1993
Dam Safety 1993 145
146 Dam Safety 1993
• the repairs using a thin riprap protection should be extended to a depth below the
solicitation level which is equal to at least 2 Hs (where Hg is the significant wave height);
• a berm, with appropriate stone size, dumped or placed at the foot of a thin riprap cover, is
rearranged gradually by wave action, acting at different levels of attack, to form a stable protection.

Main Conclusions

The main conclusions reached concerning the behaviour of the structures are as follows:
• for a majority of the structures, the maximum significant wave height, since reservoir filling,
reached 50 to 80% of the revised design wave, and for many of the structures this height corresponds to
100% of the original design wave;
• 92% of the structures (excluding the freeboard dykes) have a performance varying from
acceptable to excellent, that is 98% of the homogeneous and sand and gravel dykes (flat slopes) and 82% of
the rockfill dams and dykes (steep slopes);
• generally, the most important damage was caused by the presence of fine material (less than 300
mm) in the riprap;
• poor interlocking, poorly graded riprap having mostly large and small sizes with very little W50
sizes and the presences of voids in the riprap was also a contributing factor;
• in a few cases, the riprap was undersized or did not meet the specifications.

The reasons that the repairs done over the years were not completely successful in limiting the damage are
as follows:

• the riprap was not repaired to depths which covered the zone of most frequent attack completely;
• damage to repaired zones was due to the undermining of the repair material since the stone
size of the repair material was generally adequate.

Finally, regarding the design calculations, the following conclusions were drawn:

• the method proposed in SPM-77 and used for the Phase I calculations, with the effective fetch (T.
Saville), is appropriate for calculating waves on the reservoirs of the La Grande Complex;
• the Bertram-Taylor formula used in Phase I for the stone size calculation is a good approximation
of the minimum average weight (W50) which can resist a given wave height;
• the original Hudson formula corresponding to a uniform riprap is appropriate for stone size
calculations for repair work.

REPAIR WORK

The decision to do repair work on a given structure is based on the present condition of the
structure, the historical performance of the riprap and on the requirement that the in-place riprap
meets the dimcnsions required for the revised design.

Dam Safety 1993 147


Criteria to Select Structures Requiring Repairs

The following criteria were used as a guide to select structures to be repaired;

• if a structure was highly solicitated, significant amounts of fines (D < 300 mm) were not
tolerated in the zone of frequent attack;
• the D50 of the in-place riprap should be superior to the minimum D50 required by Bertram-Taylor,
for deep water, using the revised H,;
• for structures where the historical or present (summer 1992) performance was
unsatisfactory or acceptable but with local damage, repairs were proposed;
• for structures already repaired, where the lower limit of the repairs was visibly unstable or
inadequate in the zone of frequent attack, repairs were proposed.

These criteria were applied while taking local conditions such as shallow water areas into consideration.
Dams and structures with steep slopes, in general, were treated with special attention due to their
importance and the fact that the mode of failure of the riprap was more severe and rapid and could
cause sliding or sloughing of the crest. Damage observed to structures with flatter slopes with
adequate protection was generally limited and evolved slowly.

Criteria to be used for Repairs

a) Design Wave
• for repairs, the design wave (H D ) corresponds to the significant wave height (H,)
generated by a 1/1000 year storm, using the SMB wave generation method as presented in the SPM-77 an
d the T. Saville method for effective fetch;

b) Stone Weight
• the weight of the stone for the repairs is calculated using the original Hudson formula which
corresponds to a uniform riprap with HD = 1,0 H, and a Kd value of 3,5;
• the diameter of the stone is calculated using the relation proposed by Hudson relating the D to
W.

c) Freeboard
• the freeboard is calculated considering Hd = 1,25 H, and a runup coefficient of 0,85 to 1,35
depending on the upstream slope of the structure and the wave steepness;

d) Minimum Repair Elevation


• repairs are done to a depth of at least 2,0 H, below the level of attack.

Repair Techniques

Following the verification of the in-place stone sizes and the determination of the damaged zones, it
was possible to group the repair work into two broad categories 1) local repairs and 2) systematic
repairs.

148 Dam Safety 1993


a) Local Repairs
Structures which had adequate protection but had minor local damage associated with weak zones
caused by placement (poor interlocking, presence of voids, concentration of fines) were proposed for
repair. These local repairs would consist essentially of repairing the damaged or weak zones by
rearranging the existing stone and adding stone of appropriate size.

b) Systematic Repairs
In cases where the structures have generalized damage and the riprap is, in whole or in part, undersized
according to the revised design, systematic repairs were proposed.

Several repair techniques were envisaged and studied as follows:

• the placing of a thin layer of riprap from the crest of the structure;
• the placing of a thin layer of riprap using a backhoe or a crane mounted on a barge for the
repairs that have to be done on the lower part of the drawdown zone;
• the construction of a berm with random or selected rockfill;
• the placing of gabion mats using barges for the installation.

Of the techniques studied, two (2) methods were retained:

• the placing of a thin riprap cover. Above the waterline, it is fairly easy with a backhoe to place
the rockfill and to interlock it for stability. However, below the waterline, this operation
becomes very difficult if not impossible because of poor visibility. To overcome this difficulty, a small
berm is built at the waterline to the maximum reach of the equipment so that the resulting dumped
material adequately protects the submerged slope. The stone of this small berm is rearranged by the wave
action to form an efficient, stable surface over the zone to be protected;

• the placing of a riprap in the form of a berm on the upstream face. This technique of repair
is in fact just the dumping and pushing of the material down the upstream face of the structure from
the crest or the building of a berm with a minimum width of seven (7) metres along the upstream face.

CONCLUSIONS

The results of the study and the application of the retained criteria led to the recommendation of
doing local repairs on some thirty (30) structures and general repairs on seventeen (17) structures
of the Complex (seven (7) of these are on the Caniapiscau Reservoir).

A vast majority of the retaining structures have had an acceptable performance since reservoir filling. In
general, most of the damage that occurred was due to the presence of excessive fines in the riprap. In
certain cases, where general repairs were required, this was due to the fact that the riprap
specifications were sometimes relaxed or the riprap was undersized because the wind velocity over water
was underestimated in the original design.

Dam Safety 1993 149


The study, which correlated the maximum significant wave that the riprap sustained with the average
diameter (D50) of the rockfill that resisted, indicates that the Bertram-Taylor formula is a good
approximation for determining the minimum stone size. However, in order to avoid stone sizes which
are widely graded and which risk having significant sizes which are smaller than the W 50 required, a
uniform riprap is recommended. The original Hudson formula corresponding to this type of riprap
is appropriate for the sizing of the repair stone and has been verified by the model tests.

It is important to recall, that the very nature of the type of material used for the protection of the
structures, implies a dynamic evolution of the material and that local minor damage will occur which
should be repaired during normal maintenance operations.

ACKNOWLEDGEMENTS

The authors would like to thank Hydro-Québec for permitting this information to be published and
Bernard Boncompain who worked on the mandate in the early stages.

REFERENCES

BERTRAM, G.E. 1951. "Slope protection for earth dams". Transactions of the 4th
Congress on Large Dams, ICOLD, New Delhi, India, Vol. 1 (Q.13R.26), 209221.

HUDSON, R.Y. 1959. "Laboratory investigation of rubble-mound breakwaters". Journal


of the Waterways and Harbours Division, ASCE, 85(3), 93-121.

Shore protection manual, 1977. 3rd Ed., U.S. Army Corps of Engineers, Coastal Engineering
Research Center, Washington, D.C.

TAYLOR, K.V. 1973. "Slope protection on earth and rockfill dams". Transactions of the 11th
Congress on Large Dams, ICOLD, Madrid, Spain, Vol. III (Q.42, R.13), 215235.

150 Dam Safety 1993

View publication stats

You might also like