See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/274898992

Sizing Compressed-Air Energy Storage Tanks for Solar Home Systems

Conference Paper · June 2015

DOI: 10.1109/CIVEMSA.2015.7158620

CITATIONS READS
9 5,141

4 authors:

Andi Setiawan Ardyono Priyadi

Sepuluh Nopember Institute of Technology Sepuluh Nopember Institute of Technology
6 PUBLICATIONS 23 CITATIONS 250 PUBLICATIONS 1,309 CITATIONS

SEE PROFILE SEE PROFILE

Margo Pujiantara Mauridhi Hery Purnomo

Sepuluh Nopember Institute of Technology Sepuluh Nopember Institute of Technology
136 PUBLICATIONS 775 CITATIONS 640 PUBLICATIONS 4,079 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Ardyono Priyadi on 16 December 2015.

The user has requested enhancement of the downloaded file.

 Sizing Compressed-Air Energy Storage Tanks
for Solar Home Systems

1
A. Setiawan A. Priyadi, 2 M. Pujiantara,
Electrical Engineering Dept. and 3 M. H. Purnomo
1. Institut Teknologi Sepuluh Nopember, Institut Teknologi Sepuluh Nopember
2. Universitas Jember Email: 1 priyadi@ee.its.ac.id, 2 margo@ee.its.ac.id,
Email: andi.setiawan14@mhs.ee.its.ac.id 3
hery@ee.its.ac.id

Abstract—A generic problem of distributed solar home sys- II. P ROPOSED SHS CAES S YSTEM
tems (SHS) is the lifetime of the chemical storage battery. In
this paper, a model of compressed-air energy storage (CAES) Fig.1 shows a schematic of entire SHS CAES system. This
based SHS is developed and simulated to determine the size of system consists of a PV module equipped with a maximum
the storage tank according to the required load and operating power point tracking (MPPT) controller that drives a piston
time. Industrial air motors and permanent magnet direct current compressor run by direct current (DC) motor. The compressor
generator specifications are used for simulation model. Other increases the pressure of atmospheric air required to be stored
features include sizing of the storage tank and determining in the tank.
appropriate regulator settings to improve the system efficiency.
Downstream of the storage tank, a valve regulates the air
keywords: compressed-air, storage, modeling, renewable en- flow to run an air motor that acts as a microturbine coupled
ergy with a permanent-magnet DC (PMDC) generator. The output
of this DC generator supplies a fixed power DC load.

I. I NTRODUCTION

Inspired by successful implementations of large-scale

compressed-air energy storage (CAES), researchers are now
investigating applications of CAES for small-scale systems [1],
[2], [3], [4]. In this paper we discuss an application of small-
scale CAES to replace chemical battery storage in solar home
systems (SHSs).

SHS is a government electrification project for un-powered Fig. 1. Schematic of entire SHS-CAES
rural areas where it is difficult to use power source other than
solar power. The project is implemented by distributing small In this paper, the discussion focuses on the size of tank
solar photovoltaic (PV) systems with capacity to cover the required to store the potential energy of the compressed air
minimal electrical power requirements needed for lighting in that provides the mechanical power needed by DC generator
rural communities. that supplies a fixed power DC load for 12 hours. The design
of the rest of system (upstream of the storage tank) will be
One of the major problems of these SHS systems is discussed in future reports.
the chemical battery storage. Battery lifetime is very short
compared to that of PV solar module. Furthermore, the repair The challenge in determining the size of components,
damaged batteries cannot be undertaken using technologies especially the storage tank, lies in the properties of the compo-
that can be taught easily to local communities in rural areas. nent themselves. Furthermore, the availability of commercial
components contributes its own problems: limitation of avail-
With reference to this problem, research to develop a ability require some components to be operated outside of their
CAES-based SHS prototype was initiated by assuming that optimal or rated conditions. Modeling is therefore important
the same electrical power is required for lighting as in the to ensure that accurate results are generated by the simulation.
existing SHS. This power and the associated load hours are the
references for determining the sizes of the CAES components A. DC Load System
required.
The loads to be supplied by this CAES system are energy-
All the components needed to build the CAES system were saving DC lamps. Energy-saving lamps can typically perform
modeled. The microturbine and generator models are based on at a steady rated power across a wide range of voltage because
industrial specifications [5], [6], [7]. Other information related they use a switching power supply. The load will consequently
to CAES modeling and prototyping of small-scale CAES in draw variable input current to meet the rated operating power
[2], [3], [4], [8], [9], [10] was also employed. if operating at variable input voltage.

978-1-4799-6092-7/15/$31.00 ©2015 IEEE

 TABLE I. E XAMPLE S PECIFICATION OF T YPICAL LED D RIVER
where Lag and Rag are the inductance and resistance, respec-
Model: LD1007-12-C0700 tively, of the generator armature.
Input: Voltage range: 10.5 − 30 V ; Maximum current: 1 A
DC Output: Working voltage: 2−12V ; Constant current: 700mA ±5%
A dynamic mathematical model of the PMDC generator
Efficiency: 85% typical ; Maximum power: 8.4W wass implemented in MATLAB using the specification data
Protection: OPP, SCP-Auto recovery ; EMC Compliance: FCC, CE for MOOG C34-L60-10 [7].
Dimensions (LWH):46×34×19±0.5mm ; Weight: 65 gram typical
Operation temp.:-30o C to +50o C, Tc:70o C ; Storage Temp.:-40o C to +85o C C. Air Motor Model
A linearized models of LZB 14 AR034 (100W) [6] is used
as the turbine model that is coupled with the DC generator
Table I. shows an example of a typical 8.4 W DC light
model. Fig. 3 shows the performance diagram of this air motor.
emitting diode (LED) driver [5]. The specification indicates
that this LED driver can maintain constant DC current output
and nominal power in an input voltage range from 10.5 V to
30 V DC with a typical efficiency of 85%.

B. DC Generator Model
The PMDC generator is the simplest component to be
integrated into this SHS CAES system. It offers simplification
in modeling calculations for sizing the rest of the components.
Fig. 2 shows the dynamic model of the PMDC generator used
in this research.

Fig. 3. Performance diagram of LZB-14-AR034-11

Fig. 2. Dynamic model of PMDC generator TABLE II. S PECIFICATION OF A IR M OTOR

Model/Type LZB14-AR034-11
This generator is driven by an air motor that acts as the Data at 6.3 bars ; Min starting torque: 1.01 N-m
prime mover. The dynamic behavior of this coupled generator Stall torque: 1.25 N-m ; Free speed: 3374 rpm
Datat maximum power point
is given by [3],
Torque: 0.63 N-m ; Speed: 1680 rpm
dωram Power: 0.11 kW ; Air consumption: 3.6 l/s
Tm − (Bm + Bg )ωr − Teg = (Jm + Jg ) (1)
dt
where Tm is the mechanical input torque, Teg is the elec- These data were used to develop a model with pressure
tromagnetic torque of generator, ωr is angular speed of the (p in bar) and angular speed (ω in rpm) as the inputs. Using
rotor, Jm and Jg are the moments of inertia of the air interpolation or extrapolation with the help of a lookup table,
motor and generator, respectively, and Bm and Bg are the this model will calculate power (P m in Watt), torque (M m in
viscous coefficient of friction of prime mover and generator, Newton-meter) and air consumption (v̇ in l/s). The additional
respectively. air motor specification data, shown in Table. II. were used to
constrain the model.
The back emf of the generator, Eeg , is given by:
Eeg = Km ωr (2) D. Storage Tank Model
where Km is the speed constant of the generator. Dry air contains roughly 80% nitrogen (N2 ) which is
a diatomic gas, so it can be assumed that air is an ideal
The load torque of the coupling shaft is the generator’s diatomic gas. The relationship between the pressure, volume
electromagnetic torque, i.e., and temperature is given by Eq. (5):
Teg = Ke iag (3) pV = mRT (5)
where Ke is the torque constant and iag is the armature current where R is the specific gas constant in J/(kg.K), p is the
of the generator. pressure of the gas, V is the volume occupied by the gas,
Finally, the terminal voltage generated by the generator can m is the mass of the gas, and T is temperature measured in
be determined using Eq.4. P ascal, m3 , kg and Kelvin, respectively.
diag In this CAES system, the air motor uses flowing air mass
Vt = Eeg − Rag iag − Lag (4) from the storage tank to achieve the required speed and torque
dt
needed to turn the generator. This flowing mass (ṁ) can be B. Determining Pressure and Air Consumption of Air Motor
expressed as
V dp The air motor model was developed based on factory data
ṁ = (6) and specifications. This model requires speed and pressure
RT dt input information for the inputs to produce output results for
torque, power and air flow.
In this simulation, air mass flow, ṁ, is obtained from the
volume flow as the output of the air motor model, multiplied
by a constant ρ which is the density of air.
Integration of this mass flow rate ṁ enables calculations of
the air mass flowing out of the storage tank and the pressure
drop of the air in the storage tank, as shown in Eqs. 7 and 8.
Z
ms (t) = m(0) − ṁ dt (7)

RT
p(t) = ms (t) (8) Fig. 5. Torque and air flow as function of input pressure at 1404 rpm
V

In Eq. 7, ms (t) is the mass of air in the storage tank after The air motor model was driven with increasing values of
time t and m(0) is initial air mass. Temperature is assumed to pressure at a constant speed of 1404 rpm. The input pressure
be constant in this process. and related air consumption in this arrangement are shown in
Fig. 5.
III. S IMULATION R ESULTS AND D ISCUSSION
C. Sizing the Storage Tank
A. Performance of Loaded PMDC Generator Model
The air tank volume needed to provide sufficient stored
The PMDC generator modeled in MATLAB was linked to energy to run this CAES system was calculated by equating
the LED load model and driven with a ramp value from 0.0 Eqs. 6, 7 and 7 to give Eq. 9.
to 3.5 N-m to simulate the input torque of the generator shaft.
The results are shown in Fig. 4.
RT
V = v̇t (9)
According to the data given in Table. I, the lamp will p0 − pset
operate at constant output voltage and current if the input
voltage is between 10.5 and 30.0 V. The lowest value was where p0 and pset are the respective initial and output pressure
chosen because this requires the lowest mechanical torque to settings of the pressure regulator, v̇ is the volumetric air flow
di rate required by the air motor (determined in the previous
run the generator. Equating Eqs. 2 and 4 for dtag = 0 and
simulation), and t is backup time required to run the system.
using the data taken from [7], the speed needed to generate
Fig. 6 shows the tank size required for various backup times.
10.55V is 1404 rpm. The current is obtained by calculating
the input current of the LED driver using data in Table. I.
The torque needed by the generator to run at this speed
with this current flowing in its armature is calculated using
Eqs. 3 and 1. The torque value is 0.3264 N-m.

Fig. 6. Volume of tank as a function of backup time

D. Whole System Simulation

Simulation of entire system was performed to evaluate
other options for decreasing the storage tank size needed or
increasing the operating time for the SHS-CAES proposed in
this work.
Fig. 4. Power input and output of PMDC as a function of the input torque
Fig. 7 shows these simulation results using a 1m3 , 8 bar
storage tank. The total load of 25.2 W was divided into three
Fig. 4 shows the simulation results of the generator model loads with the same power to determine whether reducing the
run using ramp value of torque loaded with the LED lamp load could increase the operating time.
model. At a torque of 0.3264 N-m the generator produces the
29.65 W of electrical power needed by the load. Under these For this 3×8.4 W load, the pressure regulator was set
conditions, the generator performs at 61.7% efficiency. to 3.511 bar, for the other two loading options, the pressure
 Although this tank size appears very large compared to
the physical dimensions of a chemical storage battery, it still
makes sense for applications in rural areas where SHS will be
applied.
Experiments at variable load indicate that load reduction
is not always able to extend the backup time of CAES.
Investigations of the efficiency related to the regulator setting
indicate that, if it is possible to use a boost converter with
a wider input voltage range and with an adequate level of
efficiency, then reduction in the size of the required storage
tanks can be expected.
Further work is required to evaluate the components up-
stream of the storage tank and to complete this assessment
of the possibility of this system replacing conventional SHS
arrangements.

R EFERENCES
Fig. 7. Simulation of SHS-CAES backup time: 8 bar, 1m3 , using a variable [1] S. Vazquez, S.M. Lukic, E. Galvan, L.G. Franquelo, J.M. Carrasco.,
load ”Energy storage systems for transport and grid applications,” IEEE
Transactions on Industrial Electronics, vol.57, pp. 3881-3895, 2010.
[2] G. Kleiser, V. Rauth, ”Dynamic modelling o compressed air energy
regulator was set to 3.0 bar. Setting below this value are not storage for small-scale industry applications”, IJEE 2013,vol. 3, no.3,
recommended for this air motor [6]. pp. 127-137, 2013.
[3] V. Kokaew, M.M. Torbati, S.M. Sharkh, ”Maximum efficiency or power
The results shows that reducing the load does not always tracking of stand-alone small scale compressed air energy storage
increase the backup time. In this case, reducing the load to system, in Proc. Energy Procedia, vol. 42, pp. 387396, Mediterranean
1×8.4 W does not increase the backup time because the Green Energy Forum, 2013.
pressure regulator setting is not simultaneously reduced. This [4] D. Villela, V. V. Kasinathan, S. D. Valle, M. Alvarez, G. Frantziskonis,
P. Deymier, K. Muralidharan., ”Compressed-air energy storage system
increases the speed of the generator and air consumption of for stand-alone off-grid photovoltaic modules,” in Proc. Photovoltaic
the air motor. Specialists Conference (PVSC),35th IEEE, 2010.
[5] MAGTECH Industries Corp., LD1007 Series Product Specification:
High Power Constant Current LED Driver, ANZ:Z140, 2006.
[6] ATLAS COPCO AIR MOTOR, A Full Range of Compact, High Perfor-
mance Air Motors (product catlogue), avaliable on www.atlascopco.com
[7] Moog Components Group, C23 Series Specification Data Sheet,
1213 N Main Street Blacksburg 24060 USA, available on
www.moog.com/components
[8] V. Vongmanee, V.Monyakul, ”Modelling of a compressed air enenrgy
electrification by using induction generator based on field oriented
control principle,” JEET vol. 9, no. 5, pp. 1511-1519, 2014.
[9] G. Grazzini and A. Milazzo, ”Thermodynamic analysis of CAES/TES
systems for renewable energy plants”, Renewable Energy,vol. 33, no.
9, 2008.
Fig. 8. Efficiency performed around regulator setting [10] H. Niklas, O. Vhringer, C. Kruck, L. Eltrop, ”Simulation and analysis of
different adiabatic compressed air energy storage plant configurations”,
Applied Energy vol.93, pp. 541548, Elsevier Ltd., 2011.
Fig. 8 presents simulation results that show the efficiency
at selected regulator settings. A regulator setting of 3.4 bar
performed most efficiently for the required load power (29.65
W), but this load will suffer from an under-voltage condition
because the switching power supply requires a minimum
voltage of 10.5 V.
On the other hand, setting the regulator to 3.6 bar decreases
efficiency, which results in shorter backup time for this load.

IV. C ONCLUSION
The sizing of storage tanks for a SHS-CAES has been
determine by modeling all the components downstream of the
storage tank. To operate the system with SHS load of 29.65
W for 12 hours requires a tank size of 18 m3 , with an initial
pressure of 8 bar and regulator setting 3.511 bar.

View publication stats