Accepted Manuscript

Experimental study of carbon dioxide as working fluid in a closed-loop

compressed gas energy storage system

Abdul Hai Alami, Abdullah Abu Hawili, Rita Hassan, Mohammed Al-Hemyari,
Kamilia Aokal

PII: S0960-1481(18)31365-X

DOI: 10.1016/j.renene.2018.11.046

Reference: RENE 10809

To appear in: Renewable Energy

Received Date: 08 September 2018

Accepted Date: 11 November 2018

Please cite this article as: Abdul Hai Alami, Abdullah Abu Hawili, Rita Hassan, Mohammed Al-
Hemyari, Kamilia Aokal, Experimental study of carbon dioxide as working fluid in a closed-loop
compressed gas energy storage system, Renewable Energy (2018), doi: 10.1016/j.renene.
2018.11.046

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form.
Please note that during the production process errors may be discovered which could affect the
content, and all legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

1 Experimental study of carbon dioxide as working fluid in a closed-loop

2 compressed gas energy storage system
3 Abdul Hai Alami1,2*, Abdullah Abu Hawili3, Rita Hassan4, Mohammed Al-Hemyari5 and
4 Kamilia Aokal2
5 1Sustainable and Renewable Energy Engineering, University of Sharjah, POBox 27272, Sharjah,
6 United Arab Emirates
7 2Center for Advanced Materials Research, Research Institute of Science and Engineering (RISE),
8 University of Sharjah, POBox 27272, Sharjah, United Arab Emirates
9 3Sharjah Research Institute, Sharjah, United Arab Emirates
10 4Mechanical Engineering Department, University of Sharjah, POBox 27272, Sharjah, United
11 Arab Emirates
12 5Mechanical Engineering Department, American University of Sharjah, POBox 26666, Sharjah,
13 United Arab Emirates
14 *corresponding author, email: aalalami@sharjah.ac.ae, phone: +971-56-160-5355, fax: +971-6-
15 5585191
16

17 Abstract

18 This paper investigates the utilization of carbon dioxide gas available in mass pressurized storage
19 caverns as a working fluid for a modular low pressure compressed gas energy storage (CGES)
20 system. The system is made up of three 7 liter cylinders that discharge into an air turbine to convert
21 the system potential energy into kinetic and eventually into electricity through an onboard
22 generator. The operating pressures are kept low (around 3 bar) in order to maintain the adiabatic
23 operational assumption and not necessitate the usage of heat exchangers. The gas is then rerouted
24 back into the storage cavern after expansion. The reported energy conversion efficiency is 46.2 %
25 for the three active cylinders operating in tandem and 76.4 % for them operating in unison. The
26 main advantage of the proposed system is its flexibility to function under high power density for
27 the latter or high energy density requirements for the former arrangement, respectively. This level
28 of discharge control allows the system to serve in ranges that previously required independent
29 storage systems addressing narrow power/energy density needs. This charge/discharge cycle takes
30 advantage of the availability of the stored and pressurized carbon dioxide, which is denser than air
31 and hence produced higher power output and required 65% less space than its air-operated
32 counterpart.

33

34 Keywords: carbon dioxide storage; modular compressed air storage; energy storage systems

1
 ACCEPTED MANUSCRIPT

35 Nomenclature
36 A cross sectional area [m2]
37 I current [A]
38 I air turbine moment of inertia [kg.m2]
39 P pressure [Pa]
40 M Mach number
41 t discharge time [s]
42 V total volume of cylinders [m3]
43 v voltage [V]
44 w air turbine angular speed [rad/s]
45 Subscripts
46 t total (stagnation) condition
47 RMS root mean square value
48 Superscripts
49 * critical (sonic, M=1) condition
50 Greek symbols

51  efficiency

52  polytropic constant (specific heat ratio)

53

54 1. Introduction:

55 Carbon dioxide is a direct product of complete combustion of fossil fuels and biological processes
56 in living organisms. It is the most common greenhouse gas in the atmosphere and is a direct
57 contributor to adverse global climate change. With the continuous increase in global energy
58 demand, more fossil fuels are burnt leading to unprecedented increase in CO2 emissions [1].
59 Carbon capture and storage (CCS) refers to a series of technologies and processes where CO2 is
60 captured at the source and injected underground or in containers [2]. CO2 is later stored in
61 geological storage sites as supercritical fluid [3]. CCS allows low-emissions coal power generation
62 to exist, introducing a solution to the system instead of changing its structure. On the other hand,

2
 ACCEPTED MANUSCRIPT

63 CCS is not a new technology as the injection of CO2 into depleted oil wells is a common practice
64 since the 1970s [4]. Some of the challenges that face CO2 storage in geological reservoirs are:
65 public acceptance, lack of detailed regional studies, limited number of the numerical and 3D
66 models that describe the reservoirs, in addition to financial and political issues [5]. Moreover,
67 complications stemming from seismic events, geochemical changes, health effects on humans and
68 animals due to CO2 leakage and the relatively high cost of transportation are some of the main
69 barriers of applying it on a wide scale [6] [7]. Lund and Mathiesen, have reported on the role and
70 feasibility of CCS, showing that the as-is system is not feasible and more cost-effective solutions
71 shall be studied [8]. However, since CCS is one of the leading climate change control tools
72 available, it cannot be abandoned, and further development is needed [9]. Therefore, the need to
73 replace simple geological storage of CO2, the economic cost reduction and efficient storage and
74 monitoring of CCS was the subject of various research efforts [10] [11]. One of the most promising
75 alternatives of the existing CCS system is utilizing the captured CO2 for various industrial
76 processes rather than generating more for a given purpose. Table 1 below summarizes some of the
77 industrial applications that are reported to reuse the captured CO2.

78 Table 1. Industrial applications of CO2

Ref. Description Method

[12] Ca-silicates in the transformation of potassium feldspar after adding potassium

calcium chloride hexahydrate at 800 ̊C. Ca-silicate can transform the feldspar to
CO2 into stable calcium carbonate up to 84.7%. minimize CO2
emissions

[13] mineral that exists in nature such as Olivine and mafic are effective CO2 carbonation
[14] CO2 carbonation reactions factors. However, experiments showed using minerals
that Olivine dissolution is more effective than the pure Olivine where
the fluid penetration along grain boundaries allowed an increase in
the Olivine reactive surface area. Moreover, water vapor enhanced
the carbonation capacity according to experiments done.

[15] Electrochemical, bioelectrochemical and photoelectrochemical Converting CO2

methods are used to convert CO2 into useful products for the industry. to industrial
products

3
 ACCEPTED MANUSCRIPT

[16] Three CO2 utilization methods are used in china, CCR model is used CCR, BCC and
for chemical applications, biological utilization using BCC model SE-CCR DEA
and for the geological utilization SE-CCR DEA model is used.

[17] The high purity CO2 stream is going to be recycled within the CO2 utilization
methanol synthesis plant, reduce the consumption of natural gas (NG) in methanol
and methane up to 25.6% and reduce the greenhouse gas emissions synthesis plant.
(GHG) of the methanol synthesis plant fueled by NG.

[18] Post combustion CO2 captured is fed into the Fischer-Tropsch Gas- CO2 used for
to-liquid (GTL) plant steam-methane reformer SMR. Where results generation of
showed that 105.5 tones/h of CO2 is utilized in one train of Sasol Oryx liquid fuels.
GLT plant.

79

80 Coupling appropriate storage technologies with renewable energy sources is an important decision
81 befalling energy engineers, especially due to the double uncertainty in both supply and demand.
82 [19] [20]. Battery storage is the most popular choice for most technologies, but it is far from ideal,
83 as various applications require different levels of power or energy densities, and hence the values
84 of round-trip efficiency would suffer. As a consequence, alternative technologies are needed and
85 are indeed being developed to minimize the number of energy conversion stages (and hence losses)
86 in both charge and discharge directions. These technologies include, but are not limited to,
87 flywheels [21] for direct kinetic energy storage, and Pumped Hydro Storage (PHS) [22], and
88 compressed air energy storage (CAES). The latter is one of the most promising methods of bulk
89 energy storage; leading to many researchers studying them and working on their methods of
90 developing them [23]. Its ability of regulating the variable energy sources based on the demand is
91 one of the most attractive qualities [24]. This technology is being associated with the wind energy
92 storage in many occasions [25]. Many creative studies on developing the CAES system and
93 reducing its cost were done, these include but are not limited to: utilizing heat waste from the
94 compressor [26], and combining the system with a gas engine, heat exchangers and refrigeration
95 system [27]. Moreover, Long Xiang Chen et al. have shown 2% efficiency increase by adjusting
96 the throttling to keep an adiabatic system [28], and 6.26% using volatile fluid principle to create
97 an isobaric adiabatic CAES system [29]. In addition, a previous work by Alami et al. have shown
98 a Low pressure, modular (CAES) system for wind energy storage applications. Such a modular

4
 ACCEPTED MANUSCRIPT

99 system could be easily controlled for power and energy applications, or scaled up by simply adding
100 more air containers to the systems, hence its utilization in this research [30].

101 This paper investigates the utilization of carbon dioxide as a working fluid for a modular low
102 pressure compressed gas energy storage system. The system is closed-loop, drawing carbon
103 dioxide potentially from underground caverns into a number of pressurized cylinders (three in the
104 current study) where CO2 is kept at pressures 2, 2.5 and 3 bar. The cylinders exhaust into an air
105 turbine/generator system that allows the conversion of the potential energy into electrical energy.
106 The discharge from the cylinders happens either sequentially for applications requiring high
107 energy density or in unison for high power density applications. A test rig has been designed to
108 identify important experimental parameters, including carbon dioxide source, pressure regulator
109 and heater to control the pressure and temperature, respectively, of the carbon dioxide entering the
110 cylinders. The operation at the specified pressure range is consistent with the adiabatic
111 assumptions and expected ideal gas behavior of the carbon dioxide. Voltage and current
112 measurements are made for each discharge case, and the energy, power and overall system
113 efficiency are calculated for each case and compared to similar compressed air energy storage
114 (CAES) systems.

115

116 2. Theoretical analysis:

117 The system under consideration is designed to deliver at both high power and high energy
118 density requirements, depending on the modulation of its discharge time (typically between 0.01-
119 100 s) [30]. Since the mass density of CO2 is significantly higher than that of air, it is expected
120 that the output power will be higher than an equivalent air-based CAES [31]. On a Ragone plot,
121 the system with low discharge times (simultaneous discharge of pressurized modules) corresponds
122 to a storage system in the level of flywheels and super capacitors storage systems. On the other
123 hand, in the case of sequential discharge of pressurized modules, the system is considered to be a
124 high energy density similar to battery storage systems. In either case, the expected power density
125 is higher than super capacitors power, but for energy density it’s almost equal to the batteries [30]
126 [32] [33].

5
 ACCEPTED MANUSCRIPT

127 The simulation approach will assume polytropic behavior to arrive at results that best match
128 the expected response to compression/decompression cycles [34]. Also, the temperature variation
129 during compression/decompression will be modeled as external heat addition/rejection to allow
130 adiabatic and isentropic assumptions to be valid. Since the flow will be maintained subsonic, the
131 critical pressure ratio is limited to less than unity in order to avoid supersonic flow and pertinent
132 shock waves. For the reference CAES system, the critical pressure ratio is 0.5283 [30].
133 Shockwaves are expected to occur at a nozzle ratio equals 1, whereas the experimental area ratio
134 for this system is 2 [30] [35].

135 The above assumptions will be imposed using heated 150W-210W regulators retain the
136 inlet temperature at 20 ˚C, the inlet volume flow rate is 25 L/min, and the inlet pressure varies at
137 2, 2.5 and 3 bars.

138 For isentropic operation, the following pressure relation can be applied:

𝛾+1
(𝛾 ‒ 1) 2 2(𝛾 ‒ 1)
𝑃𝑡2 1+[ 2] 𝑀
2
139 𝑃𝑡1
=( (𝛾 ‒ 1)
) (1)
2
1+[ 2] 𝑀
1

140 where γ is the specific heat ratio (equals 1.28 for carbon dioxide).

141 P1=Pt1 and M1=0 since the stagnation conditions are applied on the inlet.

142 To fix the exit conditions knowing that P1=Pt1=Pt2, the following equation is used:

𝛾
𝑃2 1 (𝛾 ‒ 1)
143 𝑃𝑡2
=( (𝛾 ‒ 1)
) (2)
2
1+[ 2]𝑀
2

144 From the previous equation the outlet pressures can be determined at M=1, and the inlet
145 pressure will be varied between 2, 2.5 and 3 bar (gauge). Hence the minimum values of the outlet
146 pressure are 1.0987, 1.3737 and 1.6484 bar. The estimation of the maximum outlet velocity at
147 different pressures enables obtaining the overall and mechanical efficiencies of the system.
148 However, the maximum theoretical outlet velocity of the flow at M=1 equals the speed of sound
149 in CO2 which is 266.25 m/s [36]. The actual value of Mach number equals to 0.215, and the

6
 ACCEPTED MANUSCRIPT

𝐴
150 maximum actual velocity at the exist can be obtained from the isentropic relations (at ∗ = 2.8 and
𝐴
𝑇
151 ∗ = 1) to be 57.22 m/s.
𝑇

152 The overall efficiency of the system is affected by losses due to friction, fittings,
153 compressibility which will decrease the velocity and expansion losses. Mechanical losses will
154 occur in the bearings of the turbine, and that will affect the mechanical efficiency [30] .

155 Finally, to obtain the efficiency the following relation are used:

𝐼. 𝑣. 𝑡
156 𝜂𝑜𝑣𝑒𝑟𝑎𝑙𝑙 = 𝑉.∆𝑃
(3)

157 The mechanical efficiency is calculated using:

1 2
2
𝐼𝜔
158 𝜂𝑚𝑒𝑐ℎ = 𝑃. 𝑉
(4)

159 3. Experimental Setup:

160 High pressure CO2 cylinders with an electrical heated CO2 regulator with maximum output
161 pressure of 3.5 bar were used to fill three 7-liter steel cylinders which are attached to three analog
162 pressure gauges (7 bar max), the cylinders are connected in parallel via 1/4 inch-diameter PVC
163 pipe network and brass fittings. To better control the charge cycle, three manual air valves (10 bar
164 max) are installed at the inlet of each cylinder, with a one-way valve similarly installed at the
165 cylinder inlet to prevent the backflow of gas. The discharge cycle is controlled via a PLC (Siemens
166 LOGO 230 RC, 220 AC, 8 inputs) directing three 5/2-way solenoid valves (4V210-08, 220 AC
167 input, 1.5e8 bar operational pressure) that would either open in unison to give maximum power
168 density, or in sequence to provide maximum energy density.

169 The air turbine/electrical generator assembly (Figure 1) is manufactured by Hüttinger

170 (http://www.huettinger.de). The air turbine handles a maximum torque of 800 N cm, while the
171 electrical generator produced 24 V and 33 A at a maximum rotational speed of 3350 rpm, with a
172 15:1 gearbox that transmits the motion from the turbine output to the generator input.

7
 ACCEPTED MANUSCRIPT

173

174 (a)

175

176 (b)

177 Fig. 1 (a) schematic of the experimental setup and (b) the actual system

178 The electrical output (current and voltage) from the generator is recorded by an NI 9221 C Series
179 Voltage Input Module that is attached to an NI cRIO-9025, which records a measurement each
180 500ms. The current is measured through a connection to a 1ohm power resistor. The unit is
181 connected to a computer via an Ethernet connection where LabVIEW software is used to log, plot,
182 and save the data to Excel documents. LUTRON DT-2259 Digital Photo Tachometer
183 w/Stroboscope is used to record the angular velocity of the turbine/generator experimentally to be
184 compared with their theoretical counterparts.

8
 ACCEPTED MANUSCRIPT

185 4. Results and discussion:

186 4.1. Time constant and energy quality approximation

187 It is important to determine the expected values for discharge times at various starting pressures.
188 The voltage values are measured for the 3, 2.5 and 2 bar settings and the RMS values are found to
189 be 20.35, 17.38 and 13.1 V and the discharge times are 43, 35.5 and 31.5 seconds, respectively.
190 The discharge plots are shown in Figure 2 (a-c) for the three respective cases of discharge to the
191 gauge pressure of 1 bar.

192
30 25
Exp
Exp
3 bar 2.5 bar
25
20
V =20.35V V =17.38 V
RMS RMS
20
Voltage/ V

Voltage/ V
15
15
10
10

5
5
Fit Fit

0 0
0 10 20 30 40 0 10 20 30 40
193 Discharge time/ s Discharge time/ s

194 (a) (b)

20
18 Exp 30
2 bar
16
25
14 V RMS=13.1 V
3 bar
RMS voltage/ V

20
Voltage/ V

12
2.5 bar
10
15 2 bar
8
6 10

4
5
2 Fit
0 0
0 5 10 15 20 25 30 30 32 34 36 38 40 42 44
195 Discharge time/ s Discharge time/ s

196 (c) (d)

197 Fig. 2. Discharge times and RMS voltage for single cylinder full discharge for (a) 3 bar, (b) 2.5 bar and (c) 2 bar cases. (d)
198 Energy quality for the three cases in terms of RMS voltage and discharge duration

9
 ACCEPTED MANUSCRIPT

199

200 The voltage and discharge time for the 3 bar case is understandably highest and thus the expected
201 energy quality is also maximum. This is shown in Figure 2 (d) as the voltage and time progressively
202 increase with the increase of starting pressure. The RMS value will be considered as the benchmark
203 for voltage and current estimation for subsequent tests.

204 4.2. Simultaneous discharge

205 For high power density requirements, the simultaneous discharge from the three cylinders is
206 deployed. The maximum voltage values cannot be maintained for long periods of discharge times
207 due to the rapid decay of pressure within the cylinders. The RMS values of voltage, however, are
208 sustained for up to 100 seconds for the 3 bar starting pressure case at 22.4 V, which is far more
209 than the duty cycle of a supercapacitor [37]. The voltage variation with time shows identical trends
210 for the three simultaneous discharge from the three cylinders at 3, 2.5 and 2 bar, and is shown in
211 Figure 3 (a-c).

35 30
Exp
Exp
30 25

25 VRMS =20.2 V
VRMS =22.4 V 20
Voltage/ V
Voltage/ V

20
15
2.5 bar
15 2.5 bar
3 bar
10
10

5
5
Fit Fit

0 0
0 20 40 60 80 100 0 20 40 60 80 100
212 Discharge time/ s time [s]

213 (a) (b)

10
 ACCEPTED MANUSCRIPT

25
Exp
20
V =16.4 V
RMS

Voltage/ V
15

2 bar
10

5
Fit
0
0 20 40 60 80
214 Discharge time/ s
215 (c)

216 Fig. 3. Sequential discharge for (a) 3 bar, (b) 2.5 bar and (c) 2 bar

217

218 4.3. Sequential discharge

219 The sequential discharge is needed for high energy density, where the variation in the supplied
220 voltage out of the storage cylinders is expected to be constant. The performance depicted in Figure
221 4 (a-c) for the three respective starting pressures also show consistent trends, as each cylinder is
222 discharged around the 3.2 second mark, which is consistent with the primary tests that determined
223 the appropriate discharge time that preserves energy quality (see Figure 2). The energy density
224 discharge time is programmed via the PLC controller and can be varied to provide various levels
225 of total discharge times depending on the desired application. There is a 15 % loss in output
226 voltage, however, for longer discharge time per cylinder due to the loss of energy quality that is
227 maximum in the range 3-2.5 bar, below which the voltage decay is seen to be more rapid (see again
228 Figures 2).

229

11
 ACCEPTED MANUSCRIPT

35
35
Exp
30
30 Exp
25 V
RMS
=22.8 V 25
V =21.3 V
Voltage/ V

RMS

Voltage/ V
20 20

15 15

10 10
3 bar Fit 2.5 bar
5 5 Fit

0 0
0 2 4 6 8 10 0 2 4 6 8 10
230 Discharge time/ s Discharge time/ s

231 (a) (b)

25
Exp

20
V RMS=16.2 V
Voltage/ V

15

10

5 2 bar

Fit
0
0 2 4 6 8 10
232 Discharge time/ s

233 (c)

234 Fig. 4. Sequential discharge for (a) 3 bar, (b) 2.5 bar and (c) 2 bar

235 During experiments, the sequential discharge cycle that starts with 3 bar in Figure 4 (a) shuts the
236 solenoid of the first cylinder while the cylinder retains around 2.5 bar of pressure. Thus, the three
237 graphs of Figure 4 can be thought of as a continuation to one another in discharge time, where the
238 total discharge time for each cylinder is estimated from Figure X1. For example, if the three
239 cylinders are to be discharged down to 1 bar from a 3 bar starting pressure, the total discharge time
240 is 43 x 3 ≈ 130 second at a RMS voltage of 20.4 V according to Figure 2.

241 4.4. Kinetic energy and system efficiency analysis

242 The air turbine rotational velocity for the three experimental protocols is plotted against the
243 starting pressure in Figure 5. The tachometer records the highest, lowest and final rotational
244 velocity value for each experimental run.

12
 ACCEPTED MANUSCRIPT

245 For the current discharge conditions, the sequential discharge shows the highest velocity values
246 (around 4750 RPM) and the smallest range of variation. There is an overlap between the single
247 cylinder discharge and all-cylinders discharge, but the advantage goes to the latter case,
248 especially at higher discharge pressures.

249
5250

Sequential
4500
cylinder
discharge
Rotatinal Speed / rpm

3750 region

3000

2500 All
cylinder X Single
discharge cylinder
1500 region discharge
region
750

0
1.5 2 2.5 3 3.5
250 Pressure/ bar

251 Fig. 5. Air turbine rotational velocity

252

253 Since the system flexibility comes from controlling the discharge time, it is expected that imposing
254 less stringent energy quality requirements on the sequential operation will increase the overlap
255 between it and all-cylinder operation region. In any case, the maximum achievable kinetic energy
256 for the maximum rotational velocity is around 125 kJ/kg of the rotor mass, accredited to the high-
257 performance air turbine used for experiments.

258 The overall efficiency of the system for various operating conditions is summarized in Table 2

259 Table2. Summary of overall system efficiency at various operating conditions

Operation mode Single cylinder All cylinders Sequential

Initial pressure/ bar 2 2.5 3 2 2.5 3 2 2.5 3
Efficiency/ % 24.0 48.6 79 35.2 61.2 76.4 18 36.8 46.2
260

13
 ACCEPTED MANUSCRIPT

261 The overall efficiency is highest for the complete discharge of a single cylinder at 3 bar starting
262 pressure, followed closely by the simultaneous discharge of all cylinders at 3 bar. The higher
263 energy quality at 3 bar for all operating modes manifest itself with the reported higher efficiency
264 values. It is important to note that the low power compressor is expected to consume some of the
265 power during periods of low demand to create the closed loop operation. Some available piston-
266 type, 1/8 hp (from Grex USA http://www.grexusa.com/grexairbrush/products.php5?id=AC1810-B that operates
267 at 20 L/min and produces an outlet pressure of 4 bar) compressor will operate over an extended
268 period of time (6-8 hours) to evacuate the CO2 collection container shown in Figure 1 (a), and thus
269 is not expected to operate at its full power. Also, the compressor will not overlap with the
270 charge/discharge operation of the cylinders and thus will not affect the overall efficiency of the
271 system.

272 4.5. Ragone plots

273 To compare the system response under various load requirements, the Ragone plot is produced
274 and shown in Figure 6 for the sequential operation as well as for the simultaneous discharge of all
275 cylinders.

276

5
Sequential
operation
Specific energy/ J/kg

4 All cylinders
complete
depletion
3

2 All cylinders
80%
depletion
1

0
60 80 100 120 140 160 180 200 220
277 Specific power/ W/kg

278 Fig. 6. Ragone plot for sequential and two cases of simultaneous discharge: complete and up to 80% depletion

14
 ACCEPTED MANUSCRIPT

279 Since the sequential discharge is meant to provide better energy coverage, it reigns this range in
280 the Ragone plot. The complete discharge of all cylinders shows the expected higher specific power
281 output while being unable to cover ranges lower than 120 W/kg. It is interesting to note that if the
282 cylinders are not completely discharged, their performance in terms of energy density is
283 completely restricted, while still capable of providing the same level of power density as when
284 completely discharged. The only advantage of the latter case is shorter charging times especially
285 when operating with multiple cylinder sets (multiples of three), where depleted units are charged
286 as the operational ones are discharging if the needs arise.

287 4.6. Comparison with other storage methods

288 The widely used Lithium ion batteries are regarded as the most electrical storage devices. The
289 current gas storage system has a gravimetric energy of 0.3 to 1.5 Wh/kg and power density of 120
290 to 220 W/kg. Table 3 shows a comparison with the Li-ion batteries and supercapacitor storage
291 technologies (see https://www.supercaptech.com/battery-vs-supercapacitor).

292 Table 3. Summary of overall system efficiency at various operating conditions

Feature Li-ion Battery Supercapacitor CGES

Gravimetric energy (Wh/kg) 100 - 265 4 to 10 0.3 to 1.5
Volumetric energy (Wh/L) 220 - 400 4 to 14 >7
Power density (W/kg) 1,500 3,000 - 40,000 121 to 220
Voltage of a cell (V) 3.6 2.7 - 3 16.2 to 22.8
ESR (mΩ) 500 40 - 300 NA
Max. efficiency (%) 75 - 90 98 79
Cyclability (nb recharges) 500 - 1,000 500,000 500,000
Life 5 - 10 years 10 - 15 years >20 years
Self-discharge (% per month) 2 40 - 50 (descending) 10%
Charge temperature °C 0 to 45°C -40 to 65°C 20 °C
Discharge temperature °C -20 to 60°C -40 to 65°C 20 °C
Deep discharge pb Yes No Yes
Overload pb Yes No No
Risk of thermal runaway Yes No No
Risk of explosion Yes No No
Charging 1 cell Complex Easy Easy
Charging cells in series Complex Complex Easy
Voltage on discharge Stable Decreasing Stable
Cost per kWh 200 - 1,000 € 10,000 € 200 - 400 €
293

15
 ACCEPTED MANUSCRIPT

294 The compressed gas energy storage system stands out in terms of cost, safety and cyclability. Also,
295 the chemical, thermal and electrical stability of the system makes it a natural contender for
296 traditional storage technologies, especially when directly coupled with a charging mechanism that
297 used excess mechanical energy, for example from a remote wind farm during low demand times.
298 Compared to the similar CAES system developed by our group and detailed in reference [24], the
299 3 bar operation of using carbon dioxide as a working fluid significantly reduces the number of
300 equivalent cylinders needed to be equivalent to energy storage in batteries by 47-61%. Table 4
301 shows the number of 7 liter cylinders (ϕ= 15 cm and L=40 cm)

302 Table 4. Equivalent cylinders required for the operation of one battery

Total cylinders required

Pressure (bar) Battery capacity (AH)
CAES CGES Cylinders less
3.00 7 1483 900 61 %
3.00 4.2 839 542 65 %
3.00 1.12 226 109 48 %
3.00 0.875 181 86 47 %
303

304 The 7 liter cylinder is 15 cm diameter and 45 cm length. This means that 100 cylinders require less
305 than less than a meter cube (0.8 m3) of storage space for the 24 h operation of various capacity
306 batteries. The 7 AH battery is capable of operating compressors for air conditioning and
307 refrigerating units requires around 7m3 of storage space, which can easily be housed in a small
308 portion of a standard storage room measuring less than 2 x 3 x 3 m in dimensions. The number of
309 cylinders can also be reduced using bigger cylinders or operating at higher pressures. The latter is
310 not recommended in the case of CGES system to minimize the heat transfer issues of cyclic system
311 operation.

312 4.7. Advantages and Disadvantages

313 In general, the application of CO2 as a working fluid has enabled the operation at lower starting
314 pressures and required a smaller number of storage cylinders than its compressed air counterpart.
315 There are also some other important points of comparison that are highlighted in Table 5.

316

16
 ACCEPTED MANUSCRIPT

317 Table 5. Advantages and disadvantages of utilizing carbon dioxide as a working fluid

Advantage Disadvantage
Offers both power and energy density options for Setup, pressure accumulation, programming require
small scale operation a long lead time.
CO2 density higher than air allowing for higher Higher pressures than 3 bar will require heat
energy output at lower pressures compensation
CO2 being free of moisture System maintenance and trouble shooting can be
challenging
Can be automated with minimal human interaction Individual pressure regulators are required per
during operation cylinder
One compressor could be used to fill up the tanks. Individual solenoids are needed per cylinder
Easy to fill tanks in series.
Generators are added in parallel for higher output. Cost of modular air generator is high
Power that are able to handle starting surge
Larger cylinders could be used as lower pressures are -
used for maximum power output
Layout could be customized per energy requirement. -
Valves could be added to deter compressed gas to
main? modular systems.
Low cost per kWh compared to other -
electrochemical storage devices
High product lifecycle with comparable efficiencies -
with existing electrochemical storage devices
318

319 Conclusion

320 This work presents the design and operation of a closed-loop compressed gas energy storage (CGES)
321 system with carbon dioxide as a working fluid. The carbon dioxide is provided either from large storage
322 caverns, or by direct capture from the environment, then sent back to storage using low power compressor.
323 The main advantage of using CO2 gas as a working fluid is its higher density advantage over standard air,
324 allowing operation at higher levels of efficiency and energy output, as well as being a single component
325 gas that is free of moisture. The experimental setup featured three cylinders operating either in single,
326 sequential or simultaneous discharge cycles to drive an air turbine/generator setup. A maximum starting
327 pressure of 3 bar is used to mitigate heat transfer losses during charge/discharge cycles and enhance the
328 safety features. The overall efficiency of the system is shown to be around 79 %, with an operational
329 flexibility that provides high energy or high power density depending on the application. This flexibility
330 stems from programmable solenoid valves that allows any combination of discharge to take place, making
331 the system suitable not only for stationary applications, but for small vehicle propulsion and energy storage.
332 The system is also compared against a similar CAES system, with obvious advantages in terms of
333 efficiency, energy output and small physical footprint. Since the system operates in a closed loop, it could
334 be housed within the cavern where CO2 is stored to minimize the impact of any leakage to the surrounding

17
 ACCEPTED MANUSCRIPT

335 environment, and hence there is no expected detrimental impact on the environment in the operational
336 horizon of the proposed system.

337 Acknowledgement

338 We wish to acknowledge the University of Sharjah for their funding support through grant No: V.C./ G.R.
339 C./S.R. 356/2015

18
 ACCEPTED MANUSCRIPT

340 References
341

[1] R. S. Haszeldine, "Carbon Capture and Storage: How Green Can Black Be?," SCIENCE, vol. 325, no.
5948, pp. 1647-1652, 2009.

[2] M. L. Szulczewski, C. W. MacMinn, H. J. Herzog and R. Juanes, "Lifetime of carbon capture and
storage as a climate-change mitigation technology," PNAS, vol. 109, no. 14, p. 5185–5189, 2012.

[3] J. Gibbins and H. Chalmers, "Carbon capture and storage," Energy Policy, vol. 36, pp. 4317-4322,
2008.

[4] B. Praetorius and K. Schumacher, "Greenhouse gas mitigation in a carbon constrained world: The
role of carbon capture and storage," Energy Policy, vol. 37, pp. 5081-5093, 2009.

[5] M. D. Aminu, S. A. Nabavi, C. A. Rochelle and V. Manovic, "A review of developments in carbon
dioxide storage," Applied Energy, vol. 208, pp. 1389-1419, 2017.

[6] Z. Zhang and D. Huisingh, "Carbon dioxide storage schemes: Technology, assessment and
deployment," Journal of Cleaner Production, vol. 142, pp. 1055-1064, 2017.

[7] S. Spiecker, V. Eickholt and C.Weber, "The impact of carbon capture and storage on a decarbonized
German power market," Energy Economics, vol. 43, pp. 166-177, 2014.

[8] H. Lund and B. V. Mathiesen, "The role of Carbon Capture and Storage in a future sustainable
energy system," Energy, vol. 44, pp. 469-476, 2012.

[9] M. E. Boot-Handford, J. C. Abanades, E. J. Anthony, M. J. Blunt, S. Brandani, N. M. Dowell, J. R.

Fernandez, M. C. Ferrari, R. Gross, J. P. Hallett, R. S. Haszeldine, P. Heptonstall, A. Lyngfelt, Z.
Makuch, E. Mangano and R. T. J. Porter, "Carbon capture and storage update," Energy &
Environmental Science, vol. 7, no. 1, pp. 130-189, 2014.

[10] P. Markewitz, W. Kuckshinrichs, W. Leitner, J. Linssen, P. Zapp, R. Bongartz, A. Schreiber and T. E.

Muller, "Worldwide innovations in the development of carbon capture technologies and the
utilization of CO2," Energy & Environmental Science, vol. 5, pp. 7281-7305, 2012.

[11] N. Heinemann, M. Wilkinsona, G. E. Pickupb, R. S. Haszeldine and N. A. Cutler, "CO2 storage in the
offshore UK Bunter Sandstone Formation," International Journal of Greenhouse Gas Control, vol. 6,
pp. 210-219, 2012.

[12] X. HePing, W. YuFei, J. Yang, L. Bin, Z. JiaHua, Z. Ru, X. LingZhi, L. Tao, Z. XiangGe, Z. HongMei, L.
Chun and L. HouFang, "Simultaneous mineralization of CO2 and recovery of soluble potassium
using earth-abundant potassium feldspar," Chinese Science Bulletin, vol. 58, no. 1, pp. 128-132,
2013.

[13] S. Kwon, M. Fan, H. F. M. DaCosta and A. G. Russell, "Factors affecting the direct mineralization of
CO2 with olivine," Environmental Sciences, vol. 23, no. 8, pp. 1233-1239, 2011.

19
 ACCEPTED MANUSCRIPT

[14] R. V. Noort, C. Spiers, M. Drury and M. Kandianis, "Peridotite dissolution and carbonation rates at
fracture surfaces under conditions relevant for in situ mineralization of CO2," Geochimica et
Cosmochimica Acta, vol. 106, pp. 1-24, 2013.

[15] N. S. Spinner, J. A. Vegaa and W. E. Mustain, "Recent progress in the electrochemical conversion
and utilization of CO2," Catal. Sci. Technol, vol. 2, no. 1, pp. 19-28, 2012.

[16] J.-L. Fan, X. Zhang, J. Zhang and S. Peng, "Efficiency evaluation of CO2 utilization technologies in
China: A super-efficiency DEA analysis based on expert survey," Journal of CO2 Utilization, vol. 11,
pp. 54-62, 2015.

[17] D. Milani, R. Khalilpour, G. Zahedi and A. Abbas, "A model-based analysis of CO2 utilization in
methanol synthesis plant," Journal of CO2 Utilization, vol. 10, pp. 12-22, 2015.

[18] A. Rafiee, M. Panahi and K. R. Khalilpour, "CO2 utilization through integration of post-combustion
carbon capture process with Fischer-Tropsch gas-to-liquid (GTL) processes," Journal of CO2
Utilization, vol. 18, pp. 98-106, 2017.

[19] W. F. Pickard, A. Q. Shen and N. J. Hansing, "Parking the power: Strategies and physical limitations
for bulk energy storage in supply–demand matching on a grid whose input power is provided by
intermittent sources," Renewable and Sustainable Energy Reviews, vol. 13, pp. 1934-1945, 2009.

[20] S. V. D. Linden, "Bulk energy storage potential in the USA, current developments and future
prospects," Energy, vol. 31, pp. 3446-3457, 2006.

[21] B. Blound, H. Bernhoff and M. Leijon, "Flywheel energy and power storage systems," Renewable
and Sustainable Energy Reviews, vol. 11, pp. 235-258, 2007.

[22] C. Bueno and J. Carta, "Wind powered pumped hydro storage systems, a means of increasing the
penetration of renewable energy in the Canary Islands," Renewable and Sustainable Energy
Reviews, vol. 10, pp. 312-340, 2006.

[23] M. Budt, D. Wolf, R. Span and J. Yan, "A review on compressed air energy storage: Basic principles,
past milestones and recent developments," Applied Energy, vol. 170, p. 250268, 2016.

[24] A. Foley and I. D. Lobera, "Impacts of compressed air energy storage plant on an electricity market
with a large renewable energy portfolio," Energy , vol. 57, pp. 85-94, 2013.

[25] P. Denholm and R. Sioshansi, "The value of compressed air energy storage with wind in
transmission-constrained electric power systems," Energy Policy, vol. 37, pp. 3149-3158, 2009.

[26] H. Safaei, D. W. Keith and R. J. Hugo, "Compressed air energy storage (CAES) with compressors
distributed at heat loads to enable waste heat utilization," Applied Energy, vol. 103, pp. 165-179,
2013.

[27] E. Yao, H. Wang, L. Wang, G. Xi and F. Maréchal, "Multi-objective optimization and exergoeconomic
analysis of a combined cooling, heating and power based compressed air energy storage system,"
Energy Conversion and Management, vol. 138, pp. 199-209, 2017.

20
 ACCEPTED MANUSCRIPT

[28] L. X. Chen, P. Hu, P. P. Zhao, M. N. Xie, D. X. Wang and F. X. Wang, "A novel throttling strategy for
adiabatic compressed air energy storage system based on an ejector," Energy Conversion and
Management, vol. 158, pp. 50-59, 2018.

[29] L. X. Chen, M. N. Xie, P. P. Zhao, F. X. Wang, P. Hu and D. X. Wang, "A novel isobaric adiabatic
compressed air energy storage (IA-CAES) system on the base of volatile fluid," Applied Energy, vol.
210, pp. 198-210, 2018.

[30] A. H. Alami, K. Aokal, J. Abed and M. Alhemyari, "Low pressure, modular compressed air energy
storage (CAES) system for wind energy storage applications," Renewable Energy, vol. 106, pp. 201-
211, 2017.

[31] g. badino, "the legend of carbon dioxide heaviness," cave and karst studies, vol. 71, pp. 100-107,
2009.

[32] T. Christen and M. W. Carlen, "Theory of Ragone plots," Journal of Power Sources , vol. 91, pp. 210-
216, 2000.

[33] PeterJ.Hall and EuanJ.Bain, "Energy-storage technologies and electricity generation," Energy Policy,
vol. 36, p. 4352–4355, 2008.

[34] Z. Fairuz and IngoJahn, "The influence of real gas effects on the performance of supercritical,"
Tribology International, vol. 102, pp. 333-347, 2016.

[35] m. r. j. al-tameemi and z. yu, "Thermodynamic approach for designing the two phase motive nozzle
of the ejector for transcritical CO2 heat pump system," science direct, vol. 142, pp. 1206-1212,
2017.

[36] M. Suchenek and T. Borowski, "Measuring Sound Speed in Gas Mixtures Using a Photoacoustic
Generator," International Journal of Thermophysics, vol. 39, 2018.

[37] A. Latif, A. Sengupta, M. Kumari and B. Bhattacharyya, "Voltage decay behaviour of super
capacitor," in 2016 International Conference on Energy Efficient Technologies for Sustainability
(ICEETS), Nagercoil, India, 2016.

342

343

21
 ACCEPTED MANUSCRIPT

 This work presents the modular compressed gas energy storage (CGES) system using
carbon dioxide as a working fluid
 Three cylinders discharged sequentially or simultaneously depending on load
requirements using a PLC and solenoid valves
 The system can be charged/discharged almost infinitely, there is little need for
maintenance
 The obtained overall efficiencies reach up to 79%