Journal of Water Resources Planning and Management,
Journal Year:
2009,
Volume and Issue:
136(4), P. 433 - 443
Published: Oct. 3, 2009
Optimization
of
large
and
hydraulically
complex
water
distribution
systems
(WDSs)
is
computationally
expensive
as
simulation
models
are
required
to
evaluate
the
performance
solutions
problem
at
hand.
Metamodels
can
act
a
surrogate
or
substitute
for
these
provide
significant
speed-ups
in
optimization
process.
The
application
metamodels
field
WDS
has
been
limited
date,
little
guidance
given
terms
constructing
systems.
While
it
relatively
straightforward
obtain
satisfactory
metamodel
approximations
simple
WDSs,
this
not
necessarily
case
more
networks.
In
order
reduce
complexity
relationship
that
be
approximated
by
metamodels,
number
factors
have
considered,
including
hydraulic
model,
decision
space,
locations
which
outputs
from
model.
This
research
presents
systematic
methodology
dealing
with
demonstrates
effectiveness
approach
applying
an
actual
WDS.
A
system
Wallan,
Victoria,
Australia,
selected
demonstration
purposes.
Four
different
metamodelling
scenarios
presented
here.
results
show
that,
study,
some
skeletonization
model
achieve
suitably
accurate
metamodels.
reduction
average
daily
pumping
costs
$457
$363;
saving
21%.
net
present
value
(NPV)
over
25
years
used
objective
function,
includes
both
chlorine
costs.
current
operating
regime
corresponds
NPV
$1.56
million,
while
optimized
solution
$1.34
million;
14%.
addition
economic
benefits,
achieves
adequate
disinfection
throughout
system,
whereas
deficits
residuals
several
system.
Journal of Water Resources Planning and Management,
Journal Year:
2009,
Volume and Issue:
136(4), P. 412 - 432
Published: Oct. 12, 2009
During
the
last
two
decades,
water
resources
planning
and
management
profession
has
seen
a
dramatic
increase
in
development
application
of
various
types
evolutionary
algorithms
(EAs).
This
observation
is
especially
true
for
genetic
algorithms,
arguably
most
popular
several
EAs.
Generally
speaking,
EAs
repeatedly
prove
to
be
flexible
powerful
tools
solving
an
array
complex
problems.
paper
provides
comprehensive
review
state-of-the-art
methods
their
applications
field
management.
A
primary
goal
this
ASCE
Task
Committee
effort
identify
organized
fashion
some
seminal
contributions
areas
distribution
systems,
urban
drainage
sewer
supply
wastewater
treatment,
hydrologic
fluvial
modeling,
groundwater
parameter
identification.
The
also
identifies
major
challenges
opportunities
future,
including
call
address
larger-scale
problems
that
are
wrought
with
uncertainty
expanded
need
cross
fertilization
collaboration
among
our
field's
subdisciplines.
Evolutionary
computation
will
continue
evolve
future
as
we
encounter
increased
problem
complexities
societal
pressure
more
innovative
efficient
solutions
rises.
Journal of Water Resources Planning and Management,
Journal Year:
2008,
Volume and Issue:
134(4), P. 337 - 346
Published: July 1, 2008
Reducing
energy
consumption
of
water
distribution
networks
has
never
had
more
significance
than
today.
The
greatest
savings
can
be
obtained
by
careful
scheduling
operation
pumps.
Schedules
defined
either
implicitly,
in
terms
other
elements
the
network
such
as
tank
levels,
or
explicitly
specifying
time
during
which
each
pump
is
on/off.
traditional
representation
explicit
schedules
a
string
binary
values
with
bit
representing
on/off
status
particular
interval.
In
this
paper
new
presented.
It
based
on
controlled
triggers,
where
maximum
number
switches
specified
beforehand.
schedule
divided
into
series
integers
integer
hours
for
active/inactive.
This
reduces
potential
(search
space)
compared
to
representation.
Ant
colony
optimization
(ACO)
stochastic
meta-heuristic
combinatorial
problems
that
inspired
foraging
behavior
some
species
ants.
paper,
an
application
ACO
framework
was
developed
optimal
proposed
adapted
ant
Optimization
and
solved
schedules.
Minimization
electrical
cost
considered
objective,
while
satisfying
system
constraints.
Instead
using
penalty
function
approach
constraint
violations,
violations
were
ordered
according
their
importance
solutions
ranked
order.
tested
small
test
large
real-world
network.
Results
are
those
simple
genetic
algorithm
hybrid
uses
level-based
triggers.
Chaos An Interdisciplinary Journal of Nonlinear Science,
Journal Year:
2011,
Volume and Issue:
21(1)
Published: March 1, 2011
This
paper
explores
a
variety
of
strategies
for
understanding
the
formation,
structure,
efficiency,
and
vulnerability
water
distribution
networks.
Water
supply
systems
are
studied
as
spatially
organized
networks
which
practical
applications
abstract
evaluation
methods
critically
evaluated.
Empirical
data
from
benchmark
used
to
study
interplay
between
network
structure
operational
reliability,
robustness.
Structural
measurements
undertaken
quantify
properties
such
redundancy
optimal-connectivity,
herein
proposed
constraints
in
design
optimization
problems.
The
role
demand
toward
system
efficiency
is
studied,
an
assessment
failures
based
on
disconnection
nodes
source(s)
undertaken.
absence
conventional
degree-based
hubs
(observed
through
uncorrelated
nonheterogeneous
sparse
topologies)
prompts
alternative
approach
studying
structural
identification
cut-sets
optimal-connectivity
invariants.
A
discussion
scope,
limitations,
possible
future
directions
this
research
provided.
Journal of Water Resources Planning and Management,
Journal Year:
2011,
Volume and Issue:
138(2), P. 153 - 161
Published: May 13, 2011
A
water
distribution
system,
represented
as
a
spatially
organized
graph,
is
complex
network
of
multiple
interconnected
nodes
and
links.
The
overall
robustness
such
in
addition
to
the
reliability
individual
components,
depends
on
underlying
structure.
This
paper
presents
deterministic
network-based
approach
study
relationship
between
structure
function
systems
critically
review
application
structural
measurements
analysis
vulnerability
systems.
Benchmark
supply
networks
are
studied,
their
level
resistance
random
failures
targeted
attacks
bridges
cut-sets
explored.
Qualitative
concepts
redundancy,
optimal
connectivity,
quantified.
Among
other
measurements,
two
metrics
meshedness
coefficient
algebraic
connectivity
found
great
use
toward
quantifying
redundancy
respectively.
brief
discussion
usefulness,
scope,
limitations
proposed
real-world
presented.
Water Resources Research,
Journal Year:
2012,
Volume and Issue:
48(6)
Published: May 10, 2012
The
reliability
and
robustness
against
failures
of
networked
water
distribution
systems
are
central
tenets
supply
system
design
operation.
ability
such
networks
to
continue
when
components
damaged
or
fail
is
dependent
on
the
connectivity
network
role
location
individual
components.
This
paper
employs
a
set
advanced
analysis
techniques
study
systems,
its
relationship
with
robustness,
susceptibility
damage.
Water
modeled
as
weighted
directed
by
using
physical
hydraulic
attributes
A
selection
descriptive
measurements
utilized
quantify
structural
properties
benchmark
at
both
local
(component)
global
(network)
scales.
Moreover,
novel
measure
component
criticality,
demand‐adjusted
entropic
degree,
proposed
support
identification
critical
nodes
their
ranking
according
failure
impacts.
application
value
this
metric
demonstrated
through
two
case
in
USA
UK.
Discussion
focuses
potential
for
gradual
evolution
abstract
graph‐based
tools
more
practical
methods,
where
theoretical
framework
vulnerability
better
planning
management
decisions
presented.
Journal of Water Resources Planning and Management,
Journal Year:
2012,
Volume and Issue:
139(2), P. 175 - 183
Published: March 5, 2012
The
cost
of
electricity
used
for
pumping
in
water-distribution
systems
typically
represents
the
largest
part
total
operational
costs.
Therefore,
optimization
pump
operations
is
a
major
concern
water
utilities
around
world,
especially
recent
years
with
significantly
increasing
energy
prices.
Recently,
to
maximize
and
savings,
scheduling
frequently
done
real
time
by
integrating
relevant
software
into
company's
supervisory
control
data
acquisition
(SCADA)
system.
This,
however,
requires
solving
complex,
large,
nonlinear
problem
computationally
efficient
manner,
less
than
1
h.
To
achieve
this,
pump-scheduling
solved
this
paper
novel
hybrid
method
that
uses
linear
programming
(LP)
greedy
algorithm:
LPG.
new
methodology
applied
two
case
studies:
artificial,
benchmark
study
Anytown
network
real-life
Richmond
(WDN)
UK.
results
obtained
clearly
demonstrate
LPG
capable
problems
an
extremely
manner
while
preserving
accuracy
(i.e.,
near
optimality)
solution.
This
makes
particularly
appealing
use
real-time
applications.
