Failure of embankment dams as a result of explosive attack can have serious ramifications for the people and environment downstream of the structure. Understanding how explosives affect these dams can help owners and operators develop mitigation strategies.
By Abass Braimah and Mohammad Rayhani
Dams are considered critical infrastructure under International Humanitarian Law because of the massive effect a breach or failure could have on the population and environment. For example, the failure of South Fork Dam in 1889 caused more than 2,200 deaths in Johnstown, Pa., USA.1 A similar failure could cause greater casualties and economic losses because of the many cities in the flood plains of dams. Were Aswan Dam in Egypt to fail, it would cause fatality to a large percentage of Egypt’s population while sweeping many structures into the Mediterranean Sea.
The most common causes of dam failure are design errors, geological instability, extreme inflow, sub-standard materials/techniques and poor maintenance. However, an increase in global terrorism has imposed another potential cause of failure. The destruction of dams has the potential to cause catastrophic floods and put strain on electrical power supplies, which in turn could have adverse effects on many national critical infrastructure sectors. Thus, terrorists and other groups seeking to cause mass casualties or economic disruption or attract media attention could target dams.
The effects of explosives on dams are not well-understood, and most dam owners/engineers lack the expertise to estimate the vulnerability of dam infrastructure to such attacks. Knowledge of these effects will undoubtedly lead to implementation of mitigation strategies aimed at limiting damage to dam infrastructure and its consequential effects on downstream communities and the electrical power system. Although little test data is available to aid the dam owner/designer in identifying vulnerabilities and establishing the amount of explosives that can compromise the many critical dam infrastructure systems, data from attacks on dams during conflicts or wars are available. Investigations into these attacks provide insight into the robustness of dams to attacks with various amounts of explosives at a number of locations on the dam structure.
Attacks on dams
A literature review reveals only a few attempts by terrorists or violent protestors to attack dam infrastructure. One dam targeted by terrorists is Chingaza in Colombia, which supplies water to the city of Bogota. The dam is gravel fill with a concrete face and impounds a reservoir with a capacity of 223 million cubic meters. The Revolutionary Armed Forces of Colombia (FARC) have attacked several dams and aqueducts. In January 2002, FARC detonated an explosive device in a gate valve inside a tunnel in Chingaza Dam in an attempt to breach the structure and disrupt water supply to Bogota and flood the city of Villavicencio downstream of the dam. However, the attack was not successful and the dam was not breached.
In July 2011, the Indian Army intercepted a terrorist threat to Bhakra Nangal Dam in Himachal Pradesh, India, a concrete gravity structure that holds the distinction of being the world’s highest straight gravity dam. The dam, on the Satluj River, impounds water for a 1,325 MW hydroelectric project. Reports indicated two terrorist groups were planning strikes during the monsoon, when the water level behind the dam is at its highest, to cause maximum damage to downstream structures and people. Luckily, no such attack was carried out.
With regard to embankment dams, the below case studies present information on two attacks and the accompanying damage.
Sorpe Dam in North Rhine-Westphalia, Germany, was constructed from 1922 to 1935. It is an earthfill embankment with a watertight concrete core wall. The dam crest is 700 m long, and the height above the valley floor is 60 m, with a maximum water depth of 57 m. The dam is 10 m wide at the crest and 307 m wide at the base, with 1:2.25 and 1:2.50 slopes of the upstream and downstream faces, respectively.2
|This dry model embankment dam was built to help further understanding of the effects of explosives on embankment dams.|
Sorpe Dam was attacked during Operation Chastise (Dambusters Raid) of WWII. The air attack was carried out on May 17, 1943, by a squadron of the Royal Air Force using the Upkeep (bouncing) Bomb developed for that purpose. This bomb was 1.5 m long and 1.3 m in diameter and was filled with 3,600 kg of RDX explosive.3 Sorpe Dam suffered two hits on its crest that resulted in craters about 12 m deep. The dam structure, however, did not fail. Shortly after the Dambusters Raid, the water level in Sorpe Reservoir was lowered as a precautionary measure.
The attacks on Sorpe Dam were repeated several times in 1944 by the Allied Forces. On October 16, 1944, an attack on Sorpe Dam resulted in 11 direct hits. Although 12 m-deep and 25 to 30 m-diameter craters were formed, the embankment dam was not breached. The second wave of attacks used 5,500 kg “Tallboy” bombs, which also did not breach the dam. After the war, the dam was repaired and the craters filled. However, in the 1950s serious leakage problems were discovered and more remedial work was authorized. In 1958, when the dam was partially drained for repairs, an unexploded Tallboy was found buried in the embankment dam and removed. Sorpe Dam was repaired and put back into service and remains in operation today.
Peruca Dam is on the Cetina River in the Republic of Croatia. It is a 425 m-long and 60 m-high rockfill dam with a convex upstream axis.4 Peruca Dam was constructed between 1955 and 1960.5
During the Balkan wars, the area around the dam was occupied by the Yugoslav Army and later by Serbian Forces, who planted 20 to 30 tons of TNT in five locations in the walls of the spillway structure and inspection gallery. On several occasions, Serbian Forces threatened to destroy the dam using these explosives and closed the spillway gate, power tunnels and outlets and maintained the water level as high as possible so that maximum damage would result.4,5
In 1992, a United Nations Protection Force took control of the dam and lowered the water level to the design elevation. On January 28, 1993, Serbian Forces retook the dam and detonated the explosives in an attempt to destroy Peruca Dam and flood the villages and hydroelectric power station downstream.4
But the effects of the detonations were not sufficient to breach the dam or lead to overtopping and subsequent erosion and failure of the structure. The explosions did leave two large craters at the abutments. The elevation of the rim of the crater was only 300 mm from the water level in the reservoir. The inspection gallery was also heavily damaged at the abutments and completely destroyed at the deepest section in the middle of the dam. A total settlement of about 1.55 m of the crest was recorded about ten months after the explosion.
The explosives did not breach the dam because the three entrances of the inspection gallery were left open, thus venting the confined gas pressure from the explosion. Also, the reservoir was about 5 m lower than the level at which Serbian Forces intended to detonate the explosives. But for the actions of the protection force, the dam would have been breached, causing about 12,000 deaths and about 60,000 displaced people. The population would have been without electricity, and all farming lands in the low-lying valley would have been flooded.5
Rehabilitation of Peruca Dam began immediately after the attack. The craters at the abutments were filled and the bottom outlet was opened to begin lowering the head pond. This relieved the damaged dam from the hydrostatic pressure and seepage. The rehabilitation continued for about three years. By April 1996, the head pond was filled to the design level and the hydroelectric station was put back into regular operation.
The damage sustained by a dam and related infrastructure during an attack with an explosive device depends on the type of dam, the type of explosive device and amount of charge used, and its placement on or relative to the infrastructure. When an explosive device is detonated, a short-lived fireball is formed, followed by a blast wave that travels omnidirectionally from the center of detonation. Depending on the center of detonation relative to the ground surface, a crater is formed, accompanied by ground shock/vibrations. If the explosive device is in a container, a portion of the energy is expended in fragmenting the container and projecting the fragments at high velocities. The blast wave also has a propensity to break up frangible structures in its path and throw them at high velocities.
The fireball from an explosion can ignite proximate flammable materials, while the ground shock/vibrations are of consequence for buried infrastructure systems. The blast wave exerts high-peak and short-duration pressures on structures in its path that can lead to failure, whereas the high-velocity fragments impart high-impulse impact loads to structures in their paths. The dimensions of explosive craters depend on the depth of burial or height of burst of the explosive relative to the ground surface. Crater size decreases with an increase in height of the burst and increases with an increase in the depth of burial.
When the explosive charge is buried or close to the ground surface, the blast/shock waves from the explosion travel through the earth mass and increase pore water pressure. This can lead to instability of embankment dams through decrease in effective pressure, piping and liquefaction of the dam material. Buried explosives are confined by the soil mass, and the detonation energy is immediately imparted into the soil, causing fracture of the rock and displacement of soil.
For underwater explosions, the over-pressure can be up to two orders of magnitude higher than that in air for the same scaled distance.6 Therefore, underwater contact charges can impart greater loads to a target surface. On the other hand, due to the confining effect of the water, particle motion will be reduced so that underwater crater formation is less effective than that by contact charges at ground/air interfaces.7
Mode of failure
The primary explosion-induced mode of failure of embankment dams is cratering and accompanying overtopping and erosion. The crater dimensions depend on the amount of explosive, location of the explosive charge relative to the body of the embankment (buried in the embankment, placed on the surface, or elevated above the surface), soil type, and moisture content of the soil. Crater formation on the crest of an embankment dam can lead to instantaneous overtopping depending on available freeboard, and the accompanying erosion and scour can cause serious damage to the structure. Crater and cavity formation on the downstream face of an embankment dam due to surface-placed or buried explosive charges can lead to piping and eventual failure. In conflict situations where military-type weapons can be used, standoff artillery with the capability of burrowing into the body of the embankment and detonating can cause much more severe cratering and damage to the dam.
The size of explosive crater has been the subject of research studies,8 and the U.S. Army developed a software program — Conventional Weapons Effect Program (ConWep) — for determining crater properties given charge mass and depth of burial. However, ConWep deals with cratering on flat ground surfaces and is not applicable to embankments, where the slopes will have a marked effect on crater size.
A preliminary research effort to investigate the effect of an embankment on crater size was carried out on a dry embankment shaped from native soil.9 The model dam was constructed by digging the sandy native soil of the test site away from ground level, leaving the native undisturbed topsoil to form the crest of the dam, while the slopes were shaped by excavation. The model dam, which represents a one-fifth scale of a 10 m-high embankment dam with 3:1 upstream slope and 2.5:1 downstream slope, was 33 m long and 2 m wide at the crest and 13 m wide at the base (see Figure 1). ANFO (Ammonium Nitrate and Fuel Oil) was used as the surface explosive charge (the charge placed on the crest of the dam). Also, the same amount of ANFO was detonated on the ground surface without embankments (infinite flat ground) and the crater sizes measured.
The average crater diameter on the dam crest was typically larger than that of the craters formed with equal masses of explosive on the infinite flat ground. The presence of the dam slope reduced the confining effects, thus a larger amount of ejecta was discharged toward the upstream and downstream faces of the dam. The craters on the crest were elliptical in shape, with the long axis across the dam crest and the short axis along the centerline. The depth of crater, however, was similar for both the crater on infinite flat ground and crater on the embankment dam. Comparison of the typical crater profiles resulting from the experiment with those predicted by ConWep showed that ConWep overestimates both crater diameter and depth (see Figure 2).9
Due to the fact that the dam material used in this study was uncompacted, native sandy soil that did not represent a typical embankment construction, further tests on a properly constructed dam with, possibly, a head of water is required to investigate the impact of explosives on behavior of embankment dams. Sieve analysis of the native sandy soil of the test site of the model embankment dam performed in accordance with ASTM D42210 indicates that the soil particles are mainly sand and gravel. The fine portion of the soil was less than 5% fine (passing the 0.075 mm sieve). The soil can mainly be classified as poorly graded sand in USCS classification system.11
Another physical model test was performed on an earthfill dam with a height of 3 m and a 2 m-wide crest to investigate the crater size.12 The fill material was nonbinding sands with grain size ranging from 0.02 to 2 mm. Blasts of one or more 200 gram charges (mostly TNT) were set off on the model dam, which was erected in the open. A crater with the natural slope of 36 degrees was found for all models tested. The laws of similitude were used to scale the explosive effect for charges up to 1,000 kg. All model tests showed that any leakage, no matter how insignificant, represents primary danger because it leads, without fail, to a dam break. The exception to this observation is if the leakage is stopped and the primary damage repaired immediately.
Effect of explosive shock on pore water pressure
Explosion-induced ground motion can cause increase in pore pressure that may take hours or days (depending on the soil properties) to dissipate. The increased pore pressure leads to a reduction in the effective stress of the soil and reduced shear strength.12,13 This reduced strength can adversely affect the stability and safety of embankment dams and slopes. A number of embankment dam failures have been reported due to adjacent blasting activities, including slope failure of Calaveras Dam in California.13 This and other examples indicate that blasting in or near embankment dams can significantly increase residual (excess) pore-water pressure and reduce stability.
Ground motions caused by explosives produce localized peak accelerations that can be several orders of magnitude greater than earthquake accelerations. When a buried charge is detonated, the rapid release of energy generates a compression wave that radiates away from the explosion and produces tensile hoop strains accompanied by intense radial compressive strains in the surrounding soil. The detonation pressure of commercial explosives ranges from 105 to 108 kPa, depending on the distance from the center of detonation. A shock wave propagating from the center of detonation typically has one sharp peak of acceleration with duration of a few milliseconds for rock and tens of milliseconds for soil.14 The ground motion frequency from blasting in most soils ranges from 6 to 9 Hz, but in loose saturated sands, silts and soft clays it can be as low as 2 Hz.13
For a deeply buried charge, most of the blast energy in the surrounding material is in the form of a compressional stress wave. When this wave reaches an interface such as the water table or ground surface, it is reflected into a tension wave. Even though the blast wave is a few milliseconds in duration, it can produce oscillatory ground motions lasting several seconds at locations several hundred meters away from the center of detonation. Blast-induced ground motion is affected by many factors, including soil type, charge weight and depth of burial. An increase in the amount of explosive charge leads to higher ground shock amplitude and lower frequencies. A buried explosion creates significantly greater ground vibration, and therefore higher excess pore water pressure, than surface or near-surface explosions. Explosions in saturated soils generally produce higher peak particle velocity and hence higher excess pore water pressure than those in partially saturated soils.
The blast-induced pore water pressure response in saturated soils can be explained through transient, residual and dissipation stages. The transient response is associated with the passage of blast-induced stress waves through the soil. After passage of the stress waves, a residual increase in pore-water pressure occurs in which the fluid phase responds elastically, while the response of soil particles is in plastic range. This residual increase in pore water pressure is important for dam safety. Dissipation of the pore-water pressure may take hours to days after the blast, depending on the soil hydraulic conductivity and thickness.
Extensive data are available on blasting and on the performance of certain structures subjected to blast vibrations. However, only limited information is available on the performance of embankment dams and other hydraulic structures subjected to blast loading. Most of the work on the effects of blast-induced pore water pressure increases on embankment dams has been from proximate explosions. Few have looked at the effects of explosion on either the crest of the dam or the upstream or downstream slopes of the embankment.
Researchers have studied blast-induced liquefaction where a complete loss of shear strength occurs as a consequence of reduced effective stress from increased residual pore-water pressure. Experimental laboratory and field studies of shock waves revealed that liquefaction did not occur in saturated sands with dry densities greater than 1.60 g/cm3.15 In addition, liquefaction did not occur in any saturated soil subjected to peak particle velocities less than 7 cm/s.16 Blast-induced residual pore water pressures occurred where the peak particle velocity exceeded 5 cm/s.17
A shear or compression strain of less than 0.01% is generally considered small enough to preclude generation of residual pore water pressures because the strains are in the elastic range.18 A shear strain of 0.01% corresponds to a peak transverse particle velocity of about 1 to 3 cm/s for cohesionless soils having shear wave velocities of 100 to 300 m/s, typical of saturated soils. For blasting with a single charge of 100 kg, the residual pore water pressures could occur out to about 50 to 100 m.12
1. Johnson, W.F., “History of the Johnstown Flood,” 1889, www.jaha.org/FloodMuseum/history.html.
2. Matich, F., “The Dams Raid: Breaking the Great Dams of Western Germany,” Talk to the Royal Canadian Military Institute, 2005.
3. “Operation Chastise: The Dams Raid,” Commonwealth War Graves Commission, Berkshire, UK, 2005, www.cwgc.org/admin/files/The%20Dams%20Raid.pdf.
4. Nonveiller, E., J. Rupcic, and Z. Sever, “War Damages and Recon-struction of Peruca Dam,” Journal of Geotechnical and Geoenvironmental Engineering, Volume 125, No. 4, April 1999, pages 280-288.
5. Vajic, N., “Protection of Dams in Armed Conflicts: The Peruca Case,” Journal of Chemists and Chemical Engineers, Volume 43, No. 3, 1994, pages 108-114.
6. Braimah, A., and E. Contestabile, “Historical Review: Bombings of Dams,” Canadian Dam Association Bulletin, Volume 20, No. 1. 2009, pages 12-21.
7. Long, W. and Y. Ho, “Blasting Engineering,” Metallurgical Industry Publishing House, 1996.
8. Ambrosini, R. D. et al., “Size of Craters Produced by Explosive Charges on or Above the Ground Surface,” Shock Waves, Volume 12, 2002. pages 69-78.
9. Li, Y., E. Contestabile, A. Braimah, and D. Wilson, “Preliminary Vulnerability of an Embankment Dam due to Explosions,” CERL Report 2007-01 (CF), 2007.
10. “Standard Test Method for Particle Size Analysis of Soils,” ASTM D 422, ASTM International, West Conshohocken, Pa., USA, 2008.
11. “Standard Practice for Classification of Soils for Engineering Purposes (Unified Soil Classification System),” ASTM D 2487, ASTM International, West Conshohocken, Pa., USA, 2008.
12. Charlie, W. A. et al., “Pore Water Pressure Increases in Soil and Rock from Underground Chemical and Nuclear Explosions,” Engineering Geology, Volume 43, No. 4, 1996, pages 225-236.
13. Bretz, T. E., “Soil Liquefaction Resulting from Blast Induced Spherical Stress Waves,” WL-TR-89-100, Weapons Laboratory, US Air Force Systems Command, 1990.
14. Stagg, K.G., and O.C. Zienkiewicz, “Rock Mechanics in Engineering Practice,” John Wiley and Sons, London, 1968.
15. Lyakhov, G.M. “Shock Waves in the Ground and the Dilatency of Water Saturated Sand,” Zhurnal Prikladnoy Mekhaniki i Tekhnicheskoy Fiziki, Volume 1, 1961, pages 38-46.
16. Puchkov, S.V., “Correlation between the Velocity of Seismic Oscillations of Particles and the Liquefaction Phenomenon of Water-saturated Sand,” in Issue No. 6, Problems of Engineering Seismology, translated by Consultants Bureau, New York, N.Y., USA, 1962.
17. Long, J.H., E.R. Ries, and A.P. Michalopoulos, “Potential for Liquefaction due to Construction Blasting,” Proceedings of the International Conferemce on Recent Advances in Geotechnical Engineering and Soil Dynamics, St. Louis, Mo., USA, 1981.
18. Dobry, R. et al., “Prediction of Pore Water Pressure Buildup and Liquefaction of Sands during Earthquakes by Cycle Strain Method,” Building Science Series 138, National Bureau of Standards, 1982.
Abass Braimah, PhD, P.Eng., is a professor of infrastructure protection and international security and Mohammad Rayhani, PhD, P.Eng., is a professor of geotechnical engineering in the Department of Civil and Environmental Engineering at Carleton University in Ontario, Canada.