Professional Documents
Culture Documents
7 Case Studies in Dam Breach Analysis, Blass
7 Case Studies in Dam Breach Analysis, Blass
Charles P. Johnson
& Associates, Inc
May 9, 2012
Teton Dam Failure Liberty Reservoir
y Idaho – 1976 y Location – Sykesville,
Maryland
y 251,000 ac‐feet of water
y Dam Height ‐ 175 feet
(80 Billion Gallons)
y Construction Type –
y Breach Length – Gravity – multiple
Approximately 130 miles monoliths
(Teton Dam to American y Drainage Area – 163
Falls Dam) Square Miles
y Impacted Areas –
y Breach Travel Time – Rexburg, Idaho – 13 miles downstream Ellicott City, Patapsco
Approximately 2 Days River valley, Baltimore
y Average Speed – 4 f.p.s. City
Hydraulics and Hydrology
Methodology
y Hydrology
y GISHydro
y Calibrated using USGS regression equations for area.
Liberty Dam
y Hydraulics Drainage Area
y Volume via MGS Bathymetric Survey Map
y Discharge via broad crested weir equation with
adjustments for construction by abutments. Obtained
via field survey
y USACE HEC‐1 Modeling program used to develop
hydrographs
Breach Parameter Development Breach Routing and Flood Mapping
y Multiple monoliths (1, 2, 3, 4) y Cross Sections
y 2’ LiDAR from Carroll, Howard, Anne Arundel, and
y Sensitivity to breach width
Baltimore Counties and Baltimore City
y Breach at abutments – RULED OUT y Determined Manning’s n values from Chows’ Open Channel
y Shallow earth fill Hydraulics
y Breach from water surface elevation to bottom of y Roads
y No road crossings built into current model but plans are to
upstream side of dam update model to include crossings where applicable.
y Time to Failure y Dams
y 3 minutes ‐ Very short due to method of failure y Low head dams (Bloede, Daniels) not included because of
height relative to breach wave and lack of available flood
storage
Breach Routing and Floodplain
Mapping Breach Routing and Flood Mapping
y ARCGIS 10 y USACE HEC‐GeoRAS 4.3
y Developed DEMs from 2’ LiDAR data sets y Unsteady Flow Model – Breach Hydrograph Input
y Delineated cross section lines y Accounting for Flow Attenuation = Reduced Peak Flows
y Created georeferenced floodplain maps y More accurate water surface profiles
y Geo‐referenced Floodplain Mapping
Challenges and Solutions
y CHALLENGE: Large scope of work with limited funding
y SOLUTION: Drew on freely available information and
software to minimize survey (examples: MGS Bathymetric
Survey, HEC‐1 and HEC‐RAS software packages, MDE Dam
Safety technical resources)
y CHALLENGE: Multiple Data Sources
y SOLUTION: Verified datum of each data source and used
ARCGIS to “sew” together multiple LiDAR datasets.
y CHALLEGE: Inflow Hydrograph Accuracy
y SOLUTION: Calibrated model using sensitivity analysis to
drainage area, RCN, and time of concentration. Adjust time
of concentration (most sensitive)
If we had to do it again . . . Kentlands Dams
y Include Road Crossings (will be done as a revision to y Location – Gaithersburg,
Maryland
this model)
y Dam Heights – 24, 19,
y Calibrate hydrologic model to additional gauged 37, 24, and 15 feet
regression equations y Construction Type –
Earth embankment w/
y Consider steady flow model at Road Crossings to structural spillways
promote model stability y Drainage Area – 323
y Include second inflow hydrograph to simulate rainfall Acres
y Impacted Areas –
over downstream drainage areas Kentlands
neighborhood,
Gaithersburg, Maryland
Hydraulics and Hydrology
Methodology Breach Parameter Development
y Hydrology y Earth Embankment
y Delineated drainage areas manually
y NWS Simple DAMBRK Program
y Determined time of concentration manually using TR‐55 guidance
y Determine Runoff Curve Number using ARC GIS and GIS land use y Assumed largest flow yielded the most conservative
and soils data sets. breach parameters (width, time to failure)
y Hydraulics y Minimum time to failure of 10 minutes (0.17 hours)
y Normal pool and flood storage volumes determined using as‐built
plans and GIS data and geo‐referenced into survey datum. y Breach from water surface elevation to bottom of
y Stage Discharge determined using as‐built plans and geo‐ upstream side of dam
referenced field survey data from all five dams to establish a
common datum.
y USACE HEC‐1 Modeling program used to develop hydrographs
Breach Routing and Flood Mapping
y Cross Sections
y 2’ LiDAR from City of Gaithersburg and Field Surveyed
topography
y Modeled using HEC‐1
y Roads
y No road crossings encountered in breach area.
y Mapping
y Plotted water surface elevations for each event on base
information in AutoCAD to address impacts to adjacent
downstream properties
Challenges and Solutions If we had to do it again . . .
y CHALLENGE: Complex dam system requiring multiple y Model downstream flooding in stream valley using
levels of input to develop breach maps HEC‐RAS or other open channel modeling program.
y SOLUTION: Used conservative approach of assuming all
dams breach “in‐series” with the breach of an upstream dam y Examine dam breach parameters using Froelich
flooding the next dam downstream and breaching it. Equations.
y CHALLENGE: Significant Tailwater Effects y Extend breach modeling downstream to Muddy
y Used HY‐8 Culvert Analysis program to develop and balance Branch Tributary, beyond Darnestown Road
tailwater rating curves on each principal spillway pipe.
y CHALLENGE: Multiple datums
y SOLUTION: Used field survey of “hard points” such as riser
weirs, pipe inverts, etc. to rectify all datums into one common
datum. Adjusted table top information accordingly.
Hydraulics and Hydrology
Montgomery Auto Park SWM Pond Methodology
y Location – Silver Spring, y Hydrology
Maryland
y Manually determined drainage areas, runoff curve
y Dam Height ‐ 29 feet
numbers, and times of concentration
y Construction Type –
Earth Embankment, y Verified ICC stormwater management computations
structural outlets
y Hydraulics
y Drainage Area – 220
Acres y Volume via field survey
y Impacted Areas – local y Discharge via field survey
residential communities
along Paint Branch y USACE HEC‐1 Modeling program used to develop
Tributary, Inter‐County hydrographs – upstream stormwater facilities
Connector included in 100 year analysis
Breach Routing and Flood Mapping
y Downstream area ‐ small depression with 60” R.C.P.
outlet – MODEL AS A DAM
y Discharges to second downstream area – small
depression with 48” R.C.P. – MODEL AS A DAM
y ICC Noise Wall channels overtopping flows down
roadway
y Use HEC‐1 to model dams (part of larger dam breach
model)
y Use HEC‐RAS 4.1 to model overtopping flows on ICC
and discharge flows from 48” and 60” R.C.P.
Challenges and Solutions
y CHALLENGE: Extremely complex downstream area
y SOLUTION: Used split flows between those that overtop
and flow down ICC and those that pass through 48” and
60” culverts to downstream tributary
y CHALLENGE: Timing of analysis with major adjacent
construction project.
y SOLUTION: Owner coordinated with ICC project teams
very closely to obtain all pertinent information required
for analysis.
If we had to do it again . . . Take‐Away’s
y Perform bathymetric survey of pond bottom (pond y Each dam is unique. No two breach analyses are the
empty but inaccessible) to improve stage‐storage same.
rating table accuracy y Think about the reality of a breach scenario: How
y Assess downstream flow modeling approach. does the breach occur? Where and how does the
y Consider using 2‐D flow routing model to more water flow when it is released? Visit the site if
accurately model complex downstream area. possible
y Data sources must be rectified into common
baselines, either by data type or physical datum
y Over‐simplification can lead to inaccurate results
Questions and Comments?
jblass@cpja.com