Overtopping risk and uncertainty analysis of embankment dams

• Main Author: Goodarzi, Ehsan
• Format: Thesis
• Language: English
• Published: 2010
• Online Access: http://psasir.upm.edu.my/id/eprint/40945/1/FK%202010%2065R.pdf; http://psasir.upm.edu.my/id/eprint/40945/

The main objectives of this study are the evaluation of overtopping risk in conjunction with uncertainty for two embankment dams in the south of Iran (Fars province) and north of Iran (Mazandaran province). Dams are one of the most significant structures that contribute to the survival of the human race. The main benefits of this structure for human society are; flood controlling, generation of hydro-electric power, and supply water for different purpose such as agriculture, irrigation, recreation, and tourism attraction. Annually, natural disasters like earthquake, flood, drought, and thunders happen in different parts of the world. Malfunctioning of a dam can result in serious economic damage to the owners, and environmental problems such as pollution, sedimentation in the downstream areas and can also result in the loss of human lives. Iran is one of the ten foremost countries sensitive to the unexpected phenomena and natural disasters. Flooding is one of the most important disasters in Iran and losses during 1951-2000 exceeded more than 11,500 people. So, constructing new flood control structures and assessment safety of existing dams are essential. There is main gap with regards to apply new statistical tools and techniques for overtopping risk analysis and assessment the safety of dams. Furthermore, considering various events such as wind in conjunction with flood which increase the probability of failure is a crucial task of hydrosystem engineers. Hence, this research presents the application of risk and uncertainty analysis to dam overtopping. A flood frequency analysis of annual maximum discharge was done for the Doroudzan and Meijaran dams applying the General Extreme Value (GEV) probability distribution. Risk of overtopping were calculated for six extreme floods taking into consideration inflow hydrograph, initial water level, discharge coefficient of spillway, and the reservoir geometry as uncertain variables. The reservoir routing technique was used to compute the highest water levels and the Monte Carlo simulation (MCS) and Latin hypercube sampling (LHS) were applied for uncertainty analysis. To consider the effect of wind speed on the overtopping risk, frequency analyses were performed on the wind data and consequently extreme wind speeds, highest wind set-up and wave run-up were calculated. The maximum risk considering the effects of wind speeds in the study was 1.5 to 3.5 times larger than the risk of overflowing due to only floods using Latin hypercube sampling (LHS) and Monte Carlo simulation (MCS) approaches, respectively. The results of this study show that, initial water level, and inflows to reservoir have significant effects on the overtopping probability of the Doroudzan and Meijaran dams. Overtopping risk increases with increase in magnitude of stated parameters. Trends of computed risks for the Doroudzan dam indicate that calculated probabilities with Latin hypercube sampling were slightly higher than the Monte Carlo simulation, while in the Meijaran dam, there was no any specific trend between the MCS and LHS results. The major findings and main contributions of this study are; solving the flood model (reservoir routing) in uncertain conditions, estimating wind set-up and wave run up to solve the wind model with considering uncertainties, estimating the probability of failure due to overtopping under different hydrologic conditions, and compare the results based on two applied uncertainty models (MCS and LHS) for two mentioned dams.