Microalgae-based
wastewater
remediation
can
contribute
to
the
decarbonization
of
sector
and
is
aligned
with
circular
economy
principals,
but
successful
large-scale
facilities
are
scarce
due
limited
knowledge
interactions
between
all
relevant
variables,
as
well
poor
development
control
automation
systems.
Machine
learning
(ML)
methods
have
potentiality
fill
this
gap
by
predicting
performance
in
microalgae
photobioreactors.
Some
attempts
been
performed
under
controlled
lab
conditions
where
variability
during
time
normally
neglected.
This
prevents
direct
application
their
results
industrial
scale.
The
aim
study
was
develop
machine
models
for
prediction
nitrogen
phosphorus
removal,
biomass
production
photosynthetic
efficiency
on
a
pilot-scale
database
obtained
1.5
years
operation.
Random
forests
(RF),
boosted
trees
(BT),
multilayer
perceptron
(MLP),
support
vector
(SVM)
were
used
purpose.
lowest
errors
MLP
model,
allowing
developed
tool
be
an
alternative
mechanistic
models.
Large
ranges
data
used,
considering
factors
diurnal
cycle
whose
influence
usually
Using
global
sensitivity
analysis
(GSA),
input-input
relationships
verified,
which
reflected
goodness
fit
(R2=0.83
-
0.95).
Moreover,
using
ML
models,
multi-criteria
optimization
operating
parameters
two
variants:
a)
(HRT,
SRT,
air
flowrate);
b)
influent
nutrient
loads.
Biotechnology Journal,
Год журнала:
2024,
Номер
19(3)
Опубликована: Март 1, 2024
Abstract
Microalgae
are
considered
to
be
a
promising
group
of
organisms
for
fuel
production,
waste
processing,
pharmaceutical
applications,
and
as
source
food
components.
Unicellular
algae
worth
being
because
their
capacity
produce
comparatively
large
amounts
lipids,
proteins,
vitamins
while
requiring
little
room
growth.
They
can
also
grow
on
fix
CO2
nitrogen
compounds.
However,
production
costs
limit
the
industrial
use
microalgae
most
profitable
applications
including
micronutrient
fish
farming.
Therefore,
novel
based
technologies
require
an
increase
efficiencies
or
values.
Here
we
review
recent
studies
focused
getting
strains
with
characteristics
cultivating
techniques
that
improve
production's
robustness
efficiency
categorize
these
findings
according
fundamental
factors
determine
Improvements
light
nutrient
delivery,
well
other
aspects
photobioreactor
design,
have
shown
highest
average
in
productivity.
Other
methods,
such
improvement
phosphorus
fixation
temperature
adaptation
been
found
less
effective.
Furthermore,
interactions
particular
bacteria
may
promote
growth
microalgae,
although
bacterial
grazer
contaminations
must
managed
avoid
culture
failure.
The
competitiveness
algal
products
will
if
discoveries
applied
settings.
Chemical Engineering Journal,
Год журнала:
2024,
Номер
485, С. 149981 - 149981
Опубликована: Фев. 25, 2024
Microalgae
cultivation
on
liquid
digestate
from
the
anaerobic
co-digestion
of
agricultural
feedstocks
is
an
interesting
option
for
nutrient
removal
and
resource
recovery
coupled
to
biomass
generation.
Both
reactors
considered
in
such
a
biorefinery
system
involve
complex
bioprocesses.
Although
different
pilot-scale
systems
coupling
digestion
algae-based
bioremediation
processes
have
been
described,
no
previous
attempts
model
entire
are
available
date.
In
this
work,
plant-wide
model,
named
ADAB
(anaerobic
algae-bacteria),
presented,
two
well-established
models
(IWA
–
ADM1)
(ALBA).
The
were
modified
with
necessary
equations
extensions
develop
dedicated
interface.
Phosphorous
dynamics
integrated,
including
activity
corrections
precipitation
processes.
ALBA
was
also
integrated
thermal
modelling
simulate
outdoor
raceway
ponds
greenhouse-covered
systems.
Solid/liquid
separation
units
pre-treatment
included.
prediction
consistency
adopted
physicochemical
sub-model
(PCM)
verified
results
both
reference
literature
Visual
MINTEQ.
reduced
complexity
PCM
limits
field
application,
but
it
better
computational
performance
seems
be
particularly
suitable
agro-zootechnical
digesters.
A
scenario
analysis
co-digester
design
operating
conditions
carried
out
assess
impacts
microalgae
cultivation.
It
highlighted
importance
proper
yet
noteworthy
robustness
performance.
use
can
facilitate
more
realistic
assessment
technical,
environmental,
economic
feasibility
full-scale
microalgal
biorefineries
based
digestate.
Chemical Engineering Journal,
Год журнала:
2024,
Номер
496, С. 154202 - 154202
Опубликована: Июль 21, 2024
Microalgae-based
wastewater
remediation
is
aligned
with
circular
economy
principals
but
successful
large-scale
facilities
are
scarce
due
to
the
limited
knowledge
of
interactions
between
all
relevant
variables
and
development
control
automation
systems.
Machine
learning
(ML)
methods
have
potential
predict
performance
in
microalgae
photobioreactors,
aim
optimize
it.
Some
attempts
been
performed
under
controlled
lab
conditions
which
prevents
direct
application
their
results
industrial
scale.
This
study
aims
develop
ML
models
for
prediction
nutrient
removal,
biomass
production
photosynthetic
efficiency
using
a
database
obtained
from
1.5-y
operation
pilot-scale
membrane
photobioreactor
(MPBR)
that
treated
sewage.
A
total
14
inputs
6
outputs
were
selected.
Random
forests
(RF),
boosted
trees
(BT),
multilayer
perceptron
(MLP),
support
vector
machine
(SVM)
tested.
The
lowest
errors
MLP
model,
allowing
developed
tool
be
used
as
an
alternative
mechanistic
models.
Large
ranges
data
used,
considering
variability
factors
diurnal
cycle
whose
influence
on
usually
neglected.
Using
global
sensitivity
analysis
(GSA),
input-output
relationships
verified,
reflecting
number
showing
significant
values
Shapley
Indices.
Additionaly,
partial
dependence
plots
showed
both
linear
nonlinear
depending
selected
outputs.
Finally,
models,
multi-criteria
optimization
operating
parameters
was
two
variants:
a)
(HRT,
SRT,
air
flowrate
(Fair));
b)
influent
loads
Fair,
nitrogen
phosphorus
loading
rates).
