Probabilistic Design of Coastal Flood Defences in Vietnam Probabilistic Design of Coastal Flood Defences in Vietnam Procfechrift tor verkrijging van de graad van doctor ‘aan de Technische Universiteit Delft, lop gezag van de Rector Magnificus professor ir. Luyben, voorzitter van het College voor Promotie, in het openbaar te verdedigen op dinstlag 21 september 2010 om 10.00 war door MAI VAN CONG Master of Science in Hydraulic Engineering geboren te Ninh Bình, Vietnam. Dit prooEchrift is goedgekeurd door de promoter: Prof. Vrijling Copromotor: Dr.
van Gelder Samenstelling promotieconnnissie: Rector Magnificus, voorzitter Prof. Vijling ‘Technical University of Delft, promotor Dr. van Gelder Technical University of Delft, copromotor Prof. Stive Technical University of Delf Prof.
Hanson Lund University, Sweden Prof. Hall sweastle University, UK Ir. Pilarcayk Formerly Rijkswaterstaat Dr. “Technical University of Delft Prof.
Uijttewaal ‘Technical University of Delft, reservelid Copyright © 2010 by Mai Van Cong 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. ISBN 975-00-9025688-1 Printed by: Sieea Repro, Delft, the Netherlands Cover layout: Pham Quang Diew Cover image: Violent ocean wave ‘source: Ocean Waves Wallpapers) to my family SUMMARY Probabilistic design of coastal flood defences in Vietnam. ‘The year 2005 brought hurricane Katrina to New Orleans (U.S), it was also a historical year for Vietnam because of the disastrous impact of typhoons on sea flood defences and ‘coastal regions.
In total, eight typhoons hit the Vietnamese coast that year resulting in Inuman casualties anel enormons economic damage. In response to that, both the central and local governments of Vietnam have paid the highest-ever attention to the rehabilita- tion of sea dike systems and the improvement of the safety of its coastal regions. ‘The improvement ofthe flood defence systems of Vietnam calls for the use of the latest available knowledge at all levels. Vietnam hasa profound practical experience in the field of flood protection; however, there is a lack of up-to-date knowledge in the field of dike design and flood risk management, especially regarding to coastal flood defences.
This study focuses on coastal flood defences in Vietnam and its objectives are to evalnate the current level of safety and reliability, to develop safety standards and design practices and to formulate a proposal for the improvement of the flood defences. ‘To meet the study objectives, the research presented in this thesis has focused on firstly reviewing, developing and widening the application of probabilistic design and state of the art of tisk assesement and risk evaluation in the field of flood defences (theory part) Secondly, as a demonstration of knowledge transfer, the proposed approach has been ap- plied to the coastal flood defence system of Nam Dinh province in Vietazun whieh is con sidered as a representative coastal system in the country (application part) ‘The main contents covered in this thesis are summarized below. Chapter 2 gives an overview of the coastaÏ ares ng approaches for flood dolene ‘Vietnam and a description of the case study area. ‘The coastal zone in the Red River Delta of Vietuam, wherethe case study was catried out, is under a potential threat of uvial flooding, coastal flooding and coastal erosion, In the region, sea dikes have been used as a predominant countermeasure with two intended functions, Le, protecting low= lying areas from coastal flooding and reducing risk for the hinterland caused by coastal erosion.
However, as was also experienced during recent typhoons in the last years, the sca dikes do not seem to fimetion very well and their strength is insufficient to withstand the hydraulic londs with a sufficiently low frequency of occurrence, This is due to the underestimation of the hydranlic boundary condlitions and imperfections in the design, construction and maintenance of the dike systems. A qualitative multi-criteria analysis has been developed for proposing a suitable coastal protection strategy for the study area, It is proposed to use a combination of sea dike systems to provide flood protection, Summary and beach erosion control measures, stich as groins and/or sand nourishments to preserve a shallow foreshore, To have a better understanding of the actual hydraulic load conditions in the coastal area and to derive input information for the reliability analysis, methods for the statisti- cai analysis of extreme hydraulic load conditions have been developed and applied (Chapter 3). These include data management techniques (ie. trend analysis, stationarity tests, seasonality analysis and peaks over threshold analysis); and statistical methods to ‘model the occurrence probabilities of extreme values (the value which is exceeded by the random environmental variable with very small probability p).
Hydraulic observations for the coast Vietnam are available for a period of 35 years and on that basis an estimate needs to be derived for a design sea load, ie, waves and storm surge levels, corresponding, to a 100 to 10,000 years return period. ‘The methods of Regional Frequency Analysis (RFA) and Mixed Distribution Fitting (MDP) are used in this study for sueh a sea load, In onder to ensure that the statistically predicted values of the interested sea loads will be as close as possible to the true values, recommendations are made to compare these values with physical hindeast models which are based on historical information of ty: phoon characteristics such as wind speed, Inspection of monthly sea water level data at Nam Dinh coast reveals that there exists a quadratic rising trend, with periodicity of 6 mouths, 12 months and 224 months in the data. The 6-month periodicity reflects well the influence of two monsoons per year to the Muctuation of sea water level: the 12-month periodicity is explained by the influence of yearly typhoons in the region. The 224-month periodicity, which is approximately 18.6 years, is almost exactly equal to a tidal citele, Although sea level rise (SLR) is not a focus of this study, based on a detected linear trend of the observed water level data the rate of SLR can be estimated approximately at 0.68 em per year at Nam Dinh coast.
In addition, this study recommends that estima- tion of SLR should be based on not ouly mathematical extrapolations, but also the loug- term nature of sea level fluctuation, ‘The theoretical background of probabilistic design and reliability analysis of flood de- fences is reviewed in Chapter4. Models for safety assessment and reliability based design are developed for the specific application to coastal flood defence systems. In the safety assessment mnodel, present probabilistic calculation techniques are applied to quantify the failure probability of the system and the system components. The effect of the length of the dike system on the system reliability was deployed in the model as well.
The reliabil- ity-based design model is used to search for an optimal geometey of the dike system given an admissible failure probability. Chapter 5 demonstrates the application of the models proposed in the previous chapter to Vietnam. The ease study area is the Nam Dinh coastal flood defence system in the Red River delta of Vietnam, It is found that the existing coastal flood defeuce system is less safe than its required safety standards. According to the stancards the dikes should bbe designed to provide a target safety of 1/20 per year.
However, the total failure prob- ability ofthe dike system is estimated to be much higher, about 0. This cor responds to the occurrence of a failure once in 7 years instead of once in 20 years as in- tended. Excessive wave overtopping induced dike failure is the dominant failure mode, Summary which likely contributes about 44% to the total failure of the dike system. Instability of annour units ofthe rovetinents gives an influence of 259%.
Overllowing and geotechnical instability of inner slope have smaller, but still a considerable contribution (10 and 20%, respectively). Piping, geotechnical instability of the outer slope and toe foot instability have a negligible influence on the safety. A sensitivity analysis shows that wave height and water level are the most important loading parameters. The findings are inline with the practical experiences and observations from dike failures in recent years that oc- curred in the Nan Dink dike system.
‘The application of the reliability based design model shows that if the existing safety standard (1/20 per year) would be applied, the geometry of the current dikes would be Insufficient. For instance, the crest level of the dike required from the reliability based design is 6.80 meters, whereas the existing crest level of the dike is 5.50 meters An extensive overview of approaches for risk assessment and risk evaluation is given in Chapter 6. In general, risk is defined as the product of probability of occurrence and con= sequence of disasters Lo be considered. Regarding flood risk in this study, the conse- quences are considered in terms of loss of life (individual risk’ and societal risk!) and economic damage (economic risk).
A central question in risk management concerns the acceptance of risk by the publie and the decision makers. Presently, the acceptable flood ing risk is judged from two points of view which relate to the acceptance of immdation ‘conseqnences. The first is acceptable risk regarding loss of life, which is under judgement of society and individuals, The second is viewed under an economic cost benefit balance. Chapter 7 and Chapter 8 present applications of the theory given in Chapter 6 to the Vietnamese situation.
In Chapter 7, anapproach for estimation of the level of acceptable flood risk for Vietnam is propased. The approach is based on earlier framework of Vi- ‘ling (2000) in which acceptable risk level is based on characteristics of the activity (benefit, voluntariness) and baseline mortality statisties in a countey. In this assessment, first the mortality due to natural causes and daily activities such as traffic is involved, Based on Vietnamese statistics, baseline mortality has been determined including the risks of traffic and living in a flood-prone area in Vietnam An FN-curve for flooding in Vietnam is constructed showing the probability of accidents ‘with fatalities. Adapting the risk evaluation framework of Vrijling (2000), the national level of acceptable risk of Vietnam is established, Evaluation of the present floods on the national seale shows that acceptable societal risk of Vietnam is five to ten times higher than in the Netherlands, To obtain a risk limit for the Nam Dinh coastal region, the nationally acceptable risk bas to be distributed over all the regions in one country.
A societal risk limit for Nam Dinh is obtained and it is found that the current level of flood risk exceeds this limit by far. According to the limit the flood risk is acceptable if the low lying coastal region of Nam "Individual risk: The probability (per year) of being killed at certain location assuming pert nent presence of the population, ' Societal risk: The probability of exceedance (per year) of an accident witha certain number of fatalities. Societal risk is often shown by means of an FN curve. It displays the probability per year of accidents with N of more fatalities, Summary Dinh is protected to a safety standard with exeeodanee probabilities lower than 1/150 per year.
‘The optimal safety is evaluated based on an economic approach for risk based design in Chapter 8. Historical data on economic loss due to floods in Vietham is available since 1970. The distribution of these losses can best be approximated by a lognormal distribu tion and the yearly expected value of the damage equals approximately SUS 589410” per year. This value is comparable with the actual flood damage of Vietnam during the last ‘years (over the the last 10 vears the total flood damages on average are about 1.0% of the GDP of Vietnam per year) Economic risk analysis of the Nam Dinh dike ring shows that the safety standard of 1/50 -yoars should be applied to the present economic situation (Situation 1).