Dam Safety Concepts

Master of Science in Hydraulics and River Basin Managment geboren te Sarajevo, Bosnia and Herzegovina Dit proefschrift is goedgekeurd door de promotoren All rights reserved. No part of this book may be reproduced in any form or by any means including photocopy, without written permission from the copyright holder. SUMMARY The majority of dams constructed in the world are dams that can be categorized as embankment dams. Throughout history we can point to many failures of dams, and embankment dams in particular. Nowadays it is clear that the goal to construct stable dams has not been achieved, even with advanced technologies and construction techniques available. There are always unexpected factors which can produce unforeseen problems, and also most of the suitable sites are already utilized. This might increase the probability of failure in the future; therefore constructors will have to face more complicated geological conditions. The main reasons for failures are inconsistency between design (design hypothesis) and reality (unpredictability encountered on the site or during construction), natural processes like flash floods, rock and landslides, earthquakes and deliberate human actions. Research on failure case histories and lessons learned from them provide important answers and leads to improvement in dam safety approach. The concerns of dam safety are related to dam procedures that will avoid a dam break or diminish the probability of a dam break or any other abnormal event. Most of dam safety decisions are based on the predictions of the probability of dam failure and of resulting loss-of life. In this thesis the intention is to highlight the methodolo-gies for dam safety decisions. Previously, risk and uncertainty methods have been applied for safety assessments of hydraulic structures but they were restricted to some extent. In this thesis an improvement of the methodology proposed by Hsu et al. (2011) is suggested in three aspects. The first is the development of multivariate flood frequency analyses in which the annual maximum peak discharges and the surface runoff are modelled as a bivari-ate distribution function. The second is the treatment of the initial reservoir level as a random variable, and the third is the overtopping assessment sample zone which is divided jointly by multivariate flood frequency, wind frequency and initial stage frequency , generating eight sub domains. Damage of property due to dam break discharge is certain, but loss of life depends on the flooded area and population. Therefore, analysis of dam breaching …