Applied Mathematics and Nonlinear Sciences,
Journal Year:
2024,
Volume and Issue:
9(1)
Published: Jan. 1, 2024
Abstract
With
the
rapid
growth
of
data
centers,
optimizing
energy
consumption
has
become
a
critical
challenge.
This
paper
proposes
an
management
framework
that
integrates
Long
Short-Term
Memory
(LSTM)
networks
for
load
prediction
with
dynamic
resource
allocation
algorithm
to
optimize
in
centers.
The
LSTM
model
is
employed
accurately
predict
future
workloads
based
on
historical
data,
enabling
proactive
adjustment
utilization.
algorithm,
informed
by
LSTM-based
predictions,
dynamically
manages
server
operations
and
cooling
systems
through
techniques
such
as
voltage
frequency
scaling
(DVFS)
consolidation,
ensuring
resources
are
allocated
efficiently
real
time.
proposed
incorporates
multi-objective
optimization
approach
balance
savings
system
performance.
We
use
genetic
algorithms
fine-tune
predicted
load,
response
Experimental
results
show
our
method
achieves
significant
compared
traditional
static
approaches,
minimal
impact
service
quality.
research
highlights
potential
machine
learning-driven
reducing
environmental
footprint
centers
while
maintaining
operational
efficiency.
Energies,
Journal Year:
2025,
Volume and Issue:
18(1), P. 149 - 149
Published: Jan. 2, 2025
Nowadays,
the
number
of
transistors
on
electronic
components
is
increasing
exponentially
leading
to
an
ultra-high
heat
flux
(106~107
W/m2).
The
non-uniform
load
distribution
chip
and
local
hot
spots
in
thermal
environment
are
key
issues
a
data
center
(DC).
Microchannel
liquid
cooling
effective
method
inhibit
accumulation
chip.
Optimizing
air
crucial
approach
realizing
energy
savings.
This
study
summarizes
latest
research
management
by
microchannel
optimization
DCs.
existing
concerning
structure
universality
sink
(MCHS),
stability
flow
boiling
new
coolant,
prediction
spots,
intelligent
control
system
identified.
Furthermore,
novel
strategy
multi-scale
synergy
recommended,
which
expected
suppress
heighten
temperature
uniformity.
review
provides
valuable
insights
into
for
It
paves
way
applying
innovative
technologies
artificial
intelligence
methods
promote
efficient
operation
low-carbon
retrofit
Clean Energy Science and Technology,
Journal Year:
2024,
Volume and Issue:
2(1), P. 141 - 141
Published: March 5, 2024
The
airflow
organization
of
the
data
center
directly
affects
temperature
control
performance
and
energy
efficiency
cooling
equipment.
servers
at
bottom
rack
usually
suffer
from
insufficient
rate
poor
effect.
This
is
because
limited
distance
between
perforated
floor,
small
horizontal
velocity
supply
air
flow.
study
aims
to
improve
uniformity
in
vertical
direction
by
deflectors,
thereby
further
improving
overall
center.
size
installation
deflectors
were
optimized
according
both
experiment
numerical
simulation
results.
From
results,
it
recommended
install
deflector
with
a
width
100
mm
an
angle
45°
under
floor
for
single-side
supply.
For
double-side
supply,
should
be
installed
30°
achieve
best
distribution.
Consequently,
intake
significantly
increased,
occurrence
local
hot
spots
was
effectively
reduced.
without
conducted
ANSYS.
results
show
that,
increased
inlet
16.98%
improved
dater
conditioners
1.98%.
Sustainability,
Journal Year:
2024,
Volume and Issue:
16(16), P. 7222 - 7222
Published: Aug. 22, 2024
In
the
age
of
digitalization
and
big
data,
cooling
systems
in
data
centers
are
vital
for
maintaining
equipment
efficiency
environmental
sustainability.
Although
many
studies
have
focused
on
classification
optimization
center
systems,
systematic
reviews
using
bibliometric
methods
relatively
scarce.
This
review
uses
analysis
to
explore
classifications,
control
optimizations,
energy
metrics
aiming
address
research
gaps.
Using
CiteSpace
databases
like
Scopus,
Web
Science,
IEEE,
this
study
maps
field’s
historical
development
current
trends.
The
findings
indicate
that,
firstly,
strategies,
focal
points.
Secondly,
assesses
applicability
air-cooled
liquid-cooled
different
operational
environments,
providing
practical
guidance
selection.
Then,
air
demonstrates
that
optimizing
design
static
pressure
chamber
baffles
has
significantly
improved
airflow
uniformity.
Finally,
article
advocates
expanding
use
artificial
intelligence
machine
learning
automate
collection
analysis,
it
also
calls
global
standardization
metrics.
offers
new
perspectives
design,
optimization,
performance
evaluation
systems.
