🕑 Reading time: 1 minute
Dams play an extremely important role in flood control, power generation, water supply, and irrigation. However, they also pose significant threats to the downstream areas as a result of the water that they block. Despite the increasing safety of dams resulting from the improved engineering knowledge and better construction quality, a full non-risk guarantee is not possible, and accidents may occur owing to natural hazards or human actions.
A dam can fail due to structural deficiencies in the original design or by external events that trigger conditions that exceed dam capabilities. Dam failure results in severe flash flooding that affects the downstream communities, livestock, property, and infrastructure. International bodies and emergency management agencies have identified a lack of awareness of the effects of dam failure outside of the dam safety community.
This article describes the biggest dam failures in the history of mankind. Also, we will discuss, the design features, reasons for the dam failure, and the lessons learned from it.
1. St. Francis Dam
On 12 March 1928, one of the worst civil engineering disasters of the 20th century occurred with no warning. The St. Francis Dam, located 64 km northwest of Los Angeles, California, failed unexpectedly. Due to the failure of the dam, 47 million m3 of water was released into San Francis Quito Canyon. The result was devastating and over 400 people died due to the failure of the dam.
1.1 Design of the St. Francis Dam
The St. Francis Dam was designed and constructed by a prominent engineer, William Mulholland. The dam was initially designed as a stepped concrete gravity arch of 152 m radius and the initial design height was 53 m from the floor of the canyon. The resulting reservoir capacity would be approximately 38 million m3 of water.
Several uncalculated design modifications to the height of the dam were incorporated during construction. Due to this, the final height of the dam was 59 m with a reservoir capacity of 47 million m3. Construction of the dam began in 1924 and completed in 1926. However, the life of the dam was very short as it failed just after two years of operation.
1.2 Failure Causes
Although a definitive cause for the dam failure was never determined but various theories have been proposed. With theories such as sabotage, geology, and poor construction, several reasons for the failure have been investigated.
Eventually, it was concluded that the failure was entirely due to the geology of the site. The site location of the St. Francis Dam was unsuitable for concrete gravity dam. The varying geology between the two canyon walls and the inconsistent and unstable properties of the rock were unacceptable for a dam foundation.
Not only the geology of the site was in question but also the design and construction techniques were questionable. The failure to incorporate essential dam safety features into the design had resulted in the collapse.
A lack of uplift relief and expansion joints, along with drastic modifications of the dam during the construction had contributed to the collapse of the St. Francis Dam.
1.3 Lessons Learned
The following lessons were learned from the St. Francis Dam failure and it brought several changes in the way future dams were to be designed.
- Extensive geological surveys of potential dam sites must be an integral part of the design process.
- Uplift acting on the dam base is a major design consideration. The foundation and seepage prevention design should consider uplift force.
- New laws were enacted by the state of California that required any proposed dam design to be evaluated by an independent review panel before approval for construction.
2. Malpasset Dam
The failure of Malpasset Dam represented the first failure of an arch dam. The suddenness of the failure, given that nothing abnormal had been detected at the dam within the hours preceding the event added to the uncertainty.
2.1 Design of the Malpasset Dam
The Malpasset Dam was located in southern France. It was a double curvature arch dam with a maximum height of about 60 m and a crest length of about 223 m. The thickness of the concrete varied from 1.5 m at the crest to 6.8 m at the center of the base. The dam created a reservoir with an estimated total capacity of about 51 million m3, which was filled very slowly over a five-year period.
2.2 Failure Causes
At the time of the failure, the level of the water was 0.3 m below the spillway elevation. Incidentally, when the water level was 0.3 m, heavy rains occurred. This resulted in the rise of the water level by almost 4 m within three days.
It was decided to open the bottom outlet gate in the dam. This would permit a controlled release of water and prevent damage to a highway bridge under construction downstream of the dam. The response of the dam had been monitored intermittently during filling by surveying equipment located on the downstream face of the dam.
Within hours of the bottom outlet gate being opened on 2 December 1959, the dam failed without warning. The city of Frejus, 7 km downstream of the dam suffered heavy losses resulting from the release of the water and debris. More than 300 people died in the disaster. Blocks of concrete from the dam were washed as much as 1.5 km to downstream.
2.3 Lessons Learned
A number of valuable lessons that changed design and construction methods for future dams were deduced from the Malpasset Dam failure. Most notably, design and construction methods to mitigate the effects of uplift pressures in the foundation of the dam were developed. Further, the use of appropriately located and selected monitoring instruments increased after the dam failure.
3. Vajont Dam
On 9 October 1963, Vajont Dam, located in Italy, suffered a failure of its southern rock slope over an approximate length of 2 km. The result was devastating and five villages and 2,040 lives were lost due to the failure of the dam.
3.1 Design of the Vajont Dam
A 276 m high double-arched dam across the Vajont River valley in northern Italy was constructed between 1957 and 1960. The dam created a reservoir with an estimated capacity of about 169 million m3.
In 1959, concerns about potential slope stability problems in the reservoir were raised, which resulted in further analyses being undertaken. These studies confirmed that there was a slide problem. However, there was disagreement to the volume of material which would be involved in a slide. Researchers predicted that the range of volume of material involved in sliding was from relatively small volumes to 10 to 20 m deep volumes. The main reason behind the sliding problem was the magnitude of deep-seated movements associated with large fissure cracks.
3.2 Failure Causes
Recognition of potential slope stability problems resulted in the installation of a monitoring program in 1960. It was considered that the rate of movements of sliding could be controlled by raising and lowering the reservoir water level. Results from the monitoring program over the next three years, which relied primarily on devices to detect surface movements, confirmed a relationship between the reservoir level and the slide mass movement. However, it failed to tell about the rate at which the ultimate failure would have occurred.
On 9 October 1963, the southern rock slope of the reservoir failed over an approximately 2 km length. Movement rates of the slide mass of approximately 275 million m3 during the failure were 25 m/s. However, the typical rates recorded during the previous three years of monitoring were in the range of 1 cm/day to at most 20 cm/day on the day of the failure.
The massive slide mass came to rest approximately 360 m laterally and 140 m upward on the opposite bank of the reservoir. The top of the slide mass was 160 m above the crest of the arch dam.
At the time of the failure, the reservoir was about 66% full and contained an estimated 115 million m3 of water. The water level had been lowered in a controlled fashion by approximately 10 m in the preceding two weeks. As the slide mass plunged into the reservoir, the water was displaced over the dam crest in a stream estimated to be up to 245 m above crest level.
3.3 Lessons Learned
The massive slide into the Vajont reservoir yielded important lessons in regard to the analysis and monitoring of slope movements. The difficulty of predicting when a slide mass will accelerate or fail became evident. Also, the difficulty of estimating changes in states of stress and strength during sliding was reinforced.
4. Lower San Fernando Dam
The construction of Lower San Fernando Dam started in 1912 as part of a hydraulic fill reservoir system in San Fernando, California. The hydraulic fill was placed between 1912 and 1915. The material was excavated from the bottom of the reservoir and discharged through sluice pipes located at starter dikes on the upstream and downstream edges of the dam.
4.1 Design of the Lower San Fernando Dam
Lower San Fernando Dam was constructed at the height of 40 m as a hydraulic-fill earthen dam. The construction configuration resulted in upstream and downstream shells of sands and silts, and a central core region of silty clays.
A stratum of variable thickness between 3 and 5 m of weathered shale was placed on the top of the hydraulic fill material in 1916. Several additional layers of roller-compacted fill were placed between 1916 and 1930 and raised the dam to its final height of about 40 m. A roller-compacted berm was placed on the downstream side in 1940. The dam created a reservoir with an estimated capacity of about 25 million m3.
4.2 Failure Causes
On February 9, 1971, the San Fernando earthquake occurred with an estimated magnitude of 6.6 on the Richter scale. At the time of the earthquake, the water level in the reservoir was about 11 m below the crest. This reduced level was due to an earlier seismic stability analysis that imposed a minimum operating freeboard criterion of 6 m.
During and immediately after the earthquake, a major slide event involving the upstream slope and the upper part of the downstream slope occurred. As a result of the slide, a freeboard of about 1.5 m remained. Given the likelihood of further damage in the presence of after-shocks, 80,000 people living downstream of the dam were evacuated over a single day. As predicted, after-shocks resulted in the overflow of the dam and approximately 275 million m3 of water flowed to the downstream side. Due to the failure of dam, over 340 people died.
4.3 Lessons Learned
The disastrous slide of the San Fernando Dam had a major impact on future earth dam design and construction procedures. The availability of site-specific seismograph data permitted extensive analyses to be performed and showed the results for preparedness. Among other factors, the potential problems with hydraulic fill structures and the need for revised procedures for dynamic stability analyses of earth dams were developed after the San Fernando Dam failure.
5. Teton Dam
Teton Dam was an earth dam constructed between 1972 and 1975. The failure of the dam occurred in June 1976, when the water was being filled for the first time inside it. The flooding of the downstream regions after the failure of the dam resulted in the loss of 14 lives and caused an estimated loss of $400 million.
5.1 Design of the Teton Dam
A 90 m high zone-filled earth dam was constructed in a steep-walled canyon eroded by the Teton River in Idaho. It had a wide silt core, with upstream and downstream shells consisting mainly of sand, gravel, and cobbles.
In the main section of the dam, the impervious core was keyed into the foundation of 30 m depth to serve as a cut-off trench. Lesser cut-off trenches were excavated at both abutments through the permeable rock. Reservoir filling commenced in November 1975 at an intended rate of about 0.3 m/day. Delays in completing the construction of the outlet work combined with heavier than expected spring melt run-off resulted in a filling rate up to 1.2 m per day in May 1976.
5.2 Failure Causes
The dam failed on June 5, 1976, when the water level in the reservoir was at an elevation of 9 m below the embankment crest and 1 m below the spillway crest. Breach of the dam crest and complete failure was preceded over a period of two days by increasing quantities of seepage. This seepage was observed initially 460 m downstream and later on the downstream face of the dam.
Noticeable increase in seepage rate from the face of the dam adjacent to the abutment about 40 m below the crest occurred during the morning of 5 June. By approximately 10:30 am, the flow rate of seepage increased to about 0.4 m3/sec. This quantity continued to increase as a 1.8 m diameter tunnel formed perpendicular to the longitudinal axis of the dam. By 11:00 am, a vortex was observed in the reservoir.
The seepage flow rate increased rapidly from this time onwards, accompanied by progressive upward erosion of the tunnel crown. The dam crest was breached at about 11:55 am, with complete failure of the dam.
5.3 Lessons Learned
The failure of Teton Dam contributed to significant learnings as it was the tallest earthen dam to have failed. It provided important lessons regarding the need for instrumentation, the need for protective filters to prevent uncontrolled seepage erosion, the design of cut-off trenches, consideration of the impact of frost action, and the importance of adequate compaction control criteria and methods.
The location of the dam site was unsuitable for the construction of the dam. Thus, the main reason for the failure of St. Francis Dam was the geology of the dam site.
The dam failed due to the opening in the bottom outlet gate, leading to the sudden failure.
Vajont Dam failed due to the instability of the slope.
The increase in the seepage rate, accompanied by progressive upward erosion of the base of the dam led to its failure.