Scientific Reports,
Journal Year:
2024,
Volume and Issue:
14(1)
Published: Oct. 29, 2024
The
porous
underground
structures
have
recently
attracted
researchers'
attention
for
hydrogen
gas
storage
due
to
their
high
capacity.
One
of
the
challenges
in
storing
aqueous
solutions
is
estimating
its
solubility
water.
In
this
study,
after
collecting
experimental
data
from
previous
research
and
eliminating
four
outliers,
nine
machine
learning
methods
were
developed
estimate
To
optimize
parameters
used
model
construction,
a
Bayesian
optimization
algorithm
was
employed.
By
examining
error
functions
plots,
LSBoost
method
with
R²
=
0.9997
RMSE
4.18E-03
identified
as
most
accurate
method.
Additionally,
artificial
neural
network,
CatBoost,
Extra
trees,
Gaussian
process
regression,
bagged
regression
support
vector
machines,
linear
had
values
0.9925,
0.9907,
0.9906,
0.9867,
0.9866,
0.9808,
0.9464,
0.7682
2.13E-02,
2.43E-02,
2.44E-02,
2.83E-02,
2.85E-02,
3.40E-02,
5.68E-02,
1.18E-01,
respectively.
Subsequently,
residual
plots
generated,
indicating
performance
across
all
ranges.
maximum
-
0.0252,
only
4
points
estimated
an
greater
than
±
0.01.
A
kernel
density
estimation
(KDE)
plot
errors
showed
no
specific
bias
models
except
model.
investigate
impact
temperature,
pressure,
salinity
on
outputs,
Pearson
correlation
coefficients
calculated,
showing
that
0.8188,
0.1008,
0.5506,
respectively,
pressure
strongest
direct
relationship,
while
inverse
relationship
solubility.
Considering
results
research,
method,
alongside
approaches
like
state
equations,
can
be
applied
real-world
scenarios
storage.
findings
study
help
better
understanding
solutions,
aiding
systems.
Gas Science and Engineering,
Journal Year:
2023,
Volume and Issue:
121, P. 205196 - 205196
Published: Dec. 16, 2023
This
review
presents
a
State-of-Art
of
geochemical,
geomechanical,
and
hydrodynamic
modelling
studies
in
the
Underground
Hydrogen
Storage
(UHS)
domain.
Geochemical
assessed
reactivity
hydrogen
respective
fluctuations
losses
using
kinetic
reaction
rates,
rock
mineralogy,
brine
salinity,
integration
redox
reactions.
Existing
geomechanics
offer
an
array
coupled
hydro-mechanical
models,
suggesting
decline
failure
during
withdrawal
phase
aquifers
compared
to
injection
phase.
Hydrodynamic
evaluations
indicate
critical
importance
relative
permeability
hysteresis
determining
UHS
performance.
Solubility
diffusion
gas
appear
have
minimal
impact
on
UHS.
Injection
production
cushion
deployment,
reservoir
heterogeneity
however
significantly
affect
performance,
stressing
need
for
thorough
experimental
studies.
However,
most
current
efforts
focuses
assessing
aspects
which
are
crucial
understanding
viability
safety
In
contrast,
lesser-explored
geochemical
geomechanical
considerations
point
potential
research
gaps.
Variety
software
tools
such
as
CMG,
Eclipse,
COMSOL,
PHREEQC
evaluated
those
underlying
effects,
along
with
few
recent
application
data-driven
based
Machine
Learning
(ML)
techniques
enhanced
accuracy.
identified
several
unresolved
challenges
modelling:
pronounced
lack
expansive
datasets,
leading
gap
between
model
predictions
their
practical
reliability;
robust
methodologies
capable
capturing
natural
subsurface
while
upscaling
from
precise
laboratory
data
field-scale
conditions;
demanding
intensive
computational
resources
novel
strategies
enhance
simulation
efficiency;
addressing
geological
uncertainties
environments,
that
oil
simulations
could
be
adapted
comprehensive
offers
synthesis
prevailing
approaches,
challenges,
gaps
domain
UHS,
thus
providing
valuable
reference
document
further
efforts,
facilitating
informed
advancements
this
towards
realization
sustainable
energy
solutions.
