Journal of Energy and Power Engineering 14 (2020) 193-210
doi:10.17265/1934-8975/2020.06.003
Artificial Intelligence Integration with Energy Sources
(Renewable and Non-renewable)
Bahman Zohuri
1
, Farahnaz Behgounia
1
and Ziba Zibandeh Nezam
2
1. Ageno School of Business, Golden Gate University, San Francisco, California 94105, USA
2. Department of Physics, University campus 2, University of Guilan, Rasht, Iran
Abstract: In the past decade or so, AI (artificial intelligence) technology has been growing with such a mesmerizing speed that today
its presence in almost any industry that deals with any huge sheer volume of data is taking advantage of AI by integrating it into their
day-to-day operation. Meanwhile, seven billion people worldwide shape the world’s energy system and directly impact the
fundamental drivers of energy, both renewable and non-renewable sources, to meet the demand for electricity from them. These energy
sources can be reached from nature such as solar, wind, etc., and human-made such as NPPs (nuclear power plants) in the form of either
fission as an old technology since the Manhattan project and in the near future as fusion in the form of magnetic or inertial confinements.
Meanwhile, AI controlling nuclear reactors are about to happen. The basic idea is to apply AI with its two subset components as ML
(machine learning), and DL (deep learning) techniques to go through the mountains of data that come from a reactor, spot patterns in it,
and calling them to the unit’s human attention operators is not invadable either. Designers of such nuclear reactors will combine
simulation and real-world data, comparing scenarios from each to develop “confidence [in] what they can predict and what is the range
of uncertainty of their prediction”. Adding that, in the end, the operator will make the final decisions in order to keep these power plants
safe while they are in operation and how to secure them against cyber-attack natural or human-made disasters. In this short
communication article, we would like to see how we can prove some of these concepts; then a NPP manufacturer can pick it up and use
it in their designs of a new generation of these reactors.
Key words: AI, ML, DL, renewable and nonrenewable source of energy, fusion and fission reactors, SMRs (small modular reactors)
and generation four system, IoT (internet of things), dynamic site, return on investment, total cost of ownership.
1. Introduction
Since about 600,000 years ago, when humans, for
the first time, learned how to master the fire and used it
to his advantage, energy became an integral part of his
life. As time passed and technology advanced, the type
of energies used has changed (wind, water, and wood).
It was not until Industrial Revolution that the demand
for energy changes substantially to the point that the
existing energy sources were not enough. Some of the
factors that contributed to this increase in demand were
rapid growth of the human population, building new
cities, factories, roads, bridges, etc. An increase in
demand pushed humans to search for a new source of
Corresponding author: Bahman Zohuri, Ph.D., adjunct
professor, research fields: artificial intelligence and machine
learning.
energy. The result was the discovery of coal. History
shows that as time passed, both the demand for energy
and looking for new sources of energy continued to
grow.
Note that: Before the industrial revolution, our
energy needs were modest. We relied on the sun—and
burned wood, straw, and dried dung when the sun
failed us for heat. For transportation, horses’ muscle
and the power of the wind in our sails took us to every
corner of the world. For work, we used animals to do
jobs that we could not do with our own labor. Water
and wind drove the simple machines that ground our
grain and pumped our water (Union of Concerned
Science). Wrigley also stresses that the shift from an
economy that relied on land resources to one based on
fossil fuels is the essence of the Industrial Revolution
and could explain the Dutch and British economies’
D
DAVID PUBLISHING
Artificial Intelligence Integration with Energy Sources (Renewable and Non-renewable)
194
differential development. Both countries had the
necessary institutions for the Industrial Revolution to
occur, but capital accumulation in the Netherlands
faced a renewable energy resource constraint, while in
Britain, domestic coal mines in combination with
steam engines, at first to pump water out of the mines
and later for many other uses, provided a way out from
the constraint [1].
Energies can be classified based on many different
characteristics that they have. In general, they are
divided into two basic categories: renewable and
non-renewable. Renewable or clean energy is a type of
energy that comes from natural sources or processes
that are constantly replenished. We also study the role
of technology in renewable energies. The impact that
renewable energies have on the environment and the
advantages and disadvantages of each one, holistically,
are discussed in this paper, while more details can be
found by Zohuri and McDaniel [2].
Renewable energy, which is naturally replenished on
a human time scale, is listed here as:
(1) Sunlight,
(2) Wind,
(3) Rain,
(4) Tides,
(5) Waves, and,
(6) Geothermal heat.
Renewable energy often provides energy in four
critical areas: electricity generation, air and water
heating/cooling, transportation, and rural (off-grid) or
rather or characteristic of the countryside rather than
the town energy services [2].
According to the US Energy Information
Administration, non-renewable resources are any
resources that “do not form or replenish in a short
period”. The most common non-renewable resources
include fossil fuels like crude oil, natural gas, coal,
and uranium nuclear energy in the form of the
fission process and contrast in the near future, we can
consider the commercial aspect of nuclear fusion as
well [2].
In particular, these approaches apply to the new
generation of these NPPs (nuclear power plants) both
from fission and fusion generating energy
prospectively and apply to Generation IV and possible
futuristic Generation V of these plants. These
generations are pushing the modulization of these
reactors while getting their size and foot-print reduced
drastically in respect to the previous generation of these
reactors, notably the fission type NPP. See Fig. 1,
where the history of these types is depicted by the
invention of their generation historically.
Fig. 2 is the presentation of both renewable and
non-renewable sources of energy by their types,
selectively as well. Considering that integration of
AI (artificial intelligence) as a complement and
supportive element to its human partner is invadable
in order to operate these new generations of reactors
[3]. As we said in the abstract of this paper, the
operation complexity of these new generations of
fission/fusion reactors requires manipulation of many
data and analytics that implementation of AI with its
subsets such as ML (machine learning) and DL (deep
learning) becomes a mandatory factor to their human
partner.
The basic idea is that application of AI integrated
with ML and DL techniques would go through the
mountains of data that come from a reactor, spot
patterns in it, and calling them to the unit’s human
operators’ attention is not an invadable fact.
The nuclear industry in term of SMR (Small
Modular Reactor) of Generation Four (GEN-IV) as
well as aSMR (advanced SMR) (i.e., MSR (molten salt
reactor) and ISMR (integral molten salt reactor)) in the
past few years has gained much momentum and
recently a lot grant dollar has allocated to universities
such as North Carolina State University (e.g., a $3.4
million grant) by US DOE’s (Department of Energy’s)
ARPA-E (Advanced Research Projects
Administration-Energy) in the form of a research
consortium to explore applications of AI for the
technology associated with NPPs of advanced types.
Fig. 1 Evol
u
Fig. 2 Integ
r
Artificial In
t
u
tion of NPPs.
r
ation of AI a
n
t
elligence Int
e
n
d NPPs.
e
gration wit
h
h
Energy Sources (Renewable and No
n
n
-renewable)
19
5
5
Artificial Intelligence Integration with Energy Sources (Renewable and Non-renewable)
196
Fig. 3 Smart and centralized control centers.
Source: https://www.imaginovation.net.
DOE via its ARPA-E branch under a newly
established task team was assigned to investigate and
review the augmentation of AI and its status of this
technology for applications in NPPs, as well as identify
guidelines for AI work.
The task also was to identify the work required to
allow the nuclear industry to realize the maximum
benefit from the technology. The nuclear industry’s
state was analyzed to determine where the application
of AI technology could be of the greatest benefit.
As we stated, future NPPs are projected to be highly
modularized, with the possibility of several NPPs being
operated from one central control room.
Smart, centralized control centers are becoming
handy-dandy for those mundane days by taking over
human operator control with integrated AI-based
system in place [4]. See Fig. 3.
When coupled with AI, the data can give the grid
operators new insights for better control operations. It
offers flexibility to the energy suppliers to cleverly
adjust the supply with demand.
The advanced load control systems can be installed
with the equipment, such as industrial furnaces or large
AC units, which can automatically switch off when the
power supply is low. Intelligent storage units can also
be adjusted based on the flow of supply.
Additionally, smart machines and advanced sensors
can make weather and load predictions that can
improve the integration and efficiency of renewable
energy.
AI-based diagnostics and control systems can make
this possible by taking over the mundane day to day
oversight tasks that the operators must perform to keep
the plant running.
While still maintaining control of the plant’s plan,
the operators can use these systems to do the simpler
tasks, allowing the operators to concentrate on the
more complex ones. The diagnostics systems will
enable the operators to spend less time on actual signal
monitoring, and more time on continuing activities
such as interpretation of data compiled over long
intervals.
One of the more useful aspects of the AI-based
systems is explaining the reasons for the choices they
have made and that they are recommending to the
operators. The complete line of logic can be presented,
Fig. 4 Pyra
m
with nothin
g
that no ove
r
be pre-che
c
consensus o
AI syste
m
diversity,
m
Any syste
m
system, wh
i
arriving at
system wo
u
b
efore any
t
developed t
o
operators to
for design c
o
These ar
e
result of A
I
industry.
2. The Fu
t
The ener
g
uncertaintie
s
expert in th
world will
l
today.
The worl
d
not necessa
r
can be con
fi
Artificial In
t
m
id of AI, M
L
g
left out, an
d
r
sights have
b
c
ked and w
o
f many expe
r
m
s can help
m
aking nucle
a
m
developed
i
ch would u
s
the solution
u
ld always
b
t
hing would
o
constantly
problem co
n
o
nservatisms
e
a few bene
f
I
technology
t
ure of En
e
g
y future wil
l
s
highlighte
d
e field, we
c
l
ook a lot d
i
d
energy sy
s
r
ily in the w
a
fi
dent of so
m
t
elligence Int
e
L
and DL histo
r
d
allow the o
p
b
een made.
D
o
uld be rep
r
t system ope
provide for
a
r plants as
can be che
c
s
e a differen
t
than the fi
r
b
e checked
b
be done.
S
monitor pla
n
n
ditions, an
d
.
f
its that can
b
augmentati
o
e
r
gy
l
be very dif
f
d
in variou
s
c
an be certai
n
i
fferent in 2
0
s
te
m
will be
a
y, we woul
d
m
e of the tren
d
e
gration wit
h
r
ical trend.
p
erators to e
n
D
eductions w
o
resentative
o
rators.
redundancy
safe as they
c
ked by an
o
t
approach t
o
r
st one, and
b
y the oper
a
S
ystems ca
n
n
t functions,
d
reduce the
n
b
e mentioned
o
n to the nu
c
f
erent. For al
l
s
reports b
y
n
that the en
0
30 than it
d
transformed
,
d
like to see.
d
s highlight
e
h
Energy Sou
n
sure
o
uld
o
f a
and
are.
o
ther
o
be
any
a
tors
n
be
alter
n
eed
as a
c
lear
l
the
y
an
ergy
d
oes
,
but
We
e
d in
rep
o
co
n
we
l
wh
e
Th
e
an
d
Co
-
ma
r
co
n
inc
r
the
see
m
but
co
u
b
e
c
oil
me
n
all
co
u
to
w
en
e
is
n
po
w
A
b
u
s
ha
v
Co
m
rces (Renew
o
rts on curr
e
n
sumption, e
n
l
l as socially
.
e
n there is st
e
growing w
e
d
other non-
O
-
operation a
n
r
kets and C
O
n
side
r
dealin
g
r
easing dom
i
emergence
m
to be nece
s
not sufficie
n
u
ld temporar
oming incre
a
is over. But
m
n
tion other e
x
government
s
u
ntries alike,
a
w
ards a clea
n
e
rgy system.
T
n
ow. So, wh
a
w
er co
m
pete
?
A
variety of c
o
s
iness say th
e
v
e submitted
n
m
mission’s
s
able and No
n
e
nt global tre
n
n
vironmenta
l
.
But, that c
a
ill time to c
h
e
ight of Chi
n
O
ECD (The
O
n
d Develop
m
O
2
emissions
g
with glob
a
i
nance of na
t
of low-car
b
s
sary solutio
n
n
t. Moreover,
ily cause p
r
a
singly appa
r
m
any of the
x
ternal factor
s
s
’ power, p
r
a
cting alone
o
n
er, smarte
r
T
ime is runni
a
t we need t
o
?
” [5].
o
mpanies that
a
e
answer m
a
n
ew designs
t
s
afety engin
e
n
-renewable)
n
ds in energ
y
l
ly, and eco
n
a
n—and mus
t
h
ange the ro
a
n
a, India, the
O
rganization
f
m
ent) regio
n
is somethin
g
a
l warming.
t
ional oil co
b
on energy
n
s to the pro
b
while mark
e
r
ices to fall
r
ent that the
key policy d
r
s
) remain in
d
r
oducing an
d
or
together, s
t
r
, and more
ng out, and t
h
o
ask is that
“
a
re in the ene
r
a
y be yes.
M
t
o the Nucle
a
e
ers, and tha
t
19
7
y
supply an
d
n
omically, a
s
t
—be altere
d
a
d we are on
.
Middle East
,
f
o
r
Economi
c
n
s in energ
y
g
we need t
o
The rapidl
y
mpanies an
d
technologie
s
b
le
m
at hand
,
e
t imbalance
s
back, it i
s
era of chea
p
r
ivers (not t
o
d
oubt. Withi
n
d
consumin
g
t
ee
r
the worl
d
competitiv
e
h
e time to ac
t
“
Can nuclea
r
r
g
y
productio
n
M
anufacturer
s
ar
Regulator
y
t
agency ha
s
7
d
s
d
.
,
c
y
o
y
d
s
,
s
s
p
o
n
g
d
e
t
r
n
s
y
s
198
Fig. 5 Typi
c
already app
r
are built on
a
manufactur
e
formed par
t
approvals.
S
nuclear en
g
from indus
t
constructio
n
Like anot
h
35-year hia
t
not a sure b
this time b
e
intervening
moon landi
n
worthwhile
?
To answ
e
unresolved
brought up
b
To argu
e
producing
technologie
s
investment,
from the d
a
University
o
3. Fission
By the n
a
Artificial In
t
c
al NPP in our
r
oved some a
s
a
previously
a
e
rs, and archi
t
t
nerships to
S
warms of st
g
ineering pr
o
t
ry and go
v
n
. See Fig. 5.
h
er moon sh
o
t
us in orders
et. The expe
r
e
cause of te
c
decades. But
n
g, there is
a
?
er
this questi
o
problems as
b
y an MIT re
p
e
the first
p
a NPP wit
h
s
from tota
l
we need to
u
a
y it was
b
of
Chicago [6
and Fusio
n
a
ture of the n
u
t
elligence Int
e
backyard.
s
ready for c
o
a
pproved site
t
ecture/engin
build plant
udents enrol
l
o
grams, and
v
ernment pr
e
o
t, new react
o
is certainly
r
ts say that i
t
c
hnological
p
as with a pr
o
a
nother ques
t
o
n, we need
t
sociated wit
h
p
ort [6].
p
oint, which
h
its mod
e
l
ownership
u
nderstand t
h
b
orn in the
b
].
n
Nuclear
P
u
clea
r
proce
s
e
gration wit
h
o
nstruction if
. Utilities, re
a
eering firms
h
s, pending
f
l
in college-
l
rosy projec
t
e
dict a surg
e
o
rs’ launch a
f
possible, th
o
t
would be e
a
p
rogress ove
r
o
ject as large
t
ion: Would
i
t
o satisfy the
h
nuclear p
o
is the cos
e
rn and to
d
and retur
n
h
e beast’s n
a
b
asement o
f
P
ower Ene
r
s
s, we are ab
l
h
Energy Sou
they
a
cto
r
h
ave
f
inal
l
evel
t
ions
e
in
f
ter a
o
ugh
a
sier
r
the
as a
i
t be
four
o
we
r
t of
d
ay’s
n
on
a
ture
f
the
rgy
l
e to
ge
n
pro
(
(
2
E
co
n
rea
c
(R
e
hol
i
the
3.1
F
tec
h
Nu
c
fiss
Un
i
I
n
nu
c
too
l
of
e
as
nu
c
rel
e
ele
m
abs
rad
i
the
sh
o
unt
i
O
rea
c
ge
n
co
m
Re
a
ca
n
su
b
Ru
s
T
L
W
en
e
rces (Renew
n
erate two ty
p
cess, and the
y
1) Fission Pr
o
2
) Fusion Pr
o
E
ach of t
h
n
sideration, a
n
c
tors are
a
e
search and
i
stically hav
e
types of rea
c
Fission
D
ri
v
F
ission ener
g
h
nology. We
c
lea
r
Fission
ion reactor
i
i
versity of C
h
n
short, nucl
e
c
lea
r
fusion p
l
s. Fission p
o
e
nergy by spl
i
U-235, Pu-
2
c
lea
r
fission i
s
e
ased, causin
g
m
ents if t
h
orbing mate
r
i
oactive and
produced
w
uld make su
r
i
l a million y
e
O
ve
r
time,
s
c
tors have
g
n
eration as d
e
m
mercially,
i
a
cto
r
) and P
W
n
be found
marines, par
t
s
sians and re
c
T
oday the
G
W
Rs are prod
u
e
rgy [7]; how
e
able and No
n
p
es of ener
g
y
are:
o
cess. Drivi
n
o
cess. Drivin
g
h
ese proce
s
n
d in the for
m
a
vailable
be
Developme
n
e
described e
a
c
tors involve
d
v
en Nuclear
E
g
y is not a
n
have come
t
Reaction sin
c
i
n the
b
ase
m
h
icago aroun
d
e
a
r
fission p
o
lants are pro
d
o
we
r
plants p
r
i
tting heavy
n
2
39. From
a
s
a chain reac
g
splitting of
n
h
ey are not
r
ial. Fission
f
have a high
w
aste from
t
r
e about the s
e
ars.
s
ince Fermi
’
g
one throug
h
e
picted in Fi
g
i
n the form
W
R (Pressu
r
in most m
o
t
icularl
y
in
U
c
ently Chine
s
G
eneration
T
u
cing electri
e
ver, these g
e
n
-renewable)
g
y that are d
r
n
g Nuclear E
n
g
Nuclear E
n
s
ses is he
a
m
of fission,
c
e
yond labo
r
n
t). In this
a
ch of these
p
d
.
E
nerg
y
n
ything ne
w
t
o learn abo
u
c
e Fermi inv
e
m
ent of his
o
d
1942 [7].
o
we
r
plants (
r
d
ucing com
m
r
oduce an ex
t
n
uclei eleme
n
a
physics po
c
tion in whic
h
n
ew lighter n
u
controlled
f
uel and its p
r
half-life. W
h
t
hese power
afety of the
d
’
s invention
,
h
so much
r
g
. 1. Today,
of LWR (
r
ized Water
R
o
dern naval
U
SN (United
S
s
e naval forc
e
T
hree (GE
N
city as a so
u
e
nerations ar
e
r
iven by thi
s
n
ergy
n
ergy
a
vily unde
r
c
ommerciall
y
r
atory R&
D
section, w
e
p
rocesses an
d
w
in today’
s
u
t Controlle
d
e
nted his firs
t
o
ffice at th
e
r
eactors) an
d
m
ercial powe
r
reme amoun
t
n
t cores, suc
h
int of view
,
h
neutrons ar
e
u
clei of othe
r
by neutro
n
r
oduct are s
o
h
ile dumpin
g
plants, on
e
d
umping are
a
,
the fissio
n
r
evision an
d
they operat
e
Light Wate
r
R
eacto
r
) tha
t
vessels an
d
S
tates Navy)
,
e
s.
N
-III) of th
e
u
rce of clea
n
e
at the end o
f
s
r
y
D
e
d
s
d
t
e
d
r
t
h
,
e
r
n
o
g
e
a
n
d
e
r
t
d
,
e
n
f
Artificial Intelligence Integration with Energy Sources (Renewable and Non-renewable)
199
their life cycles, and recent accidents such as
Fukushima Daiichi Nuclear Disaster have given a
different perspective to this type of source energy
driving electricity.
Lesson learned from accidents such as Fukushima
has forced us to have a different way of thinking and to
design the new generation, where the safety is number
one priority and modularization of this generation
known as GEN-IV is the front runner for the new
fission reactor, where the modularization helps to
reduce the foot-print of this generation to smaller
real-estate. We have also been looking for an even
smaller and smaller version known as micro nuclear
reactor for space and military applications [8].
The advanced version of these new generations of
four operating at a higher temperature (i.e., above
800 °C) is an excellent candidate to be integrated with
a combined cycle of top or bottoming of open Air
Brayton and Rankine cycles. With these configurations,
the thermal output efficiency of these reactors makes
them more cost-effective and provides a better TCO
(total cost of ownership) and, consequently, an
excellent ROI (return on investment) regarding
renewable sources of energy [3, 9, 10].
Most six common suggested technical versions of
these reactors are listed below as [3, 7]:
(1) VHTR (very high-temperature reactor),
(2) MSR,
(3) SFR (sodium-cooled fast reactor),
(4) SCWR (super critical water-cooled reactor),
(5) GFR (gas-cooled fast reactor), and,
(6) LFR (lead-cooled fast reactor).
Experts are projecting worldwide; electricity
consumption will increase substantially in the coming
decades, especially in developed countries with very
high GDP (gross domestic product) around the world,
accompanying economic growth and social progress
that directly impact rising electricity prices. These
countries have focused fresh attention on NPPs. New,
safer, and more economical nuclear reactors could not
only satisfy many of our future energy needs but could
combat global warming as well. Today’s existing NPPs
online in the United States provide a fifth of the
nation’s total electrical output [7].
Taking into account the expected increase in energy
demand worldwide and the growing awareness about
global warming, climate change issues and sustainable
development, nuclear energy will be needed to meet
future global energy demand [7].
Fission NPP technology has evolved as distinct
design generations as we mentioned in the previous
section and briefly summarized here again as follows
(i.e., Fig. 1):
First generation: prototypes, and first realizations
(~1950-1970).
Second generation: current operating plants
(~1970-2030).
Third generation: deployable improvements to
current reactors (~2000 and on).
Fourth generation: advanced and new reactor
systems (2030 and beyond).
The Generation IV International Forum, or GIF, was
chartered in July 2001 to lead the world’s leading
nuclear technology by nations’ collaborative efforts to
develop next-generation nuclear energy systems to
meet the world’s future energy needs.
Eight technology goals have been defined for
Generation IV systems in four broad areas:
(1) Sustainability,
(2) Economics,
(3) Safety and reliability, and finally,
(4) Proliferation resistance and physical protection.
Many countries share these ambitious goals as they
aim to respond to the twenty-first century’s economic,
environmental, and social requirements. They
establish a framework and identify concrete targets for
focusing on GIF Research and Development (R&D)
efforts.
As we stated above, eight technology goals have
been defined for Generation IV systems in four broad
areas: sustainability, economics, safety and reliability,
proliferation resistance, and physical protection [7].
200
3.2 Fusion
D
Since th
e
matter, whe
r
toward ther
m
as a new so
u
it can be ca
t
thus control
and thus, pr
o
future dema
n
Note that
Liquid, Vap
where the
v
Plasma.
What is
g
sun and all s
interaction
t
what we ar
e
version of
producing c
l
However
,
devices on
achievemen
t
The poin
t
the energy
equals the
maintain th
e
The ratio
breakeven c
is needed fo
practical re
a
(i.e. Lawso
n
The theo
r
the nuclear
f
of light ato
m
helium. As
hydrogen, n
a
fused toget
h
energy, and
around us,
sources by
water.
Artificial In
t
D
riven Nucle
a
e
discovery
o
r
e the gas is
m
onuclea
r
fu
s
u
rce of clean
e
t
egorized as
a
ling thermo
n
o
ducing elec
t
n
d for energ
y
: Matter in
n
o
r
, and final
l
v
apo
r
starts
g
g
oing on, for
tars within t
h
t
aking place
e
trying to
a
this interac
t
l
ean energy
w
,
to achieve t
h
earth as a g
o
t
that falls u
n
t
reaching
b
r
e
being relea
s
amount of
e
plasma reac
of input to o
u
orresponds t
o
r the reactor
t
a
sons, it is de
s
n
Criteria).
r
y of fusion
fu
f
usion conce
p
m
ic nuclei t
o
an example,
a
mely deute
r
h
er to produc
e
so long as
w
then we ha
v
extracting
bo
t
elligence Int
e
ar
Energy
of
Plasma as
in ionized f
o
s
ion power h
a
e
nergy, and t
h
a
renewable
s
n
uclear reacti
o
t
ricity to me
e
y
driven elec
t
n
ature can
be
l
y, if the tem
p
g
etting to be
example, at
t
h
e Galaxy is
h
to generate
a
chieve—a c
o
t
ion on ear
t
w
ith the unli
m
h
ese controll
e
o
al, it woul
d
n
de
r
Lawson
C
e
akeven is
ba
s
ed via fusi
energy
b
ei
n
tion.
u
tput energy
i
o
a Q of 1. A
t
o generate n
s
irable for it t
fu
ndamentall
y
p
t that relies
o
form heav
we conside
r
r
ium (D) and
e
helium (He)
w
e have acce
s
v
e access t
o
o
th D and
T
e
gration wit
h
the 4th sta
t
o
rm, our atte
n
a
s been expa
n
h
at is why pe
rh
s
ource of en
e
o
n driven en
e
t our presen
t
t
ricity [4].
e
found as S
o
p
erature incr
e
ionized is c
a
t
he surface o
f
h
ot Plasma (F
i
heat, and th
a
o
ntrolled pr
o
t
h, as mean
s
m
ited source.
e
d fusion rea
c
d
be a
b
reak
e
C
riteria [11-
1
a
sically calle
d
on reaction
n
g consume
d
i
s denoted Q
,
Q of at leas
t
et energy, b
u
o be much hi
g
y
works base
d
on two ele
m
ier ones suc
h
r
two isotop
e
tritium (T),
t
, producing
c
s
s to ocean
w
o
such unli
m
T
from the o
c
h
Energy Sou
t
e of
n
tion
n
ded
rh
aps
e
rgy,
e
r
gy
t
and
o
lid,
e
ases
a
lled
f
the
i
g. 6)
a
t is
o
cess
s
of
c
tion
e
ven
1
3].
d
for
that
d
to
,
and
t
one
u
t for
g
her
d
on
m
ents
h
as
e
s of
t
o be
c
lean
w
ate
r
m
ited
c
ean
Fig
.
A
pre
s
sci
e
are
:
(
(
2
A
ma
n
at
en
g
co
m
I
n
its
a
IC
F
pla
c
rea
c
ma
t
en
c
gra
n
ap
p
I
n
ele
c
thr
o
Lo
r
lin
e
A
tor
o
pla
s
an
d
rces (Renew
.
6 Fusion in
t
A
s far as fusi
o
s
ently, two t
y
e
ntists and e
n
:
1) MCF (ma
g
2
) ICF (inert
i
A
mong these
n
y different s
u
the laborat
o
g
ineers are e
x
m
mercial sta
g
n
this part o
f
a
ssociated R
&
F
and the ki
n
c
e to demo
n
c
tors in the
t
ters are ap
ourage thos
e
n
ula
r
inform
a
p
ropriate refe
r
n
the case o
f
c
trical cond
u
o
ugh interac
t
r
entz force a
n
e
as illustrate
d
A
s Fig. 7a ill
u
o
idal field a
n
s
ma current (
R
d
the resulting
able and No
n
t
eraction of pl
a
o
n driven nu
c
y
pes of fusio
n
n
gineers, ex
p
g
netic confin
i
al confinem
e
two types
o
u
ggested nu
c
o
ry level t
h
x
periencing
a
g
e.
f
this section,
&
D reactors,
a
n
d of researc
h
n
strate the c
o
near-future
proached h
e
e
readers w
h
a
tion to refe
r
r
ences at the
f
MCF, we a
r
u
ctivity of t
h
t
ion with ma
n
d follow a
h
d
in Fig. 7a.
u
strates, one
n
d the coils (
B
R
ed) and the
twisted field
w
n
-renewable)
a
sma taking o
n
c
lea
r
energy
i
n
are under
s
p
ert in this fi
e
ement fusio
n
e
nt fusion).
of
confineme
n
c
lea
r
power f
u
h
at these s
c
a
nd trying t
o
,
we define f
i
a
nd then, we
t
h
, which ha
s
o
mmercializ
a
time frame
.
e
re holistica
l
h
o are intere
s
r
themselves
end [11-13].
r
e taking adv
h
e Plasma
t
gnetic fields
h
elical path a
l
can take ad
v
B
lue) that p
r
poloidal fiel
d
w
hen these a
r
n
sun surface.
i
s concerned
,
s
tudies by th
e
e
ld, and the
y
n
).
n
ts, there ar
e
u
sion reactor
s
c
ientists an
d
o
push to th
e
i
rst MCF an
d
t
ake a look a
t
s
been takin
g
a
tion of suc
h
.
Both thes
e
l
ly, and w
e
s
ted in mor
e
to one of th
e
antage of th
e
t
o contain i
t
by applyin
g
l
ong the fiel
d
antage of th
e
r
oduce it, th
e
d
created by i
t
r
e overlaid.
,
e
y
e
s
d
e
d
t
g
h
e
e
e
e
e
t
g
d
e
e
t
,
Artificial Intelligence Integration with Energy Sources (Renewable and Non-renewable)
201
Fig. 7(a) Magnetic helical path illustration.
Source: www.wkipedia.org.
Fig. 7(b) Concept of a toroidal fusion reactor.
Source: www.wkipedia.org.
Fig. 8 Magnetic fields in Tokamak.
Source: www.wkipedia.org.
Fig. 9 Example of a Stellarator design.
Source: www.wkipedia.org.
Fig. 10 Hot plasma, magnetically confined in a Tokamak
reactor.
Source: www.wkipedia.org.
A typical toroidal machine concept, also illustrated
in Fig. 7b, is a pinch-effect to drive the control fusion
reactor [11, 12].
Tokamak (i.e. a Russian word, means donut shape)
reactor and Stellarator were the two devices that
offered such solution as illustrated in Figs. 8 and 9.
As we mentioned, the Tokamak solution was one of
the first devices suggested for MCF (magnetic
confinement fusion) driven fusion in a controllable
202
fashion. Fi
g
where the
Experiment
a
France is
m
engineering
confinemen
t
Fig. 11
cut-away o
f
ICF (iner
t
of using a
h
laser) that i
n
by heating
deuterium (
D
that these p
e
of hydroge
n
In this fi
g
orange is
b
thermal ene
r
In this
p
adiabaticall
y
form in ord
e
we can reac
h
pellet fuel
o
(Raleigh T
a
threshold o
frequency o
f
incoming
s
compress t
h
because the
the induced
like a mirro
r
The NIF
LLNL (La
w
Livermore,
confinemen
t
(deuterium-
t
in Fig. 13.
Inertial c
o
density and
beam outsi
d
droplet of p
e
Artificial In
t
g
. 10 is a
d
ITER (Inte
a
l Reactors)
m
eg
a
intern
which wi
l
t
in the worl
d
shows mor
e
f
ITER.
t
ial confine
m
h
eat source
s
n
itiates a con
t
and comp
D
) and tritiu
m
e
llets most o
ft
n
as fuel as d
e
g
ure, the
b
l
u
b
lowoff; pu
r
r
gy.
p
rocess, lase
r
y
and as muc
h
er
to overco
m
h
ignition te
m
of
D and T
a
ylo
r
) instabi
f
plasma fr
e
f
laser-drive
n
s
hining lase
r
h
e pellet at
incident las
e
Plasma at th
e
r
[11, 13].
(National Ig
n
w
rence Liver
m
California,
t
fusion, us
t
ritiu
m
) fuel
i
o
nfinement f
u
heating of
d
e the react
o
e
llet occurs.
t
elligence Int
e
d
epiction of
rnational T
h
Tokamak
i
ational rese
a
l
l be the
l
d
(Fig. 10).
e
detailed i
n
m
ent fusion) i
s
s
uch as an
H
t
rolled nucle
a
ressing a
p
m
(T) to ign
i
ft
en contain t
h
e
picted in Fig
u
e arrows re
p
r
ple is inw
a
r
compresse
s
h
as possible
m
e instability
c
m
perature wi
t
before we r
e
lit
y
or even
e
quency to
n
fusion. At t
h
r
no longer
the ablation
er
beam will
b
e
surface of
t
n
ition Facili
t
m
ore Nation
a
is based
o
ing a laser
i
n a capsule
(
u
sion happen
s
thermonucle
a
or
chamber
s
e
gration wit
h
such illustra
t
h
ermonuclea
r
i
n the sout
h
a
rch and f
u
l
argest mag
n
n
ternal view
s
built on the
H
PL (high-p
o
ar
fusion rea
c
p
ellet of
m
i
tion temper
a
h
ese two isot
o
. 12 [13].
p
resent radia
t
a
rdly transp
o
s
the pellet
in a symmet
r
c
ircumstance
t
hin the coro
n
e
ach or pass
we reach t
o
be equal to
h
is limitatio
n
can push
surface, si
m
b
ounce
b
ack,
t
he pellet wi
l
t
y) experime
n
a
l Laborator
y
o
n laser in
e
to confine
(
target) as s
h
s
according t
o
ar
fuel by d
s
ystem, whe
h
Energy Sou
t
ion,
r
of
h
of
u
sion
n
etic
and
idea
o
wer
c
tion
m
ixed
a
ture
o
pes
t
ion:
o
rted
fuel
r
ical
s, so
n
a of
RT
o
the
the
n
, the
and
m
ply
and
l
l act
n
t at
y
) in
e
rtial
DT
h
own
o
the
ense
re a
Fig
.
Fig
.
usi
n
Sou
r
Fig
.
Fig
.
rces (Renew
.
11 ITER cu
t
.
12 Scheme
o
n
g lasers.
r
ce: www.wki
p
.
13 NIF faci
l
.
14 Fusion i
n
able and No
n
t
-away illustr
a
o
f the stages o
f
p
edia.org.
l
ity at LLNL.
n
tegrated with
n
-renewable)
a
tion.
f
inertial conf
i
HPC and AI.
i
nement fusio
n
n
Artificial Intelligence Integration with Energy Sources (Renewable and Non-renewable)
203
In conclusion, AI, in conjunction with HPC
(high-performance computing) is approaching points
to bright for fusion energy in the near future.
Due to the nature of experimental approaches and
cooperative data as well as images that are required to
be analyzed by scientists and engineers involved with
new Generation IV (GEN-IV) in case of fission process
driving energy production and fusion in case of both
MCF and ICF deal also a lot of data analytics and
image processing analysis, thus usage of a
super-computer of HPC type becomes a mandatory
requirement. See Fig. 14.
Researchers are using DL techniques on DOE
super-computers to help develop fusion energy.
As we stated above, to ensure Plasma—the fourth
fundamental state of matter—retains its heat and does
not interact with materials in the containment vessel,
researchers employ doughnut-shaped fusion devices
called Tokamaks, which use magnetic fields to trap
fusion reactions in place. However, large- or exa-scale
plasma instabilities called disruptions can interfere
with this process.
Science simulation, visualization, data, and learning
applications will greatly benefit from the massive
computational resources available with future exascale
systems. Researchers in the ALCF’s (Argonne
Leadership Computing Facility’s) Aurora ESP (Early
Science Program) are blazing the trail toward reaping
those benefits from the US DOE’s Argonne National
Laboratory’s upcoming Aurora exascale
super-computer.
4. Nanotechnology and Energy Storage
Nanotechnology creates and manipulates matter at
the molecular level, making it possible to create
materials with improved properties, such as being
lightweight and having ultrahigh strength and greater
capabilities such as electrical and heat conductivity.
Many applications are possible for the energy industry.
The oil industry already uses nanoscale catalysts for
refining petroleum. Nanoparticles with unique catalytic
capabilities are being researched to refine thick, gooey
oil sands into highly refined oil more effectively and
efficiently.
Nanotechnology may be a promising solution for the
transmission and storage of energy, particularly
electrical power and hydrogen. Nano-based materials
may create new opportunities to transport electricity
efficiently and at a lower cost over very long distance.
In this unique and yet very new industry, AI also
plays a significant role in enhancing performance and
applying it, even in the nuclear industry.
Furthermore, Aurora and the next generation of
Exascale super-computers will apply HPC and AI
technologies to even cancer research and climate.
In the end, science at the exascale level is the future
of both fission and fusion of today and the future
technology of these two processes.
5. What Are AI, ML and DL?
The past decade up to now has encountered a new
revolutionary technology that seems to have many
applications across the entire industry (Fig. 4). This
innovative technology is called AI that has been
driving business intelligence to a different level,
considering any business operation with a magnitude
of incoming data to be analyzed. These business
day-to-day operations with a share volume of data (i.e.,
Big Data) require augmentation of AI in conjunction
with HPC.
Even the energy sector, both renewable and
non-renewable sources, is in need of AI in need of their
data analytics and data predictive [14], respectfully. This
section briefly defines what the AI is and what other
components are involved with AI system to make a
business operational in a resilience model—a good
BRS (business resilience system) [15].
In a very holistic way, AI by today’s definition is
known as narrow AI (or weak AI) [16].
This kind of AI (Fig. 15) is designed to perform
narrow/simple tasks such as facial recognition, internet
searches, or driving a car in an autonomous mode.
204
To reca
p
machines,
u
humans and
In other
w
market of
machines a
s
ultimate go
a
super AI.
Such pr
o
definition i
s
program, pa
r
think and le
distinguishe
think logic
a
Sapiens or
“
“Man the
M
With this
key factors
o
It is es
s
and differe
n
comes to de
v
Witho
u
the scope o
arise, and e
x
In fact,
can cause a
m
discuss and
Different
One of
t
capabilities
from exper
i
human-like
today—fro
m
cars—rely
h
processing)
can be tr
a
processing
patterns in t
h
AI will
p
technology
t
essential.
Artificial In
t
p
, AI is i
n
u
nlike the nat
u
animals.
w
ords, AI th
a
technology,
s
smart and
a
l, to the poin
t
o
gression w
i
s
the ability
r
ticularly in t
arn very si
m
d points abo
u
a
lly and fab
r
“
Wise Man”
M
ake
r
”).
basic unders
t
o
ne should k
n
s
ential to dis
n
t phases of
v
eloping app
u
t recognizin
g
f
the related
x
pectations
m
the “broad”
d
m
isrepresent
a
develop tod
a
types of AI
a
t
he signific
a
that make it
p
i
ence, adjust
tasks. Most
m
chess-play
i
h
eavily on D
L
[16]. Using
t
a
ined to ac
c
large amou
n
h
e data.
p
lay in our
t
hat we enco
u
t
elligence Int
e
n
telligence
d
u
ral intellige
a
t is the new
is the sci
e
intelligent l
i
t
that we go
fr
i
thin the do
of a comp
u
he case of H
P
m
ila
r
to the h
u
u
t us as huma
n
r
icate physi
c
in Latin and
t
anding of A
I
n
ow about A
I
tinguish diff
e
the evolutio
lication prog
r
g
the differe
n
application
s
m
ay be far fro
m
d
efinition of
A
a
tion of the t
y
ay
a
re illustrate
d
a
nt advantag
e
p
ossible for
m
to new inp
u
AI example
s
i
ng compute
r
L
and NLP
(
t
hese techno
l
c
omplish s
p
n
ts of data
fast-paced l
u
nter in our
d
e
gration wit
h
d
emonstrated
nce displaye
d
buzzword o
f
e
nce of ma
k
i
ke human a
fr
o
m
a weak
A
main of A
I
u
te
r
algorith
m
P
C or machi
n
u
man being.
T
n
, are that w
e
c
ally (i.e., H
o
Homo Fabi
a
I
, there are ce
r
I
:
e
rent types o
n of AI wh
e
r
ams.
n
t types of A
I
s
, confusion
m
reality.
A
I is “vague
”
y
pe of AI th
a
d
in Fig. 16.
e
s of AI is
m
achines to
l
u
ts, and per
f
s
you hear a
b
r
s to sel
f
-dri
v
(
natural lang
u
l
ogies, comp
u
p
ecific tasks
and recogn
i
ife, and mo
d
d
ay-to-day li
h
Energy Sou
by
d
by
f
the
k
ing
s an
A
I to
I
by
m
or
n
e, to
T
wo
e
can
o
mo
a
n or
r
tain
f
AI
e
n it
I
and
may
”
and
a
t we
the
l
earn
f
orm
b
out
v
ing
u
age
u
ters
by
i
zing
d
ern
fe is
Fig
.
Sou
r
Fig
.
F
wi
d
wit
h
inf
o
res
u
Th
e
ge
n
hu
m
ma
k
T
int
o
un
d
ind
u
M
b
ui
l
ex
p
fro
m
sta
t
as
p
M
L
alg
o
b
as
rces (Renew
.
15 An AI in
r
ce: www.wiki
p
.
16 Differen
t
F
urthermore,
d
e range of s
m
h
humans [1
7
o
rmation, an
a
u
lting insigh
t
e
se days, AI
n
erated by
h
m
ans’ ability
k
e complex
d
T
o understan
d
o
the various
d
erstand how
u
stry’s vario
u
M
L is the
b
ra
n
l
d computer
s
p
erience. Ind
e
m
the data. It
t
istics, AI, a
n
p
redictive a
n
L
’s main id
e
o
rithms that
ed on them.
R
able and No
n
ception.
p
edia.org.
t
types of AI.
integrating
A
m
art tools th
e
7
], enables p
e
a
lyze the da
t
t
to have a
is essential
h
umans and
to absorb a
n
d
ecisions
b
as
e
d
how AI wo
r
sub-domain
those domai
n
u
s fields.
n
ch of AI th
a
s
that auto
m
e
ed, ML is
is a research
n
d computer
s
n
alytics or st
a
e
a is that
i
learn from
d
R
ecent progr
e
n
-renewable)
A
I within tec
e
se days, wh
i
e
ople to rethi
n
t
a, and finall
y
better-infor
m
since the a
m
machines
f
n
d interpret
e
d on that da
t
r
ks, one need
s
s of AI (i.e.
,
n
s could be
a
a
t holisticall
y
m
atically imp
r
all about th
e
field at the i
n
s
cience and i
s
a
tistical lear
n
i
t is possib
l
d
ata and mak
e
ss in ML ha
s
hnology as
a
i
le partnerin
g
n
k and gathe
r
y
, utilize th
e
m
ed decision
.
m
ount of dat
a
f
a
r
outpace
s
the data an
d
t
a.
s
to deep div
e
,
Fig. 3) an
d
a
pplied to th
e
y
addresses t
o
r
ove throug
h
e
knowledg
e
n
tersection o
f
s
also know
n
n
ing. Indeed
,
l
e to creat
e
e prediction
s
s
been drive
n
a
g
r
e
.
a
s
d
e
d
e
o
h
e
f
n
,
e
s
n
Artificial Intelligence Integration with Energy Sources (Renewable and Non-renewable)
205
by developing new learning algorithms and theory and
the ongoing explosion in online data availability and
low-cost computation.
With big data growth, ML has become a significant
and key technique in solving problems. ML finds the
natural pattern in data that generates insight to help
make better decisions and predictions. It is an integral
part of many commercial applications ranging from
medical diagnosis, stock trading, energy forecasting,
and many more.
Consider the situation when we have a complicated
task or problem involving a large amount of data with
lots of variables but with no existing formula or
equation. ML is part of a new employment dynamic,
creating jobs that center around analytical work
augmented by AI.
ML provides smart alternatives to analyzing vast
volumes of data. ML can produce accurate results
and analysis by developing fast and efficient
algorithms and data-driven models for real-time data
processing.
DL is the subset of ML that, on the other hand, is the
subset of AI. DL is inspired by the structure of the
human’s brain. DL algorithms attempt to draw similar
conclusions as humans would by continually analyzing
data with a given logical structure. To achieve this, DL
uses a multi-layered structure of algorithms called
neural networks. Just as humans use their brains to
identify the patterns and classify the different types of
information, neural networks can be taught to perform
the same data tasks.
Whenever humans receive new information, the
brain tries to compare it with known objects. The same
concept is also used by deep neural networks. By using
the neural network, we can group or sort the unlabeled
data based on similarities among the samples in the
data. Artificial neural networks have unique
capabilities that enable DL models to solve tasks that
ML models can never solve.
One of the main advantages of DL lies in solving
complex problems that require discovering hidden
patterns in the data and/or a deep understanding of
intricate relationships between a large number of
interdependent variables. When there is a lack of
domain understanding for feature introspection, DL
techniques outshine others, as you have to worry less
about feature engineering. DL shines when it comes to
complex problems such as image classification, NLP,
and speech recognition.
6. The Role of AI in Improving the
Renewable Energy Sector
With recent movement toward decarbonization and
reduction of CO
2
on the earth, demands for new and
clean energy production sources drive electricity
production, including energy storage for on- and
off-line support of grid and electricity network. As part
of clean energy producing electricity, any source of
renewable energy source looks very appealing, and as
we all know, one of these sources happens to be solar
power for the time being.
But how can AI help in improving renewable energy
supply? Moreover, how can it be integrated into the
energy sector? We will discuss this matter here in this
section.
Global energy demands are growing every year. And
fossil fuels will not be able to fulfill our energy needs in
the future. Carbon emissions from fossil fuels have
already hit an all-time high in 2018 due to increased
energy consumption.
On the other hand, renewable energy is emerging out
as a reliable alternative to fossil fuels. It is much safer
and cleaner than conventional sources. With the
advancements in technology, the renewable energy
sector has made significant progress in the last decade.
However, there are still a few challenges in this
sector that can be addressed with the help of emerging
technologies.
Technologies like AI and ML/DL can analyze the
past, including historical data, optimize the present and
incoming data, and predict/forecast the future. AI in the
renewable energy sector can resolve most of the
Artificial Intelligence Integration with Energy Sources (Renewable and Non-renewable)
206
challenges because we are dealing with massive
incoming data from all aspects and types of renewable
sources of energy, including nuclear energy as a clean
source, in the form of fission presently and fusion in the
near future [4].
One of the significant challenges of producing
renewable energy is the unpredictability of the weather.
Solar and wind are the leading sources of renewable
energy, and the power generation largely depends on
the weather.
Although we have efficient technologies in place for
weather forecasting, there are going to be sudden
changes in the climate that can affect the energy flow.
The supply chain of renewable energy is prone to such
vulnerabilities. Therefore, it needs to be smoothened
enough to cope up with unexpected changes.
Secondly, the recent developments in energy storage
technology are quite promising. But they are yet to be
tested thoroughly.
The demand for renewable energy will only increase
in the future. And that is why renewable energy
companies should invest in ML, AI, the IoT (Internet of
Things), and other emerging technologies to improve
productivity and overcome the shortfalls.
Note that: The IoT is a means to connect devices and
appliances to the internet. This is the technology that
enables you, for example, refrigerator to order more
milk, when you have a (key performance indicator) in
place with some smartness of SLA (service level
agreement) built into it and this, where AI, ML, and DL
play the rule to do the ordering. An excellent example
of such an appliance in the present market that acts as
semi-supervised AI is the Alexa device of
Amazone.com [4].
Even the large consumers of renewable energy, like
supermarkets, factories, offices, and railways, can use
AI technology to make data-driven decisions.
Now the question is how artificial energy (AI)
technology, with its two other components, namely
ML and DL [8] can improve the renewable energy
sector.
We may be able to answer the question with the
following debate.
AI is taking on many new roles in
society—becoming our coworker or our partner,
serving as a virtual assistant in our homes, operating
our cars, and more [3].
The electric grid is one of the complex machines on
earth. However, it is evolving rapidly with the addition
of variable renewable energy sources.
Due to the inherent variability of wind and solar, the
current grid faces many challenges in accommodating
renewable energy diversity.
The utility industry needs smart systems that can
help improve renewables’ integration into the existing
grid and make renewable energy an equal player in the
energy supply.
Here is, what we could holistically state that, how AI
technology can improve the reliability of renewable
energy and modernize the overall grid [4, 8].
6.1 Smart, Centralized Control Centers
The energy grid can be interconnected with devices
and sensors to collect a large amount of data.
When coupled with AI, the data can give the grid
operators new insights for better control operations. It
offers flexibility to the energy suppliers to cleverly
adjust the supply with demand.
6.2 Improved Integration of Microgrids
AI can help with the integration of microgrids and
managing distributed energy. When the
community-level renewable energy generation units
are added to the primary grid, it becomes hard to
balance the grid’s energy flow.
The AI-powered control system can play a vital role
in solving the quality and congestion issues.
6.3 Improved Safety and Reliability
While the biggest goal of AI in renewable energy is
to manage the intermittency, it can also offer improved
safety, efficiency, and reliability.
It can he
l
patterns an
d
health.
For exam
p
collect the
d
wear and t
e
health of th
e
maintenanc
e
6.4 Expand
The inte
g
suppliers e
x
service mo
d
The AI-p
o
data related
on energy c
o
The data
w
services an
d
help retail s
u
6.5 Smart
G
The inte
g
storage) ca
n
to the rene
w
This sma
r
of data coll
e
decisions o
n
This will
the local e
n
exchange w
i
6.6 Cyber-
A
The elect
target for
t
other types
o
elements of
ICS risk e
q
personnel, a
Modern
n
malware ga
i
deceiving
Artificial In
t
l
p you under
s
d
identify th
e
p
le, the AI-p
o
d
ata from wi
n
e
ar. The sys
t
e
equipment
a
e
is needed.
the Market
g
ration of A
I
x
pand the m
a
d
els and enco
u
o
wered syste
m
to energy c
o
o
nsumption.
w
ould help s
d
launch ne
w
u
ppliers to ta
r
G
rid with
I
nte
l
g
ration of AI
n
provide a s
u
w
able energy
i
r
t grid will
be
e
cted from se
v
n
energy allo
c
also help mi
c
n
ergy needs
i
th the prima
r
ttac
k
Preven
t
ric grid has
t
he nation-st
a
of
attack, an
d
the electric
g
q
uipment da
m
nd environm
e
n
etwork att
a
i
ning a footh
o
an emplo
y
t
elligence Int
e
s
tand the en
e
e
devices’ en
e
o
wered predi
c
n
d turbine s
e
t
e
m
will mo
a
nd alert the
o
I
can help r
e
a
rketplace by
u
raging high
e
m
s will be a
b
o
llection and
uppliers opti
m
w
service m
o
r
get new con
l
ligence Stor
a
with IES (i
n
u
stainable an
d
i
ndustry.
e
able to anal
y
v
eral sensors
c
ation.
c
rogrids to e
f
while conti
n
r
y grid.
t
ion
been identif
i
a
te, terrorist
,
d
power plant
s
g
rid. Sophis
t
m
age, and
b
e
ntal damag
e
a
cks begin
w
o
ld on a corp
o
y
ee into
d
e
gration wit
h
e
rgy consum
p
e
rgy leakage
c
tive analysi
s
e
nsors to mo
n
nito
r
the ov
o
perato
r
whe
n
e
newable en
introducing
er
participati
o
b
le to analyz
e
provide ins
i
m
ize the exi
s
o
dels. It can
sume
r
mark
e
a
ge
n
telligent en
d
reliable sol
u
y
ze a vast am
o
and make ti
m
f
ficiently ma
n
n
uing the p
o
i
ed as a stra
t
,
hacktivist,
s
remain esse
n
t
icated attac
k
b
ring injuri
e
e
.
w
ith a piec
e
o
rate networ
k
d
ownloading
h
Energy Sou
p
tion
and
s
can
n
ito
r
erall
n
the
ergy
new
o
n.
e
the
i
ghts
s
ting
also
e
ts.
ergy
u
tion
o
unt
m
ely
n
age
o
we
r
t
egic
and
n
tial
k
s on
e
s to
e
of
k
and
an
att
a
co
n
att
a
sel
e
On
c
att
a
eit
h
or
e
M
sof
t
int
r
sec
u
ma
n
T
ma
l
an
d
eff
e
7.
C
T
life
wo
r
thr
o
gro
w
inc
r
en
v
iss
u
are
so
m
per
s
S
ke
y
thr
o
7.1
Sta
n
B
rces (Renew
a
chment. Th
e
n
nection to a
a
cke
r
uses t
h
e
ct additiona
l
c
e deep eno
u
a
ckers ultim
a
h
e
r
stealing i
n
e
ven damagi
n
M
odern, sop
h
t
ware protec
t
r
usion dete
c
u
rity updat
e
n
agement sy
s
T
o recap, th
e
l
ware from h
a
d
by curling
e
ct on our ne
t
C
onclusio
n
T
hink about
h
and how th
a
r
ldwide. Glo
b
o
ugh 2040,
w
ing prospe
r
r
easing popu
l
v
ironmental s
u
es the world
often extre
m
m
etimes emo
t
s
pectives on
m
S
ome of the r
o
y
trends that
w
o
ugh 2040 ar
e
Energy
P
o
w
n
dards
B
y 2030, th
e
able and No
n
e
malware t
y
command a
n
h
is remote c
o
l
machines t
h
u
gh into thei
a
tely launch
n
formation, s
h
n
g equipmen
t
h
isticated at
t
t
ions, includ
i
c
tion syste
m
e
programs,
s
tems.
e
cyber-sec
u
a
ving any res
i
in it, smart
t
.
n
s
h
ow access t
o
a
t translates
t
b
al energy de
reflecting
i
r
ity and
b
ett
e
l
ation world
w
ecurity are t
w
will face in t
h
m
ely compl
e
t
ionally lade
n
m
any proble
m
o
les that eve
r
w
ill play in o
u
e
:
w
ers Moder
n
e
world’s ec
o
n
-renewable)
y
pically tunn
n
d control se
o
nnection to
h
rough layer
s
r
targeted n
e
their end-
g
h
utting down
t
.
t
acks routine
i
ng firewalls
m
s, anti-vir
u
and stro
n
u
rity tool pr
e
i
dence withi
n
malware ha
s
o
energy affe
t
o billions of
mand will c
o
i
ts fundame
n
er
living sta
n
w
ide “Energ
y
w
o of the m
o
h
e coming d
e
e
x topics. T
h
n
. There are
m
m
s involved
”
r
yone needs
t
ur
global ene
r
n
Economie
s
o
nomic mid
d
20
7
els a remot
e
rver, and th
e
compromis
e
of firewalls
.
e
twork, thes
e
g
ame attack
:
entire plants
,
ly defeat al
l
, encryption
,
u
s systems
,
n
g passwor
d
e
vents smar
t
n
our networ
k
s
an advers
e
cts your ow
n
other peopl
e
o
ntinue to ris
e
n
tal link t
o
n
dards for a
n
y
security an
d
o
st importan
t
e
cades. Thes
e
h
ey are als
o
m
any differen
t
”
[9].
t
o consider a
s
r
gy landscap
e
s
and Livin
g
d
le class wil
l
7
e
e
e
.
e
:
,
l
,
,
d
t
k
,
e
n
e
e
o
n
d
t
e
o
t
s
e
g
l
208
likely expa
n
people. Thi
s
living stand
a
developing
businesses
air-conditio
n
7.2 Global
Non-OECD
Despite e
f
likely incre
a
be in non-
O
demand wil
same amou
n
7.3 Electric
i
Nations
Human
a
reliable su
p
demand wil
l
by a near
d
countries.
7.4 Electri
c
400%
The mos
t
be electrici
t
growing ab
o
and wind to
by 2040,
h
Artificial In
t
n
d from 3
bi
s
growth will
a
rds, resultin
g
countries
a
and acce
s
n
ed homes.
Energy Ne
e
Nations
f
ficiency gai
n
a
se by nearl
y
O
ECD countr
i
l likely incr
e
n
t of energy
u
i
t
y
D
emand
N
a
ctivity con
t
p
plies of e
l
l
rise by 60
%
d
oubling of
p
c
ity from Sol
a
t
rapidly ex
p
t
y from sol
o
ut 400%.
T
global electr
i
h
elping the
t
elligence Int
e
i
llion to mo
r
coincide wit
h
g
in rising e
n
a
s people
d
s
s cars,
a
e
ds Rise ab
o
n
s, global en
e
y
25%. Nearl
y
i
es (e.g., Chi
n
e
ase about 4
0
u
sed in Amer
i
N
early Doub
l
t
inues to
be
l
ectricity.
G
%
between 2
0
p
owe
r
dema
n
a
r and Win
d
p
anding ene
r
a
r
and win
d
T
he combine
i
city supplie
s
CO
2
intens
i
e
gration wit
h
r
e than 5
b
i
l
h
vastly impr
o
n
ergy use in
m
d
evelop mo
d
a
ppliances,
o
ut 25%, Le
d
e
rgy demand
y
all growth
n
a, India),
w
0
%, or abou
t
i
ca today.
l
es in Non-O
E
e
dependen
t
G
lobal electr
i
0
16 and 2040
n
d in non-O
E
d
Increases a
b
r
gy supplies
d
, together
w
d share of
s
s
is likely to t
r
i
ty of deli
v
h
Energy Sou
l
lion
o
ved
m
any
d
ern
and
d
by
will
will
w
here
t
the
E
CD
t
on
i
city
, led
E
CD
b
out
will
w
ith
s
ola
r
r
iple
v
ered
ele
c
7.5
of
N
N
val
u
wh
i
car
b
lik
e
wit
h
7.6
M
o
M
co
n
fue
l
20
3
rol
e
dri
v
ind
u
7.7
A
c
c
A
en
e
car
b
a
n
glo
b
wil
l
lev
e
rces (Renew
c
tricity to fal
l
Natural Ga
s
N
eeds
N
atural gas’s
u
able energy
i
le also h
b
on-intensiv
e
e
ly to increa
s
h
about half
i
Oil Plays a
d
ern Produc
t
M
ore electric
n
ventional en
g
l
s use by th
3
0. However,
e
in the worl
d
v
en by com
m
u
stry.
D
ecarboniz
a
c
elerate
A
s the worl
d
e
rgy efficie
n
b
on-intensiv
e
n
early 45%
b
al GDP.
G
l
likely peak
e
l.
able and No
n
l
more than 3
s
Expands R
o
abundance
source to m
e
elping the
e
sources of
e
s
e more than
i
ts electricity
Leading Ro
l
t
s
cars and ef
f
g
ines will lik
e
e world’s li
g
the oil will
c
d
’s energy
m
m
ercial transp
a
tion of the
W
d
’s economy
n
cy gains
e
energy so
u
decline in
G
lobal energ
y
by 2040 at a
b
n
-renewable)
0%.
o
le to Meet a
and versatili
e
et a wide va
r
world sh
i
e
nergy. Natu
r
any other e
n
generation g
l
e in
A
iding
M
f
iciency imp
r
e
ly lead to a
p
g
ht-duty ve
h
c
ontinue to
p
m
ix, with gro
w
p
ortation and
W
orld’
s
Energ
y
nearly doub
and a s
h
u
rces will
c
the carbon
y
-related C
O
b
out 10% ab
Wide Variet
y
t
y make it
a
r
iety of need
s
i
ft to les
s
r
al gas use i
s
n
ergy source
,
rowth.
M
obility an
d
r
ovements i
n
p
eak in liqui
d
h
icle fleet
by
lay a leadin
g
w
ing deman
d
the chemica
l
y
System Wil
l
les by 2040
,
ift to les
s
c
ontribute t
o
intensity o
f
O
2
emission
s
ove the 201
6
y
a
s
s
s
,
d
n
d
y
g
d
l
l
,
s
o
f
s
6
Artificial Intelligence Integration with Energy Sources (Renewable and Non-renewable)
209
7.8 Nuclear Driving Electricity
Although by definition of renewable and
non-renewable energy sources, NPPs are considered
non-renewable type source. But being a clean source of
energy that does not produce, and carbon monoxide or
dioxide can be considered a renewable source of
energy-producing electricity. Their new generation
with a small foot-print known as SMRs may allow
sitting, where a large one is unsuitable, or a cluster of
small reactors could substitute for a large one.
AI along with a slew of other advanced technologies
as part of its components (see Fig. 3) such as ML and
DL, and ANN (artificial neural network), all together,
they all have demonstrated a fantastic and huge
potential to transform the energy and utility sectors.
With momentum behind the move of
decarbonization, decentralization, and the rollout of
novel technologies, utilities, IPPs (independent power
producers), and other energy companies are employing
AI to manage the imbalance in demand and supply
caused by the growing share of renewable energy
sources.
One of the benefits of AI and its augmentation of it
in energy, either renewable or non-renewable, would
be tracking all the data associated with the above roles
through its sub-components of ML and DL. This way,
we have a huge RoI (Return on Investment) for the
money that would go into such a cyber smart system
through the IoT process.
“In combination with other technologies like Big
Data, cloud, and Internet of Things (IoT), AI can
support the active management of electricity grids by
improving the accessibility of renewable energy
sources”, said Swagath Navin Manohar, Research
Analyst, Energy & Environment.
With all the information behind the scene, we can
use the collected sheer of data around these roles to
forecast the 2040 demand level for energy to be
prepared to meet such demand rather than shock us and
our source of energy driving electricity.
With these data, then collecting information from
them, we would be kneadable enough to be powerful
enough for our appropriate and smart decision making
to produce the right amount of energy at the right time
to meet the demand [10].
With past decade AI has gained huge momentum
and over the next decade, AI is expected to boost
efficiencies across the renewable energy sector by
automating operations in the solar and wind industries.
It will also allow utilities and IPPs to launch new
business and service models.
Information technology and applied science
engineering play an essential role in society, from
improving decision-making to advancing humanity’s
knowledge of the world and the universe.
Supercomputing, or HPC, enables scientists and
engineers to push the edge of what is possible for US
science and innovation. Using HPC-based modeling
and simulation, they are able to study systems that
otherwise would be impractical or impossible to
investigate in the real world due to their complexity,
size, fleeting nature, or the danger they pose.
Exascale computing will provide the capability to
tackle scientific discovery and national security
challenges at levels of complexity and performance
that previously were out of reach.
In conclusion, the collaboration between AI/ML and
DL with HPC augmentation and its partner human is
essential.
Furthermore, “In addition to making the electricity
system intelligent and flexible, AI algorithms help
utilities and energy companies understand and
optimize consumer behavior and manage energy
consumption across different sectors”, noted Manohar.
“Meanwhile, complex ML algorithms combined with
Artificial Intelligence Integration with Energy Sources (Renewable and Non-renewable)
210
real-time weather data from satellites, ground-based
observation, and climate models can be used to forecast
the electricity generated by Renewable Energy Source
(RES) like wind, solar, and ocean”.
AI-based applications can create further revenue
opportunities for the energy and utility sector by:
Empowering software applications to analyze
large data sets, identifying patterns, detecting
anomalies, and making precise predictions.
Aiding the development of smart applications that
can autonomously make accurate decisions based on
learning. This drives AI’s integration with a wide range
of applications.
Enabling customer-centric solutions that
understand evolving customer needs and make
automatic recommendations.
Using predictive analytics to improve equipment
O&M and predict downtime, which can extend the
lifetime of the equipment.
Facilitating active customer participation in
demand-response programs using game theory
algorithms and leveraging block chain to protect data.
References
[1] Wrigley, E. A. 1988. Continuity, Chance and Change:
The Character of the Industrial Revolution in England.
Cambridge University Press, 1st edition.
[2] Zohuri, B., and McDaniel, P. J. 2021. Introduction to
Energy Essentials: Insight into Nuclear, Renewable, and
Non-Renewable Energies. London: Academic Press
Publishing Company. (to be published)
[3] Zohuri, B. 2019. Small Modular Reactors as Renewable
Energy Sources, 1st ed. Berlin: Springer Publishing
Company.
http://energy.mit.edu/research/future-nuclear-power/.
[4] Zohuri, B. 2017. Plasma Physics and Controlled
Thermonuclear Reactions Driven Fusion Energy, 1st ed.
Berlin: Springer Publishing Company.
[5] Zohuri, B., and McDaniel, P. J. 2017. Combined Cycle
Driven Efficiency for Next Generation Nuclear Power
Plants: An Innovative Design Approach, 2nd ed. Berlin:
Springer Publishing Company.
[6] Massachusetts Institute of Technology. 2003. The Future
of Nuclear Power: An Interdisciplinary MIT Study. ISBN:
0-615-12420-8.
http://energy.mit.edu/research/future-nuclear-power/.
[7] Zohuri, B., and McDaniel P. J. 2019. Thermodynamics in
Nuclear Power Plant Systems, 2nd ed. Berlin: Springer
Publishing Company.
[8] Zohuri, B. 2020. Nuclear Micro Reactors, 1st ed. Berlin:
Springer Publishing Company.
[9] Zohuri, B., and McDaniel, P. J. 2019. Advanced Smaller
Modular Reactors: An Innovative Approach to Nuclear
Power, 1st ed. Berlin: Springer Publishing Company.
[10] Zohuri, B., and McDaniel, P. J. 2018. Combined Cycle
Driven Efficiency for Next Generation Nuclear Power
Plants: An Innovative Design Approach, 2nd ed. Berlin:
Springer Publishing Company.
[11] Zohuri, B. 2017. Inertial Confinement Fusion Driven
Thermonuclear Energy, 1st ed. Berlin: Springer
Publishing Company.
[12] Zohuri, B. 2017. Magnetic Confinement Fusion Driven
Thermonuclear Energy, 1st ed. Berlin: Springer
Publishing Company.
[13] Zohuri, B., and Mossavar, F. R. 2020. A Model to Forecast
Future Paradigms: Volume 1: Introduction to Knowledge
Is Power in Four Dimensions, 1st ed. Apple Academic
Press.
[14] Zohuri, B., and Moghaddam, M. 2017. Business
Resilience System (BRS): Driven through Boolean, Fuzzy
Logics and Cloud Computation: Real and near Real Time
Analysis and Decision-Making System, 1st ed. Berlin:
Springer Publishing Company.
[15] Zohuri, B., and Zadeh S. 2020. Artificial Intelligence
Driven by Machine Learning and Deep Learning, 1st ed.
Nova Science Pub Inc.
[16] Zohuri, B., and Mossavar, F. R. 2020 “Artificial
Intelligence versus Human Intelligence: A New
Technological Race.” Acta Scientific Pharmaceutical
Sciences 4 (5): 50-8.
[17] Sullivan, P. 2013. “The Future of Energy.” Georgetown
Journal of International Affairs 14 (1): 3-6.
