Annals of Advances in Chemistry,
Journal Year:
2024,
Volume and Issue:
8(1), P. 001 - 007
Published: March 11, 2024
The
combustion
properties
of
energetic
materials
have
been
extensively
studied
in
the
scientific
literature.
With
rapid
advancement
data
science
and
artificial
intelligence
techniques,
predicting
performance
solid
rocket
propellants
(SRPs)
has
become
a
key
focus
for
researchers
globally.
Understanding
forecasting
characteristics
SRPs
are
crucial
analyzing
modeling
mechanisms,
leading
to
development
cutting-edge
materials.
This
study
presents
methodology
utilizing
neural
networks
(ANN)
create
multifactor
computational
models
(MCM)
burning
rate
propellants.
These
models,
based
on
existing
data,
can
solve
direct
inverse
tasks,
as
well
conduct
virtual
experiments.
objective
functions
(direct
tasks)
pressure
(inverse
tasks).
research
lays
foundation
developing
generalized
forecast
effects
various
catalysts
range
SRPs.
Furthermore,
this
work
represents
new
direction
science,
contributing
creation
High-Energetic
Materials
Genome
that
accelerates
advanced
FirePhysChem,
Journal Year:
2024,
Volume and Issue:
unknown
Published: June 1, 2024
The
present
study
was
devoted
to
setting
a
universal
T-independent
predictive
model
of
equivalent
in-service-time
(EIST)
for
homogenous
solid
propellant
(HSP)
surpass
the
limits
van't
Hoff
law
particularly
when
high
aging
temperatures
and/or
extended
durations
are
employed
in
artificial
plans.
To
achieve
this
objective,
four
double
base
rocket
propellants
(DBRP)
underwent
4
months
at
323.65
K,
338.65
353.65
and
368.65
with
sampling
conducted
every
20
days.
Fourier
Transform
Infrared
spectrometry
(FTIR)
showed
that
homolytic
scission
O-NO2
bonds
hydrocarbon
chains
nitrate
esters
main
processes
occurring
during
chemical
decomposition.
With
heating
temperature
increase,
decomposition
becomes
more
predominant.
Furthermore,
scatter
plot
from
Principal
Component
Analysis
(PCA)
FTIR
spectra
obtained
each
showed,
respectively,
over
than
88.9%,
94.3%,
97.4%,
98.6
variances
were
described
by
first
principal
component.
This
latter
value
found
97.6%
PCA
applied
all
spectra.
Using
PCA/FTIR
approach
recently
developed,
EIST
assessed
investigated
samples.
Subsequently,
an
individual
set
temperature,
which
used
establish
model.
final
computed
relative
deviation
5.3%
compared
those
experimental
way.
Moreover,
two
similar
DBRPs
aged
different
have
been
validate
model,
associated
mean
absolute
percentage
error
(MAPE)
4.6%.
comprehensive
statistical
analysis
highlighted
excellent
goodness-of-fit
metrics
decrease
increase
natural
temperature.
Antibiotic
resistance
is
a
critical
global
public
health
challenge
driven
by
the
limited
discovery
of
antibiotics,
rapid
evolution
mechanisms,
and
persistent
infections
that
compromise
treatment
efficacy.
Combination
therapies
using
antibiotics
nanoparticles
(NPs)
offer
promising
solution,
particularly
against
multidrug-resistant
(MDR)
bacteria.
This
study
introduces
an
innovative
approach
to
identifying
synergistic
drug–NP
combinations
with
enhanced
antimicrobial
activity.
To
carry
this
out,
we
compiled
two
groups
data
sets
predict
minimal
concentration
(MC)
zone
inhibition
(ZOI)
various
combinations.
CatBoost
regression
models
achieved
best
10-fold
cross-validation
R2
scores
0.86
0.77,
respectively.
We
then
adopted
machine
learning
(ML)-reinforced
genetic
algorithm
(GA)
identify
NPs.
The
proposed
was
first
validated
reproducing
previous
experimental
results.
As
proof
concept
for
discovering
combinations,
Au
NPs
were
identified
as
highly
when
paired
chloramphenicol,
achieving
minimum
bactericidal
(MBC)
71.74
ng/mL
Salmonella
typhimurium
fractional
inhibitory
index
6.23
×
10–3.
These
findings
present
effective
strategy
providing
combating
drug-resistant
pathogens
advancing
targeted
therapies.
Process Safety Progress,
Journal Year:
2025,
Volume and Issue:
unknown
Published: March 9, 2025
Abstract
Identification
of
high‐energy
compounds
used
in
the
pharmaceutical
industry
has
been
made
easy
via
differential
scanning
calorimetry,
but
when
designing
new
synthetic
routes,
working
with
or
limited
amounts
material,
inability
to
isolate
an
intermediate,
calorimetry
and
other
thermal
hazard
data
may
not
be
readily
available.
Here
we
report
a
machine
learning
model
that
uses
first
principles
as
baseline
for
predicting
decomposition
energies
materials
without
having
know
products
beforehand.
The
depends
on
bond
dissociation
simulate
breaking
then
subsequent
reconstruction
various
likely
products,
summing
energy
consumed
released.
A
light
gradient
boosting
was
trained
using
along
several
features
showed
significant
improvement
accuracy
compared
another
omitting
model,
especially
molecules
any
obvious
high
functional
groups.
Another
two
models
were
predict
shock
sensitivity
explosivity
potential
accuracy.
These
boosted
allow
informed
decisions
regarding
hazards
before
synthesizing,
isolating,
purchasing
them.
The Journal of Physical Chemistry A,
Journal Year:
2024,
Volume and Issue:
128(41), P. 9045 - 9054
Published: Oct. 8, 2024
Thermal
resistance
of
energetic
materials
is
critical
due
to
its
impact
on
safety
and
sustainability.
However,
developing
predictive
models
remains
challenging
because
data
scarcity
limited
insights
into
quantitative
structure–property
relationships.
In
this
work,
a
deep
learning
framework,
named
EM-thermo,
was
proposed
address
these
challenges.
A
set
comprising
5029
CHNO
compounds,
including
976
constructed
facilitate
study.
EM-thermo
employs
molecular
graphs
direct
message-passing
neural
networks
capture
structural
features
predict
thermal
resistance.
Using
transfer
learning,
the
model
achieves
an
accuracy
approximately
97%
for
predicting
thermal-resistance
property
(decomposition
temperatures
above
573.15
K)
in
compounds.
The
involvement
descriptors
improved
prediction.
These
findings
suggest
that
effective
correlating
from
atom
covalent
bond
level,
offering
promising
tool
advancing
design
discovery
field
