Professional Documents
Culture Documents
David T. Williams, Ph.D., P.E., P.H., CFM, D.WRE National Technical Director, Water Resources, PBS&J dtwilliams@pbsj.com Michael DePue, P.E., CFM, D.WRE Associate Vice President, Water Resources, PBS&J mdepue@pbsj.com
Illinois Association for Floodplain and Stormwater Management Conference March 12-13, 2008, Tinley Park, Illinois
This paper is partially derived from short course lectures developed by the staff and engineers at the USACE Hydrologic Engineering Center (HEC)
Lawn Lake Dambreak (10 minutes to full breach), U/S of Estes Park
HEC-1 (COE) HEC-RAS (COE) DAMBRK (NWS) FLDWAV (NWS) Simplified DAMBRK (NWS) TR-61 (also known as WSP2, NRCS) TR-66 (Simplified Dam Breach Routing Procedure) BREACH (HR Wallingford, England)
General Comments
In HEC-RAS, a dam breach analysis is performed on an inline weir and a levee breach is performed on a lateral weir. The breach editor can be accessed from either the unsteady flow analysis editor or the breach button on the inline or lateral weir editor. To compare a breached with a not breached run, the project should have two plans, but they can (and in many cases should) have the same geometry.
Breach Progression
The Breach Progression editor defines the breach growth rate. By default, the breach growth is assumed to be linear from start to maximum size (Full Formation Time). A non-linear breach growth curve can be developed by entering a Time Fraction (from zero to 1.0) and a Breach Fraction (from zero to 1.0) and is plotted in the graphic next to the table. Adjusting the curve in makes a more realistic breach growth; however, the curve can sometimes be used in order to improve stability. If the program is having stability problems during part of the breach growth, the rate of growth during that time can be reduced by adjusting the curve.
Determining the size and growth rate for breaches is a very inexact science.
The use of several different procedures, to get a range of values, is generally recommended.
The following are good references and give guidance on breach analyses
Bureau of Reclamation, U.S. Department of the Interior (Dam Safety Research Report)
developed best-fit and envelope curves for both breach parameters and peak outflow from breached earthfill dams as a function of a breach formation factor.
MacDonald and Langridge-Monopolis (Journal of Hydraulic Engineering, Vol. 110, No. 5, May, 1984, p. 567-586)
Froehlich (Water Resources Engineering, 1995 ASCE Conference, San Antonio, Texas, August 14-18, 1995, p. 887-891)
Embankment Dam Breach Parameters Revisited developed a best-fit regression equation for peak flow based on a breach formation factor.
Victor M. Ponce (his website), Case Studies of Dambreach Geometry and Hydraulics shape factor representing dam-breach geometry is formulated and a relationship between Froude number based on peak discharge and the new shape factor is identified. http://ponce.sdsu.edu/dambreachgeometry.html
When performing a dam breach analysis, it is useful to develop a range of values for the various breach parameters. For instance, the user might create a table that has a high, medium, and low estimate for each breach parameter. By combining the parameters in various combinations, a range of dam break hydrographs can be created that [hopefully] represents a maximum and minimum hydrograph limit, as well as a most probable estimate.
Breach Sensitivity
High Max Bottom Width Side Slopes Time to Breach Medium Low
200
150
100
1.5 15 min
1.25 45 min
How important the breach parameters are depends largely on how far downstream from the dam the study area is located. At the toe of the dam, the computed peak flow often varies wildly within the assumed limits of the breach parameters. However, the study area further downstream, the breach parameters have less and less effect on the flow hydrograph.
The total volume of water in each of the different hydrographs is basically the same (being the stored water behind the dam). As the hydrographs move downstream, a sharp hydrograph (such as a dam break) will attenuate much more quickly than a flat hydrograph. How close they get to each other depends on the distance they are routed, steepness of the stream, roughness of the river and floodplain, and amount of floodplain storage available for attenuating the hydrograph.
Hydrographs from different assumed breach parameters can converge to produce the same peak flow (and for that matter, the same rise and fall) over a surprisingly small distance. In the above example, Fread found that the hydrographs have substantially converged within five miles and are indistinguishable by mile 13. Levee breaches in general and dam breaks in particular, have rapidly varying water surfaces and flows. Because of this, a small time step is required to maintain stability and accuracy; often on the order of a few seconds for a dam break problem.
Cross sections that are spaced too far apart below a dam break can cause stability and accuracy problems. The Courant condition should be kept close to or less than one (e.g., if the velocity is 20 ft/sec and a 10 sec. time step is used, the cross section spacing should be no closer than 200 feet). Dam breaks are usually associated with large amount of sediment, debris and turbulence. The Mannings n should be increased (particularly at cross sections near the dam) to account for these effects.
The reported breach time, especially for piping initiated breaches, often includes a significant period of time when the flow is, relatively speaking, a trickle. This time should not be included in the Critical Breach Time unless the user makes appropriate adjustments to the non-linear growth time. If the dam break has stability problems, turning the Mixed Flow option on may help (even if the flow is not going supercritical). This may reduce the accuracy of the water surfaces, particularly immediately downstream of the dam. This error is usually minor compared with the error caused by the unknowns in the breach parameters.
A weighting applied to the finite difference approximations available in HEC-RAS Computation Options and Tolerance menu Theoretically can vary from 0.5 to 1.0 Practical limit is from 0.6 to 1.0 1.0 = most stable; 0.6 = most accurate The default in RAS is 1.0
Start at 1.0 and reduce towards 0.6 as long as the model stays stable and produce reasonable results
Small time steps Appropriate cross section spacing downstream of dam breach (Courant <1.0) Increased Mannings n values just downstream of dam Additional iterations (calculation options) Total breach time may not all be Critical time Adjust non-linear breach growth, especially if including non-critical breach time. Mixed Flow (if necessary, even if it is not going supercritical) Adjust theta weighting factor, starting at 1.0 and lower as much as possible with 0.6 as the lowest