DAM BREAK ANALYSIS & DISASTER MANAGEMENT PLAN

11 DAM BREAK ANALYSIS & DISASTER MANAGEMENT PLAN 11.1 INTRODUCTION 11.1.1 Dam Break Phenomenon The construction of dams in rivers can provide considerable benefits such as the supply of drinking and irrigation water as well as the generation of electric power and flood protection; however the consequences which would result in the event of their failure could be catastrophic. They vary dramatically depending on the extent of the inundation area, the size of the population at risk, and the amount of warning time available. Dam break may be summarized as the partial or catastrophic failure of a dam leading to the uncontrolled release of water. Such an event can have a major impact on the land and communities downstream of the breached structure. A dam break may result in a high flood wave traveling along a valley at quite high speeds. The impact of such a wave on developed areas can be sufficient to destroy infrastructure, such as, roads, railways and bridges, and, to damage buildings. With such destructive force comes an inevitable loss of life, if advance warning and evacuation were not possible. Additional features of such extreme flooding include movement of large amounts of sediment (mud) and debris along with the risk of distributing pollutants from any sources, such as chemical works or mines in the flood risk area. Though, there have been great advancements in design methodologies, failures of dams and water retaining structures may still occur. Failure of the Malpasset concrete dam in France in 1959 led to 433 casualties and eventually prompted the introduction of dam safety legislation into France. In October 1963, 2000 people died in Italy, when a landslide fell into the Vaiont reservoir creating a flood wave some 100 m high that overtopped the dam and flooded into the downstream valley. In July 1985, about 90% of the 300 people living in Stava near the Stave Dam in Italy also died when 151

this mine tailing dam failed. More recently, in May 1999, a dam failed in Southern Germany causing 4 deaths and over 1 billion Euro of damage. In Spain 1997, failure of a dam on the Guadalquivir river, not far from Sevilla, caused immense ecological damage from the release of polluted sediments into the river valley. Similarly, in Romania earlier this year, failure of a mine tailings dam released lethal quantities of cyanide into the river system, polluting the environment and a major source of drinking water for both Romania and Hungary. In India, the breaching of Kodaganar Dam (Tamil Nadu) in year 1977 caused a huge loss of property in downstream area. About 2000 people died due to breaching of Machhu II dam (Gujarat) in year 1979 and the flood wave of order of 10 m caused a heavy devastation in Morvi town and near by villages. In year 2005, the failure of Nand Gavan dam in Maharashtra and Pratapura dam in Gujarat caused severe flooding in downstream area. The above instances of dam breaks establish that hazard posed by dams, large and small alike, is very real. As public awareness of these potential hazards grows, and tolerance of catastrophic environmental impact and loss of life reduces, managing and minimizing the risk from individual structures is becoming an essential requirement rather than a management option. 11.1.2 Need for Dam break modeling The first European Law on dam break was introduced in France in 1968 following the earlier Malpasset Dam failure that was responsible for more than 400 injuries. Since then many countries have also established requirements and in others, dam owners have established guidelines for assessment. In India, Risk assessment and disaster management plan has been made a mandatory requirement while submitting application for environmental clearance in respect of river valley projects. Preparation of Emergency Action Plan after detailed dam break study has become a major component of dam safety programme of India. The extreme nature of dam break floods means that flow conditions will far exceed the magnitude of most natural flood events. Under these conditions, flow will behave differently to conditions assumed for normal river flow modelling and areas will be inundated, that are not normally considered. This makes dam break modelling a separate study for the risk management and emergency action plan. 152

The objective of dam break modelling or flood routing is to simulate the movement of a dam break flood wave along a valley or indeed any area downstream that would flood as a result of dam failure. The key information required at any point of interest within this flood zone is generally: i) Time of first arrival of flood water ii) Peak water level extent of inundation iii) Time of peak water level iv) Depth and velocity of flood water (allowing estimation of damage potential) v) Duration of flooding The nature, accuracy and format of information produced from a dam break analysis will be influenced by the end application of the data. For example: Emergency Planning To reasonably prepare an emergency plan, it will be necessary for the dam break analysis to provide: i) Inundation maps at a scale sufficient to determine the extent of and duration of flooding in relation to people at risk, properties and access routes ii) Identification of structures (bridges etc.) likely to be destroyed iii) Indication of main flow areas (damage potential of flow) iv) Timing of the arrival and peak of the flood wave v) Identification of features likely to affect mobility / evacuation during and after the event including impact on infrastructure and the deposition and scour of debris and sediment. Development Control Development control will focus mainly on the extent of possible inundation resulting from different failure scenarios. Consideration may also be given to the characteristics of the population at risk. 11.1.3 Present Dam Break Modeling Study The present study for the Lower Siang H.E. Project comprises of: 1. Prediction of outflow hydrograph due to dam breach 2. Routing of dam breach flood hydrograph through the downstream valley to get the maximum water level and discharge along with time of travel at different locations of the river downstream of the dam 153

3. Routing the PMF through the reservoir and downstream valley without dam breach to get the maximum discharge and water level at different locations of the river downstream of the dam 4. Channel routing the PMF through the downstream valley in the virgin condition of Siang river i.e. without Lower Siang dam to get the maximum discharge and water level at different locations of the river downstream of the dam. These studies will give assessment of inundation area due to flood caused by dam breach over and above the inundation due to the heaviest flood that may occur without dam being constructed. 11.2 DAM BREAK MODELING PROCESS 11.2.1 Introduction to Dam Break Modeling Generally, dam break modeling can be carried out by either i) scaled physical hydraulic models, or ii) mathematical simulation using computer. A modern tool to deal with this problem is the mathematical model, which is most cost effective and reasonably solves the governing flow equations of continuity and momentum by computer simulation. Mathematical modeling of dam breach floods can be carried out by either one dimensional analysis or two dimensional analysis. In one dimensional analysis, the information about the magnitude of flood, i.e., discharge and water levels, variation of these with time and velocity of flow through breach can be had in the direction of flow. In the case of two dimensional analysis, the additional information about the inundated area, variation of surface elevation and velocities in two dimension can also be assessed. One dimensional analysis is generally accepted, when valley is long and narrow and the flood wave characteristics over a large distance from the dam are of main interest. On the other hand, when the valley widens considerably downstream of dam and large area is likely to be flooded, two dimensional analysis is necessary. In the instant case, as the Saing valley is long and the flood wave characteristics over a large distance from the dam are of main interest, one dimensional modeling was adopted. 154

11.2.2 Hydrodynamic Modeling The essence of dam break modeling is hydrodynamic modeling, which involves finding solution of two partial differential equations originally derived by Barre De Saint Venant in 1871. The equations are: i. Conservation of mass (continuity) equation ( Q/ X) + (A + A 0 ) / t - q = 0 ii. Conservation of momentum equation ( Q/ t) + { (Q 2 /A)/ X } + g A (( h/ X ) + S f + S c ) = 0 where, Q = discharge; A = active flow area; A 0 = inactive storage area; h = water surface elevation; q = lateral outflow; x = distance along waterway; t = time; S f = friction slope; S c = expansion contraction slope and g = gravitational acceleration. 11.2.3 Selection of Model Selection of an appropriate model to undertake dam break flood routing is essential to ensure the right balance between modeling accuracy and cost (both in terms of software cost and time spent in developing & running the model). In the instant case, MIKE 11 model developed by Danish Hydraulic Institute has been selected for the present study because of its wide acceptability in India and abroad. 11.2.4 MIKE 11 Model The core of the MIKE 11 system consists of the HD (hydrodynamic) module, which is capable of simulating unsteady flows in a network of open channels. The results of a HD simulation consist of time series of water levels and discharges. MIKE 11 hydrodynamic module is an implicit, finite difference model for unsteady flow computation. The model can describe sub-critical as well 155

as supercritical flow conditions through a numerical description, which is altered according to the local flow conditions in time and space. Advanced computational modules are included for description of flow over hydraulic structures, including possibilities to describe structure operation. The formulations can be applied for looped networks and quasi two-dimensional flow simulation on flood plains. The computational scheme is applicable for vertically homogeneous flow conditions extending from steep river flows to tidal influenced tributaries. The following three approaches simulate branches as well as looped systems. i) Kinematic wave approach: The flow is calculated from the assumption of balance between the friction and gravity forces. The simplification implies that the Kinematic wave approach can not simulate backwater effects. ii) Diffusive wave approach: In addition to the friction and gravity forces, the hydrostatic gradient is included in this description. This allows the user to take downstream boundaries into account, and thus, simulate backwater effects. iii) Dynamic wave approach: Using the full momentum equation, including acceleration forces, the user is able to simulate fast transients, tidal flows, etc., in the system. Depending on the type of problem, the appropriate description can be chosen. The dynamic and diffusive wave descriptions differ from kinematic wave description by being capable of calculating backwater effects. The solution algorithm for the different flow descriptions is identical in the inner programme structure, implying that the user does not have to distinguish between the different computational levels, when running the program. In the instant case, dynamic wave approach was adopted for a better simulation. Hydrodynamic module utilizes a space staggered grid consisting of alternating h and Q points, i.e., points where water levels (h) and discharges (Q) are computed sequentially. Topographic data are entered at the h points, and discharge relations are evaluated at Q points. During simulations, the complete non-linear equations of open channel flow are solved numerically at the grid points at specified time intervals for the given boundary conditions. 156

11.2.4.1 Solution technique In order to obtain a stable solution to the finite difference scheme, two conditions viz. (i) Velocity condition and (ii) Courant condition have to be satisfied. (i) Velocity condition: (V. t/ x) 1-2 (ii) Courant condition: C r = [(V+ (g.d)). t]/ x 10-15 C r is the Courant number, v is the cross-sectional mean velocity, g is the acceleration due to gravity, d is the mean depth, t is the time step, x is the space step (the distance between adjacent h-points) The most important considerations determining the selection of space and time steps for a particular model application are the expected wave lengths and duration of the wave period, and the ability to adequately resolve the channel topography. The space step length must be chosen ensuring a sufficient number of points along the channel axis to resolve the expected waves. The wavelength is determined by the wave period and the speed of propagation. A second concern is the adequate resolution of rapid changes in topography along the channel axis, and this may require extra grid points. The time step must be selected so that all expected significant wave periods are adequately resolved in time. As the duration of tidal waves is generally shorter than flood waves, the time step of a hydrodynamic model, which simulates tidal flows requires a shorter time step than that used in flood wave computations. The solution to the combined system of equations at each time step is performed in a computational grid consisting of alternative Q-point and h-point, i.e. points where the discharge Q and water level h respectively, are computed at each time step. A typical layout of channel section with computational net is shown in Fig. 11.1. 11.2.4.2 Boundary conditions in general The boundary conditions in MIKE 11 are distinguished between external and internal boundary conditions. Internal boundary conditions are (i) links at nodal points, (ii) structures, (iii) internal inflows, and (iv) wind friction. External boundary conditions may consist of (i) constant values for h or Q, (ii) time varying values for h or Q, and (iii) relation between h and Q. Generally, model boundaries should be chosen at points, where either water level or discharge measurements are available so that the model is used for predictive purposes. It is 157

important that the selected boundary locations lie outside the range of influences of any anticipated changes in the hydraulic system. The structure description combines a wide range of elements covering weirs, narrow crosssections, flood plains, reservoirs operations, etc., and which can be regarded as an internal boundary condition. The description is obtained by replacing the momentum equation with an h-q-h relation or an h-q relation. The grid to be used to describe a structure will consist of h-point on both side, and a Q-point at the structure. Lateral inflows can also be accommodated in MIKE 11 Hydrodynamic module (HD). The lateral inflows are specified at h-points, and are included in the continuity description. 11.2.4.3 Topographical requirement and discretization MIKE 11 HD is a physical modeling system, and hence, data related to the detailed physical characteristics of the study area must be obtained, if realistic results are to be expected. Topographic data are necessary to provide an adequate geometrical and topographical description of the river system, flood plains, and all important structures. First, the layout of the channel network is determined, and all significant channels identified, including the locations of the main channel confluences and bifurcations. Flood cells subject to inundation must be delimited, and the network of discharge exchange between the flood cells and the main river channels need to be identified. Cross-sections are required at regular intervals along the river. These must extend up to the river bank to encompass any natural or man-made river embankments. In the model schematization, the available cross-sections are placed at h-points. The cross-sections should be representative of the entire channel reach between the adjacent Q-points. Hence, channels which exhibit highly irregular cross-sectional variations require denser grid, and hence, have greater data requirements. The equations of one dimensional flow assume a horizontal water level surface across the channel section. Where flow occurs over wide flood plains, which are separated from the main river channel by natural levees or man made embankments, a purely one dimensional description is no 158

longer adequate. The description of such areas, called flood cells, is readily accommodated by MIKE 11 module through linking of the individual cells via an appropriate discharge formulation. For example, where the flood cell boundary constitutes a road or flood embankment, the discharge relation describing the exchange of flows between the cells by overtopping of the embankment is a simple weir formulation. Where flood cells are interconnected by road culverts, the standard culvert formulation of MIKE 11 HD may be used. Longitudinal profiles along the flood cell embankments, which lie directly adjacent to the main river are also required to establish the locations and levels at which over bank spilling may occur. Similar profiles are required along the embankments or a road, which separate the individual flood cells, if over bank spilling is possible. Topographic input for flood cells consists of a flooded area / water level elevation relation, such that the storage characteristics for each cell may be identified. Possible sources of data for the topographical input include contour maps, hydrographic charts, aerial photographs, satellite imagery, etc. However, in the majority of cases, sufficiently detailed information can only be obtained from controlled field surveys. It is essential that all topographical data levels relate to a common fixed reference level, preferably mean sea level. Besides topographic data, hydrometric data are necessary to enable the model to be calibrated against actual events, and thus, provide a basis for verification of the chosen schematization. Hydrometric data are also required at the model boundaries for any subsequent operation of the model. The main types of hydrometric data required are water levels and discharges. 11.2.5 MIKE 11 Model set-up The Dam Break Module in MIKE 11 simulates the outflow hydrograph resulting from the failure of a dam. The model set-up consists of a single or several channels, reservoirs, dam break structures and other auxiliary dam structures such as spillways, bottom outlets etc. As the flood propagation due to the dam break will be of highly unsteady nature, the river course needs to be described accurately through the use of as many cross-sections as possible, particularly where the cross-section is changing rapidly. Further, the cross-sections should extend as far as possible to cover the highest modeled water level, which normally will be in excess of the highest recorded 159

flood level. If the modeled water level exceeds the highest level in the cross-section for a particular location, MIKE 11 will extrapolate the processed Data as a vertical wall, and this will give conservative results. 11.2.5.1 River channel set-up The river channel set-up for dam break modeling is the same as for the HD model except that the dam break structure is located in a separate reservoir branch, which contains 3 calculation points, i.e., two h-points and one Q-point. If a spillway is added to the dam, it can be described as a separate branch with 3 calculation points. The dam and spillways are located at a Q-point. The river set-up with a dam and, with dam and spillway are shown in Fig. 11.2 and Fig. 11.3 respectively. 11.2.5.2 Description of reservoir and appurtenant structures a) Reservoir To obtain an accurate description of the reservoir storage characteristics, the reservoir is normally modeled as a single h-point in the model. This will usually correspond to the upstream boundary of the model, where also the inflow hydrograph is also specified. The description of the reservoir storage is entered in the processed data. The surface storage area of the dam is described as a function of the water level and it is entered as additional flooded area. The lowest water level given for the reservoir should be somewhere below the final breach elevation of the dam. The cross-sectional area is set to a large finite value and is used only for calculating the inflow head loss into the breach. The inflow head loss can be calculated as : 2 H = (V s /2g) C i [1-(A s / A res )] Where, V s = Velocity through the breach C i = Inflow head loss coefficient A s = Flow area through the breach, and A res = Cross-sectional area of the reservoir In order to obtain a reasonable head loss description it is only necessary that A res >> A s so that [1-(A s /A res )] = 1. The hydraulic radius is set to any non-zero value. 160

The total surface area of the reservoir is calculated as: A total = b.2 x + Additional flooded area Since the total surface area is already described by the additional flooded area, the first term should be equal to zero. Therefore, the width b should be set to zero. b) Dam At the Q point, where the dam break structure is located, the momentum equation is replaced by an equation, which describes the flow through the structure. As the momentum equation is not used at the Q point, the x step is of no relevance. The maximum x for the river branch, where the dam is to be placed should, therefore, be greater than the distance between two cross-sections in the reservoir branch, so that no cross-section is interpolated between the actual cross-sections. c) Spillways and other structures At the node, where two branches meet (Fig 3) the surface flooded area is taken as the sum of the individual flooded areas specified at the h-points. Therefore, if the reservoir storage has already been specified at the reservoir h-point, the spillway h-point should not contain any flooded areas. Both the width b, and the additional flooded area should be set to zero and other parameters such as the cross-sectional area and hydraulic radius should be the same as for the reservoir. 11.2.5.3 Boundary conditions for dam break modeling The boundary conditions must be specified at both upstream and downstream limits of the model. The upstream boundary will generally be an inflow into the reservoir at the first reservoir h- point. The downstream boundary will generally be a stage-discharge relationship at the last cross section of the set up. 11.2.6 Specifications of Dam Break Structures The following information relating to dam break structures need to be specified: (i) Geometrical specifications (ii) Breach characteristics (iii) Failure moment, and (iv) Failure mode 161

11.2.6.1 Breach development Earth and Rockfill dams usually do not collapse instantaneously, but they develop breaches, which increase gradually. The failure time may vary between a few minutes up to a few hours, depending on amongst other, the dam geometry and the construction material. The development of the breach determines the breach outflow hydrograph, and an accurate description of the breach development is, therefore, required in near field dam breach studies. In the far-field studies, an accurate flood routing procedure is of more importance, because the outflow variation is rapidly damped out as the flood propagates downstream. 11.2.6.2 Failure modes The dam break module of MIKE 11 allows selection of one of various breach development modes. Either linear failure mechanism or an erosion based formulation may be selected. The linear failure mode assumes a linear increase in the breach dimensions in time between specified limits. In the erosion based mode, the increase in breach dimensions is calculated from the prevailing hydraulic conditions in the breach, and from the given geometrical data. For both modes, limits of the final breach width and level are specified. These may be determined, for example, by the original valley embankments. a) Linear failure modes The necessary data required to fully specify a linear dam failure are shown in Fig. 11.4. In addition, the user specifies the duration of the breach development and whether the failure is to commence at a given time, or is initiated by overtopping of the dam. This facility has applications in simulating the cascading failure of several dams located on the same river. b) Erosion based failure The enlargement of the breach in earth fill dams from erosion of the dam core material may also be determined from sediment transport considerations. Erosion based breach formulations are based on sediment continuity equation for the breach. Numerous sediment transport formulae are available, of which two have been implemented in the breach formulation, being those of Engelund- Hansen (1967), and Meyer-Peter and Muller (1947). Modeling of the variation of the width of the breach is more difficult to relate to the classical theories of sediment transport. Due to the development of a wall boundary layer along the often very steep side walls of the breach, the 162

theories for bed load and suspended load do not apply. As an approximation, the sediment transport at the sloping walls is assumed to be proportional to that in the central part of the breach. The coefficient of proportionality (side erosion index) is of the order of 0.5 1.0. 11.2.7 Initial Conditions Though in many cases, dam failure may occur on a dry river bed downstream of the dam, but such conditions are not possible in MIKE 11, which require a finite depth of water, in order to ensure the continuity of the finite difference algorithm. Therefore, before a dam break is actually simulated, it is necessary to create a steady state hot-start file, which can be used for all subsequent dam break simulation. This file is created by: (i) Giving a lateral inflow at the first h-point in the river (ii) Setting the inflow into the reservoir to zero, and (iii) Specifying the dam break structure to fail by overtopping, ensuring that the dam crest level is greater than the specified reservoir level. Initial conditions (water level and discharge) must be specified in HD parameter file, including the reservoir level, at which the dam break simulation should commence. The set-up should be run until a steady state condition is reached (i.e., Q=constant=lateral inflow up to the downstream boundary). 11.2.8 Dam Break Simulations The dam break simulation may be carried out using the hotstart file generated as mentioned above, specifying the upstream boundary as the inflow hydrograph. The time step depends upon the slope of the river bed and should be selected of the order of 0.5 to 5 minutes according to the slope. 11.3 DESCRIPTION OF THE PROJECT Lower Siang (9 X 300 MW) is lower most project proposed on river Siang, as a part of series of projects in the Siang basin in Arunachal Pradesh, India. A number of projects have been proposed in the Siang basin with a view to exploit the vast hydro potential. The proposed dam site is situated near Bodak village, about 23km upstream of Pasighat. The project envisages construction of 86m high concrete dam (above river bed level) on Siang River with a gross storage capacity of 1421 163

Mcum at FRL 230m and submergence area as 5151 Ha. The reservoir length at FRL is approximately 77.5 km and 28.5 km along main Siang River and Siyom River respectively. The Probable Maximum Flood (PMF) for Siang Lower HE Project is 60115 cumec. The catchment of Siang Lower project lies between latitude 27 59 30 N and 31 10 0 N and 82 0 0 E and 97 20 0 E longitudes. The catchment area upto the proposed dam site of Lower Siang project is about 2,50,594 sq.km, out of which about 2,36,556 sq.km lies in Tibet and the remaining i.e. about 14,038 sq.km lies in India. The catchment plan of river Siang draining in Indian territory is given in Fig. 11.5. The upstream elevation along the dam axis is given in Fig. 11.6. 11.3.1 Salient Features of the Project The salient features of Lower Siang HE Project are given below; I. Location i) State Arunachal Pradesh ii) River Siang iii) Dam site Near Bodak village about 23 km upstream of Pasighat iv) Nearest airport Guwahati II. Hydrology i) Catchment area at dam site 2505946 sq.km ii) PMF 60115 cumec III. Reservoir i) FRL El. 230 m ii) MWL El. 234.40 m iii) MDDL El. 220.0m iv) Gross storage at FRL 1421 M Cum IV. Concrete Gravity Dam i) Total length 709.49 m ii) Non-overflow 549.49 m iii) Overflow 160 m iv) Top of dam El. 235 m v) Height of dam above 111 m deepest foundation 164

V. Spillway i) Type Gated Ogee Spillway iii) No. of radial gates 8 iii) Size of each radial gate 20m (W) x 22.6m (H) iv) Crest level of spillway EL 208 m v) No. of under sluices 14 vi) Size of each sluice radial gate 7m (W) x 12m (H) vv) Sill level of radial under sluice gate EL 168 m 11.4 INPUT DATA AND MODEL SETUP 11.4.1 Input Data Requirement Dam break flood analysis requires a range of data to depict accurately to the extent possible the topography and hydraulic conditions of the river course and dam break phenomenon. The important data required are; (i) Cross sections of the river from dam site and up to location downstream of the dam to which the study is required (ii) Elevation-surface area relationship of the reservoir (iii) Rating curve of spillway and sluices (iv) Salient features of the all hydraulic structures at the dam site and also in the study reach of the river (v) Design flood hydrograph (vi) Stage-discharge relationship at the last river cross section of the study area (vii) Manning s roughness coefficient for different reaches of the river under study (viii) Rating curve of all the hydraulic structures in the study reach of the river For the present study, the following data has been used; 11.4.1.1 River cross sections For dam break studies of Lower Siang HE Project the Siang river for a length of 51.31 km downstream of the dam site, has been represented in the model by cross sections taken at regular interval varying from 5 km to 10 km. To get the discharge and water level at every 1 km interval the dx maximum in the MIKE11 model setup has been specified as 1000 m. In the case of extreme 165

floods the flood water spreads beyond the normal course of the river, where the resistance to flow will be high due to presence of bushes, vegetation etc. Considering the above the Manning's roughness coefficient for the entire study reach of the river has been taken as 0.040. 11.4.1.2 Reservoir and dam The reservoir has been represented in the model by a separate reservoir branch and its elevation-surface area relation, which has been specified at Chainage 0 km of the reservoir branch, is given in Table-11.1. Since the reservoir has been represented by elevation-surface area at chaingae 0 m of the reservoir branch, the length of the reservoir branch can be assumed arbitrarily to any value as the same does not play any role in the calculations performed by MIKE11 model. Further the dam can be placed anywhere in the reservoir branch within the assumed length of reservoir branch of model set up. For the present case the length of the reservoir branch has been assumed as 5000 m. The dam has been placed at Chainage 2500 m of the reservoir branch and dam breach parameters specified therein. Table-11.1: Elevation-Area relationship of the reservoir Elevation (m) Surface area (sq m) 149.4 0 160 1756400 170 4221900 180 8101700 190 13792200 200 21693100 210 28245300 220 39793600 230 51159200 240 62189000 250 77561000 11.4.1.3 Spillway and Sluices The spillway with 8 radial gates of size 20m (W) x 22.6m (H) each has been represented in the model by its rating curve as given in Table-11.2. The same has been specified at chainage 2500 m of the spillway branch of the model setup. The sluice with 14 radial gates of size 7m (W) x 12m 166

(H) each and invert level at EL 168 m has been specified as control structure at chainage 2500 m of sluice branch of model set up. The gate opening for all 14 gates of the sluice has been correlated with the reservoir water level with gates opening varying from 0.1 m to 12 m for reservoir level at EL 230 m (FRL) and EL 234.4 m (MWL) respectively. Table-11.2 Rating curve of spillway Reservoir level (m) Discharge through 8 spillway gates fully opened (cumec) 208 0 210 980 212 2754 214 5026 216 7687 218 10672 220 13934 222 17440 224 21162 226 25079 228 29170 230 33419 232 37812 234 42335 235 44641 11.4.1.4 Probable Maximum Flood Hydrograph (PMF) The probable maximum flood hydrograph which has been used as the upstream boundary of the dam break model set up and applied at chainage 0 km of the reservoir branch is given in Table-11.3. Table-11.3 : Probable Maximum Flood Hydrograph (PMF) Time (hour) Discharge (cumec) Time (hour) Discharge (cumec) Time (hour) Discharge (cumec) Time (hour) Discharge (cumec) 0 30500 51 52591 102 37348 153 32894 1 31652 52 51088 103 37187 154 32877 2 31665 53 50374 104 36974 155 32860 167

3 32459 54 49659 105 36760 156 32833 4 32470 55 49472 106 36466 157 32807 5 32486 56 49670 107 36318 158 32780 6 32512 57 50337 108 36171 159 32767 7 32650 58 50116 109 35998 160 32753 8 32788 59 49941 110 35826 161 32727 9 32926 60 49952 111 35611 162 32700 10 33063 61 50022 112 35504 163 32674 11 33288 62 50092 113 35396 164 32661 12 33512 63 50231 114 35282 165 32648 13 33960 64 51188 115 35167 166 32638 14 33909 65 52543 116 34994 167 32627 15 34858 66 54096 117 34908 168 32617 16 35359 67 55879 118 34821 169 32611 17 35860 68 57833 119 34718 170 32606 18 36863 69 60115 120 34615 171 32592 19 37814 70 59641 121 34478 172 32577 20 38765 71 58890 122 34409 173 32562 21 40180 72 58047 123 34340 174 32555 22 40441 73 56861 124 34265 175 32547 23 40702 74 55119 125 34190 176 32540 24 41137 75 53238 126 34095 177 32533 25 41572 76 52038 127 34048 178 32526 26 42164 77 51600 128 34001 179 32522 27 42755 78 51162 129 33933 180 32518 28 43938 79 50724 130 33865 181 32521 29 45399 80 50287 131 33806 182 32524 30 46860 81 49898 132 33751 183 32530 31 48670 82 49509 133 33696 184 32536 32 50690 83 48731 134 33641 185 32542 33 53276 84 47797 135 33586 186 32541 34 53085 85 46843 136 33530 187 32539 35 52855 86 45140 137 33473 188 32538 36 52730 87 44289 138 33417 189 32537 37 52445 88 43437 139 33388 190 32536 38 52240 89 42668 140 33360 191 32027 39 52034 90 41899 141 33310 192 31518 40 51622 91 41075 142 33262 193 31009 41 52424 92 40663 143 33213 194 30500 42 53504 93 40251 144 33188 168

43 55051 94 39819 145 33163 44 56597 95 39386 146 33121 45 58673 96 39025 147 33078 46 58340 97 38694 148 33036 47 57694 98 38362 149 33015 48 56794 99 38096 150 32994 49 55894 100 37829 151 32961 50 54094 101 37508 152 32927 11.4.1.5 Downstream boundary For the present case the study reach of the river is about 51.13 km downstream of the Siang Lower dam axis. In order to have no influence of the downstream boundary in the study reach of the river the same has been applied at a location 60 km downstream of the dam site. The downstream boundary (stage-discharge relationship) worked out using Manning s equation is given in Table-11.4. Table-11.4 Stage-discharge relationship - downstream boundary of MIKE11 model set up Stage (m) Discharge (cumec) 107.71 0.00 108.69 18.83 109.67 104.22 110.65 309.58 111.64 703.76 112.62 1376.49 113.60 2680.25 114.58 4937.07 115.56 8452.61 116.54 13649.41 117.52 20909.84 118.51 30795.54 119.49 45646.69 120.47 67584.30 121.45 96556.71 122.43 133342.50 123.41 178543.29 123.74 195388.52 The MIKE11 model set up for the dam break studies is given in Fig.11.7 169

11.5 DAM BREAK AND OTHER HYDRODYNAMIC SIMULATIONS 11.5.1 Selection of Dam Breach Parameters Estimation of the dam break flood will depend on time of failure, extent of overtopping before failure, size, shape and time of the breach formation, etc., which are called dam breach parameters. The breach characteristics that are needed as input to the existing dam break models are i) Initial and final breach width; ii) Shape of the breach; iii) Time duration of breach development, and iv) Reservoir level at time of start of breach. The predominant mechanism of breach formation is, to a large extent, dependent on the type of dam and the cause due to which the dam failed. A study of the different dam failures indicate that concrete arch and gravity dams breach by sudden collapse, overturning or sliding away of the structure due to inadequate design or excessive forces that may result from overtopping, earthquakes and deterioration of the abutment or foundation material. As per the UK Dam Break Guidelines and U.S. Federal Energy Regulatory Commission (FERC) Guidelines, in the case of concrete gravity dams, the breach width should be taken 0.2-0.5 times the crest length of the dam. The breach development time for gravity dam should be about 0.2 hour. The breach depth can been taken corresponding to the relatively weaker locations in the dam such as galleries, sluices etc. 11.5.2 Critical conditions for Dam Break Study The critical condition for a dam break study is when the reservoir is at FRL and design flood hydrograph is impinged. For the Lower Siang HE Project the FRL is at EL 230 m and MWL is at EL 234.4 m. For the reservoir routing it has been assumed that all spillway gates are fully opened when the PMF impinges in the reservoir with initial reservoir level at FRL. All the 14 gates of the sluices have been assumed to open gradually from gate opening of 0.1 m for reservoir level at FRL and gate opening of 12 m for the reservoir level at MWL. The maximum water level reached in the reservoir routing is 233.60 m, which occurs 72 hours after the impingement of PMF. The time series of reservoir level during routing the PMF through the reservoir is given in Fig. 11.8. The PMF, discharge through spillway and sluices is given in Fig. 11.9. 170

From the reservoir routing of the PMF it has been concluded that the spillway and sluice capacity is quite adequate to negotiate the PMF safely through the reservoir. Further, since the top of the dam is at El 235 m, no overtopping of the dam will occur. Hence for the hypothetical case of dam break simulation and also to get the maximum dam breach flood peak it would be appropriate to assume the dam breaches when the reservoir level is at EL 233.40 m. Hence the dam has been assumed to breach 72 hours after the impingement of PMF when the reservoir level is at maximum EL of 233.60 m. 11.5.3 Breach Parameters selected for Sensitivity Analysis of Dam Break Simulation Considering the criteria for selection of breach parameters and critical condition for the dam break study as discussed in para 11.5.1 and 11.5.2, two different cases of breach parameters as given in (Table-11.5) have been identified for sensitivity analysis of dam break simulations. In all these two cases, the initial breach elevation has been taken corresponding to the top of dam (EL 235 m). The final bottom elevation of the breach has been taken as 168 m corresponding to invert level of sluices and 194.925 corresponding to invert level of power intake. The breach side slope has been taken as zero as applicable for concrete gravity dam. The time of failure has been taken as 10 minutes for the instantaneous failure of gravity dam. Table-11.5 Breach parameters considered for sensitivity analysis Case No. Breach Elevation (m) Initial Final Breach Width (m) Breach Development Time (Minutes) 171 Max. Discharge through breach (cumec) Remarks 1. 235 168 144 10 102861 Four non-overflow blocks of 36 m each with cumulative breach width of 144 m considered to break up to the invert level of sluice at EL 168 m. 2. 235 194.925 225 10 72652 Non-overflow blocks between left and right spillway with a total width of 225 m considered to break up to invert level of power intake at EL 194.925 m As case-1 generates the maximum discharge through the breach due to 67 m depth of breaching section, the same has been finalized for detailed outputs of dam break simulation.

11.5.4 Dam Break Simulation (breach width 144 m, breach depth 67 m, breach development time 10 minutes) Taking the above breach parameter and critical condition of para 11.5.2 the dam break simulation has been simulated. The time base of PMF used for hydrodynamic simulations in MIKE11 model set up varies from 1-8-2009 00:00:00 hours to 9-8-2009 02:00:00 hours. In the simulation the dam has been assumed to breach at 4-8-2009 00:00:00 hours i.e. 72 hours after the impingement of the PMF, to get the maximum discharge through breach. The discharge through the breach in this case is 102861 cumec, which occurs on 4-8-2009 at 00:10:00 hours i.e. 10 minutes after the start of breaching. The combined discharge of the breach, spillway and sluices just downstream of the dam has been calculated as 137304 cumec which gets attenuated to 114035 cumec at about 50 km downstream of the dam. The time series of discharge through breach is given in Fig. 11.10. The dam break flood hydrograph just downstream of dam comprising of cumulative discharge of breach, spillway and sluices is given in Fig. 11.11. RESERVOIR 2500 is the breach location as per MIKE11 model set up. The peak of the dam breach flood just down stream of the dam is 137304 cumec which includes about 60000 cumec due to PMF. Hence the contribution of reservoir storage in the dam breach flood peak is about 77000 cumec. The maximum discharge, water level and their time of occurrence at different locations of the Siang river downstream of the dam are given in Table 11.6 and 11.7 respectively. From the Table-11.8, it can be seen that the flood peak of 137304 cumec gets attenuated to 114035 cumec during its translation for about 50 km reach of the river. The travel time of dam break flood for this reach is about 3 hours and the velocity of the flood wave is about 17 km/h. Table-11.6 : Maximum discharge due to dam breach flood (breach width 144 m and breach depth 67 m) Note : SIANG 494.00 means location of Siang river 494 m d/s of Siang Lower dam. The same way all other locations should be read. The breach has been assumed to start on 4-8-2009 at 00:00:00 hours Chainage (m) d/s of Siang Maximum discharge (cumecs) Time of occurrence Lower dam (Date:hours:Minutes:seconds) SIANG 494.00 137304 4-8-2009 00:39:59 SIANG 1482.00 137076 4-8-2009 00:39:59 SIANG 2470.00 136815 4-8-2009 00:39:59 172

SIANG 3458.00 136567 4-8-2009 00:39:59 SIANG 4446.00 136284 4-8-2009 00:39:59 SIANG 5381.67 136023 4-8-2009 00:39:59 SIANG 6265.00 135797 4-8-2009 00:39:59 SIANG 7148.33 135668 4-8-2009 00:50:00 SIANG 8031.67 135687 4-8-2009 00:50:00 SIANG 8915.00 135624 4-8-2009 00:50:00 SIANG 9798.33 135634 4-8-2009 00:50:00 SIANG 10700.83 135628 4-8-2009 00:50:00 SIANG 11622.50 135462 4-8-2009 00:50:00 SIANG 12544.17 135524 4-8-2009 00:50:00 SIANG 13465.83 135392 4-8-2009 00:50:00 SIANG 14387.50 135112 4-8-2009 00:50:00 SIANG 15309.17 135083 4-8-2009 00:50:00 SIANG 16207.00 135151 4-8-2009 00:50:00 SIANG 17081.00 134703 4-8-2009 01:00:00 SIANG 17955.00 134798 4-8-2009 01:00:00 SIANG 18829.00 134738 4-8-2009 01:00:00 SIANG 19703.00 134615 4-8-2009 01:00:00 SIANG 20635.00 134351 4-8-2009 01:00:00 SIANG 21625.00 133844 4-8-2009 01:00:00 SIANG 22615.00 133337 4-8-2009 01:09:59 SIANG 23605.00 133116 4-8-2009 01:09:59 SIANG 24595.00 132612 4-8-2009 01:09:59 SIANG 25563.89 131790 4-8-2009 01:09:59 SIANG 26511.67 131417 4-8-2009 01:20:00 SIANG 27459.45 130870 4-8-2009 01:20:00 SIANG 28407.22 129993 4-8-2009 01:30:00 SIANG 29355.00 129704 4-8-2009 01:30:00 SIANG 30302.78 129087 4-8-2009 01:30:00 SIANG 31250.55 128643 4-8-2009 01:39:59 SIANG 32198.33 128309 4-8-2009 01:39:59 SIANG 33146.11 127661 4-8-2009 01:50:00 SIANG 34100.50 127439 4-8-2009 01:50:00 173

SIANG 35061.50 126784 4-8-2009 01:50:00 SIANG 36022.50 126464 4-8-2009 02:00:00 SIANG 36983.50 125870 4-8-2009 02:00:00 SIANG 37944.50 125455 4-8-2009 02:09:59 SIANG 38905.50 124859 4-8-2009 02:09:59 SIANG 39866.50 124392 4-8-2009 02:20:00 SIANG 40827.50 123689 4-8-2009 02:20:00 SIANG 41788.50 122928 4-8-2009 02:30:00 SIANG 42749.50 121623 4-8-2009 02:39:59 SIANG 43678.89 120613 4-8-2009 02:39:59 SIANG 44576.67 119567 4-8-2009 02:50:00 SIANG 45474.45 118560 4-8-2009 02:50:00 SIANG 46372.22 117746 4-8-2009 03:00:00 SIANG 47270.00 116838 4-8-2009 03:09:59 SIANG 48167.78 116119 4-8-2009 03:09:59 SIANG 49065.55 115427 4-8-2009 03:20:00 SIANG 49963.33 114644 4-8-2009 03:20:00 SIANG 50861.11 114035 4-8-2009 03:30:00 Note : The dates shown are relative dates as used in MIKE11 model set up. Table-11.7 Maximum water level due to dam breach flood (breach width 144 m and breach depth 67 m) Note : SIANG 988.00 means location of Siang river 988 m d/s of Siang Lower dam. The same way all other locations should be read. * Cross sections of Siang river interpolated by MIKE11 model Chainage (m) d/s of Bed level (m) Maximum water Time of occurrence Siang Lower dam level (m) (Date:hours:Minutes:seconds) SIANG 0.00 149.40 210.71 4-8-2009 00:50:00 SIANG 988.00 * 209.78 4-8-2009 00:50:00 SIANG 1976.00 * 208.73 4-8-2009 00:50:00 SIANG 2964.00 * 207.66 4-8-2009 00:50:00 SIANG 3952.00 * 206.03 4-8-2009 00:50:00 SIANG 4940.00 158.83 202.83 4-8-2009 00:50:00 174

SIANG 5823.33 * 200.68 4-8-2009 00:50:00 SIANG 6706.67 * 198.48 4-8-2009 00:50:00 SIANG 7590.00 * 196.24 4-8-2009 00:50:00 SIANG 8473.33 * 193.94 4-8-2009 00:50:00 SIANG 9356.67 * 191.74 4-8-2009 00:50:00 SIANG 10240.00 140.76 189.56 4-8-2009 00:50:00 SIANG 11161.67 * 188.54 4-8-2009 00:50:00 SIANG 12083.33 * 187.72 4-8-2009 00:50:00 SIANG 13005.00 * 186.58 4-8-2009 00:50:00 SIANG 13926.67 * 184.97 4-8-2009 00:50:00 SIANG 14848.33 * 182.83 4-8-2009 00:50:00 SIANG 15770.00 149.96 179.99 4-8-2009 01:00:00 SIANG 16644.00 * 177.78 4-8-2009 01:00:00 SIANG 17518.00 * 175.87 4-8-2009 01:00:00 SIANG 18392.00 * 174.16 4-8-2009 01:00:00 SIANG 19266.00 * 172.60 4-8-2009 01:00:00 SIANG 20140.00 145.67 171.21 4-8-2009 01:09:59 SIANG 21130.00 * 169.36 4-8-2009 01:09:59 SIANG 22120.00 * 167.48 4-8-2009 01:09:59 SIANG 23110.00 * 165.61 4-8-2009 01:09:59 SIANG 24100.00 * 163.90 4-8-2009 01:20:00 SIANG 25090.00 138.19 162.58 4-8-2009 01:20:00 SIANG 26037.78 * 161.25 4-8-2009 01:20:00 SIANG 26985.55 * 159.90 4-8-2009 01:30:00 SIANG 27933.33 * 158.51 4-8-2009 01:30:00 SIANG 28881.11 * 157.08 4-8-2009 01:30:00 SIANG 29828.89 * 155.62 4-8-2009 01:39:59 SIANG 30776.67 * 154.12 4-8-2009 01:39:59 SIANG 31724.45 * 152.57 4-8-2009 01:39:59 SIANG 32672.22 * 151.01 4-8-2009 01:50:00 SIANG 33620.00 137.66 149.52 4-8-2009 01:50:00 SIANG 34581.00 * 147.99 4-8-2009 02:00:00 SIANG 35542.00 * 146.47 4-8-2009 02:00:00 SIANG 36503.00 * 144.94 4-8-2009 02:09:59 175

SIANG 37464.00 * 143.41 4-8-2009 02:09:59 SIANG 38425.00 * 141.89 4-8-2009 02:20:00 SIANG 39386.00 * 140.37 4-8-2009 02:20:00 SIANG 40347.00 * 138.86 4-8-2009 02:30:00 SIANG 41308.00 * 137.43 4-8-2009 02:39:59 SIANG 42269.00 * 136.20 4-8-2009 02:50:00 SIANG 43230.00 121.97 135.32 4-8-2009 02:50:00 SIANG 44127.78 * 134.60 4-8-2009 03:00:00 SIANG 45025.55 * 133.88 4-8-2009 03:09:59 SIANG 45923.33 * 133.16 4-8-2009 03:09:59 SIANG 46821.11 * 132.45 4-8-2009 03:20:00 SIANG 47718.89 * 131.73 4-8-2009 03:20:00 SIANG 48616.67 * 131.01 4-8-2009 03:30:00 SIANG 49514.45 * 130.29 4-8-2009 03:39:59 SIANG 50412.22 * 129.56 4-8-2009 03:39:59 SIANG 51310.00 114.71 128.83 4-8-2009 03:50:00 Note : The dates shown are relative dates as used in MIKE11 model set up. 11.5.5 Maximum discharge and water level in SIANG river due to occurrence of PMF without dam breach To know the maximum discharge and water levels at different locations of Siang river downstream of the dam due to occurrence of PMF, when reservoir is at FRL, but without any dam breach, the simulation has been simulated in MIKE11 model. The maximum discharge and water level obtained at the different locations along the river reach is given in Table-11.8 and 11.9 respectively. The velocity of flood wave in this case is also about 17 km/h. Table 11.8 Maximum discharge due to occurrence of PMF without dam breach The PMF has been impinged in to the reservoir on 1-8-2009 at 00:00:00 hours. The PMF peak is 60115 cumec occurring on 3-8-2009 21:00:00 hours as per MIKE11 simulation time. Chainage (m) d/s of Lower Maximum discharge (cumecs) Time of occurrence Siang dam (Date:hours:Minutes:seconds) SIANG 494.00 58172 4-8-2009 00:00:00 SIANG 1482.00 58171 4-8-2009 00:00:00 SIANG 2470.00 58169 4-8-2009 00:00:00 176

SIANG 3458.00 58168 4-8-2009 00:00:00 SIANG 4446.00 58167 4-8-2009 00:00:00 SIANG 5381.67 58166 4-8-2009 00:00:00 SIANG 6265.00 58165 4-8-2009 00:00:00 SIANG 7148.33 58163 4-8-2009 00:00:00 SIANG 8031.67 58162 4-8-2009 00:00:00 SIANG 8915.00 58161 4-8-2009 00:00:00 SIANG 9798.33 58159 4-8-2009 00:00:00 SIANG 10700.83 58158 4-8-2009 00:00:00 SIANG 11622.50 58156 4-8-2009 00:00:00 SIANG 12544.17 58154 4-8-2009 00:00:00 SIANG 13465.83 58152 4-8-2009 00:00:00 SIANG 14387.50 58150 4-8-2009 00:00:00 SIANG 15309.17 58148 4-8-2009 00:00:00 SIANG 16207.00 58146 4-8-2009 00:00:00 SIANG 17081.00 58143 4-8-2009 00:00:00 SIANG 17955.00 58139 4-8-2009 00:00:00 SIANG 18829.00 58134 4-8-2009 00:00:00 SIANG 19703.00 58128 4-8-2009 00:00:00 SIANG 20635.00 58119 4-8-2009 00:00:00 SIANG 21625.00 58110 4-8-2009 00:00:00 SIANG 22615.00 58112 4-8-2009 01:00:00 SIANG 23605.00 58119 4-8-2009 01:00:00 SIANG 24595.00 58124 4-8-2009 01:00:00 SIANG 25563.89 58130 4-8-2009 01:00:00 SIANG 26511.67 58134 4-8-2009 01:00:00 SIANG 27459.45 58138 4-8-2009 01:00:00 SIANG 28407.22 58140 4-8-2009 01:00:00 SIANG 29355.00 58140 4-8-2009 01:00:00 SIANG 30302.78 58139 4-8-2009 01:00:00 SIANG 31250.55 58136 4-8-2009 01:00:00 SIANG 32198.33 58131 4-8-2009 01:00:00 SIANG 33146.11 58123 4-8-2009 01:00:00 SIANG 34100.50 58111 4-8-2009 01:00:00 177

SIANG 35061.50 58096 4-8-2009 01:00:00 SIANG 36022.50 58077 4-8-2009 01:00:00 SIANG 36983.50 58077 4-8-2009 02:00:00 SIANG 37944.50 58089 4-8-2009 02:00:00 SIANG 38905.50 58097 4-8-2009 02:00:00 SIANG 39866.50 58103 4-8-2009 02:00:00 SIANG 40827.50 58105 4-8-2009 02:00:00 SIANG 41788.50 58102 4-8-2009 02:00:00 SIANG 42749.50 58093 4-8-2009 02:00:00 SIANG 43678.89 58076 4-8-2009 02:00:00 SIANG 44576.67 58053 4-8-2009 02:00:00 SIANG 45474.45 58024 4-8-2009 02:00:00 SIANG 46372.22 57990 4-8-2009 03:00:00 SIANG 47270.00 58000 4-8-2009 03:00:00 SIANG 48167.78 58004 4-8-2009 03:00:00 SIANG 49065.55 58003 4-8-2009 03:00:00 SIANG 49963.33 57997 4-8-2009 03:00:00 SIANG 50861.11 57984 4-8-2009 03:00:00 Note : The dates shown are relative dates as used in MIKE11 model set up. Table-11.9 Maximum water level due to occurrence of PMF without dam breach Note : SIANG 988.00 means location of Siang river 988 m d/s of Lower Siang dam. The same way all other locations should be read. * Cross sections of Siang river interpolated by MIKE11 model Chainage (m) d/s of Lower Siang dam Bed level (m) Maximum water level (m) 178 Time of occurrence (Date:hours:Minutes:seconds) SIANG 0.00 149.40 194.90 4-8-2009 00:00:00 SIANG 988.00 * 194.23 4-8-2009 00:00:00 SIANG 1976.00 * 193.41 4-8-2009 00:00:00 SIANG 2964.00 * 192.37 4-8-2009 00:00:00 SIANG 3952.00 * 190.87 4-8-2009 00:00:00 SIANG 4940.00 158.83 188.33 4-8-2009 00:00:00 SIANG 5823.33 * 186.36 4-8-2009 00:00:00 SIANG 6706.67 * 184.49 4-8-2009 00:00:00