AE6412-Debrief-Fall09

AE6412-Debrief-Fall09 -...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: AE
6412
Final
Project
Comments
 
 1. To
remind
all
this
was
supposed
to
be
a
turbulent
combustion
final
project
so
 the
approach
to
deal
with
“turbulence”
is
fundamental
to
this
effort.
 2. Some
of
the
presentation
and
reports
were
of
the
quality
I
expected
for
a
 semester
long
effort
but
I
was
disappointed
by
some
reports
and
some
 presentations.

 3. In
general,
many
of
you
failed
to
clearly
identify
what
was
the
“objective”
of
 the
research
project
and
“why”.
The
link
of
these
objectives
to
turbulent
 combustion
was
in
some
cases,
only
indirect.

 4. In
the
same
manner,
the
conclusions
did
not
clearly
spell
out
exactly
what
 was
accomplished
relative
to
the
objectives
and
what
were
the
key
 observations
relevant
for
turbulent
combustion.
 5. Some
of
the
numerical
work
was
grossly
ad
hoc
in
the
nature
of
turbulence
 initialization
and/or
defining
the
basis
of
the
study.
It
appears
that
in
some
 studies,
just
two
or
three
calculations
were
performed
and
some
plots
made
 but
there
was
no
attempt
to
verify
if
any
were
correct.
This
was
very
 troublesome.
 6. Some
of
the
reports
were
poorly
written
or
looked
like
cut
and
paste
of
 sentences
and/or
material
with
very
little
attempt
to
merge
them
into
a
 single
cohesive
document.
This
has
serious
implications
if
you
are
planning
 to
write
thesis
and/or
journal
papers.
 7. The
following
comments
are
for
all
the
presentations.
If
some
comments
 were
already
addressed
in
the
original
report
then
a
brief
“rebuttal”
should
 be
included
at
the
BOTTOM
of
the
report,
starting
on
a
new
page.
 8. ALL
CHANGES
should
be
done
in
a
different
COLOR
to
indicate
where
 changes
are
being
made
–
even
if
substantial
revision
is
made.
 9. The
revised
document
should
be
emailed
ONLY
–
this
is
due
by
Dec
11th
5
pm
 without
fail.

 
 
 
 Acharya:
Acoustically
Forced
Turbulent
Flame
 
 1. Estimate
the
u’/SL
range
if
you
assume
the
forcing
amplitude
contributes
to
 this
turbulent
fluctuation
–
what
does
acoustic
forcing
do
to
the
turbulent
 fluctuations?
 2. What
are
the
uncertainties
and
errors
in
your
analysis?
 3. If
ST/SL
v/s
u’/SL
is
plotted
using
1
above
is
there
any
trend
that
is
 consistent
with
conventional
closures?
 4. Can
you
estimate
the
wrinkling
at
the
acoustic
wavelength
and
compare
it
to
 the
characteristic
length
scales
–
Kolmogorov
or
integral
scale?
 5. ST/SL
v/s
phase
–
is
there
any
effect
of
turbulence
on
the
peak
at
200
 degree?
 6. Summary
conclusions
should
try
to
identify
what
are
the
expected
results
of
 such
a
study
if
there
was
NO
turbulence
–
i.e.,
compare
and
contrast
results
 with
and
without
turbulence.
 
 Foley:
Premixed
swirl
stabilized
combustion
 
 1. Equations
1
and
2
are
not
proper
equations
 2. There
is
no
estimate
of
u’/SL
and
Re
for
the
test
case
 3. Quality
of
figures
(line
plots)
is
very
bad
and
is
not
proper.
 4. Text
is
rather
poorly
written
and
really
should
be
revised.
Many
statements
 are
made
and
they
appear
to
be
just
picked
up
from
lectures
or
notes
but
 without
proper
integration.
Statement
such
as
Gibson
scale
relevance
etc
are
 not
relevant
unless
there
is
an
actual
attempt
to
estimate
it
and
show
that
the
 data
shows
no
scales
below
this
scale.
 5. Discussion
regarding
small‐scale
turbulence
in
the
text
and
presentation
is
 incorrect.
What
is
meant
by
local
isotropy?
All
turbulent
flows
exhibit
 isotropy
at
the
small‐scales
–
you
need
to
revise
your
writing
and
review
the
 concept
of
local
isotropy
–
there
seems
to
be
a
serious
misconception
or
lack
 of
understanding.

The
Borghi
diagram
is
still
valid
for
swirling
flow
as
long
 as
the
range
of
scales
are
between
integral
and
Kolmogorov
scale
–
this
range
 is
inertial
range
and
is
considered
isotropic.
The
large
coherent
structures
 are
not
considered
part
of
this
range.
You
need
to
explain
or
revise
these
 discussions
since
they
are
incorrect.
 6. What
is
the
physics
behind
the
longitudinal
correlation
on
the
flame
analysis?
 Does
it
tell
you
anything
useful
in
terms
of
understanding
turbulent
 combustion?
 7. You
need
to
explain
why
swirling
flow
flame
is
similar
or
dissimilar
to
a
 premixed
flame
without
swirl
–
that
is,
what
does
swirl
do
to
the
turbulent
 flame
speed?
 
 Girgis:
Strain‐vorticity
alignment
in
a
2D
TML
 
 1. What
does
this
study
has
to
do
with
turbulent
combustion?
Can
you
justify
 this
effort?
Only
2
simple
2D
calculations
appear
to
have
been
performed
 with
global
kinetics,
which
should
not
have
taken
so
much
time.
If
you
had
 initialized
the
TML
with
random
turbulence
in
addition
to
the
coherent
mode
 and
played
with
the
amplitudes
then
it
could
be
considered
a
reasonable
 attempt
 2. The
presentation
and
explanation
was
very
poor
and
there
appears
to
be
lack
 of
interest
or
understanding
of
the
material
in
the
presentation.
Regardless
 of
whether
this
is
your
research
area
(this
is
issue
is
irrelevant
for
a
course),
 this
is
an
academic
advanced
course
with
a
significant
assessment
based
on
 the
project
and
your
presentation.

 3. The
initialization
is
2D
coherent
modes
and
yet
the
solution
shows
many
 smaller
structures
–
these
look
incorrect
and
unphysical.
There
appears
to
be
 no
attempt
to
check
if
the
results
mean
anything
correct.
 
 Gottiparthy:
Effect
of
turbulence
on
detonation
 
 1. What
are
new
results
here
in
terms
of
the
effect
of
turbulence?
Is
the
inflow
 turbulence
chosen
here
realistic
for
this
application?
Need
to
explain
why
 coherent
structures
grow
if
the
initialization
is
random
–
i.e.,
if
there
were
no
 initial
turbulence
will
they
form
and
grow?
 2. How
exactly
is
partially
premixed
mixture
defined?
 3. Fig
6
and
7
need
to
quantify
the
key
differences
and
similarities
 4. What
is
meant
by
the
increase
in
detonation
velocity
“sometimes”
–
need
to
 justify
the
range
of
increase/dicreases
–
can
it
be
correlated
to
the
formation
 of
coherent
structures?
Do
CS
grow
due
to
turbulence
or
partial
premixing
 effect?
Are
the
local
hot
spots
numerical
or
are
they
well
resolved
to
be
 physical?
 5. Clearly
indicate
where
the
heat
release
effect
is
in
terms
of
its
location
 relative
to
the
detonation
front
and
with
this
address
point
4
above.
 
 Prabakar:
Turbulent
Global
Consumption
Speed
 
 1. Syn
gas
composition
and
the
range
of
laminar
flame
speed
relevant
for
this
 study
should
be
clearly
noted
 4. The
results
seem
to
focus
on
ignition
but
this
is
a
much
more
complex
issue
–
 did
you
attempt
to
determine
the
ignition
delay
time
for
your
calculations?
It
 is
not
clear
if
your
simulation
is
actually
igniting
since
by
the
time
the
 products
appear
the
TML
is
grown
unphysical.

 5. It
seems
that
the
reactions
seem
to
start
everywhere

‐
this
does
not
make
 sense
and
is
not
an
ignition
event.
This
again
points
to
errors
in
the
setup
 and/or
simulation.
 6. There
was
a
lot
of
comments
about
vorticity‐strain
alignment
but
the
 analysis
was
limited
to
lots
of
vorticity
and
enstrophy.
There
was
not
much
 difference
in
the
results
with
and
without
heat
release
(probably
because
 there
was
not
much
heat
release)
–
and
the
discussion
did
not
explain
what
 heat
release
was
expected
to
do
to
the
alignment
(which
strictly
speaking
 was
not
computed).
Also,
there
was
no
clear
discussion
of
the
result
of
scalar‐ gradient‐strain
alignment
–
again
such
analysis
requires
computing
the
 strain
the
in
the
plane
of
the
flame
which
means
computing
the
tangent
to
the
 surface.
Not
sure
if
this
was
done
in
this
manner.
If
this
was
done
then
it
 should
be
clearly
noted
and
explained.
 7. Strain
is
maximum
in
the
braid
region
in‐between
the
vortices
and
not
in
the
 vortex
region;
also
regions
of
high
vorticity
does
not
imply
mixing
and
the
 flame
is
typically
located
in
the
interface
region
that
may
or
may
not
be
in
 high
strain.
Strain‐vorticity‐scalar‐gradient
interaction
is
important
and
 complex
and
the
effect
of
heat
release
on
this
has
been
discussed
in
many
 papers
in
the
past
but
very
few
appear
to
have
to
be
reviewed.
 8. The
explanation
that
the
flame
is
burning
on
one
side
and
the
shear
is
getting
 pushed
up
is
not
clear
–
a
reference
is
cited
but
is
that
case
same
as
yours?

 2. Some
additional
background
on
the
relevance
of
ST
should
be
given.
For
 example,
since
Syngas
has
both
CO
and
H2
exactly
how
is
SL
defined
and
then
 how
is
ST
related
to
the
SL.
Need
to
show
how
is
SL
computed.

 3. Is
there
any
data
to
confirm
the
effect
of
stretch
effect
on
syngas
flames?
Can
 you
or
did
you
verify
that
your
OPPDIFF
calculations
were
correct
–
i.e.,
by
 checking
some
other
well
established
flame
(e.g.,
methane).
 4. What
kinetics
model
was
used
for
syn
gas?
 5. Inflow
turbulence
generation
–
since
there
is
some
estimate
for
ST/SL
v/s
 u’/SL
some
more
discussion
on
the
increase
in
ST/SL
for
fixed
u’/SL
with
 increase
in
H2
should
be
given.
I.e.,
exactly
what
is
the
physics
here?
u’/SL
 can
be
fixed
by
changing
fuel
mixture
and/or
turbulence
when
the
H2
 content
is
changed
–
however,
does
it
ensure
that
the
flame
is
still
thin
and
 satisfies
the
flamelet
criteria?
 
 Renganathan:
Le
number
effect
on
turbulent
premixed
flame
 
 1. The
writeup
contained
very
little
analysis
and
discussion
of
the
work
actually
 performed
but
contained
a
lot
of
summary
from
published
work

‐
in
some
 case
these
are
just
listed
but
their
relevance
to
the
actual
work
being
 performed
was
not
clarified.
 2. Exactly
what
was
the
objective
of
this
study?
your
presentation
did
not
 follow
the
report
and
lacked
clear
explanations.
 3. What
is
ksgs‐LES
being
done
for
a
2D
test
case?
Does
this
even
make
sense?
 4. Inflow
turbulence
does
not
make
sense
especially
the
manner
of
 initialization.
The
write‐up
was
incomplete
with
very
little
explanation
of
 what
exactly
was
being
done
and
why?
If
the
turbulence
was
set
only
at
t
=
0
 and
in
a
non‐uniform
grid
it
will
not
survive
long.
In
fact,
with
a
non‐uniform
 grid
the
turbulence
will
decay
quite
differently.

 5. It
was
mentioned
that
turbulence
is
being
reset
every
time‐step.
However,

 then
the
simulation
will
not
satisfy
the
governing
equations
since
the
 initialization
is
arbitrary.
The
solution
therefore
will
be
continuously
 readjusting
and
will
not
show
anything
physical
in
the
statistical
sense.
The
 explanation
provided
for
this
is
incomplete
or
incorrect.
 6. The
details
of
the
setup,
grid,
etc.,

are
not
provided.
IF
the
grid
is
so
refined
 why
is
the
flame
looking
so
thick?
Is
the
domain
of
application
in
the
flamelet
 regime?
If
so,
did
you
verify
this
is
correct?
 7. Exactly
what
is
the
meaning
of
using
a
no‐slip
walls
–
with
a
unclustered
grid
 and
for
a
flame‐turbulence
study?
It
is
not
clear
if
the
cited
reference
is
using
 no‐slip
wall
since
it
would
change
the
SL
near
the
wall
and
for
this
narrow
 region
probably
impact
the
entire
flame
structure
unless
there
is
a
huge
grid
 resolution
near
the
wall.
It
is
not
clear
if
the
Bell
citation
was
reviewed
 carefully
for
this
study.
 8. Was
there
any
attempt
to
verify
if
the
results
were
correct?
Did
you
make
an
 estimate
for
the
SL
for
this
flame
and
compare
with
what
is
being
predicted?
 Some
comments
on
the
effect
of
Le
number
is
made
but
there
seem
to
be
no
 
 Smith:
Turbulent
Stirring
in
LEM
 
 1. Given
the
range
of
Re‐Sc
studied
here
you
need
to
define
the
number
of
 stirring
v/s
diffusion
being
carried
out
–
what
is
the
difference
between
the
 diffusion
and
stirring.

 2. Exactly
how
is
scalar
dissipation
defined?
It
can
be
directly
defined
from
the
 scalar
field
in
the
LEM
as
well
as
from
the
fluctuating
field
–
is
there
any
 difference?
 3. What
is
the
effect
of
initialization
on
the
statistics?
For
example,
if
the
initial
 scalar
gradient
is
changed
to
a
block
initialization
will
the
statistics
converge
 for
all
the
mapping?
 4. Since
your
results
look
different
from
Kerstein
is
there
any
way
to
verify
why
 there
are
these
differences?
 5. Your
presentation
included
slides
that
were
not
in
the
report
so
this
needs
to
 be
corrected.
 
 Pasamurti:
Le
effects
in
Premixed
H2‐air
flame
 
 1. Is
this
is
a
LES?
Given
that
it
is
2D
study
–
it
cannot
be
a
LES.
If
it
is
DNS
it
is
 OK
if
all
scales
are
being
resolved
–
in
any
case,
the
resolution
and
the
scales
 etc
must
be
discussed.
 2. The
Maccormack
scheme
is
second
order
in
time
–
why
are
you
saying
it
is
 first
order
in
time?
 3. The
definition
of
Le
and
how
it
is
defined
in
the
setup
needs
to
be
clarified
–
 is
this
even
a
valid
approach?
If
there
are
different
species
and
if
the
Le
is
 fixed
for
all
what
is
the
diffusion
coefficient
for
each?
Is
the
thermal
 diffusivity
different
for
each
species?
In
the
code
thermal
and
molecular
 diffusivities
are
used
as
a
mixture
property
so
is
this
definition
of
Le
 consistent?
 4. Do
your
results
show
consistency
with
published
work
in
the
Le
effects?
 Most
Le
studies
use
single
step
kinetics
–
so
has
there
been
any
study
similar
 to
yours?
Your
presentation
about
the
LE
effects
appears
to
be
uncertain
or
 unclear.
Review
this
and
revise
your
writeup
to
show
that
you
understand
 this
properly.
 5. Re‐plot
all
curves
with
proper
labels.
Not
“Normalized
ST”
but
ST/SL.
What
is
 meant
by
time
in
the
x‐axis?
can
you
explain
the
increase
in
ST
with
time?
 You
mention
this
is
similar
to
the
Poinsot
test
case
–
if
so
do
your
results
 agree
at
all?
What
are
the
similarities
and
differences
with
their
result?
 6. The
initialization
is
again
not
well
explained
–
does
it
show
proper
physics
of
 flame‐turbulence
interactions?
 7. Clearly
note
what
were
the
key
results
from
your
study
and
what
is
relevant
 for
turbulent
mixing
layers.
 
 basis
for
it
in
this
study.
Since
this
is
a
methane
flame
with
reduced
1‐step
 kinetics
where
does
differential
diffusion
effect
comes
in?

 
 
 Sanyal:
Differential
Diffusion
in
Autoignition
 
 1. Since
this
is
a
2D
DNS
(I
assume?)
–
the
range
of
scales
that
are
being
 resolved
should
be
discussed.
 2. Why
do
you
have
the
second
term
in
the
diffusion
velocity
in
terms
of
the
 temperature
gradient?
Is
this
Soret
(or
Dufort
–
I
forget
which)
important?
 3. Clarify
what
is
meant
by
fixing
the
Le
number
for
multispecies
flows
–
is
 there
any
reference
you
can
cite?
If
this
is
based
using
the
mixture
thermal
 diffusivity
–
exactly
what
is
the
diffusion
coefficients
for
the
species
you
are
 using?
 4. Did
you
try
comparing
the
ignition
delay
for
this
case
against
the
classical
 PSR
type
of
result
that
is
reported
in
the
books?
For
zero
or
low
turbulence
 this
ignition
delay
should
be
similar.

 5. Clarify
the
definition
of
the
turbulence
time
scale
and
how
is
strong
 turbulence
quantified?
 6. Can
you
verify
your
results
in
terms
of
physics
and
show
that
these
results
 are
as
expected?
Revise
your
conclusions
to
highlight
what
it
tells
you
about
 the
effect
of
turbulence
on
autoignition.
Also,
note
or
discuss
how
these
 results
are
affected
by
the
Le
number
effect
if
at
all.
 
 Nagarajan:
Linear
stability
analysis
of
turbulent
mixing
layer
 
 1. Clearly
explain
what
is
meant
by
reduction
of
growth
rate
for
turbulent
 mixing
layer
since
it
is
for
the
stability
mode
and
not
for
the
actual
mixing
 layer??
 2. Since
it
is
a
2D
analysis
you
need
to
show
or
prove
that
the
2D
stability
 analysis
without
the
wake
and/or
turbulence
matches
classical
growth
rates
 and
predicts
the
proper
stability
curve.
 3. Figure
legends
for
the
various
curves
are
needed.
Figure
titles
should
clearly
 say
what
you
are
showing.
 4. Effect
of
combustion
and
heat
release
on
the
growth
rates
should
be
 discussed
against
cases
without
heat
release
and
again
compared
or
 discussed
in
terms
of
what
is
available
in
literature.
 ...
View Full Document

Ask a homework question - tutors are online