Summary of doctoral dissertation: Wave Overtopping at sea dikes with crown-walls in the northern coastal delta of Vietnam

Chia sẻ: _ _ | Ngày: | Loại File: PDF | Số trang:27

Thêm vào BST

Báo xấu

21
lượt xem 6
download

Download Vui lòng tải xuống để xem tài liệu đầy đủ

The thesis aims to investigate the influence of low crown-walls on overtopping discharge and the behaviour of overtopping flow at sea dikes. By doing so, the reliability of overtopping estimation is increased to improve the dike design guidelines currently applied in Viet Nam.

Chủ đề:

Bình luận(0) Đăng nhập để gửi bình luận!

Lưu

Nội dung Text: Summary of doctoral dissertation: Wave Overtopping at sea dikes with crown-walls in the northern coastal delta of Vietnam

MINISTRY OF EDUCATION AND MINISTRY OF AGRICULTURE TRAINING AND RURAL DEVELOPMENT WATER RESOURCES UNIVERSITY NGUYEN VAN THIN WAVE OVERTOPPING AT SEA DIKES WITH CROWNN-WALLS IN THE NORTHERN COASTAL DELTA OF VIETNAM Specialization: Hydraulic Engineering Code No : 62-58-40-01 SUMMARY OF DOCTORAL DISSERTATION HA NOI, 2014
This scientific work has been accomplished at Water Resources University Advisor 1: Assoc. Prof. Nguyen Ba Quy Advisor 2: Prof. Ngo Tri Vieng Reviewer No. 1 Prof. Dinh Van Uu Reviewer No. 2 Prof. Tran Dinh Hoi Reviewer No. 3 Prof. Tran Dinh Hoa This Doctoral Thesis will be defended at the meeting of the University Doctoral Committee in room No………… on ……. This dissertation is available at: - The National Library - The Library of Water Resources University
INTRODUCTION 1. Rationale Viet Nam is a nation that is seriously affected by climate change and sea level raise. There are many centres of economy and culture along the coastline. In the north, sea dikes are relatively low with narrow crests, steep seaward and landward slopes; most dikes are directly exposed to wave attack. Historical records show that wave overtopping often causes damage to dike crests and landward slopes. One of the effective measures to reduce overtopping is the use of (low) crown-walls, because heightening dikes or constructing outer berms are costly and not feasible, especially dikes with narrow margins. Till now, studies into the influence of low crown-walls and interaction between wave-wall and overtopping are limited. A better understanding of the influence of crown-walls and promenade on overtopping is necessary to improve the present dike design guidelines in Viet Nam. These facts lend the foundation for this thesis that is to investigate “Wave Overtopping at sea dikes with crown-walls in the northern coastal delta of Vietnam”. 2. Research objectives The thesis aims to investigate the influence of low crown-walls on overtopping discharge and the behaviour of overtopping flow at sea dikes. By doing so, the reliability of overtopping estimation is increased to improve the dike design guidelines currently applied in Viet Nam. 3. Scope of the study Investigation of wave overtopping at sea dikes with low crown-walls in the north of Viet Nam. 1
4. Research contents Review on overtopping on sea dikes with low crown-walls; Physical model to investigate the effects of low crown-walls on wave overtopping at sea dikes; Numerical and physical examination of the wave-wall interaction and overtopping flow on sea dikes with low crown-walls; Case study-Overtopping at Giao Thuy dike, Nam Dinh province 5. Approach and study methods 5.1. Approach To obtain the objectives, the author carried out a literature study on wave overtopping at sea dikes with low crown-walls to select an approach that is inherent and also creative, suitable for Viet Nam. 5.2. Study approaches Literature review; Physical and numerical modelling; Application. 6. Implications Building low crown-walls on seadikes to heighten the crest level, reduce wave overtopping is considered technically and economically viable in Viet Nam. There exists limited research into overtopping at sea dikes with low crown-walls, especially the interaction between wave and wall. Therefore, better understanding of the wall influence on wave overtopping discharge and the overtopping flow characteristics will help improve the reliability of design for this type of sea-dikes in Vietnam. 2
7. New contributions - Insights into the influence of low crown-walls on wave overtopping and the merit of wall promenade through a detailed examination of the wave- wall interaction; - Empirical equations to determine the overall influence factor of low crown-walls for regular waves (Figure 2-12); - Relationship between splash height and wave parameters and wall geometry (Figure 2.13); - Proposal on the use of a new sea-dike cross-section with crown-walls and promenade, appropriate for the northern coastal detla of Viet Nam (Figure 4.8). 8. Thesis contents In addition to the Introduction, Conclusions and Recommendations, the thesis consists of 04 chapters Chapter 1: Literature study on overtopping on sea dikes with low crown-walls; Chapter 2: Physical model to investigate the effects of low crown-walls on wave overtopping at sea dikes; Chapter 3: Wave-wall interaction and overtopping flow on sea dikes with low crown-walls; Chapter 4: Case study-Overtopping at Giao Thuy dike, Nam Dinh province. CHAPTER 1 REVIEW ON OVERTOPPING ON SEA DIKES WITH LOW CROWN-WALLS 1.1 Introduction to research into overtopping at sea dikes Due to significant changes of climate and environment, the frequency and intensity of natural hazards gradually increase especially storms, tide and sea level rise. As a result, overtopping at sea dikes remains as a risk to countries with 3
sea. There exists limited research on overtopping at sea dikes with low crown- walls and no one is comprehensive yet. Therefore, studies on overtopping at sea dikes are essential in Viet Nam and elsewhere 1.2 Causes, failure mechanisms of sea dikes and measures 1.2.1 Causes of damage to sea dikes There are many failure mechanisms of sea dikes but historical records show that wave overtopping mainly causes damage to the dike crest and landward slope and dike breaching as a consequence. 1.2.2 Damage mechanism due to wave overtopping There are many mechanisms leading to dike failure, from local damage to overall collapse; the reasons, influence factors, consequences are very various. Analysis indicates that dike failures due to wave overtopping are the most common. 1.2.3 Measures to reduce overtopping at sea dikes in the north Nowadays, there are several ways to reduce overtopping at sea dikes, ‘hard’ and ‘soft’ solutions such as submerged breakwaters, concrete blocks, seaward berms, high crests and mangrove … However, the conditions of construction space and economy are limited low crest-walls are popular and effective in reducing overtopping at sea dikes in Viet Nam. 1.3 Sea dikes with low crown-walls in the northern coastal delta Sea dikes with low crown-walls (W/Hs ≤ 0.5) located near the seaward edge are very popular in Viet Nam. This is considered as a simple and effective method to increase the crest level, reduce overtopping in the present situation. Crown- walls are applied where there is no more land to enlarge the dike cross-sections or budget is constrained; or it is not allowed to heighten the crest level in order to reserve residence and tourism areas. 4
1.4 Research into overtopping at sea dikes with low crown-walls 1.4.1 TAW 2002 In TAW (2002), the influence of crown-walls on overtopping discharge is not clear because it is an unknown variable; crown-walls increases the equivalent slope that the discharge becomes greater. However, the discharge is then corrected by a factorv. This overall influence factor only takes into account the seaward inclination but not the interaction between wave-wall, overtopping flow and the wall dimensions. 1.4.2 Viet Nam Till now, there is limited research into overtopping at sea dikes with low crown- walls in Viet Nam. Recent works do not cover all aspects of overtopping at Vietnamese dikes. Tuan et al. (2009) proposed a new method to assess the effect of crow-walls on overtopping. However, the influence of promenade was not discussed (S = 0). Tuan (2013) investigated the influence of crown-walls and promenade on overtopping. Though, he did not consider the interaction between wave-wall and the behaviour of overtopping flow when the wall exists. Furthermore, reduction effect was not determined for regular waves. 1.5 Conclusions of chapter 1 Overtopping is a danger to sea dikes. Damage due to overtopping is the most important. Crown-walls are effective to heighten the crest level and reduce overtopping. Studies on overtopping from TAW (2002) to Tuan (2013) are not complete. In line with Tuan (2013), the author performed tests with physical model to investigate the influence of crown-walls on regular waves, especially the interaction between wave-wall and the flow behaviour at walls. The thesis used numerical models from NLSW to RANS-VOF to consider the wall effects with regard to the interaction between wave-wall and the flow behaviour at walls. 5
The obtained results will provide insight into the characteristics of overtopping at dikes with low crown-walls, partly improve the dike design in Viet Nam. CHAPTER 2 PHYSICAL MODEL TO IVESTIGATE THE EFFECTS OF LOW CROWN– WALLS ON WAVE OVERTOPPING AT SEA DIKES. 2.1 Study objective Consideration of the influence of low crown-walls on overtopping discharge, the interaction between wave-wall and the behaviour of overtopping flow at sea dikes with walls. 2.2 Model similitude For the similarity between model and prototype, three criteria are required geometry, kinematic and dynamic. For the similitude of wave, model has to be geometric similarity; the scale has to follow the Froude criterion. In short wave tests with geometrically undistorted models, the Froude criterion is automatically satisfied. 2.3 Experiments with regular waves 2.3.1 Wave flume The Holland wave flume is 45 m long, (effective length of 42 m), 1.2 m high, and 1.0 m wide. The wave maker, which is equipped with an advanced automated system of Active Reflection Compensation, is capable of generating regular and irregular waves (JONSWAP) up to 0.3m in height and 3.0 s in peak period (Figure 2.1). 6
Figure 2.1 Overview of the wave flume. Hình 2.1 Toàn cảnh máng sóng sử dụng thí nghiệm 2.3.2 Dike model and parameters The dike model dimensions and testing parameters are selected according to a model length scale of 1/10. The dike slopes were smooth and impermeable, 70 cm height with a seaward slope of 1/3. The low crown-walls were 4, 6 and 9 cm. The walls were made detachable to allow varying wall height (W) and promenade width (S) with regard to test scenarios. The wall could be moved back and force to change the promenade from 0 to 10 and 20 cm. The foreshore was 24.5 m long and 1/100 steep (Figure 2.5). Figure 2.5 Test set-up for regular waves. 2.3.3 Test programme A water depth of d = 0.60m was chosen for testing. A group of 3 wave gauges was positioned at the dike toe and another gauge was 24.5 m away from the toe. A camcorder with high resolution was mounted normal to the flume to capture 50 frames/second in order to observe the interaction between wave-wall and overtopping flow. Each test consisted of 10 regular waves and stopped before these got disturbed by reflected ones (Table 2.1). 7
Table 2.1 Test programmes with regular waves Number Wave parameter Rc (m) W (cm) S (cm) of tests H (m) T (s) 40 0.16- 0.24 1.5 – 2.5 0.10 0; 4; 6; 9 0; 10; 20 2.3.4 Test procedures and measurement parameters Preparation time was from June to August 2012 and test duration lasted from August till September 2012. Measurement parameters includes: wave height H, wave period T, mean overtopping discharge q, splash height Hb, flow thickness on wall crest Ht, crest freeboad Rc 2.4 Data analysis 2.4.1 Influence of crown-walls on wave overtopping discharge The overall influence factor by crown-walls is the product of component factors due to wall height and wall promenade, respectively. 1 1 1 𝑊 1 𝑆 1 = . = (1 + 𝑐1. . ) . (1 + 𝑐2. . ) (2-11) 𝛾𝑣 𝛾𝑤 𝛾𝑠 𝑅𝑐 𝜉 𝐻 𝜉 Determine the overall influence factor 𝛾v (measured) and 𝛾v (computed), perform regression analysis to establish the relationship between these parameters, derive two coefficients c1 =1.26, c2 = 1.44 (Figure 2.12) 1 𝑊 1 𝑆 1 = (1 + 1,26. . ) . (1 + 1,44. . ) (2-12) 𝛾𝑣 𝑅𝑐 𝜉 𝐻 𝜉 2.4.2 Wall influence on splash height The splash height was estimated by image analysis with Matlab. The author 𝐻𝑏 𝑆.𝐻 established the relationship between and , which reads y = 1.544e-30.9x 𝐻 𝑔.𝑊.𝑇2 with R2= 0.624. Based on this curve, one may roughly predict splash height with regard to wave and wall characteristics (Figure 2.13). 8
Figure2.10 Overall influence factore of low walls v (measured - computed) 2.500 2.000 1.500 y = 1.544e-30.9x R² = 0.624 Hb/H 1.000 .500 .000 .000 .005 .010 .015 .020 .025 .030 .035 .040 S.H/g.W.T 2 𝐻𝑏 𝑆.𝐻 Figure 2.13 Relationship between and 𝐻 𝑔.𝑊.𝑇 2 𝐻𝑏 𝑆.𝐻 Hình 2.13 Biểu đồ quan hệ 𝐻 với 𝑔.𝑊.𝑇 2 9
2.5 Conclusions of chapter 2 The thesis successfully established empirical formulas estimating the overall influence factor by crown-walls on mean overtopping discharge for regular waves (2-12) and functions of splash height with regard to wave and wall (Figure 2-13). CHAPTER 3 INTERACTION BETWEEN WAVE – WALL AND OVERTOPPING FLOW ON SEA DIKES WITH LOW CROWN-WALLS 3.1 Problem definition At different levels of detail, the thesis modelled overtopping flow at sea dikes with low crown-walls using several programmes including NLSW (Non-Linear Shallow Water) and RANS-VOF (Reynolds Averaged Navier Stokes – Volume Of Fluid). NLSW modelling is simple and fast in calculation, e.g. 1000 waves can be simulated in 5 to 10 minutes. This program can estimate relatively well mean overtopping discharge at dikes with mild slope and have no crown-wall. It has several shortcomings when applied to structures with complicated shape, e.g. dikes with crown-walls. The thesis used NLSW (Tuan and Oumeraci, 2010) to simulate and compute with regular waves. The obtained results were compared to experiments in wave flume. RANS–VOF modelling (COBRAS-UC, numerical wave flume) is able to simulate the interaction between wave – wall and flow at structures of any shape (vertical walls, hollow walls …), from wave generation at boundary to wave propagation as in physical flumes. However, the calculation efficiency is low. It takes many hours to simulate some seconds in real time when run on a normal computer so that it is difficult to apply to regular waves. Therefore, NLSW of Tuan and Oumeraci (2010) was used to validate the estimated values of discharge. The computations were compared to measurements with irregular waves. Both numerical and physical flumes were 10
deployed to assess the interaction between wave – wall, i.e. overtopping flow characteristics. 3.2 NLSW modelling (Tuan and Oumeraci, 2010) 3.2.1 Basic formulations The model by Tuan and Oumeraci (2010) is based on the flux-conservative form of the NLSW equations solved with a high order total variation diminishing (TVD), Roe-type scheme: U  F ( x,U )   S ( x, U ) (3-1) t x where conserved vectors U , F ( x,U ) and source term vector S ( x,U ) are defined as follows: h  U ( x)    (3-2)  uh   uh  F ( x, U )   2  (3-3)  u h  gh / 2  2 0  S ( x, U )    (3-4)  gh( Sbx  S f  Sr )  in which g is the gravitational acceleration, h is the flow depth, u is the horizontal flow velocity, Sbx and Sf are bed slope and friction slope, respectively. Note that Sr is the surface roller slope term, added by Tuan and Oumeraci (2010) to account for the influence of wave breaking through the drastic motion of surface rollers in the surfzone on the mean flow. 3.2.2 Wave overtopping of irregular waves NLSW cannot model a vertical wall because the shallow water limit is violated, a pragmatic manipulation of the wall geometry is necessary. The author used two 11
pragmatic approaches: equivalent wall and equivalent freeboard (Figure 3.1 and 3.2). Mean water levels were used in combination with wave signal recorded by gauges, which were positioned in front of the model dike toe (the closer these gauges to the toe, the higher accuracy the results of NLSW). Figure 3.1 Conversion of crown-walls into slope (TAW-2002) Figure 3.2 Conversion of crown-walls into equivalent freeboard method On the landward side, the outflow boundary used a water level constant and very low with regard to crest level (in order to prevent any influence on overtopping). The simulation time was the same as the physical model experiments (1000.Tp ~ 10 minutes PC). In general, the computed results agree reasonably well with the experimental data, R2= 0.88 and 0.87 for the first and second approaches, respectively. The mean prediction error is 39.8% with a standard deviation of  56.2%. However, discrepancies still exist for some particular cases of very low overtopping rates at high walls and walls without promenades. This is because 12
the strong interaction between wave-wall could not completely be resolved in NLSW by using pragmatic wall schematizations. Figure 3.3 Wave overtopping computation with wall schematization according to (TAW 2002): measured versus computed Figure 3.4 Wave overtopping computation with equivalent freeboard approach: measured versus computed 13
3.3 RANS-VOF modelling (COBRAS-UC, numerical simulation) 3.3.1 Numerical wave flume Numerical wave flumes are capable of modelling the wave-structure interaction that is very comparable to physical flumes. They are applicable for almost any complex geometric and structural configurations. They have been validated against experimental data giving reliable result. 3.3.2 Basic formulations In the model, the mean turbulent flow is based on 2DV RANS equations  ui 0 (3-12) xi  ui  ui 1 p 1    ui   uj   gi     uiu j  (3-13) t x j  xi  x j  x j  and the turbulence closure is the (k-) transport equations, which relate the fluctuating components of the flow to the turbulence kinetic energy k and the dissipation rate of turbulence  k k    k   ui  uj   t      uiu j  (3-14) t x j x j   k x j   x j      t      ui  2 (3-15)  uj       C1 uiu j  C2 t x j x j     x j  k x j k where ui is mean velocity in the i- the direction (i, j =1, 2 for a two dimensional flow), p is mean pressure,  is fluid density, gi is gravitational acceleration in the i-th direction,  uiuj is Reynolds stresses modelled according to the nonlinear eddy viscosity. The empirical coefficients are k = 1.0,  =1.3, C1 = 1.44, C1 = 1.92;= / and t = Cdk2/ (Cd = 0.99) are kinematic and eddy viscosity, respectively.COBRAS-UC resolves the flow on a non-uniform rectangular grid. The arbitrary free surface is tracked using the Volume of Fluid (VOF) method. 14
3.3.3 Overtopping discharges of irregular waves Each train consists of at least 1000 waves (1000.Tp = 2200s) which can be simulated in 75hours on a 3.1GHz processor - 4GB RAM PC. Because of this low efficiency, only 14 scenarios were considered. These are combinations of a wave condition at the generator (Hm0 = 0.10 m, Tp = 2.2 s and water depth of 0.55 m) and several seaward slopes and crown-walls (height of W= 6 and 9cm; with and withour promenade). Figure 3.7 Wave overtopping of irregular waves: COBRAS-UC vs. NLSW model The study compares values of mean overtopping discharge obtained from measurements with physical models and computation with COBRAS-UC, NLSW (two ways of varying equivalent walls). Apparently, COBRAS-UC works more properly than NLSW with mean errors of 60.1% (a standard deviation of  63.2%) and 129.4% ( 100.6%) for PA1 and PA2, respectively. The results from COBRAS-UC and measurements match relatively well with a mean error 39.7% (a standard deviation of  24.5%). However, COBRAS- 15
UC may produce an error up to 63% in the cases of small discharge (Figure 3.7). 3.3.4 Wave overtopping of regular waves (discharges and wave-wall interaction) 3.3.4.1 Mean overtopping discharge With 40 experiments performed on 10 dike models, regular wave trains were generated to evaluate the mean overtopping discharge in for each test. The CPU time for each test was around 6 hours on a standard PC (or about 1 hour CPU time equals to 10 seconds of flow time). Discharges predicted by COBRAS-UC were compared to the experimental data. Generally speaking, COBRAS-UC appears to reliably estimate overtopping rates at low crown-walls of various configurations (R2 = 0.95). The mean error is 23.4% with a standard deviation of 30.2 %. Considerable discrepancy can be found for a few cases of high walls (W=9cm, open triangles). In comparison with irregular waves, the overall agreement is relatively good (Figure 3.8). Figure 3.8 Wave overtopping of regular waves: measured vs computed by COBRAS-UC 16
3.3.4.2 Interaction between wave – wall and flow At first, the wave – wall interaction, which results in wave overtopping at sea dike, is assessed using captured video images. A wave overtopping event can be characterised through four successive phases: contact splash, fall-over, major green overtopping and withdrawal. These phases are briefly described as follows. The process of wave overtopping starts as the wave tongue collides with the wall and results in a violent jet (splash) into the air (Figure 3.13). In the second phase, the splash collapses as it reaches a certain height and falls over the wall. At the end of this phase, a green flow over the wall is formed (Figure 3.14). If wave continues to thrust wave overtopping in the third phase simply follows the earlier buffer flow in the form of green overtopping to obtain a maximum depth above the wall crest (Figure 3.15). Green wave overtopping continues until wave retreats on the seaward slope in the last phase (Figure 3.16). Snapshots of the flow computed by the model at all computed instances were also reviewed to determine the corresponding maximum splash height for comparison with the physical experiments. Numerically, COBRAS-UC is able to capture a fragmented splash but not that with small individual drops (tiny sprays) like in the physical models. Physically, the approximation of the free surface tracking method (VOF), exclusion of the surface tension as well as disregard of air entrainment an aeration processes are possible causal factors. Practically, coarse grid resolution relative to the (extremely small) size of individual drops employed in the numerical model may also play a role. These issues require much more sophisticated studies as well as computer capacity to be resolved. At present, this effect on wave overtopping with respect to the dike design is neglected. The four-phase process of the wave-wall interaction for Case REW6S20_4 is now examined in detail with COBRAS-UC. The model considerably under-predicts the water surface profile in the splashing area around the wall during the first two stages (Figures 3.13 and 3.14). In the mean time, the surface profile in the last two stages, where mainly simply green overtopping take places, is very well predicted by the model (Figures 3.15 and 3.16). 17
Figure 3.13 Wave splash at wall t= 27.1s Figure 3.14 Overtopping flow on wall ( Measured) crest t=27.3s ( Measured) Figure 3.15Overtopping flow Figure 3.16 Run-down t=27.8s t = 27.5s ( Measured) ( Measured) At a lower level of detail, Figures 3.17 and 3.18 respectively show the model prediction of the maximum splash height (occur in phase , Figure 3.13) and the maximum green flow depth above the wall (occur in phase 3, Figure 3.15) compared with the experimental data. Under the same wave condition and wave height, the maximum splash height decreases as the promenade width decreases (Figure 3.17). Also, within the wall geometry and wave parameters considered herein, a higher wall (relative to the wave height) would cause a higher splash height. 18