feb 2003hec-ras version 3.11 slides adapted from hec unsteady flow course unsteady flow course
TRANSCRIPT
Feb 2003 HEC-RAS Version 3.1 1
HEC-RAS Version 3.1HEC-RAS Version 3.1Unsteady FlowUnsteady Flow
Slides adapted from HECSlides adapted from HEC Unsteady Flow CourseUnsteady Flow Course
Feb 2003 HEC-RAS Version 3.1 2
Unsteady Flow Documentation
Technical/theoretical - Chapters 2 and 5 from EM 1110-2-1416
http://www.hnd.usace.army.mil - click on TECHINFO, then Engineering Publications, then Engineering Manuals, scroll for manual
HEC-RAS User’s Manual - Chapter 8 (data input and window operations)
HEC-RAS Application Guide - Chapter 17 (example application)
Feb 2003 HEC-RAS Version 3.1 3
When to use Unsteady Flow
Tidal/estuary fluctuation Off-channel storage Dam breach routing Channels with flat slopes Levee overtopping Hydraulic structures affected by changing
backwater Large amounts of storage behind roads or
culverts
Feb 2003 HEC-RAS Version 3.1 4
Steady vs. Unsteady
Difference in handling boundary friction and other losses
Difference in numerical solution algorithm Difference in handling non-flow areas Difference in flow and boundary condition data
requirements Difference in calibration strategy Difference in application strategy
Feb 2003 HEC-RAS Version 3.1 5
2
o f
( /A)Q Q h gA( ) 0S St x xα
Momentum Equation:
Continuity Equation:
Q A + = 0x t
Unsteady Flow Equations
Feb 2003 HEC-RAS Version 3.1 6
2
o f
( /A)Q h + gA( - S + S ) = 0x x
α
Energy (momentum) Equation:
Continuity Equation:
Q = VA
Steady Flow Equations
Feb 2003 HEC-RAS Version 3.1 7
Distance vs Time Solution Grid
X = distance, feet t = time, seconds
t
x1,1 2,1
1,2 2,2
Feb 2003 HEC-RAS Version 3.1 8
Finite Difference Term
Q Q (2,2) - Q(1,2) Q(2,1) - Q(1,1) ----- = ----------------------- + ( 1 - ) --------------------- X X X
= theta weighting factor 0.5 < < 1.0
Feb 2003 HEC-RAS Version 3.1 9
Steady – Compute exact hydraulic properties at a section for each trial water surface elevation from the elevation/station points, n-values.
Unsteady – Hydraulic properties are pre-computed for all possible water surface elevations at each cross section (hydraulic table)
Pre-Computation of Hydraulic Properties
Feb 2003 HEC-RAS Version 3.1 10
Data Requirements(Flow and Boundary Conditions)
Steady: Discharge (Q) at each cross section.
Unsteady: Inflow hydrograph(s) which are routed by the model.
Feb 2003 HEC-RAS Version 3.1 11
Prepare hydrographs
(boundary conditions)
Upstream flows
Tributary (local flows)
Ungaged/unmodeled flows
Downstream (rating curve?)
Feb 2003 HEC-RAS Version 3.1 12
HEC-RAS Main Window
Feb 2003 HEC-RAS Version 3.1 13
Entering Geometric Parameters
Feb 2003 HEC-RAS Version 3.1 14
Cross Section Table Properties
Feb 2003 HEC-RAS Version 3.1 15
Pre-processing Geometry
For unsteady flow, geometry is pre-processed into tables and rating curves Cross sections are processed into tables of
area, conveyance, and storage Bridges and culverts are processed into a
family of rating curves for each structure Weirs and gated structures are calculated
on the fly during unsteady flow calculations Pre-processor results can be viewed in
graphs and tables
Feb 2003 HEC-RAS Version 3.1 16
Cross Section Properties Plot
0 1000 2000 3000 4000 5000 6000 7000650
660
670
680
690
700
Property TableRS = 138154.4
Conveyance/1000 (cfs) Storage (cu ft)
Ele
vatio
n (f
t)
Legend
Conv. Channel
Conv. Valley
Conv. Total
Storage
Feb 2003 HEC-RAS Version 3.1 17
Geometry Preprocessor
What does it do? Processes geometric data into a series of
hydraulic tables and rating curves. Why do we use it for unsteady flow? Instead of calculating hydraulic variables for
each cross-section during each iteration, the program interpolates the hydraulic variables from the tables.
Feb 2003 HEC-RAS Version 3.1 18
Conveyance Calculations
Manning Equation
1/2 Q = K Sf K = Conveyance Sf = friction/energy slope
Feb 2003 HEC-RAS Version 3.1 19
Conveyance Calculations
nch n1 n2 n3
A1 P1 A2 P2 A3 P3 Ach Pch
Klob = K1 + K2
Kch
Krob = K3
Feb 2003 HEC-RAS Version 3.1 20
Cross Section Example
Feb 2003 HEC-RAS Version 3.1 21
Geometry Preprocessor
Feb 2003 HEC-RAS Version 3.1 22
Hydraulic Property Plot
Feb 2003 HEC-RAS Version 3.1 23
Cross Section Properties Table
Feb 2003 HEC-RAS Version 3.1 24
Conveyance Subdivisions
Feb 2003 HEC-RAS Version 3.1 25
Conveyance Subdivisions
Feb 2003 HEC-RAS Version 3.1 26
Conveyance Subdivisions
Feb 2003 HEC-RAS Version 3.1 27
Conveyance Subdivisions
Boundary andInitial Conditions
Objectives: Know boundary condition options Know initial condition requirements Sources of data for both
Feb 2003 HEC-RAS Version 3.1 29
Unsteady Flow Data
External Boundaries required Upstream and Downstream ends of the river Typically flow or stage hydrograph upstream Typically rating or “normal depth” downstream
Internal Boundaries can be added Add flow within the river system Define gate operation
Initial Conditions - at the start of simulation
Feb 2003 HEC-RAS Version 3.1 30
Unsteady Flow Data Editor
Feb 2003 HEC-RAS Version 3.1 31
Boundary Conditions
Editor shows required external boundaries
Boundary Type shows available options
Upstream options: Stage Hydrograph Flow Hydrograph Stage & Flow
Hydrograph
Feb 2003 HEC-RAS Version 3.1 32
Boundary Conditions - continued
Downstream Boundary Options: Stage Hydrograph Flow Hydrograph Stage & Flow Hydrograph Rating Curve Normal Depth
Feb 2003 HEC-RAS Version 3.1 33
Flow Hydrograph
Read from DSS Select DSS file Select Pathname
Enter in Table Select time interval Select start date/time Enter flow data - or
cut & paste
Feb 2003 HEC-RAS Version 3.1 34
Sources of Time-Series Data
Historic Records (USGS) Stage Hydrographs Flow Hydrographs
Computed Synthetic Floods Rainfall-runoff modeling Peak Discharge with assumed time
distribution Others?
Feb 2003 HEC-RAS Version 3.1 35
Normal Depth
Enter Friction (energy) Slope
Program uses Manning’s equation to compute stage
Provides semi-dynamic downstream boundary
Feb 2003 HEC-RAS Version 3.1 36
Initial Conditions
Requires an initial flow for all reaches
Restart file can be read from DSS
Enter steady-flow at upstream boundary
Can add a flow-change location
Pool elevation for storage areas
Feb 2003 HEC-RAS Version 3.1 37
File and Options Menus
Feb 2003 HEC-RAS Version 3.1 38
Unsteady Flow Simulation Simulation Manager
1. Define a Plan
2. Select which programs to run
3. Enter a starting and ending date and time
4. Set the computation settings
5. Press the Compute button
Feb 2003 HEC-RAS Version 3.1 39
Output Selection
Unsteady Flow Output Stage and Flow
Hydrographs Log File Output
Post Processor Detailed output
– Max Stage– Selected Time Intervals
Feb 2003 HEC-RAS Version 3.1 40
Stage and Flow Hydrographs User Selected Locations
Feb 2003 HEC-RAS Version 3.1 41
Viewing Unsteady Flow Results
All of the output that was available for steady flow computations is available for unsteady flow (cross sections, profile, and 3D plots and tables).
Stage and flow hydrographs Time series tables Animation of cross section, profile and 3-
dimensional graphic
Feb 2003 HEC-RAS Version 3.1 42
Stage and Flow Plot
Stage
Feb 2003 HEC-RAS Version 3.1 43
Unsteady Flow Rating Curve
Feb 2003 HEC-RAS Version 3.1 44
Log File Output
can be generated during computations information about progression of
simulation can make a large, large file
are you sure you want to open it?
Feb 2003 HEC-RAS Version 3.1 45
Post Processor
Can be run after the unsteady simulation is completed
Provides profiles for the maximum stage and at regular intervals
All regular graphics and tables can be used to view the post process results
Graphics can “animate” the simulation
Feb 2003 HEC-RAS Version 3.1 46
Profile Animation
Feb 2003 HEC-RAS Version 3.1 47
Accuracy/Stability/SensitivityObjective
For students to have a better understanding of model accuracy, stability, and sensitivity.
To become familiar with the available parameters within HEC-RAS that will allow you to develop a stable and accurate model.
To learn how to detect, find, and fix model stability problems.
Feb 2003 HEC-RAS Version 3.1 48
Overview
Model Accuracy Model Stability Factors Affecting Accuracy and Stability
Cross section spacing Computational time step selection Practical delta t, hydrograph rise time / 20
Common Stability Problems Detecting Stability Problems Model Sensitivity
Feb 2003 HEC-RAS Version 3.1 49
Model Accuracy
Accuracy can be defined as the degree of closeness of the numerical solution to the true solution.
Accuracy depends upon the following: Assumptions and limitations of the model (i.e. one
dimensional model, subcritical flow only for unsteady flow) Accuracy of the geometric Data (cross sections, Manning’s
n values, bridges, culverts, etc…) Accuracy of the flow data and boundary conditions Numerical Accuracy of the solution scheme
Feb 2003 HEC-RAS Version 3.1 50
Numerical Accuracy
If we assume that the 1-dimensional unsteady flow equations are a true representation of flow moving through a river system, then only an analytical solution of these equations will yield an exact solution.
Finite difference solutions are approximate. An exact solution of the equations is not
feasible for complex river systems, so HEC-RAS uses a finite difference scheme.
Feb 2003 HEC-RAS Version 3.1 51
Model Stability
An unstable numerical model is one for which certain types of numerical errors grow to the extent at which the solution begins to oscillate, or the errors become so large that the computations can not continue.
Feb 2003 HEC-RAS Version 3.1 52
Feb 2003 HEC-RAS Version 3.1 53
Factors Affecting Model Stability and Numerical Accuracy
Cross Section Spacing Computation time step Theta weighting factor Solution iterations Solution tolerances
Feb 2003 HEC-RAS Version 3.1 54
Calculation Options and Tolerances
Feb 2003 HEC-RAS Version 3.1 55
Cross Section Spacing
Cross sections should be placed at representative locations to describe the changes in geometry.
Additional cross sections should be added at locations where changes occur in discharge, slope, velocity, and roughness.
Cross sections must also be added at levees, bridges, culverts, and other structures.
Feb 2003 HEC-RAS Version 3.1 56
Cross Section Spacing - Slope
Bed slope plays an important role in cross section spacing. Steeper slopes require more cross sections Streams flowing at high velocities may
require cross sections on the order of 100 feet or less.
Larger uniform rivers with flat slopes may only require cross sections on the order of 1000 ft or more.
Feb 2003 HEC-RAS Version 3.1 57
Cross Section Spacing - How do you know if you have enough XS:
Use the HEC-RAS cross section interpolation.
Make a new plan and run the model.
Compare the before and after.
If no significant difference, then OK!
Feb 2003 HEC-RAS Version 3.1 58
Theta Weighting Factor
Theta is a weighting applied to the finite difference approximations when solving the unsteady flow equations.
Theoretically Theta can vary from 0.5 to 1.0. However a practical limit is from 0.6 to 1.0
Theta of 1.0 provides the most stability. Theta of 0.6 provides the most accuracy.
The default in HEC-RAS is 1.0. Once you have your model developed, reduce theta towards 0.6, as long as the model stays stable.
Feb 2003 HEC-RAS Version 3.1 59
Common Stability Problems Too large of a time step. Not enough cross sections Model goes to critical depth – RAS is limited
to subcritcal flow for unsteady flow simulations
Bad downstream boundary condition (i.e. rating curve or slope for normal depth)
Bad cross section properties, commonly caused by: levee options, ineffective flow areas, Manning’s n values, etc..
Feb 2003 HEC-RAS Version 3.1 60
Common Stability Problems - Continued
Cross section properties that do not go high enough, or are way to high (curves are spread to far apart).
Bad bridge/culvert family of rating curves. Wide and flat lateral weirs/spillways – send
to much flow over a given time step. Gated spillways that are opened or closed to
fast.
Feb 2003 HEC-RAS Version 3.1 61
Detecting Stability Problems
How do you know you have a stability problem? Program completely blows up during run Program goes to maximum number of
iterations for several time steps in a row. Program has oscillations in the computed
stage and flow hydrographs
Feb 2003 HEC-RAS Version 3.1 62
Detecting Stability Problems - Continued
What do you do when this happens? Note the simulation time when the program either blew
up or first started to oscillate. Turn on the “Detailed Output for Debugging” option and
re-run the program. View the text file that contains the detailed log output of
the computations. Locate the simulation output at the simulation time when the solution first started to go bad.
Find the river station locations that did not meet the solution tolerances. Then check the data in this general area.
Feb 2003 HEC-RAS Version 3.1 63
Turning on Detailed Output for Debugging
Feb 2003 HEC-RAS Version 3.1 64
Viewing Detailed Log Output
Feb 2003 HEC-RAS Version 3.1 65
Model Sensitivity Numerical sensitivity:
Computation time step – try a smaller value to see if the output changes significantly.
Theta – start at 1.0, after you have a working model then try to reduce it towards 0.6.
Weir/Spillway stability factors – if you are using stability factors, try to reduce them to the lowest value you can get away with.
Weir/Spillway exponential decay factors – in general I would leave them alone, they will not effect the sensitivity of the output much.
Feb 2003 HEC-RAS Version 3.1 66
Model Sensitivity - Continued Physical Parameter Sensitivity:
Manning’s n Values – What if the true n values were 10% higher or Lower?
Cross Section Spacing – Test by interpolating Cross Section Storage – What if there is really more or
less storage in the cross sections (I.e. ineffective flow areas, etc…)
Weir/Spillway coefficients – For lateral weirs/spillways the coefficient selected can have a great impact on the results.
Bridge/Culvert Parameters – normally only effect the locally computed stages, unless it is a flat area in which the bridge causes great backwater.
The End