Dam failure and catastrophic flooding impacts

Dam Break

Excutive Summary

Analytical tools and techniques used in approximating dam breakages are usually evaluted and comparison made in orders to ascertain there effectiveness. During dam construction, it is necessary to evaluate the potentiality of dam failure modes, breach the necessary paramenters related to failuere modes, and define routing and the map of the consequent discharge hydrograph. This paper outlines how mapping of propable inudation emanating form the dam failures needs numerous elements i.e. likely dam failures, break parameters, hydrologic scenario that are connected to failure models, including routing and mapping of the consequent discharge hydrograph. The unsteadily flowing routing models i.e HEC-RAS are oftenly applied in computation and rfelction of the dam's downstream consequences emaneting from dam failures. Approximation of the break elements i.e. time of development and width is carried out externally to the model in use.

Introduction

In the recent past most agencies are enganging themselves in reviewing and updating safety measures concerned with dam safety and risk management. Studies relating to Dam break, involves updates on likely maximum flood inflow hydrographs and routing of most probable dam break floods to evaluate downstream impacts . The approximation of these estimate dimensions and developments is essentail in investigation of the dam safety. The flows in the downstream location and other inundated areas of the dam is directly proportional to the breach parameters, other factors tat influence include; the breach location, time of formation and the size (USACE, 1980). All these parameters are essentail in dam failure analysis.

The geometric design of a dam break ought to be approximated in order to reflect the resultant flood wave and down stream impacts . In this case an example of models that can be used to demonistrate dam breach outflow hydrograph computing and downstream routing include HEC-HMS (HEC, 2006b), HEC-RAS (HEC,2006a), NWS-DAMBRK (Fread, 1988b) amon others. The models requires the probable breach properties to be approximated outside the model. During investigation of dam break, breach dimenisions and time taken in develoment must be put in consideration in estimation of the failure scenarios (Washington State Dept. Of Ecology 1992). The necessitaties include various failure modes and different hydrologic events . The breach parameters related to hydrologics of the level of the likely maximum flood level may be different from that of sunny day failures broght about, likely, by a seismic event. Hence, anumber of breach parameters ought to be made for each connection of hydrologic event and failure scenario

Dam Break Simulation with HEC-RAS model:

Dam break tests are performed with HEC-RAS model for dam safety study and also flood damage analyses for circumstances that involves likely levee breaks. These studies in most cases need to ascertain the following circumstances and their associated uncertainties.

Hydrologic Scenarios (USACE, 1980).

Sunny day failure and,

Large hydrologic event

Failure Modes, they include (USACE, 1980).;

Overtopping

Piping

Breach definition input data and information (USACE, 1980).

Failure mode and initiation specification

Breach growth progression

Breach side slopes

Breach development time

Ultimate breach bottom elevation

Maximum breach bottom width

• Failure initiation

Selected time

Reservoir stage

Reservoir stage and duration

Data Required for Breach Simulation with HEC-RAS model .

The model employs the following parameter to define dam break (Froehlich, 1995);

Formation Time: critical breach development period

Trigger Condition: pool elevation, or clock time pool elevation and duration

Weir and Pipe Flow Coefficients: weir coefficients are applied in computation of weir flow / overtopping, whereas orifice coefficient is aplied in computation force flow./piping (Washington State Dept. Of Ecology 1992).

Location: transverse location of the centerline of the breach in the dam

Failure Mode: overtopping or piping

Shape and Progression: bottom width, bottom elevation, left and right side slopes. Linear or non-linear time formation

The dam break charactericts can be estimated in various ways which include;

Comparative analysis: this study involves making comparisons between historical failures of dams of similar sizes, water volume and materials. This is done through regression equations derived from past dam failures in order to estimate the peak outflow or breach size and time of development (Froehlich, 1995). Also computer-based models i.e computer programs that are used to develop models thus physically breaching the system by use erosion equations, priciples of hydrolics and soil mechanics equations.

Failure Location:

The location of the breach failure depends on several factors i.e. shape, type, structural elements of the dam and failure. All the factors relating to the dam, specifically historical record of the seepage and problems concerned with the foundation, ought to be put in consideration so that to place the breach in the most likely point of each failure mode. If the likely point is not identified, the centreline on the breach ought to be set to the centerline of the downstream mainchanel

Failure Mode

HEC-RAS and HEC-HMS modes are designed with hydraulic algorithms that simulates the hydrolics of piping failure and overtopping . In this case the failure mode is brought a bout by the mechanism which starts and develops the breach. Overtop failure begins at brim of the dam and spreads to maximun extends (USACE, 1980). On the other hand, piping failures begins from anywhere on elevations or locations and develops to the minor extents. The final breach size and growth or development time is more essentail in approximation of the outflow hydrograph as compared to the failure mode.

Breach development time.

Models such as HEC-RAS and HEC-HMS requires same input data as compared to the data a breach development time requires (Washington State Dept. Of Ecology 1992. This time can be discussed under:

Overlapping failure: it refers to the duration between which the breach has just eroded back to the upstream part of the dam, it is usually at the top to the time when it is completetly formed. For example, major erosion has ceased, but not up to the time the resevoir pool is emptied.

Piping Failure: it refers to the time starting at the point a major flow and material are flowing through the pipe failures to the stage the breach is fully developed.

Generally, Dam break concept focuses on description and hydraulics of overlapping breaches. The final size of the break ought to be defined. It is the size assumed to remain after the breach has stopped developing (Froehlich, 1995). The breach migt cease to develop when the whole dam has been eroded or when the reservoir has been completely been drained and hence there is luck of water to cause more erosion. Ultimately, the erosion stops at the bottom elevation, it can be either at the bed rock or reservior pool buttom, the size of the breach is reflected by the width of the bottom or the side slopes of the dam (Mac Donald, 1984).It is important to describe the beach initiation scenario called trigger mechanism, and the time it takes for the breach to develop to the maximum size. The time of growth is elapsed from the important flow through the breach to the period at which the final size is obtained. In the case of over lapping failure, the breach parameters ought to define the initiation process and weir process that controls water release from thm storage.

A number of options can be employed in identification of the breach initiation period. One of the methods involves specification of the duration of which the pool evaluation exceeds a specific threshold. This data may be influenced by geotechnical investigation. Another method involves, specification of pool elevations, the breach starts to develop soon at the point when the elevation is reached (Washington State Dept. Of Ecology 1992. Finally, the third method involves specification of the duration of simulation irrespective of the pool elevation. It is very important for "Sunny days" failures that are not connected with hydrologic events.

In the case shapes, the breach shape grows / changes in shape from the initiation stage to the ultimate configuration. The most simple growth rate is linear; in this case, its dimensions develop at a regular rate (Mac Donald, 1984). Routes are there that stimulates a breach which initially develops at fast pace then the pace drops as it approaches the end of the development time.

Determination of the size and growth rate of dam breaks

Estimation of the growth rate and size for dam breaks is not an easy task. Hence, simulations models facilatates the user to estimate the effects of a range of parameters from the data obtained from the examination. There are number of equations involved:

Regression Equations

There are number of regression equations that have been in the pasted applied in dam safety research obtained in literature and still are being put in consideration in making conclusion concerning guidance for the corps of engineers portfolio risk assessment studies i.e.Von Thun and Gillette, Macdonald and Langridge-monopolis and Froahlich (Froehlich, 1995). From these equations we obtain an standard literature review of this subject .

According to Froelich equation, Froelich used: Zoned earthen, rockfall data set and 63 earthen to come up with a set of equation s that were to be applied in estimation of the mean failure time, width and side slopes of dam break. During actual application, he calculated the height of the break by assuming that the break gegins from the top of the dam to the natural ground elevationbeing at the centreline of the break point unlike MacDonald and Langridge -- Monopolis (Froehlich, 1995)

Von Thun and Gillette (Von Thun, 1990)

He used a total of 57 dams from both the MacDonald and Langridge -- Monopolis and Froehlich researcg to come up with thier methodology. These methods advocates application of break slopes of 1.2H:1.2V; excepting dams that have cohesive layer of soil, whereas the side slopes of thedams ought to be arranged in the manner 0.6V:1.0V to 0.34:1.1V (Von Thun, and Gillette, 1990). Von Thun and Gillette come up with two dismillar set of equations to be used in break development time considering embankment matter.

NWS-BREACH (Fread, 1988a)

This model depends on the sediment mechanics and locomotion connections to simulate the dam break growth. In this model, the sediment transport method is employed to calculate the erosion pace and the size of the break in regard to the data pertaining the soil aspects of the dam and hydrograph inflow (Washington State Dept. Of Ecology 1992). Development of the break with time is computed by sediment transport equation, drastic fall of the dam as result of excessive hydrostastic force and width enlargement by slope stability. Sediment transport in this concept is in application with either cohesive or non-cohessive materials within the dam environment. Other models that can be employed in detemination of the dam break include SIMBA and HR-BREACH though they are still under development.

DAM BREACH FLOW SIMULATION USING HEC-RAS

While dams provide many desirable benefits to society, they also are a major hydro-modification to ecosystems, can be safety and boating hazards, and may degrade water quality of the river. In the Great Lakes District, dams that exceed a height of 6ft 2m and a pool volume 50 ac-ft are inspected and require a state permit to ensure they are properly maintained. For dams that fail inspections, dam owners are faced with four options: (1) do nothing, (2) modify the dam to such an extent that it is not subject to the regulations, (3) rehabilitate the dam to meet the regulatory guidelines of the permit, and (4) remove the dam. The chosen option often depends on the outcome, its economics, and the environmental and political pressures associated with the option (Gee and Brunner, 2007).

The modeling scheme in HEC-RAS is illustrated on graphical illustration of the breach dimension, location and the development which is connected with the computed upstream and downstream water surface profiles gives practical information for the expert and the customers of the learn. HECRAS applies dam breach parametric quantity formulated outwardly (applying the methods described above or any others considered suitable) to calculate the temporary development of a break in an lined up structure (dam) (Froehlich, 1995). The movements through that arrangement are calculated examining breach flow, overtopping flow, spillway discharges, gated flow, and sinking outcomes as a result of downriver backwater. Those numerous flow sections are used as an inner border line state for unfirm flow shaping of the pool and the downriver reach. The pool may be examined using either uncomplicated level pool direction-finding or as a wavering flow reach using cross sections. Variations in breach outflow hydrographs caused by different breach parameters will reduce as the submerge wave is routed through the downriver reach.

Figure 1. HEC-RAS Dam Breach Model (USACE, 1980)

SOME COMPARISONS

Currently, the breach parameter assessment techniques outlined above have been used in five circumstances. Two are theoretical failures at real schemes. One of these has computed cross sections in the pool and the downriver reach so that the results of in-pool and downstream course-plotting can be analyzed. Three actual remarkable failures are also repeated; one of which was a designed research. All of the cases talked about here are dominating failures. In all cases the breach progress was understood to be linear in time. The theoretical cases were presumed to initiate breaching when the overtopping was about 0.3 m; the described beginning time of failure was applied in the notable cases. Only the applications to the two historic breakdowns are offered here because of space limitations; complete effects can be found in (Gee and Brunner, 2007).

APPLICATIONS TO SOME HISTORIC OVERTOPPING FAILURES

Oros Dam

Oros dam (Brazil) was being constructed when it broke down by overtopping in March of 1960 (CEATI). Oros dam was constructed of a clay core with sand and rock shoulders and the height was about 35.5m. The experimental techniques of MacDonald, Froehlich and VonThun were used to this construction along with the BREACH process theory. The quantity of water on the rampage was projected to be 660*106 m3 (1).

The breach hemorrhage hydrographs calculated using these parameters are shown on Fig. 2 with the approximated outpouring hydrograph. The outflow hydrograph was derived from the record of water height during the occurrence (CEATI).

Figure 2. Breach Hydrographs for Oros Dam

Banqiao Dam (China)

Banqiao Dam broke down by looking down from a large aggressive state of weather with winds 64-72 knots (11 on the Beaufort degree) and rainfall and thunder and lightning in 1975 (CEATI). The Banqiao dam was built or erected using clay core containing shale. The upstream and downstream work was standardized earth. It can be presumed that, as a result of building processes (mostly non-mechanized), that the foundation was weakly compressed. The height of the Bangiao dam was about 24.5 meters high with a crest altitude at 116.34m. Crest width was 6m and length 2020m (Michael 1988). The upstream slope was 3H:1V and downstream 2.5H:1V. The capacity of the architectural plan for the channel that carries excess water over or around a dam or other obstruction and channel works was 1742 m3/sec; the approximated apex inflow was about 13,000 m3/sec when breaching took place. The approximated breach parameters are shown in Table 3.

The breach outflow hydrographs calculated applying these parameters are illustrated on Fig. 3 (Mohamed, 2002)

Field examinations for this occurrence consist of the pool altitude time history.These observations can be equated with the pool reduction in quantity calculated using HECRAS with the approximated breach parameters