Journal of Energy Storage 13 (2017) 48–57

Contents lists available at ScienceDirect

Journal of Energy Storage

journal homepage: www.elsevier.com/locate/est

Compressed air energy storage integrated with floating photovoltaic

plant
R. Cazzanigaa , M. Cicua , M. Rosa-Clota , P. Rosa-Clota , G.M. Tinab,* , C. Venturab
a
Koiné Multimediasrl, Italy
b
DIEEI - University of Catania, Italy

A R T I C L E I N F O A B S T R A C T

Article history:
Received 29 March 2017 Floating photovoltaic (FPV) systems are an emerging technology suitable for large plants, especially, on
Received in revised form 10 June 2017 fresh water basins. We suggest integrating a CAES system to FPV using the pipes, necessary for the
Accepted 11 June 2017 buoyancy of the modular raft structure, as a compressed air reservoir. The huge basin thermal inertia
Available online xxx allows for an isothermal compression– expansion cycle, which promises high storage efficiency.
© 2017 Elsevier Ltd. All rights reserved.
Keywords:
Floating PV
CAES
Energy storage

1. Introduction and trends pertaining to Grid-Connected renewable systems [9]

Most renewable energy technologies require the use of large
land areas in order to generate electricity. The impact on areas used
for agricultural purposes varies according to the technology.
Specifically, for photovoltaic (PV) systems, large surface areas
are needed because of the low density power of solar energy. For
medium and large size power plants, the demand for large
installation surfaces can be met with land-based PV plants but
using vast tracts of land for solar farms will increase competition
for land resources. In this regard, two possible solutions can be
investigated: 1) using the concept of agrivoltaics or co-developing
the same area of land for both solar PV power and conventional
agriculture [1] and 2) intensifying the research in the use of PVs on
[2,3], (see Fig. 1) or under [4] (see Fig. 2) water surfaces (e.g. lakes,
artificial basins, sea). The two water solutions are named floating
and submerged.
There are many ways to classify such systems according, for
example, to: size, mooring, PV technology, the presence of tracking
or not, and so on. Regarding the used of PV modules we can have:
flat and rigid PV modules (mainly Si crystalline made) [5,6] or thin-
film flexible floating PV (T3F-PV) array [7,8].
A critical issue in using PV systems is related to the variability
and limited control of its production that can be solved by using a
suitable storage system. In [9] and [10] a literature review and a
sensitivity analysis of the recent developments and future scenario
* Corresponding author.
E-mail address: giuseppe.tina@dieei.unict.it (G.M. Tina).
and PV Systems [10], in European countries with high penetration
of distributed generation, we see that storage systems are used not
only to reduce the inadvertent stress on the electricity grid, but
also to allow renewable systems to provide ancillary services to the
regional or national grid.
The main storage technology used for both stand-alone and
grid-connected PV systems is based on batteries, but others
solutions such as water/seawater pumped storage, [10] and
compressed air energy storage [11] can be considered since from
the life cycle assessment used to compare ESSs (Energy Storage
System) of different nature reported in [12] it emerges that the
traditional CAES (Compressed Air Energy Storage) and PHS
(Pumped Hydro Storage) have the highest ESOI indices

respectively 240 and 210. (ESOI refers to the total amount of
energy stored over the lifetime of a storage technology unit,
divided by the amount of energy used in producing that unit, i.e.
kWh/kWh) Among the batteries Li-ion has the highest ESOI, that is
10, evidently much lower than CAES and PHS.
In this paper, we consider the structure of the floating plant
built with modular rafts supporting the PV panels, which,
assembled, constitute the PV floating platform [3]. Rafts for FPV
plants may also use pontoons (a flotation device with buoyancy
sufficient to float itself as well as a heavy load), drums, or extruded
plastic blocks, so the term “pontoon” is very often used [2]; in this
paper as the tubes of pontoon are used to store air, hereinafter we
will use the term “pontoon”.
The pipes necessary for the pontoon buoyancy are normally in
polyethylene or in some light cheap material, but we suggest using
steel and using the pipes themselves as cylinders to store

http://dx.doi.org/10.1016/j.est.2017.06.006
2352-152X/© 2017 Elsevier Ltd. All rights reserved.
 R. Cazzaniga et al. / Journal of Energy Storage 13 (2017) 48–57 49

Table 1
Nomenclature.

Symbol Unit Description

Cv J/(mol K) Molar specific heat at constant volume
De,P m External diameter of a pipe
Di,P m Internal diameter of a pipe
EPV,R kWh Daily PV energy in a raft
Estor kWh Stored energy in the raft
FB N Buoyancy force
FB,MAX N Max buoyancy force
Lcomp kWh Compression work
Fig. 1. Raft structure for a floating PV plant. LP m Length of a pipe
ma kg Mass of air
mair,R kg Mass of the compressed air in the pipes
compressed air. We propose using the CAES technology and mP,R kg Mass of pipes
exploiting the isothermal approach by profiting from the large mPV,R kg Mass of the PV system and carpentry
thermal inertia of the water basin where the floating plant is mT,R kg Mass of the floating system
located (the nomenclature is reported in Table 1). n Number of moles
NPV,R – Number of PV modules in a raft
Pp,f Pa Final pressure inside the pipes
2. CAES versus battery storage Pp,i Pa Initial pressure inside the pipes
PPV,R kWp Peak power of PV array in a raft
A battery storage system can easily be installed on the platform. R J mol-1 K-1 Ideal gas constant
Tca,i K Initial temperature compressed air inside the pipes
Batteries are standard systems that work perfectly and are fully
Tca,f K Final temperature compressed air inside the pipes
developed on an industrial scale. However, they are costly and have tcomp s Compression time
a limited lifetime as well as problems of waste disposal or recycling Ve,R m3 External volume of pipes
at the end of their life cycle. Therefore other methods, such CAES Vi,R m3 Internal volume of pipes in the pontoon
systems, are preferable and very promising. VTP m3 Total volume (tank and pipes)
Wair,R N Weight of the compressed air in the pipes
Several studies exist comparing batteries cost versus other
WP,R N Weight of pipes
technologies. We do not want to enter into this dispute so we quote WPV,R N Weight of the PV system and carpentry
a study from Sandia National Laboratories of a few years ago [13] WT,R N Weight of the floating system
(see Table 2) where a general comparison is given, limiting our DTw 
C Temperature drop of the water of the basin
attention to energy storage systems whose storage duration can l W/mK Thermal conductivity of steel
uB 
Buoyancy angle
range from a few hours to several days and more.
Although the cost of CAES energy storage is very low, a real
estimate of costs should be made based on an experimental
development and a long test period, which are not available at
present.
Polyethylene) pipes is estimated at $1000 per pontoon, to which
We only want to stress that, in our case, the air compressed
we must add the costs of the compressor and pneumatic and/or
reservoir is an integral part of our platform, so that its cost should
hydraulic system. At the present status of the research we cannot
be only partially charged to the storage system: the increase in cost
give a reliable estimate of the real final price.
due to the use of steel pipes rather than HDPE (High Density
CAES is widely discussed in the literature and it is an approach
which appears to be suitable for large and medium storage [14],
with several possible solutions than can be classified according to
their thermodynamic properties [15]:

1. Isothermal CAES (ICAES) which has the constraint to be a low

speed process and which needs a large surface in order to
exchange heat efficiently with an external thermal bath [16]
2. Adiabatic CAES (ACAES), which implies large and insulated high
pressure reservoir. Adiabatic CAES can in turn work with or
without a thermal energy storage (TES) [10]
3. Mixed approaches where the drop in temperature during the
adiabatic expansion is partially compensated by a thermal
exchange with external sources or even supported by the use of
fossil fuel to better exploit the compression-expansion process
[17] or an Isobaric Adiabatic Compressed Air Energy Storage (IA-
CAES) system is used [18]

3. Isothermal compressed air energy storage (ICAES)

In this paper we focus on ICAES and study the possibility of

using the floating pipes of the pontoons, which are the basic
elements of the floating system, as a reservoir for high pressure air.
The full process can be summarized in the block scheme given
in Fig. 3 where the flux of electric energy is highlighted by red lines
and the air flux by white arrows.
Fig. 2. Submerged PV plant.
50 R. Cazzaniga et al. / Journal of Energy Storage 13 (2017) 48–57

Table 2
Cost comparison of storage technology from [13].

Power Subsystem Cost [$/kW] Energy Storage Subsystem Cost [$/kWh] Round-trip Efficiency [%] Cycles
Advanced Lead-acid Batteries 400 330 80 2000
Sodium/ 350 350 75 3000
sulphur Batteries
Lead-acid Batteries with Carbon-enhanced Electrodes 400 330 75 20000
Zinc/ 400 400 70 3000
bromine Batteries
Vanadium Redox Batteries 400 600 65 5000
Lithium-ion Batteries (Large) 400 600 85 4000
CAES 700 5 N/A (70) 25000
Pumped hydro 1200 75 85 25000

The working principle is the following:
 The floating PV platform supplies electric energy through PV
modules and inverters to the grid or to the air compressor
 The air compressor increases the pressure inside the pipes from
0.1 MPa up to 20 MPa
 The compressed air, stored in the floating plant pipe, can be used
when necessary to produce energy through a turbine.
To implement this storage scheme in practice we have to
modify the proposed pontoon structure [3], see Fig. 4.
In the following two paragraphs we describe the buoyancy
analysis of structure and the main issues about the electrical
design of a PV system.
3.1. Bouyancy analysis of PFV structure
For the sake of clarity we concentrate on a single pontoon (the
modulus of a large PV floating plant). It is constituted by two 12 m
long pipes, Lp, with an external diameter of 0.4 m, De,p. We suggest
building these pipes in steel and using a 5 mm thickness (internal
diameter, Di,p of 0.395 m) in order to be able to support internal
pressure, Pp,f, up to 20 MPa and more.
Within these parameters, the mass of pipes in a pontoon, mP,R, is
about 1215 kg (the exact mass depends on details of the caps and
on the steel used) and the internal volume, Vi,R is 2.87 m3.
To this mass we have to add the mass of the PV structure
(panels, copper cables, beams and catwalks), mPV,R, which
altogether can reach 500 kg, plus the compressed air, which at
pressure, Pair, equal to 20 MPa, has a mass, mair,R, of 682 kg.
Therefore the total mass of the pontoon, mT,R = mP,R + mPV,R + mair,R,
ranges from 1700 to 2400 kg (depending on the pressure on the
pipes).
The external volume of pipes, Ve,R is slightly more than 3 m3,
which guarantees a sufficient Archimedes thrust.
The maximum buoyancy, FB,MAX, defined as the highest
Archimedes thrust that happens when the object is completely
submerged, is much greater than the total weight of the pontoon,
WT,R. So there is an adequate margin of buoyancy, DFB, Eq. (1). This
ensures, for example, that one or more human operators can safely
work on an isolated pontoon.
DF B ¼ F B;MAX
W T;R ¼ 29:6
23:5 ¼ 6:4kN ð1Þ
It is important to notice that, when the pontoon floats and so FB
is equal to W T,R, the steel pipes are immersed for 2/3 of the
transversal section in the basin water, as can be seen from the
buoyancy angle, uB, of 83.8 quoted in Fig. 5, for 0.1 MPa internal
pressure, whereas at 20 MPa qB is 60.5 so that more than 2/3 of the
pipes surface is in contact with the water.
This value is important since we want to be sure that an efficient
heat exchange exists between the water basin and the air inside the
pipes.
3.2. The electrical part
A pontoon, with the dimension specified in item 3.1, is equipped
with a set of 8 to 10 PV panels, each with a peak power of 250 to
320 Wp. So the PV power installed in a pontoon, PPV,R, ranges from
2 to 3.2 kWp and the daily energy, EPV,R, produced during a sunny
day at Mediterranean latitude can be in the range of 12–18 kWh/
day.
For the conversion from DC to AC, different solutions can be
adopted, such as: a string converter (one for each pontoon), or a
central inverter (a single inverter can be used for 10 or more
pontoons).
The pontoon has large buoyancy so that it can support a small
transformer and a small compressor. However, in a large platform
built with several pontoons connected together, it is possible to

Fig. 3. Block scheme for ICAES. Fig. 4. The pontoon with the steel pipes in black.
 R. Cazzaniga et al. / Journal of Energy Storage 13 (2017) 48–57 51

Fig. 5. Floating rack and buoyancy parameters (buoyancy angle and forces acting on the pontoon are reported).

dedicate one pontoon solely to support the needed electrical and
mechanical apparatus. In this way a single unit with the same
structure, but with polyethylene pipes, would be used for electrical
and mechanical operations and could support up to 2 tons of
mechanical and electrical apparatus.
If a storage system is required (e.g. stand alone PV system or PV
system connected to an isolated grid), a battery storage system can
be installed on the platform. However, because of its cost and other
limitations, as mentioned above, other solutions that adequately fit
the peculiarity of a floating system should be investigated. Here a
CAES system is studied.
4. The isothermal CAES
It is well known that the main problem in CAES is the managing
of the thermal balance.
Most of the efforts of companies and research institutes
working in this field are devoted to this problem and it is very
important to choose between isothermal compression, with a large
amount of heat exchange, and adiabatic compression, with the
problem of storing the heat and recovering it in an efficient way
[11].
The very large thermal bath, formed by the water basin,
suggests the use of the isothermal approach. In this context, two
possible solutions are investigated:
 full range isothermal compression: from 0.1 to 20 MPa
 reduced range isothermal compression: from 10 to 20 MPa
4.1. Full range isothermal compression: from 0.1 to 20 MPa
In this case, if we assume that the final pressure is Pf = 20 MPa
and that the internal volume is Vi,R = 2.87 m3, the mole number n is
23550 and the molar mass of air is 0.029 kgma = 682 kg.
If the compression is perfectly isothermal, it allows us to store
the energy, Estor, given by the work of the mechanical compression,
Lstor, so we can write:
   
Pp;f Pp;f
Estor ¼ Lstor ¼ nRT ca;f ln ¼ Pp;f V p;f ln ð2Þ
Pp;i Pp;i
where: R is the gas constant, Pp,i, Pp,f are the initial and finale
pressure, respectively, and Tca is the temperature of the com-
pressed air (293 K).

Fig. 6. Storage energy versus pressure.

52 R. Cazzaniga et al. / Journal of Energy Storage 13 (2017) 48–57
In this case the 1st law of thermodynamics affirms that the full
work done by mechanical forces is transferred to the thermal bath
since the variation of temperature is zero.
DQ = nCvDT + Lstor
In Fig. 6 the behaviour of the energy versus the pressure in the
two steel pipes is given by expressing energy in kWh, and the
numerical values are given for the pressure of 10 MPa and 20 MPa.
The isothermal compression allows the storage of many kWh of
energy. For example, if the compression process is pushed up to
20 MPa, 84 kWh are necessary and all this energy is transferred to
the heat reservoir (water basin) guaranteeing the isothermal
process.
For this reason, it is worth expressing the value of this energy in
Cal and, as shown in Fig. 6, the amount of heat that has to be
transferred is high (for 20 MPa it reaches 72,577 Cal). This is the
critical point of the process: in the absence of an effective heat
transfer, when the pressure increases from 0.1 MPa to 20 MPa, the
gas temperature increases from 300 K up to 1360 K.
In the adiabatic process, the thermodynamic energy balance
would be completely different:
DQ ¼ 0 ¼ nC V DT þ Lstor
Therefore all the work done by the mechanical forces goes
towards increasing the internal energy of the gas. However, in an
adiabatic process the strong increase in temperature gives rise to
irreversible processes that can hardly be controlled.
In an isothermal process we can exploit the large thermal bath
formed by the water basin and the fact that our cylinders have a
large surface of about 30 m2 and are in steel, with a good
conductivity value l = 40–50 W/(mK) (depending on the carbon
content).
Assuming a drop in temperature, DT W, of only 5  C the heat
transfer takes place in a very short time: tcomp, 289 s to pass from
0.1 MPa to 20 MPa of pressure. This time is very short compared
with the typical period of daily solar energy harvesting which
Fig. 7. Compression process (the line point tract on the right plot refers to the right y-axis).
ranges between 15,000 and 20,000 s, so that in theory an
approximated isothermal process is possible.
The same happens for the expansion process, which takes place
in an even longer period if energy is requested during the night or
(3)
on cloudy days.
However this process is far from being realistic: we are not
working with a given mass of gas inside a great cylinder which is
slowly compressed to the required final size, but we are taking air
from the external atmosphere and we are pushing it up to 20 MPa
with a multistage pumping (or expansion) system.
The compressor and the pneumatic motor must be treated with
caution and a good heat exchanger is important in order to reduce
the irreversible process as much as possible.
Multistage turbines are suggested by some authors [19], other
researchers suggest the possibility of using a mixed system of
hydraulic-air compression [20]. In this case an efficient hydraulic
pump acts on the fluid which compresses the air with an efficient
heat exchange between fluid and gas.
Research groups in USA have explored the isothermal solution,
but their work is not conclusive.
About the thermal inertia of the water bath, supposing the basin
depth to be only two meters, the water volume below the pontoon
ð4Þ is 48 m3 so that the transfer of 72,653 Cal should increase its
temperature by only 1.5  C. This temperature increase is compen-
sated during the expansion phase so that the net balance is zero,
apart from the unavoidable efficiency loss. This value is small but
not negligible and, even if the basin depth is often more than 2 m, it
is prudent to limit the analysis to a lower pressure and to an
average temperature variation below 1  C.
This temperature variation should be compared with the
natural temperature increase of the upper layer of a basin during a
sunny day in the absence of water circulation.
The main problem with this proposal lies in the fact that, during
both expansion and compression, the turbine or pneumatic motors
have to deal with a strong difference in pressure and unavoidable
changes in gas temperature.
 R. Cazzaniga et al. / Journal of Energy Storage 13 (2017) 48–57
For this reason, we suggest a different more limited approach.
4.2. Reduced range isothermal compression: from 10 to 20 MPa
Since the main problem lies in the reversibility of the
isothermal compression (expansion), we suggest starting with a
higher pressure from the very beginning but performing a limited
compression. To be specific, let’s start with the pipes at a pressure
of 10 MPa and let’s compress the air inside so that we halve the
volume. In this case most of the work stored when filling the pipes
from atmospheric pressure is lost but a sizeable amount is still
available.
We can summarize the results in Fig. 7, where adiabatic and
isothermal transformations are compared and the related param-
eters explicitly given: adiabatic curves are in black whereas
isothermal transformations are represented in grey.
The process has a simple thermodynamic parametrization,
which will be discussed for a unit volume of 1 m3.
In Fig. 7 the process is studied starting from a tank of 0.5 m3
(connected to a pipe of 0.5 m3) with an initial temperature of 300 K.
During the compression process, the volume is halved and in
the isothermal compression (grey line, left plot) the pressure rises
from 10 MPa to 20 MPa and the work done is of 1.95 kWh (left plot).
If the compression is perfectly adiabatic, the temperature rises
by 69  C and the pressure arrives at 25 MPa; the work will be
2203 kWh.
In Fig. 8 the expansion process is analysed.
The isothermal process gives back the stored energy.
The adiabatic process takes place after a fairly long period of
storage and we assume that the system has reached an equilibrium
with the external temperature (in the example 300 K). In this case
the temperature during a quick expansion goes down by 73  C and
the work done is of only 1.67 kWh.
Of course, a perfect isothermal process does not exist. However,
as our compression takes place over a rather long period of time
(several h), the process is, to a good extent, reversible.
Fig. 8. Expansion process. (the linepoint tract on the right plot refers to the right y-axis).
53
In Fig. 9 some further details are given about the process: the
isothermal expansion process is analysed from a pressure of
20 MPa and a volume of 0.5 m3 to a pressure of 0.5 MPa and a
volume of 20 m3.
It is quite evident that we are interested only in the first step of
the process (black circles in the grey curve which give the work
obtained during the first phase of the expansion). So we lose the
opportunity to fully exploit the energy and we get only 2 kWh,
whereas the full expansion up to 50 m3 (and to the atmospheric
pressure of about 15 kWh) could give more than 12 kWh.
What we gain is a reduced volume and a simpler management
of the required heat exchange mechanism.
If we use the full volume of a floating pontoon (approximately
3 m3) complemented by an external tank of equal volume, we see
that the energy which can be stored on a pontoon is approximately
12 kWh and so it is suitable for the day-night storage in a sunny
location.
5. Hybrid water/air CASES
The main problem now is how to manage the compression from
10MPa to 20 MPa. In this regard, we suggest a hybrid water/air
CASES. The diagram in Fig. 10 describes the full process scheme
where the flux of electric energy is highlighted by red arrows, the
air flux by white arrows and the water flux the blue arrows. The
steps of process are:
 the floating PV platform supplies electric energy through PV
modules and inverters to the grid or to the water pump;
 the water pump compress water, up to 20 MPa in a suitable
storage tank where the pressurized water is used to increase the
air pressure in pipes from 10 to 20 MPa;
 the compressed air is stored in the floating plant pipes and when
necessary the air push the water and power an hydro-turbine to
recover the stored energy.
54 R. Cazzaniga et al. / Journal of Energy Storage 13 (2017) 48–57
Fig. 9. Detail of a general expansion process (Values refer to the blue circles, pressure, and red dots, Lexp). (For interpretation of the references to colour in this figure legend,
the reader is referred to the web version of this article.)
The system is shown in Fig. 11, where tank A is connected to the
external water (sea or basins). A hydraulic pump (schematically
represented by the black circle on the bottom) pumps water into
Tank B, thus rising the piston (black thick line) and increasing the
air pressure in Tank B and in the pontoon pipes.
The different operation phases are reported in Fig. 11, the tank is
empty and so in the tubes there is the minimum pressure and
Fig. 12, where in the upper scheme the tank is half filled and in the
lower scheme the thank is filled so there is the maximum pressure
inside the tubes.
Thanks to the hydraulic pump, this system has a high
mechanical efficiency (90%) and, due to the slow compression, it
Fig. 10. Block scheme of hybrid water/air CASES.
can approximate the isothermal compression quite well. We
estimate that the loss due to the irreversible process is below 10%,
so that the full energy efficiency should be around 80%.
Under the hypothesis of 1 MWp floating plant, the volume of
the connected pipes is approximately 1000 m3, the stored energy
2000 kWh and we need a compression tank of equivalent volume
(or a few compression tanks of the same total volume).
This pressurized volume is constituted by vertical cylinders of
suitable basis where water is pumped in order to increase the
pressure from 10 to 20 MPa and which are connected to the pipes of
the plant through compressed air joints. These tanks can be sunk to
the bottom of the basin or on the ground near the plant.
The concept of using a hydraulic compressor for pumping air
was already proposed in [21] more than three years ago and
recently the same method was applied to a process for compress-
ing air in an almost isothermal way using water spray [15].
Fig. 11. Scheme of the hydraulic compression (air inside the tube is at 10 MPa).
 R. Cazzaniga et al. / Journal of Energy Storage 13 (2017) 48–57 55

Table 3
Monthly energy balances between generation and load demand in a pontoon
equipped with CAES storage.

GPV EPV ELoad,d+ ELoad,d- ELoad,s ELoad,d

kWh/m2 kWh kWh kWh kWh kWh

Jan 73.9 102.7 0.1 80.84 105.1 185.9
Feb 90.9 127.2 0.1 48.95 119 167.9
Mar 137.5 200.1 14.2 12.5 173.4 185.9
Apr 170.4 257.4 73.7 0 179.9 179.9
May 198.9 307.6 114.7 0 185.9 185.9
Jun 210.7 326.6 140.5 0 179.9 179.9
Jul 213.7 331.1 137.6 0 185.9 185.9
Aug 192.8 301 109.2 0 185.9 185.9
Sep 156.9 240.4 54.7 0 179.9 179.9
Oct 121.5 177.8 11.8 17.54 168.4 185.9
Nov 79.9 111.3 0.1 71.51 108.4 179.9
Dec 65.5 90.7 0.2 99.3 86.6 185.9

respectively the exceeding and the missing load energy (difference

Fig. 12. Two steps of the hydraulic compression: tank half filled (upper scheme),
tank filled (lower scheme).
between the demand and the supplied energy).
From these data it is possible to conclude that there is a good
balance among generation, demand and storage in the proposed
system: in fact the load from March to October is almost fully
supplied, although in the winter season the low PV generation does
not allow to fully satisfy the load demand.
In Fig. 13 the daily energy balance for two sunny days in
December and June is plotted, where PPV is the photovoltaic power
and PDis the load power.
Of course, if the proposed generation/storage system is used for
the electrification of remote places or islands, probably a mix
renewable (wind, solar, biomass) and conventional (diesel engine)
generation system would be used, so in this case the objective
wouldn’t be just the daily balance between the energy demand and
the electrical energy generation but the optimal technical and
economical trade-off among the different energy sources [22].

Even if the context is quite different, we consider these works as
a confirmation that this theoretical approach is promising and that
it requires above all experimental testing.
6. A FPV system with storage: day-night energy management
In the first instance, we would like to verify the ICAES system
using a day-night charging/discharging approach and assuming
that we store 12 kWh isothermally for each pontoon.
In this case the heat transfer takes place in a very short time and
the increase in the temperature of the water reservoir does not
exceed 0.3  C.
Under the hypothesis that the pontoon supports eight standard
PV panels, each with a power of 250 W, that the tilt angle of the PV
is 8 and the pitch among modules is 1.5 m, the power peak of the
pontoon is WP = 2.0 kWp. We want to analyse the energy behaviour
of the PV floating pontoon with CAES storage using the
meteorological data of Palermo (Italy), Lat. 38.0 N
Long. 13.4
E, as a typical Mediterranean place. The yearly global horizontal
irradiation in Palermo is 1860 kWh/m2, specifically the minimum
and maximum daily global irradiation in kWh/m2 are respectively
1.90 and 7.27.
In this example, we assume that the system is stand-alone, that
it supplies a local load whose daily demand is equal to 6 kWh, and
that the demand is flat during the day and along the year.
Table 3 shows the monthly energy balances, where GPV is the
specific irradiance on the plane of the PV module, EPV is the energy
produced by the PV system, ELoad,d is the energy required by the
user, whereas ELoad,s is the energy supplied. ELoad,d+ and ELoad,d- are
7. Conclusions
Substituting polyethylene pipes with steel cylinders allows us
to use the pontoons of a floating PV plant as reservoirs for CAES
purpose. Several arguments favour this solution:
 The pontoon’s cost is only slightly increased and the system
looks simple from the structural point of view.
 It does not require any major modification and it works both for
fixed and tracking plants.
Two main ideas are developed in this context both based on the
isothermal compression/expansion process.
The first consists in compressing external air inside the steel
pipes. The energy that can be stored in this case reaches the very
high value of 84 kWh per pontoon and we assume that, thanks to
an efficient heat exchange with the unlimited heat reservoir at our
disposal (the water basin), this process is to a certain extent
reversible. Our estimation is that about 60% of the stored energy is
recoverable. Problems about the compressor system and pneu-
matic motors still exist.
The second process starts from the steel pipes connected to a
tank of equal volume and containing air at a pressure of 10 MPa.
The energy is stored by compressing air bya factor of two (from
10 MPa to 20 MPa) without any exchange with the exterior.
A hydraulic pumping system is used both for compression and
expansion.
In this case the stored energy is much less but it is still enough
to provide the storage needed for the day-night autonomy of the
floating PV pontoon. A great advantage of this solution is that, even
56 R. Cazzaniga et al. / Journal of Energy Storage 13 (2017) 48–57
Fig. 13. Daily curves on: 4th December (a), and 21 st June (b).
in the hypothesis that the heat exchange is not efficient and that
irreversible processes take place, the maximum losses do not
exceed a few per cent so that we estimate efficiency above 80%.The
working principles are illustrated in this paper but experimenta-
tions are needed in order to confirm these design ideas.
References
[1] H. Dinesha, J.M. Pearce, The potential of agrivoltaic systems, Renew. Sustain.
Energy Rev. 54 (2016) 299–308, doi:http://dx.doi.org/10.1016/j.
rser.2015.10.024.
[2] K. Trapani, M. Redón Santafé, A review of floating photovoltaic installations:
2007–2013, Progress in Photovoltaics, Res. Appl. 23 (4) (2015) 524–532.
[3] M. Cazzaniga, M. Rosa-Clot, P. Rosa-Clot, G.M. Tina, C. Ventura, Floating
Photovoltaic plants: performance analysis and design solutions, Renew.
Sustain. Energy Rev. (2017), doi:http://dx.doi.org/10.1016/j.rser.2017.05.269 In
Press.
[4] M. Rosa-Clot, P. Rosa-Clot, G.M. Tina, P. Scandura, Submerged photovoltaic
solar panel: SP2, Renew. Energy 35 (8) (2010) 1862–1865, doi:http://dx.doi.
org/10.1016/j.renene.2009.10.023.
[5] R. Cazzaniga, M. Rosa-Clot, P. Rosa-Clot, G.M. Tina, Floating tracking cooling
concentrating (FTCC) systems, Proceedings of 38th IEEE Photovoltaic
Specialists Conference (PVSC) (2012), doi:http://dx.doi.org/10.1109/
PVSC.2012.6317668.
[6] Y.-G. Lee, H.-J. Joo, S.-J. Yoon, Design and installation of floating type
photovoltaic energy generation system using FRP members, Sol. Energy 108
(2014) 13–27, doi:http://dx.doi.org/10.1016/j.solener.2014.06.033.
[7] K. Trapani, D.L. Millar, The thin film flexible floating PV (T3F-PV) array: the
concept and development of the prototype, Renew. Energy 71 (2014) 43–50,
doi:http://dx.doi.org/10.1016/j.renene.2014.05.007.
[8] K. Trapani, D.L. Millar, H.C.M. Smith, Novel offshore application of
photovoltaics in comparison to conventional marine renewable energy
technologies, Renew. Energy 50 (2013) 879–888, doi:http://dx.doi.org/
10.1016/j.renene.2012.08.043.
[9] C. Bussar, et al., Large-scale integration of renewable energies and impact on
storage demand in a European renewable power system of 2050-Sensitivity
 R. Cazzaniga et al. / Journal of Energy Storage 13 (2017) 48–57
study, J. Energy Storage 6 (2016) 1–10, doi:http://dx.doi.org/10.1016/j.
est.2016.02.004.
[10] G. Manfrida, R. Secchi, Seawater pumping as an electricity storage solution for
photovoltaic energy systems, Energy 69 (1) (2014) 470–484, doi:http://dx.doi.
org/10.1016/j.energy.2014.03.040.
[11] J. Mason, V. Fthenakis, K. Zweibel, T. Hansen, T. Nikolakakis, Coupling PV and
CAES power plants to transform intermittent PV electricity into a dispatchable
electricity source, Prog. Photovoltaics Res. Appl. 16 (2008) 649–668, doi:http://
dx.doi.org/10.1002/pip.858.
[12] S. Few, A. Gambhir, G. Offer, J. Nelson, N. Brandon, Energy Storage for Balancing
Intermittent Renewables: Outlook for a Range of Technologies upto 2030, in
proceedings of UKES-2015, University of Birmingham, Birmingham, 2015.
[13] S. Schoegun, Energy Storage Systems Cost Update, Sandia Laboratories, 2011.
[14] G. Venkataramani, P. Parankusam, V. Ramalingam, J. Wang, A review on
compressed air energy storage
A pathway for smart grid and polygeneration,
Renew. Sustain. Energy Rev. 62 (2016) 895–907, doi:http://dx.doi.org/10.1016/
j.rser.2016.05.002.
[15] M. Budt, D. Wolf, R. Spanc, J. Yan, A review on compressed air energy storage:
basic principles, past milestones and recent developments, Appl. Energy 170
(2016) 250–268, doi:http://dx.doi.org/10.1016/j.apenergy.2016.02.108.
[16] A. Odukomaiya, et al., Thermal analysis of near-isothermal compressed gas
energy storage system, Appl. Energy 179 (2016) 948–960, doi:http://dx.doi.
org/10.1016/j.apenergy.2016.07.059.
57
[17] H. Safaei, D.W. Keith, R.J. Hugo, Compressed Air Energy Storage (CAES) with
compressor distributed at head loads to enable waste heat utilization, Appl.
Energy 103 (2013) 165–179, doi:http://dx.doi.org/10.1016/j.
apenergy.2012.09.027.
[18] Y. Mazloum, H. Sayah, M. Nemer, Dynamic modeling and simulation of an
isobaric adiabatic compressed air energy storage (IA-CAES) system, J. Energy
Storage 11 (2017) 178–190, doi:http://dx.doi.org/10.1016/j.est.2017.03.006.
[19] G. Marano, Application of dynamic programming to the optimal management
of a hybrid power plant with wind turbines, photovoltaic panels and
compressed air energy storage, Appl. Energy 97 (2012) 849–859, doi:http://dx.
doi.org/10.1016/j.apenergy.2011.12.086.
[20] P. Troy McBride, A., Bell, P. Dax Kepshire, www.sustainx.com, 2013. (Accessed
April 2016).
[21] D. Shaw, J.J. Yu, C. Chieh, Design of a hydraulic motor system driven by
compressed air, Energies 6 (7) (2013) 3149–3166, doi:http://dx.doi.org/
10.3390/en6073149.
[22] G. Singh, P. Baredar, A. Singh, D. Kurup, Optimal sizing and location of PV, wind
and battery storage for electrification to an island: a case study of Kavaratti,
Lakshadweep, J. Energy Storage 12 (2017) 78–86, doi:http://dx.doi.org/
10.1016/j.est.2017.04.003.