hydro-informatics software, models and simulation · • deterministic vs. stochastic models • in...
TRANSCRIPT
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 1
1
Hydro-InformaticsSoftware, Models and Simulation
• Dr. B.Pirzadeh
• Department of Civil Engineering, University of Sistan & Baluchestan
2
MODELS :
WHAT ARE WE TALKING ABOUT?
Software, models and simulation
2
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 2
Software, models and simulation
MODELS IN HYDRAULICS AND HYDROLOGY
VOCABULARY: «MODEL» or «SOFTWARE»
DIFFERENCE BETWEEN:
A MODELLING SOFTWARE also called A
SOFTWARE CODE
and
A MODEL
N.B.: A SCALE MODEL IS ALWAYS A MODEL
3
4
Software, models and simulation
44
• What is Model
• A model of a system is a representation of the construction and working of the
system
• Similar to but simpler than the system it represents
• Close approximation to the real system and incorporate most of its salient features
• Should not be so complex that it is hard to understand or experiment with it
• Physical Model
• A physical object that mimics some properties of a real system
• e.g. During design of buildings, it is common to construct small physical models
with the same shape and appearance as the real buildings to be studied
• Through prototyping process
• Prototyping is the process of quickly putting together a working model (a prototype)
in order to test various aspects of a design, illustrate ideas or features and gather
early user feedback
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 3
55
Software, models and simulation
5
6
Software, models and simulation
• Mathematical Model
• A description of a system where the relationship between variables of the system
are expressed in a mathematical form
• e.g. Ohm's law describes the relationship between current and voltage for a resistor;
Hooke's Law gives the relationship between the force applied to an unstretched
spring and the amount the spring is stretched when the force is applied, etc.
• Through virtual prototyping
• Deterministic vs. stochastic models
• In deterministic models, the input and output variables are not subject to random
fluctuations, so that the system is at any time entirely defined by the initial
conditions chosen
– e.g. the return on a 5-year investment with an annual interest rate of 7%,
compounded monthly
• In stochastic models, at least one of the input or output variables is probabilistic or
involves randomness
– e.g. the number of machines that are needed to make certain parts based on the
probability of machine failure
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 4
7
Software, models and simulation
FSpring = -k∙x
Hooke’s Law
x= -FSpring/k
spring constant The amount spring
is stretched
Fspring
Fspring
8
H
L
H
23
CLHQ
Coefficient of Discharge
Rectangular Weirs
Software, models and simulation
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 5
9
• What is Simulation
• A simulation of a system is the operation of a model of the system, as an
imitation of the real system
• A tool to evaluate the performance of a system, existing or proposed, under
different configurations of interest and over a long period of time
• e.g. a simulation of an industrial process to learn about its behavior under different
operating conditions in order to improve the process
Software, models and simulation
10
Software, models and simulation
The term modeling refers to the development of a mathematical
representation of a physical situation. While, simulation refers to the
procedure of solving the equations that resulted from model
development.
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 6
A MODEL – WHAT FOR?
- To understand physical phenomena
- To predict consequences of exceptionnal events and
human activities
Software, models and simulation
11
How can models help to understand physical
phenomena?
Hypotheses → model → comparison with observed
If positive, the hypotheses could be right !!!
Engineering:
Supposing the model is correct,
To predict consequences of events or human activities
Software, models and simulation
12
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 7
13
Model concept
Software, models and simulation
REPRESENTATION
OF THE REALITY -
« SYSTEM »
INPUT OUTPUT
rainfall
Watershed basin
discharge
1313
TWO CLASSES OF MODELS:
• Correlative models & transfer functions,
• Mechanistic models that can simulate physical
phenomena
Software, models and simulation
14
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 8
THE CLASS OF CORRELATIVE MODELS &
TRANSFER FUNCTIONS
Software, models and simulation
output
input
MODEL =
A FUNCTION OR A PROCESS
APPROACHED BY CALIBRATION
(FITTING) OF NON-PHYSICAL
PARAMETERS
CRITERIA ARE OBSERVED
INPUTS & OUTPUTS
fitting
15
EXAMPLES:
- Simple correlations,
- ARMA/ARIMA and similar,
- ANN – Artificial Neuron Networks,
- GA – Genetic Algorithms,
- « Black Boxes », etc.
IN ALL CASES TRANSFER FUNCTIONS ARE DEFINED
SOLELY UPON PAST-OBSERVED INPUT/OUTPUT DATA
Software, models and simulation
16
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 9
CORRELATIVE MODELS & TRANSFER FUNCTIONS
• AVANTAGES:
AUTOMATIC FITTING OF PARAMETERS
STATISTICS OF QUALITY OF FITTING
MODEST COMPUTER RESOURCES AND
SHORT COMPUTATIONAL TIME → REAL-
TIME APPLICATIONS
Software, models and simulation
17
DISAVANTAGES & LIMITATIONS:
VALIDITY LIMITED BY THE RANGE OF THE
INPUT/OUTPUT DATA SAMPLE,
INTERPRETATION IN TERMS OF PHYSICAL PROCESSES
DIFFICULT OR IMPOSSIBLE.
THESE MODELS ARE NOT PREDICTIVE !!!
Software, models and simulation
18
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 10
MECHANISTIC (or PHYSICAL or SIMULATION )MODELS:
Differential or integral equations that are formulation of physicallaws governing modelled reality
plus
Algoirithms necessary to solve numerically the equations
and
Representation of the topography, of the geometry, of hydraulicparameters, of hydraulic structures and their operations rules, ofthe land occupation, etc.
Software, models and simulation
19
MECHANISTIC MODELS
AVANTAGES:
These models describe in deterministic way the evolution ofhydraulics and hydrology phenomena.
The description (simulation) is conform to physical laws (equations).
These models are predictive and allow for studying ofconsequences of engineering projects.
Software, models and simulation
20
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 11
MECHANISTIC MODELS
TYPICAL DISADVANTAGES :
- the effort needed to set up a model,
- computational efforts can be important,
- quality and accuracy of the results depend upon theaccuracy and how detailed are topography and hydraulics
data,
- simulation software is to be acquired on the market andeducation effort to apply is an important investment.
- Software, models and simulation
21
Software, models and simulation
I(t)
Q(t)S(t)
MECHANISTIC
MODELS: ARE BASED
ON PHYSICAL LAWS
EXAMPLE:
« RESERVOIR » MODEL
OF A WATERSHED
BASIN
22
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 12
i
Software, models and simulation
ONLY ONE PARAMETER k !!!
IF CALIBRATED, WILL THE
MODEL BE PREDICTIVE ?
I(t)
Q(t)S(t)
CONTINUITY LAW:
dS/dt = I(t) - Q(t)
CONCEPTUALISATION:
S = k Q, dQ/dt = (I - Q)/k
SOLUTION:
Q = Q0 exp(-(t-t0)/k)
23
24
Software, models and simulation
ESSENTIAL QUESTION IS:
GIVEN THE NEEDS, WHAT IS THE VALUE OF THE MODEL???
BASIC QUALITY CRITERIA ARE :
- WILL THE MODEL HELP TO ANSWER THE QUESTIONS &
ENGINEERING NEEDS?
- HOW A MODEL IS BUILT, SET UP?
- WHAT PHENOMENA CAN IT SIMULATE ?
- HOW HIGH IS ITS COST ?
2424
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 13
25
Software, models and simulation
THE VALUE OF A MODEL IS ALSO MADE BY
THE MODELLER, BY HIS KNOWLEDGE AND
EXPERIENCE OF :
hydraulics and physics in general ,
flow patterns, hydraulics and topography of the river,
objectives, purpose of the model to be built,
available simulation software systems.
IN THIS ORDER OF IMPORTANCE !!!!
2525
26
Software, models and simulation
WHAT ABOUT THE SOFTWARE ?
SOFTWARE IS NECESSARY BUT
-FAR FROM BEING SUFFICIENT ALONE,
-IN GENERAL THERE IS NO SINGLE SOFTWARE
SOLUTION:
SEVERAL SOLUTIONS ARE POSSIBLE
2626
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 14
27
Software, models and simulation
De Saint-Venant hypotheses of 1-D flow:
Uniform velocity v at every
point of the cross- section
Free surface horizontal
across the section
Vertical distribution of the
pressure is hydrostatic
Head- losses can be represented
by a Chèzy, Strickler or Manning
coefficientDo you believe this ???
And yet...
v m/s
2727
28
Software, models and simulation
How to schematise the flow in the river and its inundated plains?
Storage « pockets »
of horizontal free
surface
Main bed
Dykes, roads,
canals, ditches,...
Fields,...
2828
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 15
29
Software, models and simulation
Flooded area Main bed
Is the situation 1-D?
Is not the reality rather like:
Rising flood Receding flood
2929
30
Software, models and simulation
Possible modelling options:
De Saint-Venant 1-D type
+ storage «pockets »De Saint-Venant 1-D type
+ interconnected « cells »
De St-Venant 1-D type
interconnected network
Implication: to know enough about theory, algorithms & software !!!
3030
UNSA - DESS HYDROPROTECH Février 2002
Dr. J. A. Cunge 16
31
Software, models and simulation
Above: 1-D discretisation; Below: quasi- 2-D (river + cells)
3131
Importance of Grid Resolution
Breach B Breach B Breach B Breach B Breach B Breach B Breach B Breach B Breach B
Breach B Breach B Breach B Breach B Breach B Breach B Breach B Breach B Breach B
Breach B Breach B Breach B Breach B Breach B Breach B Breach B Breach B Breach B
20m 10m 5m
32
Courtesy Kostya Vasiliev, HLCROW