With
advances
in
vehicle
emission
control
technology,
updating
source
profiles
to
meet
the
current
requirement
of
apportionment
has
become
increasingly
crucial.
In
this
study,
on-road
and
non-road
particles
were
collected,
then
chemical
composition
individual
particle
was
analyzed
using
single
aerosol
mass
spectrometry.
The
data
grouped
an
adaptive
resonance
theory
neural
network,
identify
signatures
establish
a
spectral
database
mobile
sources.
addition,
deep
learning-based
model
(DeepAerosolClassifier)
for
classifying
established.
objective
accomplish
apportionment.
During
training
process,
achieved
accuracy
98.49%
on
validation
set
93.36%
test
set.
Regarding
interpretation,
ideal
spectra
generated
verify
its
accurate
recognition
characteristic
patterns
spectra.
practical
application,
used
perform
hourly
at
three
specific
field
monitoring
sites.
effectiveness
measurement
validated
by
combining
traffic
flow
spatial
information
with
results.
Compared
other
machine
learning
methods,
our
highly
automated
while
eliminating
need
feature
selection
enables
end-to-end
operation.
Thus,
future,
it
can
be
applied
refined
online
particulate
matter.
Environmental Science & Technology,
Год журнала:
2025,
Номер
unknown
Опубликована: Фев. 25, 2025
The
use
of
biodiesel
as
an
alternative
to
diesel
remains
controversial
because
its
potential
impact
on
organic
matter
emissions.
Herein,
intermediate-volatility
compounds
(IVOCs)
from
five
agricultural
machines
(AMs)
using
three
types
fuels
(pure
(B0),
mixed
with
5
and
20%
(B5
B20),
respectively)
were
collected
analyzed.
following
results
obtained:
(1)
Reasonable
ratios
reduced
IVOC
average
emission
factors
IVOCs
(EFIVOCs)
B5-fueled
AMs
485.8
±
272.9
mg/kg
fuel,
24.2%
reduction
that
B0-fueled
AMs.
(2)
chemical
compositions
secondary
aerosol
formation
potentials
(SOAFPs)
changed
during
fueled
by
different
blend
ratios.
highest
EFIVOCs
values
for
B5-
B20-fueled
within
the
Bin12
interval.
Consequently,
SOAFPs
n-alkanes
lower
than
those
B0.
(3)
oxygen
contents
B5
B20
increased
their
combustion
efficiency,
long-chain
carbons
in
pyrolyzed
fragmented.
However,
high-viscosity
inhibited
engine
performance
more
volatile
was
condensed
adsorbed
unburned
fuel.
In
future,
some
technologies
should
be
developed
reduce
viscosity
biodiesel.
Sustainability,
Год журнала:
2024,
Номер
16(6), С. 2292 - 2292
Опубликована: Март 9, 2024
In
tandem
with
the
advancement
of
urban
intelligent
technology,
construction
remote
monitoring
platforms
and
databases
for
non-road
mobile
machinery
is
gradually
improving
in
various
provinces
cities.
Employing
platform
enables
a
detailed
big
data
analysis
actual
operational
state
machinery.
This
method
yields
precise
on
activity
levels
types.
Importantly,
it
addresses
issue
reduced
accuracy
emission
inventories,
which
often
arises
from
conventional
practice
using
standard
recommended
values
Guide
to
determine
during
compilation
inventories.
Based
management
system
machinery,
value
level
was
obtained,
inventory
Shandong
Province
established.
The
PM,
HC,
NOx,
CO
main
including
forklifts,
excavators,
loaders,
off-road
trucks,
road
rollers,
were
measured.
findings
indicate
that
typically
exceeded
guideline’s
values.
Among
them,
annual
use
time
port
terminal
ground
handling
equipment
longest,
an
average
working
4321.5
h
per
equipment,
more
than
six
times
value.
all
types
loader
emissions
accounted
highest
proportion,
reaching
43.13%
total
pollutants.
With
tightening
national
Stage
II
III,
significant
reduction
mechanical
observed,
primarily
manifested
as
91%
decrease
NOx
emissions.
platform,
new
compiling
proposed
this
paper.
calculated
can
reflect
real
situations
provide
reference
basis
control
sustainable
policy
measures
