Energies,
Journal Year:
2023,
Volume and Issue:
16(19), P. 6830 - 6830
Published: Sept. 26, 2023
This
paper
presents
the
concept
of
an
innovative
control
a
central
heating
system
in
multifamily
building
based
on
original
thermodynamic
model,
resulting
architecture
system,
and
originally
designed
manufactured
wireless
temperature
sensors
for
thermal
zones.
The
novelty
this
solution
is
developed
layers
system:
distributed
measurement
correction
analysis,
which
existing
infrastructure
local
HVAC
controller.
approach
allows
effective
use
measured
data
from
zones
finally
sending
value
calculated
settings
to
Moreover,
analytical
layer,
model
was
also
implemented
that
calculates
necessary
amount
energy
subsystem
located
building.
algorithmic
strategy
presented
extends
functionality
significantly
improves
efficiency
existing,
classic,
reference
algorithm
by
implementing
additional
loops.
Additionally,
it
enables
integration
with
demand-side
response
systems.
successfully
tested,
achieving
real
savings
12%.
These
results
are
described
case-study
format.
authors
believe
can
be
used
other
buildings
thus
will
have
positive
impact
maintain
comfort
reduce
CO2
emissions.
Sustainability,
Journal Year:
2025,
Volume and Issue:
17(5), P. 1955 - 1955
Published: Feb. 25, 2025
This
study
enhances
thermodynamic
efficiency
and
demand
response
in
an
office
building’s
HVAC
system
using
machine
learning
(ML)
model
predictive
control
(MPC).
study,
conducted
a
simulated
EnergyPlus
8.9
environment
integrated
with
MATLAB
(R2023a,
9.14),
focuses
on
optimizing
the
of
building
Jeddah,
Kingdom
Saudi
Arabia.
Support
vector
regression
(SVR)
deep
reinforcement
(DRL)
were
selected
for
their
accuracy
adaptability
dynamic
environments,
exergy
destruction
analysis
used
to
assess
efficiency.
The
models,
MPC,
aimed
reduce
improve
response.
Simulations
evaluated
room
temperature
prediction,
energy
optimization,
cost
reduction.
DRL
showed
superior
prediction
accuracy,
reducing
costs
by
21.75%
while
keeping
indoor
increase
minimal
at
0.12
K.
simulation-based
approach
demonstrates
potential
combining
ML
MPC
optimize
use
support
programs
effectively.
Energies,
Journal Year:
2025,
Volume and Issue:
18(6), P. 1408 - 1408
Published: March 12, 2025
Building
HVAC
systems
face
significant
challenges
in
energy
optimization
due
to
changing
building
characteristics
and
the
need
balance
multiple
efficiency
objectives.
Current
approaches
are
limited:
physics-based
models
expensive
inflexible,
while
data-driven
methods
require
extensive
data
collection
ongoing
maintenance.
This
paper
introduces
a
systematic
relearning
framework
for
supervisory
control
that
improves
adaptability
reducing
operational
costs.
Our
approach
features
Reinforcement
Learning
controller
with
self-monitoring
adaptation
capabilities
responds
effectively
changes
operations
environmental
conditions.
We
simplify
complex
hyperparameter
process
through
structured
decomposition
method
implement
strategy
handle
over
time.
demonstrate
our
framework’s
effectiveness
comprehensive
testing
on
testbed,
comparing
performance
against
established
methods.
Applied Sciences,
Journal Year:
2025,
Volume and Issue:
15(8), P. 4291 - 4291
Published: April 13, 2025
Heating,
ventilation,
and
air
conditioning
(HVAC)
systems
account
for
up
to
40%
of
the
total
energy
consumption
in
buildings.
Improving
modeling
HVAC
components
is
necessary
optimize
efficiency,
maintain
indoor
thermal
comfort,
reduce
their
carbon
footprint.
This
work
addresses
lack
a
general
methodology
data
preprocessing
by
introducing
novel
approach
feature
extraction
selection
based
on
physical
equations
expert
knowledge
that
can
be
applied
any
data-driven
model.
The
proposed
framework
enables
forecasting
temperatures
individual
components.
validated
with
real-world
from
system
involving
fan
coil
unit
inertia
deposit
powered
geothermal
energy,
achieving
coefficient
determination
(R2)
0.98
mean
absolute
percentage
error
(MAPE)
0.44%.
