Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...

THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735 4721

COST BASED OPTIMIZATION OF INDUSTRIAL BULK

COMPRESED NATURAL GAS FILLING FACILITY OPERATIONS
by

Ismail CETINER a and Ahmet COSKUN b*

a
Naturelgaz Sanayi ve Ticaret A. S., Istanbul, Tutkey
b
Department of Mechanical Engineering, Faculty of Engineering,
Suleyman Demirel University, Isparta, Turkey
Original scientific paper
https://doi.org/10.2298/TSCI200430206C

In line with the increase in the world population, natural gas, which has an increas-
ing share in fossil fuels, is nowadays transported throughout pipe-lines in the form
of liquefied natural gas and compressed natural gas. Natural gas is preferred to
be transported with compressed natural gas in terms of optimum cost. Compressed
natural gas is reduced to a volume of 1/250 at 200 bar pressure in filling facilities
and is transported to multi-element gas containers or gas tankers where pipe-lines
do not reach. Although the highest cost for these plants seems to be gas transpor-
tation costs, the design, infrastructure and operational gaps, especially in plant
management of the filling facilities constitute the costs that are not significantly
visible. In parallel with the costs incurred, in this study, a pre-cooling process was
actively applied for cost-based improvement in a bulk compressed natural gas fill-
ing facility, while operational optimization was aimed passively. The filling process
of the facility in 2016 was examined according to real data and pre-cooling was
made in 2017 by adding a “chiller” to the filling process to increase the filling rate
to tankers. Thanks to the precooling in 2017, the filling amount made to tankers
increased by 7.23%. In 2018, the filling process was analyzed in detail according
to the data of 2017 and the factors affecting the filling rate were determined. Ac-
cording to these factors, the filling operation has been optimized on a cost basis.
Filling operations in 2018 have been optimized for factors varying from month to
month and even day, such as; temperature conditions, filling method, the structure
of gas tankers and filling platforms, ie the effect of the material used, personnel
effect and the filling rate of machines like chiller, compressor. After optimization,
the amount of filling made in 2018 increased by 4.36% compared to 2017.
Key words: compressed natural gas, filling rates, cost-based optimization,
chiller integration

Introduction
The life of oil, which was once thought to be the only energy source in the world
and could change the borders of countries due to political interests, is now at calculable levels.
Unlike oil, the fact that the amount of natural gas reserves found every year meets 10% of the
world’s annual natural gas consumption, drags the world towards the gas age.
Currently, international natural gas supply chain is developing rapidly and encoun-
tering an opportunity as well as challenge. Specifically, the natural gas transportation cost ac-
counts for one-third of the total natural gas industry cost [1]. In the industry, while the market
*
Corresponding author, e-mail: [email protected]
 Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
4722 THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735

war is continuing between liquefied natural gas (LNG) and petroleum-derived fuels, natural
gas is compressed and natural gas became an alternative to oil in the transportation sector. Af-
ter the recent commercial approaches, compressed natural gas (CNG) was highly preferred in
the industrial market in the world and was born as an alternative to fuels used in the industry.
Although the highest cost for CNG filling and selling companies seems to be gas transporta-
tion costs, the design, infrastructure and operational gaps in facility management constitute the
costs that are not significantly visible. In addition, for industrial bulk CNG filling plants, it was
found that, the intercooling of compressors is not enough during the periods that the seasonal
temperatures are high. This situation causes additional costs to the facility and at the same time
it causes CNG filling to stop.
There are some studies on CNG in the literature. Bielaczyc et al. [2] presented a
model for the optimization of the CNG fuel station. As a result of the study, it was observed
that non-methane HC emissions were much lower when running on CNG, while HC emis-
sions were somewhat higher. Dogan [3] in his study, designed the flow of the high pres-
sure gas compressor which is located in CNG fuel filling stations as 180 Lph and manu-
factured it. The CNG compressor compresses the natural gas that it receives from the city
network line with a pressure of 0.022 bar to 200-250 bar pressure and the compression
process takes place in four stages. At the end of his work, he shared the pressure, flow and
other tests he performed after manufacturing the compressor required at the stations for the
filling of vehicles using CNG. In his study, Dogan [4] produced a four-stage, reciprocating
type CNG compressor manufactured as a domestic type. The CNG enters the compressor at
0.022 bar pressure and exits at 200 bar pressure. Due to the high temperature during the com-
pression of natural gas, intermediate cooling step compression was made in the compressor.
During the study, the irreversibilities arising from the compressor and the cooling system were
found by moving from the Second law of thermodynamics. Efficiency in the compressor was
calculated by modelling isentropic, polytropic and isothermal state changes. In addition, the
comparison was made by compressing both CNG and air using the same compressor in the
study. Farzaneh-Gord et al. [5] thermodynamically modeled the fast filling process at CNG
fuel filling stations, then aimed to minimize the filling time and finally focused on the optimal
volume ratio of the reservoir tanks. Kagiri et al. [6], in their study, proposed to optimize the op-
eration at CNG fuel filling stations in order to minimize energy costs and aimed to program the
activities of compressors at CNG fuel filling stations in order to manage to minimize the cost
of electricity purchased and used in their research. They also suggest that the effect of abrasion
and wear will decrease in compressors that will operate at these times. Kagiri et al. [7], in their
study, stated that the economical operation of CNG fuel filling stations will reduce the cost of
fuel delivery and it will benefit consumers and it will be determined that the timing of the com-
pressors in the current CNG fuel filling stations will be optimal and the energy costs will be re-
duced according to the electricity tariff used. Kagiri et al. [8], in their study, proposed optimum
energy management strategies for CNG refueling station operations. It is stated that the energy
consumption of the compressors is the main item in the total operating costs of the fuel filling
stations. It is stated that the developed model reduces the compressor cycle and prolongs its life
by providing potential savings of 59.3% in the daily electricity costs of the station. Khadem
et al. [9] stated that one of the most important parameters in the design of CNG filling stations is
a detailed rapid filling process modelling and in their study, they developed new mathematical
modelling to analyze the rapid filling process in CNG stations. Beronich et al. [10] proposed
equations of state for CNG applications. It is stated equations of state for CNG applications is
highly dependant on the composition of the gas and the conditions (e.g. pressure). Niazmand
Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735 4723

et al. [11] carry out a numerical analysis for reciprocating compressors used in CNG stations
is studied thermodynamically for ideal model and real one. Effects of important design factors
on entropy generation, isentropic and exergy efficiencies were investigated. Nanthagopal et al.
[12] examined the hydrogen enriched compressed natural gas (HCNG) usage in case of internal
combustion engines. It has been concluded that HCNG fuels pave the road for the use of hydro-
gen vehicles in the future due to expensive after treatment technologies. Kaleemuddin and Rao
[13] in their study, base diesel engine converted into spark-ignition mode to employ gaseous
fuels (LPG and CNG). It has been proposed that CO2 emission is lower with CNG operation and
thus can be a eco-friendly operation. Sremec et al. [14] examined the effects of excess air ratio
and engine speed on the performance and emissions of CNG fuelled spark ignition engine. As
a result of the study, due to the reason of the fact that natural gas has higher octane rating than
gasoline, it has been possible to use high compression ratios. Saadat-Targhi et al. [15] found
that reciprocating compressors are the heart of CNG fuel filling stations and the main cost of
the station is due to compressor operation. It was stated that the results of the current study were
confirmed against previous experimental data and theoretical studies, and the simulation results
of the mass-flow rate were compatible with the experimental data. The study revealed that com-
pressor energy consumption for one cycle is 61.86 kWh and the average energy consumption
for filling a vehicle is 0.25 kWh/kg.
There are more studies about auto CNG stations that filling vehicles in the literature.
Past studies are limited to the perspective of saving money in filling facilities that load gas on
vehicles, changing the electricity tariff used and filling in when electricity can be purchased at
a lower cost. No studies have been found in the literature for bulk CNG filling facilities that
are established or operated to meet industrial needs. With this study, the gap in the literature
regarding the optimization of industrial bulk CNG filling plant operations will be eliminated. In
this study, pre-cooling was done by adding a chiller to the process to increase the filling rates to
tankers in an operated bulk CNG filling facility. In addition, the points that need to be improved
in the filling process were determined and the operation was optimized on a cost basis. With
this study, the result of active-passive based improvements on a bulk CNG filling facility is
expected to be beneficial for operators.
Material and method
When the filling data of 2016 is examined in a bulk CNG filling facility in operation, it
is seen that in order to increase the performance of the facility, it is aimed to increase the amount
of filling made to tankers. For this purpose, the chiller was added to the process and the filling of
2017 was completed by pre-cooling. The CNG filling facility, which is established and operated
on a land of approximately 8000 m², was designed in a simple structure, fig. 1. The main equip-
ment of the facility is consists of machinery, equipment, etc. necessary for the operation of the
energy or facility constitutes infrastructure elements, such as: gas regulation and the metering
unit, compressor station, filling platforms, administrative and technical offices, transformer, fire
pump, and generator.
The natural gas entering the facility first passes through the regulation and measure-
ment unit of the local gas distribution company with a measurement capacity of 7000 m³/h. The
gas exiting the metering unit enters the compressor station where the compression process is per-
formed. The filling facility has two CNG compressors, with a capacity of 6300 m³ per hour and
3000 m³ per hour. These compressors are three-stage, compress the natural gas supplied in the
35-70 bar pressure range up to 200 bar pressure and press it towards the filling platforms as
shown in fig. 2 to be filled with tankers.
 Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
4724 THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735

M
an
a
of gem
fic e
e nt
Compressor
Compressor – three stages
unit
Re eas tio

Compressed natural gas

gu ur n
m sta

la ing
tio
n

ea
ar
g
lin
Fil

Filling valve
p
um

Terminal 1 Terminal 2 Terminal 3 Terminal 4

ep
Fir

Figure 1. Filling facility lay-out plan [16] Figure 2. Schematic representation of the
precooled filling process [16]

In the CNG filling facility in operation four different capacity tankers as: consisting
of 150 L liquid volume tube bundles, 60 units, 120 units and 145 units tube tankers connected
in series, and, 108 tubes each of which consists of 90 L liquid volume tube bundles, connect-
ed in series are used. The capacities of the tankers used are 9000 L, 9720 L, 18000 L, and
21750 L, respectively.
The CNG filling facilities have high initial investment costs, so, in order to maintain
its profitability and operate efficiently for years, operation costs are very important. The most
important of the operational costs in these facilities is the amount of transportation at once. One
of the steps that can be taken to increase the amount of transportation at once is to use a higher
capacity tanker. Delivering to the point where the gas will be supplied with higher capacity
tankers will increase the amount of transportation at once. Another method is to select the
working pressures of CNG tubes in tankers from the ones that 250 bar, not 200 bar. Even those
with higher working pressure can be selected if possible. In this way, it is possible to load at
higher pressure, hence more gas. When CNG facilities are examined, it is seen that high pres-
sure is needed to fill the gas and it is seen that compressors that provide this condition are used.
Although the compressors have certain set values, the compressor stops when the compression
end pressure of the gas in the compressor reaches the working pressure values of CNG tubes. It
is observed that, after a certain period of time, the CNG tanker pressure has decreased slightly
compared to the working pressure.
The difference pressure, ΔP, between the compressor outlet pressure and the tanker
pressure should be interpreted. One of the reasons why ΔP is high is that gas denseness and
pressure losses on the line, such as valves, quick coupling equipment, elbows and reductions,
which will narrow the flow or make flow difficult, while gas filling at high flow continues.
Another reason is that the temperature of the CNG increases rapidly as the pressure of the tube
increases during the filling of the tubes. At this point, it is necessary to lower the temperature of
the filled gas to reduce, ΔP. The effect of gas temperature on gas pressure is very high during
the CNG filling stage. The most effective solution be done in this regard is to integrate a chiller
into the system after the compressor outlet line and to ensure that CNG cools before entering
the tanker.
In order to reduce ΔP, purchasing smaller capacity compressors to slow down the
flow or having new tankers built to increase the number of simultaneous fillings is difficult to
Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735 4725

decide in terms of cost for operators. Therefore, Chiller

the most suitable method to be implemented

should be the chiller cooling system, which is
relatively less costly than other investments.
Schematic view of the pre-cooled CNG filling
Cooling water
Compressor Heat
– three stages
facility which is shown in fig. 3, added to the

Compressed natural gas

exchanger

filling process within the scope of the study in By-pass

valve
order to increase the low fillings in hot periods
and reduce the fluctuation during the year. The Filling valve

CNG, whose temperature rises at the exit from

Terminal 1 Terminal 2 Terminal 3 Terminal 4

the compressor, is thought to be cooled before it Figure 3. Schematic representation of the pre-
cooled filling process [16]
starts to fill the tankers.
A shell and tube heat exchanger was added to the entrance of each platform and the
cooling water added to the process by chiller pulled heat from CNG, allowing the natural gas
to cool. In this way, with the decrease in ΔP, there will be more filling and the amount of trans-
portation will increase.
In order to measure the filling performance of the facility, it is necessary to compare
the amount of gas filled to tankers of different capacities with reference to the ideal gas equa-
tion. In this study, the Redlich-Kwong empirical equation developed by Kwong et al. [17],
was taken as reference in explaining the behavior of non-ideal gases at critical temperatures.
Pressure is calculated:
Ru T a
= P − (1)
Vm − b T Vm (V + b )
where a and b constants are found as:
R 2u Tc2.5
1 R 2 T 2.5
a = = 0.42748 u c (2)
(
9 3 2 −1 )
Pc Pc

3
2 − 1 R u Tc R T
b= = 0.08664 u c (3)
3 Pc Pc
where P [atm] is the gas pressure, Ru [atmLg–1mol–1K–1] – the universal gas constant, T [K] – the
gas temperature, Vm [Lg–1mol–1] – the unit mol volume of gas, Tc [K] – the critical temperature of
the gas, Pc [atm] – the critical pressure of gas, a – the corrective constant for the bond strength
of gas molecules, and b – the volume represents a corrective constant.
These constant values vary depending on the gas analyzed and are calculated accord-
ing to the critical pressure and critical temperature values of the gas. The reduced pressure and
temperature are determined:
P T
Pr = T = (4)
Pc r Tc
Tr
Pr < (5)
2
The Redlich-Kwong equation is sufficient to calculate the gas phase properties in
cases where the reduced pressure of the gas is less than half of its reduced temperature. The
critical pressure and critical temperature values of natural gas vary depending on the type and
 Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
4726 THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735

amount of gases in the composition. For this reason, in order to calculate the critical pressure
and critical temperature of the natural gas used in the compression process, it is necessary to
know the percentages of the mixed gases contained in the mixture and calculate the effects of
the critical pressures and temperatures of each gas on the mixture. For this, the values obtained
by multiplying the mix percentages and the critical temperature, critical pressure and molar
masses of the gases show the effects of the gases on the mixture. In this way, the average critical
temperature of natural gas is calculated as 197.58 K, average critical pressure is 46.10 atm and
the average mole weight is 17.38 kg/kmol.
The compressibility factor and constant, h, are expressed:
BP b
=h = (6)
Z V
1 A2 h
Z
= − (7)
1 − h B (1 + h )
where variables A² and B here are expressed:
Pr
A2 = 0.42748 (8)
Tr2.5

Pr
B = 0.08664 (9)
Tr
where A² and B [atmL–1] are the refers to variables depending on the reduced pressure and tem-
perature of the gas, h is the volume dependent corrective constant, and Z – the compressibility
factor of the gas. The volume of gas filled into the tanker is calculated:
Vb = KV (10)
Correction coefficient is calculated:
P T
K= + +Z (11)
Pb Tb
where K is the correction coefficient, V [L] – the filled tanker volume, P [bar] – the mea-
sured absolute pressure of the tanker after filling, Pb [1 atm] – the pressure in ideal conditions,
T [K] – the measured temperature of the tanker after filling, and Tb [298 K] – the ideal tempera-
ture.
In line with the stated equations, if the working pressure of the tankers is resolved
according to the reference temperature of 200 bar, that is, 197.378 atm and 298 K. The com-
pressibility factor is calculated as 1.2571 and the average natural gas amount to be filled in
1 liter unit volume is 0.24399 m³.
In addition the stated natural gas filling amount calculations, the electricity costs relat-
ed to the motor powers of the compressor and chiller are calculated in comparison the amount
that is overfilled in the existing facility:
Econsumed = ELf t (12)
where Econsumed is the total energy consumed by the chiller and compressor in return for extra
filling, E – the total load of the chiller and compressor, Lf – the load factor, and t – the time.
Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735 4727

Results and discussion

Cost based optimization of filling facility
The gas coming to the CNG filling facility is compressed by the help of a compressor
and delivered to the filling platforms, where CNG is loaded to the tankers with the help of fill-
ing valves. The data of the refills made by the facility from 2016-2018 were recorded and were
analyzed with MINITAB program. The filling quantity is shown in fig. 4 as annual-average.
According to the annual average data of the refills made without precooling in 2016, 0.21365 m³
per litre filling was made for 9000 litres capacity tankers. Filling was done at the level of
0.23028 m³ per litre for 9720 litres tankers, 0.23711 m³ per litre for tankers with 18000 litres
capacity and 0.21999 m³ per litre for 21750 litres tankers.
The overall performance of the plant is 0.22789 m³ per liter for all fillings. Accord-
ing to the Redlich-Kwong empirical equation, the amount that should be under the reference
conditions of 197.38 atm 15 °C is 0.24939 m³ per litre, and according to this reference value,
the overall performance of the facility in 2016 was calculated as 91.38%. In the refills without
precooling, the change of tanker refills on a monthly basis according to the tanker variety is
shown in fig. 5.
0.25 0.26
Volume of gas filled to the tanker [m3L–1]
Volume of gas filled to the tanker [m3L–1]

0.24939 3
Average fills for December 0.24796 m /L
Average fills for August 0.21712 m3/L Tanker capacity [L]
Reference
value 9000
9720
0.25
0.24 18000
0.23711 21750

0.24 St. Dev 0.02099

0.23028
0.23
0.22789 0.23
Average
value for
0.21999 2016
0.22 0.22
0.21365
0.21
0.21
9000 9720 18000 21750 1 2 3 4 5 6 7 8 9 10 11 12
Tanker capacity [L] Months
Figure 4. Annual average performance of Figure 5. Monthly change of tanker refills
pre-cooled refills (2016) in without pre-cooled refills (2016)

Accordingly, the first noticeable thing is that the refills made in the summer months
are much less than the winter months. While the average of tanker level was 0.24796 m³ per
litre in December when the fillings were highest, this rate remained at 0.21712 m³ per litre in
August when the fillings were the lowest. It was observed that the difference between 2 months
was 14.2%.
In 2017, tanker refills were made by pre-cooling by using a chiller. The annual av-
erage performance of the refills made in 2017 after the pre-cooled process is shown in fig. 6.
Accordingly, 0.24582 m³ per litre filling made for 9000 litres capacity tankers, 0.24582 m³
per litre filling made for tankers with 9720 litre capacity, 0.24606 m³ per litre filling made
for tankers with 18000 litre capacity, and 0.23963 m³ per litre filling made for tankers with
21750 litres capacity.
The overall performance of the facility is 0.24377 m3 per liter for all fillings. Ac-
cording to the Redlich-Kwong empirical equation, the overall performance was completed at
97.98% with an increase of 6.6% compared to 2016. The use of chiller also affected the overall
distribution of the process. The tanker filling volumes of the pre-cooled filling facility in 2017
are shown in fig. 7, and it has been observed that the filling quantity has improved compared to
 Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
4728 THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735

0.27

Volume of gas filled to the tanker [m3L–1]

Volume of gas filled to the tanker [m3L–1]

0.24939 Tanker capacity [L]

Reference
9000
value
0.248 0.26 9720
18000
21750
0.24606
0.246
0.24582 0.24582 0.25 St. Dev 0.01688

0.24437 0.24
0.244 Average
value for
2017 0.23
0.242
0.22
0.240 0.23963 Average fills for December 0.25315 m3/L
3
Average fills for August 0.22744 m /L
0.21
9000 9720 18000 21750 1 2 3 4 5 6 7 8 9 10 11 12
Tanker capacity [L] Months
Figure 6. Annual average performance of Figure 7. Monthly change of tanker
pre-cooled fillings (2017) filling in the pre-cooled filling (2017)

2016. While the process standard deviation value of 2016 was 0.02099, this value decreased to
0.01688 in 2017.
According to another data obtained, the tanker level was 0.25315 m³ per litre in De-
cember when the filling was high, while this ratio remained at 0.222744 m³ per litre in July
when the filling was low. The difference between two months decreased by 2.9 points compared
to 2016 and reached 11.30%. Precooling has reduced the distance between the endpoints of
the process, that is, reducing the standard deviation, making the process more stable. It also
increased refills per liter and reduced operating costs of the facility.
Filling in the facility was made without precooling in 2016 and tankers were loaded
with 0.22789 m³ gas per liter. After the chiller was added to the process in 2017, the fillings
were precooled and 0.24377 m3 of gas per liter was loaded into the tankers. In 2017, there was
a 7.23% increase in efficiency in filling in tankers compared to 2016. Loading more gas on the
tankers increased the amount of transportation at once. Therefore, it is seen that less shipping
will be needed to sell the same amount throughout the year. The facility data to be used in the
reimbursement period calculation is presented in tab. 1 in order to calculate the reimbursement
period of the investment according to the savings from the annual shipment and the cost of
chillers. In this case the purchase price of chillers is taken from the facility as the first invest-
ment cost. Transfer unit charge is also taken from the agreement made between the facility and
the transportation company. The electricity cost is taken from the energy market regulatory
authority (EMRA).
The average tanker filling in 2016 was calculated as 3910.13 m³. A similar perfor-
mance would have been achieved in 2017 if the process had not been modified. However,
pre-cooling was added to the process in 2017, resulting in an efficiency increase of 7.23%.
Based on this increase, the average tanker filling in 2017 was 4192.13 m³. Transporting more
gas at one time, that is, increasing the amount of transportation at once will also decrease the
number of annual shipments. According to this, according to the data of 2016, 3201 shipments
were made, while in 2017, 2985 shipments were made.
One of the important expenses for such businesses is transportation costs. To make
the pre-cooled filling with chiller integration the system, the facility has saved 216 shipments
per year. Based on this, the annual saving fee can be obtained by multiplying, the number of
shipments saved, the average shipping distance (km) of the facility and the price paid for the
shipment per km. Thus, with the addition of a chiller to the facility, the annual saving amount
was calculated as 18348.44 $ with the pre-cooling process.
Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735 4729

Table 1. Facility data to be used for the return on investment

Analysis of data
Total filling amount (2017) 12516587 m³
Average unit filling amount (2017) 0.24437 m³ per litre
Average tank capacity (2017) 17158 L
Average transfer distance (2017) 197.55 km
Transfer unit fee (2017) 0.43 $/km
Chiller purchase price 34,212 $
Average unit filling amount (2016) 0.22789 m³ per litre
The unit filling amount according to reference conditions 0.24939 m³ per litre
Compressor power 355 kW
Compressor flow 6300 m³ per hour
Chiller power 93 kW
Load factor 75%
Electricity charge including all taxes 0.0656 $/kW

Thanks to pre-cooling, tankers were filled with 7.23% more gas. The filling amount
realized with this efficiency increase is 12516587 m³ per year. If there was no pre-cooling pro-
cess, the filling amount to the tankers would be 11672654 m³. Therefore, it is determined that
843933 m³ of extra filling is made annually. To make this filling, the chiller and compressor
were operated more, and it was calculated that the chiller and compressor worked 134 hours
more per year. The energy consumed in excess per year was calculated as 45024 kWh according
to the eq. (12). Accordingly, annual electricity expense was calculated as 2965.3 $. The pay-
back period is calculated as two years and three months by proportioning the chiller purchase
price to the annual net savings. Maintenance and repair costs, interest and inflation rates are not
included in the calculation of the pay-back period.
Operational optimization of the facility filled with pre-cooling
Optimization according to environmental factors
Environmental factors are temperature-dependent factors. Temperature differences
between months, shifts filled and start time of filling are the functions of the temperature. The
filling rates that take place depending on the average air temperature taken for Antalya from the
web page of the General Directorate of Meteorology are shown in fig. 8. As can be seen from
the figure, filling rates decrease when the air temperature is high. It is observed that the filling
rates in the cold months are higher than in the hot months.
When the data of the whole year, which is another environmental factor, for the filling
effect of the shift is examined. The filling rate was 103.2% in the second shift, namely at night,
while the filling rate remained at 101.58% during the day shift. It was determined that the fill-
ings made in the night shift between two shifts yielded 1.62% efficiency compared to the day
shift.
To fill the night shift instead of daytime, tanker refills should be delayed by 6-10 hours
depending on the situation. Since tankers are one of the most expensive equipment, and they are
at a minimum amount in the facilities, it is not possible to plan the filling after approximately
 Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
4730 THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735

8 hours in order to deliver gas. Instead, it is more feasible to select the time zone that will start
filling rather than the shift at the maximum level of filling, and the efficiency will be higher. For
this reason, the change of filling rates at the facility according to the filling hours is presented
in fig. 9.
102.6
104 102.47

Filling rate [%]

Upper limit 102.4
Filling rate [%]

103 102.28
102.2
102.26
102 Average 102.0
lling 101.89
rate of 101.8
101 2017
101.6 101.61
100 Lower limit 101.48
101.4 101.36
101.31
99 101.2
1 2 3 4 5 6 7 8 9 10 11 12 00.00-06.00 06.00-09.00 09.00-12.00 12.00-15.00 15.00-18.00 18.00-21.00 21.00-23.59
10.0
10.7
11.6
12.9
15.4
16.4
20.5
20.6
25.2
25.3
28.4

Months
Average air temperature [˚C]
Filling start time

Figure 8. Filling rates depending on the Figure 9. Filling rates varying according
temperature varying from month to month to filling start time

In 2018, the highest filling rates were reached in the slice, which started between
21.00 and 06.00, and the lowest filling rate was in the filling between 12.00 and 18.00.
Optimization according to the filling method
Running the compressors by cascade method is called a sequential filling. Tanker
filling is started first with a compressor with a capacity of 6300 m³ per hour (large compressor).
High flow means high gas temperature, so the sequential filling method is applied to slow down
the flow and temperature, and filling continues with a compressor (small compressor) with a ca-
pacity of 3000 m³ per hour when the tanker pressure reaches 180 bar. The filling rates according
to the ordered filling pressures are shown in fig. 10.
The sequential filling pressure determined by the facility for the process and applied
in 2018 was determined as 180 bar. However, according to real data, if the small compressor
can be started after at 160 bars or before, instead of after 180 bars, more filling rate can be ob-
tained. It should be remembered that the filling time will increase when switching to lower-or-
der pressurization. In such cases, if the tanker pressure drops below 200 bar, it can be refilled.
This method to be applied is an additional filling application.
102.6
In order to analyze the results of addi-
102.56 tional filling at the facility, 57 test fillings were
Filling rate [%]

102.52
102.4
102.26
made using tankers with the same capacity. Af-
102.2
102.12
Average
lling ter the filling of the tankers has been complet-
102.0
ed, some of them wait for 1 hour, some for 1.5
rate of
2018
101.8
hours and the rest for 3 hours, and the effect of
101.6
101.56
holding times on the filling rate is examined. It
101.4 is possible to provide close to 1% improvement
in the filling rate between the tanker waited for
101.23
101.2
100 120 140 160
Sequential filling pressure [bar]
180
1 hour and the tanker waited for 3 hours. The
important point is to be able to create the filling
Figure 10. Filling rates varying according to
sequential filling & ordering operation that will enable tankers to
be kept waiting.
Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735 4731

One of the most important factors that 103.0

will affect tanker refills is the volume loaded 102.74

Filling rate [%]

102.5
with gas. For this reason, by increasing the 102.26

volume, ΔP decreases and filling rates will

102.15 60% Average
102.0 lling
101.87
increase. The effect of the number of tankers 28% rate of
2018
10%

filled on the filling rate is presented in fig. 11. 101.5

The rate remained at 100.45% for 1 tanker 101.0

filling, while the rate reached 102.74% after 4 2%
tankers filling at the same time. As a result, a 100.5 100.45

difference of 2.29% was achieved between the 1 2 3

Number of tankers filled at the same time
4

2 operative options. This increase, compared to

Figure 11. Filling rates varying according to
the increase in filling rates with the pre-cooling the number of tankers filled at the same time
application the CNG filling facility was lower.
The filling operation in 2018 was very well optimized. The number of tankers filled
at the 4-platform facility at the same time reached 3.45 on average. As can be seen from the
graphic, three or four tankers were filled simultaneously in 88% of the fillings. In this way, an
increase in filling rates was achieved.
The effect of compressor capacity on filling rates is shown in tab. 2. According to this,
it was found that, working with compressors with lower capacity in order to slowing down the
flow and decrease the filling temperature, increases the filling rates more.
Table 2. Filling rates according to compressor capacities
Compressor capacities [m³h–1] Filling rates [%] End-of-filling temperature [°C]
3000 102.85 14.15
6300 101.14 17.02

While the filling with the small compressor with a capacity of 3000 m³ per hour of the
two compressors used in the facility resulted in a filling rate of 102.85%, the filling with a large
compressor with a capacity of 6300 m³ per hour was completed with a filling rate of 101.14%.
While at the filling with large compressor, the end-of-filling temperatures result in 17.02 °C,
this value remained at 14.15 °C when the filling made with a small compressor. The fact that the
high filling flow increased the temperature at the time of filling, also affected the filling rates.
Apart from the pre-cooling process carried out in 2017, it is thought that approaching
the filling and shipment planning more precisely will increase the filling rates. For this reason,
awareness of the operation was targeted as passive improvement in 2018. Factors affecting fill-
ing performance and root causes were determined and the effect of these factors on filling rate
was analyzed. The determining factors are environment (month, shift, filling start time), method
(sequential filling, additional filling, multiple filling) and machine (compressor, chiller). The
data received from the facility were analyzed with the help of a statistical analysis program.
Filling operations in 2018 were carried out with the awareness of eliminating the effect of the
identified root causes partially or completely.
Annual average performances of pre-cooled and optimized refills are presented in
fig. 12. In 2018, filling was completed at the level of 0.25569 m³ per litre for 9000 litres ca-
pacity tankers, 0.25584 m³ per litre for 9720 litres capacity tankers, 0.25594 m³ per litre for
18000 litres tankers, and 0.225267 m³ per litre for 21750 litres capacity tankers. The overall perfor-
mance of the facility is 0.25504 m³ per liter for all fillings. According to the Redlich-Kwong empir-
ical equation, the filling rate was completed by 102.26%, increasing by 4.28% compared to 2017.
 Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
4732 THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735

The difference between the endpoints of the summer and winter months was calcu-
lated as 11.3% in the filling in the facility in 2017. In 2018, the operation was optimized and
the difference between the endpoints of the summer and winter months decreased by 3.47%
compared to 2017 and reached 7.83%. As a result, the overall distribution of the process in 2018
has improved compared to 2017, and the monthly change of tanker refills at the pre-cooled and
optimized CNG facility is presented in fig. 13. While the process standard deviation value of
2017 was 0.01688, this value decreased to 0.00746 in 2018.
0.270

Volume of gas filled to the tanker [m3L–1]

Volume of gas filled to the tanker [m3L–1]

0.256
Tanker capacity [L]
0.25569 0.25584 0.25594
9000
0.255 0.25504
Average 0.265 9720
value for 18000
0.254 2018 21750
St. Dev 0.00746
0.253 0.25267
0.260

0.252
0.255
0.251

0.250 0.250
Average fills for January 0.26578 m3/L
0.24939 3
Average fills for July 0.24648 m /L
0.249 Reference
9000 9720 18000 21750 value
1 2 3 4 5 6 7 8 9 10 11 12
Tanker capacity [L] Months
Figure 12. Annual average performance Figure 13. Monthly change of tanker refills
of pre-cooled and optimized refills (2018) in pre-cooled and optimized refills (2018)

Overall evaluation of pre-cooled and

no-pre-cooled optimized CNG facility
In addition investing in CNG filling facilities, accurate identification of the process,
accurate measurement and detailed analysis of the results also provide opportunities for im-
proving the operation. In order to clarify this situation, the filling was carried out only in July
2018, without precooling. It is aimed to determine the efficiency to be achieved only by optimi-
zation of operation without pre-cooling. The variation of the filling volumes made in July 2016
and August of 2018, according to the tanker capacity is shown in fig. 14.
In July 2016, the overall average of the process was 0.22001 m³ per litre, and the
performance of the process was 88.22% compared to the reference value. In 2017, the process
was actively improved and after adding precooling to the process, the overall average was
9000 9720 18000 21750
0.222744 m³ per litre and overall performance
2016 2017 2018 2018* 2016 2017 2018 2018*
Optimized process

0.25 increased to 91.2%. In 2018, pre-cooling was with precooling

0.25058 (August, 2018)
Volume of gas filled to the tanker [m3L–1]

not used in the fillings, only through passive im- 0.24648 (July, 2018)
Optimized process
0.24 with precooling
provement, filling optimization was carried out.
0.23 The overall average in July 2018 was 0.24648 0.22744 (July, 2017)
Process with

0.22
m³ per litre and the overall performance of the precooling
0.22001 (July, 2016)

process resulted in 98.83%. In August 2018, the Process with

precooling
0.21
facility was operated both pre-cooled and opti-
Year
July, 2016
2016 2017 2018 2018* 2016 2017 2018 2018*
July, 2017
Tanker capacities [L]
mized. Thus, the overall average was 0.25058 July, 2018
August, 2018*

Figure 14. Change of filling quantities made m³ per litre and the overall performance of the
in different processes according to tankers process was calculated as 100.47%.
Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735 4733

Conclusion
The results obtained from the study will be very useful for the organizations who want
to establish an industrial bulk CNG filling plant and who are operating. In addition, this study
will contribute to the gap in the literature regarding the optimization of industrial bulk CNG
filling plant operations.
It was found that the industrial bulk CNG filling plant achieved a 91.38% filling rate
in 2016, with the pre-cooled process. In order to increase the filling rates of the facility within
the scope of the study, the work was started and the chiller was added to the filling process.
Chiller was added between the compressor outlet and the tanker filling valve in the process, and
all the fillings were pre-cooled using chiller in 2017.
At the end of the 2017 when the datas examined, it was observed that the filling rate
was 97.98% after the pre-cooling process. Thus, the facility’s volume of transport increased
by 7.23%. All these improvements have saved the facility a net savings of 15393.11$ per
year, after calculating the electricity costs spent on this business, according to 2017 data. The
34212.3$ investment for chiller repaid itself in about 2 years and 3 months.
Filling operations in 2018 were carried out with the awareness of eliminating the
effect of the identified root causes partially or completely. According to this, in 2018 with an
increase of 4.28% compared to 2017, the filling rate of the facility was completed by 102.26%.
Cost based optimized of the operation giving the facility flexibility in operating costs in re-
sponse to changing conditions.
Within the scope of the study, the effect of temperature on the filling rate in the envi-
ronmental factor was examined. Shifting the fillings to the night shift as much as possible or
paying attention not to fill at noon at the peak of the Sun will allow tankers to load more gas.
There is no option such as not to refill in the summer, but during these periods, approaching
the operation more precisely, loading more gas and this will provide the company with cost
flexibility.
Increasing the volume filled with gas will decrease the differential pressure of ΔP and
the filling rates will be higher. For this reason, instead of filling a single tanker to increase the
volume, it is necessary to plan the filling of two tankers or the number of the tanker as much as
the number of platforms in the facility at the same time considering the shipment status.
All these methods to be applied are one of the most effective options for increasing
the filling rates. Therefore, the goal of the facilities should be to perform filling operations with
a low capacity compressor, by predicting critical times when customers can order gas, or to
lower the sequential filling pressure. If there is only one compressor in the facility, a sequential
pressurization effect can be created by reducing the inlet valve of the compressor and reducing
the capacity of the compressor after the appropriate pressure determined, but this time the effi-
ciency of the machine used will be reduced. If there is more than one compressor of the same
capacities in the facilities and there is no problem in their operation at the same time, the filling
can be done by applying the cascade system. It will be very useful to turn off the compressors in
sequence after the appropriate pressures and complete the filling with the last single compres-
sor to increase the filling rate. Cascade application can be done manually or automatically by
planning a frequency inverter or a similar automation system.
In addition investing in CNG filling facilities, accurate identification of the process,
accurate measurement and detailed analysis of the results also provide opportunities for im-
proving the operation. In order to elaborate on this situation, in July 2018, the refills were made
only with operation optimization without pre-cooling. In July 2016, the overall average of the
process was 0.22001 m³ per litre, and the performance of the process was 88.22% compared to
 Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
4734 THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735

the reference value. After adding precooling to the process in July 2017, the general average
was 0.22744 m³ per litre and the overall performance increased to 91.2%. In 2018, the process
did not use pre-cooling, only filling optimization was performed. In July 2018, the general aver-
age was 0.24648 m³ per litre, while the overall performance of the process resulted in 98.83%.
In 2018, the filling optimization of the pre-cooled process was carried out and in August, the
general average was 0.25058 m³ per litre, while the general performance reached 100.47%.
In 2017, with active improvement by using pre-cooling, 97.98% filling performance
was achieved throughout the year. However, in July 2018, without pre-cooling and only through
operational optimization, filling performance of 98.83% was achieved. According to these data,
it is seen that operational optimization has increased more than the efficiency increase provided
by pre-cooling. However, under normal circumstances it will not be possible to provide this
operational preciseness. It is considered that it is more convenient to make gas filling with
pre-cooling. At the filling plants, providing the gas supply must always be more important than
focusing on increasing the efficiency. In terms of the results were obtained and continuing sup-
ply, adding chillers to the process seems to be the most appropriate method in bulk CNG filling
facilities, to achieve the targeted increase in filling rates. In brief, adding chillers in accordance
with the investment plans of the bulk CNG filling facilities planned and operated, will always
increase their competitiveness in the sector.
Nomenclature
E – total load of the chiller and compressor, [kW] T – gas temperature, [K]
Econsumed – total energy consumed, [kWh] Tb – ideal temperature, [K]
K – correction coefficient Tc – critical temperature of the gas, [K]
Lf – load factor Tr – reduced temperature
P – gas pressure, [atm] t – time, [h]
Pb – pressure in ideal conditions, [atm] Vm – unit mol volume of gas, [lg–1mol–1]
Pc – critical pressure of gas, [atm] Vb – volume of gas filled into the tanker, [l]
Pr – reduced pressure V – volume of filled tanker, [l]
Ru – universal gas constant, [atmlg–1mol–1K–1]

References
[1] Zhanga, H., et al., Optimal Design and Operation for Supply Chain System of Multi-State Natural Gas
under Uncertainties of demand and Purchase Price, Computers & Industrial Engineering, 131 (2019), 3,
pp. 115-130
[2] Bielaczyc, P., et al., An Assessment of Regulated Emissions and CO2 Emissions from a European
Light-Duty CNG-Fueled Vehicle in the Context of Euro 6 Emissions Regulations, Applied Energy, 117
(2014), Mar., pp. 134-141
[3] Dogan, B., High Pressure CNG Compressor Design and Manufacturing (in Turkish), Sinop University
Journal of Science, 1 (2016), 1, pp. 7-15
[4] Dogan, B., Exergy Analysis in a Four-Stage Piston Type CNG Compressor (in Turkish), Sinop University
Journal of Science, 2 (2017), 1, pp. 159-169
[5] Farzaneh-Gord, et al., Selecting Optimal Volume Ratio of Reservoir Tanks in CNG Refueling Station with
Multi-Line Storage System, International Journal of Hydrogen Energy, 41 (2016), 48, pp. 23109-23119
[6] Kagiri, C., et al., Optimal Energy Cost Management of a CNG Fuelling Station, IFAC PapersOnLine, 50
(2017), 2, pp. 94-97
[7] Kagiri, C., et al., Optimization of a Compressed Natural Gas Station Operation Minimize Energy Cost,
Energy Procedia, 142 (2017), Dec., pp. 2003-2008
[8] Kagiri, C., et al., Optimized Response to Electricity Time-of-Use Tariff of a Compressed Natural Gas
Fuelling Station, Applied Energy, 222 (2018), July, pp. 244-256
[9] Khadem, J., et al., Mathematical Modelling of Fast Filling Process at CNG Refueling Stations Consid-
ering Connecting Pipes, Journal of Natural Gas Science and Engineering, 26 (2015), Sept., pp. 176-184
Cetiner, I., et al.: Cost Based Optimization of Industrial Bulk Compresed ...
THERMAL SCIENCE: Year 2021, Vol. 25, No. 6B, pp. 4721-4735 4735

[10] Beronich, E. L., et al., Prediction of Natural Gas Behaviour in Loading and Unloading Operations of
Marine CNG Transportation Systems, Journal of Natural Gas Science and Engineering, 1 (2009), 1,
pp. 31-38
[11] Niazmand, A., et al., Exergy Analysis and Entropy Generation of a Reciprocating Compressor Applied in
CNG Stations Carried out on the Basis Models of Ideal and Real Gas, Applied Thermal Engineering, 124
(2017), Sept., pp. 1279-1291
[12] Nanthagopal, K., et al., Hydrogen Enriched Compressed Natural Gas – A Futuristic Fuel for Internal
Combustion Engines, Thermal Science, 15 (2011), 4, pp. 1145-1154
[13] Kaleemuddin S., Rao, G. A. P., Conversion of Diesel Engine into Spark Ignition Engine to Work with
CNG and LPG Fuels for Meeting New Emission Norms, Thermal Science, 14 (2010), 4, pp. 913-922
[14] Sremec, M., et al., Influence of High Compression Ratio and Excess Air Ratio on Performance and Emis-
sions of Natural Gas Fuelled Spark Ignition Engıne, Thermal Science, 22 (2018), 5, pp. 913-922
[15] Saadat-Targi, M., et al., Thermodynamic Analysis of a CNG Refueling Station Considering the Recipro-
cating Compressor, Journal of Natural Gas Science and Engineering, 29 (2016), Jan., pp. 453-461
[16] Cetiner, I., Cost Based Optimization of Industrial Bulk CNG Filling Facility Operations, Thermal Sci-
ence, On-line first, https://doi.org/10.2298/TSCI200430206C, 2019
[17] Redlich, O., Kwong, J. N. S., On the Thermodynamic of Solutions, Chemical Reviews, 44 (1949),
1, pp. 233-244

Paper submitted: April 30, 2020 © 2021 Society of Thermal Engineers of Serbia
Paper revised: June 9, 2020 Published by the Vinča Institute of Nuclear Sciences, Belgrade, Serbia.
Paper accepted: June 19, 2020 This is an open access article distributed under the CC BY-NC-ND 4.0 terms and conditions