Serious problem of dam seismic safety was tried to be solved by reinforced concrete belt-elements incorporated in the upper part of upstream zone & clay core. But due to large construction settlements of this zone & core this solution was useless & expensive

1. The problem of seismic stability and safety of 300 m high Nurek rockfill dam
For this project the method of analyses of seismic resistance of Nurek dam under action of the Maximum Credible
Earthquake (MCE) the return period of 1 in 10,000 years, was chosen the accelerogram with the peak acceleration of
1.1g registred in canyon rock ledges on the left abutment of the arch dam under seismic shaking Pocoina 1971 in San
Fernando (California, USA) with a magnitude of 6.6. According to US specialists a very high acceleration of the
accelerogram was explained by the influence of landscape, such as increased vibration of rock ledgess. Except for this
effect the accelerogram was recalculated due to its moving from the raised surface of canyon rock ledges down to the
smooth surface thus leading to the decrease of the acceleration from 1.12 to 0.73g in a frequency range of 1 to 15 Hz
and from 0.64 to 0.55g in the begining of recording low frequency.
It must be noted that for the credibility of this seismic (dynamic) analysis the earthfill dam (hydraulic type) San
Fernando Low, practically was destroyed by the earthquake of 1971. In seismic analysis the non-instrumental
accelerogram was used modified the earthquake Pocoima with the maximum acceleration up to (0.5-0.6)g, close to the
accelerograms that were recommended by the Institute of Physics of the Earth (Academy of Sciences of the USSR)
and other institutes engaged in dynamic analyses of seismic resistance of Nurek dam. So it was concluded that the
choice of ground motion with high accelerations Pocoima 1.1g for analysis of seismic resistance of Nurek dam
without considering the topographical difference, seismological and other conditions at the point of obtaining
seismological readings particularly taking into account the projections of the canyon and place of location of Nurek
dam (with an inverted canyon 300 m below the dam crest, with 1300 m wide, comparable with the length of seismic
waves its exceeding) is considered unsubstantiated.
Photo 1. Downstream view of 300 m high Nurek rockfill dam & hydropower plant (2700 MW)
Fig. 2. Cross-section of 300 m high Nurek rockfill dam with central clay core:
1 – sandy loam core; 2 – shells of coarse gravel & pebble; 3 - filters (d=0-5 mm & d=0-50 mm); 4 – oversize rockfill surcharge
stones on upstream slope (20-40 m thickness), downstream slope (5-10 m thickness) & downstream cofferdam; 5 – downstream
cofferdam of oversize rockfill; 6 – upstream cofferdam; 7 - concrete block; 8 – consolidation grout & grout curtain; 9 – anti-
seismic belts; 10 – control gallery 2x2 m; 11- control gallery 4x4 m on dam crest; 12 – contour of first-stage dam construction; 13
– temporary clay blanket of first-stage dam construction

2. For the definition of the characteristics of shear strength of the gravel-pebble materials of shoulders of Nurek dam,
landslide tests of 35 m high embankment were performed using explosions of gravel-pebble materials of the upstream
slope of the dam with trial vertical horizontal acceleration (Ah) and (Av) measured in the embankment control points
(reaching Ah = 2.5-3.0 m/sec,Av = 5.6 m/sec; angle of embankment slope to the horizont (30 grades in water and 20
grades above the water). The angle of soil shear ψ defined by the expression: ψ = arctan[Ah(α/g) + Av(α/g)], where:
Ah and Av are components of conventional acceleration (amplitude of harmonic accelerograms,equivalent to actual
or instrumental action); α= s/(γs-γa),where:γs & γa - density of soil and water (for the unsaturated slope γa=0 & α=1).
Presented expression is analogous to Mononobe formula (Japan) for the limitt angles (at equilibrium) of the slope of
the sandy materials under seismic action. Defined by this formula the ψ angles a number of assumptions was made
according to which the duration of the oscillations were activated with explosions took sufficient for performing large
displacements of the embankment and to decrease the inclination of the slope until equilibrium condition; cohesion of
soil materials under large movements caused by seismic action disappears; to define the angle ψ it was taken into
account only horizontal acceleration component of Ah. When two components of acceleration Ah and Av are taken
into account the definition of ψ angle was performed using these accelerograms.
However the assumed assumption is unfounded. In reality reading acceleration during test embankment under
explosion takes only 2-3 peaks and lasting (4-6 sec), which is not sufficient for the realization of the large
displacements and take the slope to a limit equilibrium state, in terms of continuity of action of the harmonic force.
The assumption in which under great movements caused by seismic actions cohesion in the soil materials disappears,
it is not correct and not checked by the test results. The method discussed in principle does not establish the overall
shear strength (cohesive) soils and can not give any information about the presence or absence of cohesion in soils
because the formula above mentioned for the definition of ψ angle not gives the possibility to define the angle of
internal friction phi (φ) and cohesion C separately.
With the values of θ (tetta)=30 grades, Ah=0.3g, Av=0.6g; γa (gamma)=0 and α (alpha)=1 (the unsaturated slope),
mentioned above, we got the shear angles of gravel-pebble soils equal to ψ =40.70
under very adverse conditions of
Av in downward direction; for ψ (ksi) =46.7 grades Av=0 (which is significantly higher than ψ (ksi)=41.6 grades,
obtained earlier. Under the same conditions Av=0 and ψ(ksi)=65.8 grades with Av in the upward direction. It can be
concluded that the acceleration Av has much more influence on the angles ψ(ksi) than Ah and it can not ignored.
Thus in the safety assessment of Nurek dam the unfavorable upward direction of Av must be considered.
From the above it can make a general conclusion that the right approach and realization field test investigations of
the shear strength properties of the gravel-pebble materials, processing and expansion of test results and also
authenticity of these results are presented doubtful. The reinforcement antiseismic elements dispersed in Nurek dam
are installed with great technical difficulties and it simply resulted in extension of the construction period. Moreover
under corresponding indices quality of work placement in the construction and compaction of these dam materials can
not be made as necessary with sufficiently high indices of mechanical properties (high density and porosity, low
deformability etc.) without using poor means of less costs and time.
In Fig. 3 & Photo 2 the location scheme of antiseismic latticed belts in form of reinforced concrete structures with a
height of 3 m, with a distance between them 18-21 m below the dam zone area is presented. The lower belt is located
65 m below the dam crest. Two intermediate belts are located in the middle of the upstream shoulder, the lower and
upper belts extend to the upstream slope area and superior also extends to the upstream slope. The three reinforced
concrete belts do not reach the upstream transition zone of clay core. This method of discretized bearing structures in
the upper part of the upstream dam shoulder covers the 3 m high zones with 18 to 21 m distance between them that
leads to considerable heterogeneity in these zones during dam construction, which will be taken into account in the
analyses of seismic resistance of the dam. Meanwhile the choice of the size and location of antiseismic belts in the
dam body were made very sensible during seismic analysis by the linear spectral theory by the deficit of supportive
forces (or moments) for the most dangerous sliding surface of analysis defined using the revision of the seismic slope
safety of the outstanding common methods: by the dynamic method with peak ground motion acceleration of 1.1g of
earthquake Pocoima in 1971 and using in dynamic analyses with the ideal plastic model with associate plastic flow by
Druker-Prager. The limit permissible values the residial (plastic) deformations of soils in the dam the results of the
seismic analysis showed that the plastic deformations (up to 3-5 % in contact zones) are spreading in contact zones
towards the upstream slope and crest accepted residual deformations of (Fig. 2).
Analogously conducted credible estimates of seismic resistence Nurek dam with antiseismic belts. Obviously, that
these estimates of homogeneous dam by the linear spectral method and dynamic theory the seismic resistance of this
test dam can not be get a reliable value for the assessment of seismic forces and safety of Nurek dam, taking into
account the presence in its upstream shoulder the heterogeneous zones of antiseismic belts. With these belts it can be
seen only worsening working conditions during dam construction in thse zones. It is necessary to note that in analysis
of seismic safetyof upstream slope of the Nurek dam were considering three antiseismic belts with their bottoms on
elevations of 274, 256 and 235 m and were obtained the minimum stability coefficients correspondingly equal to
1.06; 1.05 and 1.05 not enough to provide the seismic safety of Nurek dam according to Soviet design norms.
According to these norms the downstream dam slope should be inclined more to reach the allowable stability
coefficient of 1.32.

3. There is a large quantity and quality dispersion between seismic analyses results of Nurek dam: by the dynamic
method under the same the ground motion of Pocoima (Amax=1.1g) with the use of dynamic models of different soils.
As shown in Fig. 2(a, b) at the begining of dam design the its dynamic analysis was made with the use for the soil
model with the ideal plastic law of Druker-Prager and it was recommended to strengthen the upsream shoulder below
the dam crest with antiseismic belts (Fig. 2-b). Later the dynamic analysis of Nurek dam was performed using
Zaretsky plastic soil model with accounting plastic strengthening effect (ICOLD Bulletin 122, 2001). This analysis
showed another picture of dam seismic (dynamic) response (Fig. 4): with observed seismic displacements in the
downstream slope near the dam crest, in the middle of the core and in the significant part of the upstream shoulder
between elevations 235-274 m (Fig. 4).It means that antiseismic belts are not working and have no influence on the
the dam seismic stability contrary to results of previous seismic dam analysis. These results indicate to the importance
of the proper choice of the soil dynamic model and defining their parametors in dynamic analysis and also the
inefficiency of proposed antiseismic measures to ensure the seismic resistance of Nurek dam.
The conclusions about the low seismic resistance of Nurek dam inspite of its antiseismic belts is the result of
increase not based on the analyses of seismic actions, accepted in the design stage,the significant reduction of the
characteristics of shear strength of dam materials and using simpler methods to calculate seismic resistance.
From the results of the dynamic analyses of the stress-strain state under increased accelerations in dam foundation
with use of two separate dynamic models is seen that the reinforcing belts in the zone under the dam crest do not
influence on the volumetric displacements in this zone with the deep sliding surface. (Fig. 4).
Fig. 2. Isolines of plastic deformations (%) in the Nurek dam before (a) & after construction of antiseismic belts (b)
according to the results of seismic (dynamic) analysis using the ideal plastic model of Druker-Prager for dam soils
Fig. 3. Location scheme of 3 antiseismic belts in the upstream shoulder of Nurek dam:
1 – upper belt (el. 214 m); 2 – upper belt in the zone 3; 3 – rip-rap (rockfill) upstreamslope protection; 4 – gravel-pebble
soil of both dam shoulders; 5 – clay core; 6 – fine & coarse filters; 7 – upper control gallery; 8 – gravel
Fig. 4. Scheme of distribution of plastic deformations ( %) in Nurek dam after construction of antiseismic belts
according to the results of seismic (dynamic) analysis using the ideal plastic model of Druker-Prager for soils

4. Photo 2. View from the left bank on construction of the lower antiseismic belt (el. 235 m) in Nurek dam
Fig. 5. Histogram of tension stresses (MPa) in reinforcement of antiseismic belts in upstream shoulder of Nurek dam
Nurek reservoir induced seismicity
The Nurek reservoir extends 70 km upstream with a maximum width of 5 km, depth of 215 m, a volume 10.5 cubic
km. Reservoir filling has taken place in several stages, the water level rising as construction increased the dam height.
Nurek is the only reservoir with induced seismicity where a detailed knowledge of the seismological conditions
existed prior to reservoir impoundment. This has made it possible to determine the extent of changes in the rate of
seismic activity, related to the filling of the reservoir and furthermore to study changes in other aspects of the seismic
regimen a threefold increase in the rate of seismicity related to filling, changes in local mechanisms, a decrease in
focal depths, and a migration of activity towards to reservoir.
The filling of a large reservoir changes the stress regimen and decreases effective stresses by increasing the pore
pressure. The net effect on fault zones of increased or decreased stability, depends strongly on the orientation of
preexisting stresses and the geometry of the reservoir fault system. The Tadjik depression is a broad zone of horizontal
compression and focal mechanisms are generally of a thrust type. The induced seismicity at Nurek occurs on a series
of thrust or reverse faults showing strike-slip motion. A gradual rise in pore pressure was likely the dominant factor in
triggering the earthquake southwest of the reservoir in 1971 at distances of 10-15 km.
Earthquakes as large as magnitude 4.6 have been induced by the filling of Nurek reservoir. Raising the water level
above its previous primes the area for increased seismicity. Once primed the timing of activity is controlled by
changes in reservoir filling: if filling remains high, seismicity remains low. Decreases in filling rate cause increased
seismicity: the more rapid and sustained, the more intense the increases. The largest earthquakes occurred after the
daily rate of filling reportedly was decreased by more than 0.5 m/day.