Available online at www.sciencedirect.

com

ScienceDirect
Energy Procedia 110 (2017) 597 – 603

1st International Conference on Energy and Power, ICEP2016, 14-16 December 2016, RMIT
University, Melbourne, Australia

Development of a computational model for optimal sourcing of

LNG
Israt Mustary, Harun Chowdhury*, Bavin Loganathan, Firoz Alam
School of Engineering, RMIT University, Melbourne 3000, Australia

Abstract

Recent data indicates that there is an increasing global demand of natural gas in many countries. They want to enter into the
international LNG market to meet their shortfalls. In order to minimize the transport cost the potential country needs to select the
optimum strategy for importing LNG. However, for the prospective buyers of LNG, there is no such computational model available
for sourcing LNG from the optimizing exporting locations at present. Therefore, the primary objective of the research is to develop
a universal computational model which can be used to estimate total LNG price including transportation cost for any potential
importing country for any given exporting terminal. A detailed transportation costing model was formulated to calculate the
shipping cost based on the geographical locations of LNG terminals, types of shipping vessel and other shipping associated costs.
A case study was performed to utilize the developed model for a prospective LNG importing country like Bangladesh.

© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
© 2017 The Authors. Published by Elsevier Ltd.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of the 1st International Conference on Energy and Power.
Peer-review under responsibility of the organizing committee of the 1st International Conference on Energy and Power.
Keywords: computational model; optimal sourcing; LNG; shipping.

1. Introduction

Natural gas is a conventional but cleanest fossil fuel and considered as an enviro-friendly energy source. It is usually
transported from the production facility to end users via pipelines. However, long distance transportation via pipeline
is not only expensive but also sometimes impossible. Therefore, in order to store or transport natural gas to a distant
location, it is converted to liquid form called Liquefied Natural Gas (LNG). As the energy demand is increasing day

* Corresponding author. Tel.: +61399256103; fax: +61399256108.

E-mail address: harun.chowdhury@rmit.edu.au

1876-6102 © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of the 1st International Conference on Energy and Power.
doi:10.1016/j.egypro.2017.03.191
598 Israt Mustary et al. / Energy Procedia 110 (2017) 597 – 603

by day, more and more countries are investing on exporting and importing LNG. Fig. 1 shows the LNG Trade volumes
in million tons per annum (MTPA) from 1994 to 2014 [1]. Recent data indicate that until 2014, 19 countries exported
LNG whereas 29 countries imported. New countries like Jordan, Egypt, Pakistan and Poland joined the list in early
2015 [1-2].

Fig. 1. LNG Trade volumes from 1994 to 2014 [1].

90 100
80 90
70 80
70
60
Capacity (MTPA)

Capacity (MTPA)

60
50
50
40
40
30
30
20 20
10 10
0 0
Eq.…

South…
Malaysia

Nigeria

UAE
Peru

China

Argentina
Qatar

India

Italy

Malaysia
Australia

Indonesia

PNG

Kuwait
Algeria

Oman
Yemen
Russia

Mexico
Egypt
US

Taiwan
UK

Brazil

France

Others*
Trinidad

Angola

Japan

Spain

Turki

Chili

Singapore
Brunei

Norway

(a) (b)

Fig. 2. Countries in 2014 (a) LNG Imports; (b) LNG Exports (adapted from [1]).

*where “Other” includes Belgium, Canada, the Dominican Republic, Greece, Israel, Lithuania, the Netherlands, Portugal, Puerto Rico, Thailand,
the UAE and the US
Christiansen et al. [3] reviewed research on ship routing and scheduling and provided four basic models. Dinwoodie
et al. [4] studied to synthesize maritime specialists' perceptions of changing patterns of maritime oil freight flows to
2050. Lin & Tsai [5] developed a model for carriers provide daily pickup and delivery service to customers at major
ports along the Pacific Rim. They investigate the ship routing and freight assignment problem for daily frequency
operation of liner shipping. Halvorsen-Weare & Fagerholt [6] studied a real-life ship routing and scheduling problem
from the LNG business, with both inventory and berth capacity constraints at the liquefaction port. Maxwell and Zhu
[7] examined the empirical relationship between U.S. LNG imports, the Henry Hub price of natural gas relative to
U.K. and Asia gas prices, and a proxy for LNG transportation costs using monthly data from 1997 to 2007.
Biresselioglu et al. [8] developed a mathematical model for Turkey’s future LNG supply security strategy. They
formulate and solve a mixed integer programming model that determines the optimal sourcing strategy for Turkey’s
increasing LNG demand. Their model demonstrates a number of alternative policy options for LNG supply. However,
none of these models can be used for calculating the optimum sourcing strategy for other countries who want to buy
LNG with a minimal cost from the international market places. Therefore, the primary objective of the research is to
develop a computational model to determine the suitable exporting ports for minimum transportation cost for a given
import location in a country. In this study, the global shipment model by sea routing has been implemented. A typical
LNG transportation network is illustrated in Fig. 3.
 Israt Mustary et al. / Energy Procedia 110 (2017) 597 – 603 599

Fig. 3. A typical LNG transportation network by ship [source: http://www.synfuels.com/].

2. Methodology

In order to develop the mathematical model the following steps were carried out:
x Identification of parameters that affect the costing model
x Collection of data
x Development of a computational costing model
x Determination of required output parameters
x Utilization of the developed costing model for optimal sourcing strategy by a case study

2.1. Identification of Parameters that Affect the Costing Model

Initially, a list of parameters were identified that can affect the costing model. These are: (a) Parameters associated
with LNG terminals of exporting countries: Annual production capacity, geographic location (longitude and latitude),
LNG price per unit volume excluding the shipment cost; (b) Parameters associated with LNG terminals of importing
countries: Annual capacity, geographic location (longitude and latitude); (c) Parameters associated with LNG
transportation equipment and cost from source to destination: Ship Capacity (cm), Ship Speed (km/h), Fuel
Consumption of the ship in Route (ton per day), Fuel Consumption of the ship in Port (ton per day), Fuel Price ($/ton),
Port days (for loading), Port days (for unloading), Chartering fee ($/day), Brokerage fee ($), Port Charges $/day at
delivery and receiving terminal, Canal Charges (e.g. Panama or Suez Charge) $/mcm, Cost of Insurance.

2.2. Collection of Data

Recently available data from different sources have been used for the research model development. Sources are
listed below:
x International Gas Union (IGU)
x International Energy Agency (IEA)
x ICC Commercial Crime Services
x Energy Global LNG Service
x Waterborne LNG
x HIS Inc.
x British Petroleum (BP)
600 Israt Mustary et al. / Energy Procedia 110 (2017) 597 – 603

x ExxonMobil
x Royal Dutch Shell
x BG Group
x Chevron
x PERTAMINA
x Petronas

2.3. Measurement of Distances

Distance between the source and the destination terminals are important. Google Map and Google Earth were used
to pin point the locations of LNG terminals and the data for Global Positioning System (GPS) Coordinates: Latitude
and Longitude including satellite images were collected. Fig. 4(a) shows the satellite image of Gorgon LNG Plant in
Western Australia. Commercially available Netpas software was used to determine the exact distance between the
source and destination terminals. Fig 4(b) shows the distance between Gorgon LNG terminal in Australia and
Matarbari LNG terminal in Bangladesh in a screen shot of the software. The path shown in the figure represents the
actual sea route between these terminals.

(a) (b)

Fig. 4. (a) Satellite image of Gorgon LNG Plant; (b) Screen shot of the Netpas software.

2.4. Development of the computational costing model

A basic costing model for LNG price from source to the destination is illustrated in Fig. 5. It shows that the total
cost can be calculated if three components of cost (i.e. gas production cost, liquefaction cost and shipping cost) are
known. In this model, gas production cost and liquefaction cost for unit volume [US Doller ($) per gigajoule (GJ)]
have been collected from the above mentioned data sources. Hoverer, the transportation cost was calculated depending
on the location of the sources and the destinations and route preferences.

Fig. 5. Basic costing model.

 Israt Mustary et al. / Energy Procedia 110 (2017) 597 – 603 601

The model also requires the input values for the properties of LNG carrier as shown in Table 1; however, the
chartering fees shown in this table are assumed values.

Table 1. Properties of LNG carriers.

Ship Type Average Capacity (cubic meter) Average design ship speed (km/h) Chartering Fee ($/Day)
Small 19,000 28 50,000
Small (Med-max) 75,000 32 60,000
Small Conventional 138,000 36 70,000
Large Conventional 150,000 37 80,000
Q-flex 210,000 37 90,000
Q-max 265,000 37 100,000

In order to calculate the total LNG price, several databases have been created. For example, a dataset of LNG
carrier is shown in Table 1. Similarly, a database of distances from the selected destination port to all the possible
sourcing ports has been created. Additionally, LNG pricing data (production cost or the selling price at the source) for
each source are required to calculate the total LNG price at the destination. Fig. 6 shows the list of input and output
variables used in this computational model. Several input data used in this model have been assumed, however, these
values can be easily updated with more appropriate values to determine more realistic and reliable output and the
model can be easily configured for any destination by updating the input data.

Fig. 6. Screen capture of the computational Excel model.

3. Case study: Bangladesh

Bangladesh is a developing country with rapid industrialization and infrastructure development activities.
Currently, based on purchasing power parity (PPP) Bangladesh is the 28th largest economy in the world. Bangladesh,
the country of 160 million people, wishes to become a middle income country by 2021 and a fully developed nation
by 2041. In order to accelerate the industrialization and economic growth, Bangladesh needs to increase its power
generation capacity. Due to limited energy available in Bangladesh, it needs to import raw energy (e.g., coal and gas)
602 Israt Mustary et al. / Energy Procedia 110 (2017) 597 – 603

from abroad. Recently Bangladesh is planning to set up a LNG import terminal with a capacity of 3.5 MTPA at Matar
Bari in Moheshkhali Island of Cox's Bazar district, Chittagong. The terminal, which is to be set up on build-own-
operate basis, will supply gas to power plants. Proposed project will be implemented by early 2018. The locations of
the LNG receiving terminals of Bangladesh and selected 5 different LNG export terminals were collected from Google
Map and the distances between the 5 sources and destination terminals were determined by Netpas software. The data
obtained for case study for Bangladesh are tabulated in Table 2.

Table 2. Data for case study: Bangladesh.

Sourcing Port Capacity (MTPA) GPS Coordinate Distance (km)
US: Kenai LNG 1.5 60.677520, -151.396978 14063
Indonesia: Bontang LNG T1-5 5.4 0.083339, 117.486686 4999
Algeria: Arzew-GL2Z(T1-6) 8.2 35.815078, -0.245145 12282
Qatar: QatarGas III 7.8 25.932497, 51.612528 6090
Australia: Gorgon LNG T1-2 10.4 -20.795315,115.478228 5522

4. Results and discussion

The developed Excel model was used to determine the shipping cost per unit for 6 different types of ship for 5
sourcing destinations located over a range of distance as shown in Fig. 7. It is clear for the figure that the shipping
cost increases with the increase of distance for all ship types. Additionally, the shipping cost decreases with the
increase of ship size. Smaller capacity ship has more cost than the larger one. Furthermore, as the ship size increases
the rate of shipping cost increment decreases with increasing distance. Therefore, it can be concluded that the larger
ship is more economical for LNG transportation and the cost of transportation in minimal if the selected sourcing
terminal is closer from the destination. However, the LNG production cost need to be calculated in order to determine
the total LNG cost which is not shown in this study.
$3.00
y = 0.0002x + 0.1966
$2.50 Small R² = 1
Shipping cost per unit ($/GJ)

Small (Med-max)
$2.00 Small Conventional
Large Conventional
Q-flex
$1.50 Q-max

$1.00

$0.50

$-
0 2000 4000 6000 8000 10000 12000 14000 16000
Distance (km)

Fig. 7. Shipping cost per unit ($/GJ) as a function of distance.

Fig. 8 shows the Shipping cost per unit ($/GJ) as a function of ship capacity. It can be clearly seen that the shipping
cost significantly drops with increasing ship capacity. There is a negative power correlation exits between unit
shipping cost and ship capacity.
 Israt Mustary et al. / Energy Procedia 110 (2017) 597 – 603 603

$3.00

Shipping cost per unit ($/GJ)

$2.50

$2.00

$1.50
y = 6578.5x-0.799
$1.00
R² = 0.9832
$0.50

$-
- 50,000 100,000 150,000 200,000 250,000 300,000

Ship capacity (m3)

Fig. 8. Shipping cost per unit ($/GJ) as a function of ship capacity.

5. Conclusions

A computational model has been developed to estimate the shipping cost between the source and destination LNG
terminals. This model can be utilize as a software tool for making optimal strategy for any prospective country to
source LNG from appropriate locations by minimizing the overall cost. This model can be used to estimate the CIF
(Cost, insurance and freight) price of LNG from one destination to another. Although several data of the input variables
have been assumed for the evaluation of the case study for Bangladesh, the values of the variables can be easily
updated to obtain the best possible outputs. Any country who wants to import LNG to meet their demand can use this
model to determine optimal sourcing destination. On the other hand, exporting countries can also use this model to
find prospective buyer of LNG.

References

[1] IGU. World LNG Report. Paris: IGU Publications; 2015.

[2] BP. BP Statistical Review of World Energy. London: BP plc; 2014.
[3] Christiansen M, Fagerholt K, Nygreen B, Ronen D. Ship routing and scheduling in the new millennium. European Journal of Operational
Research 2013;228(3):467-483.
[4] Dinwoodie J, Tuck S, Rigot-Müller P. Maritime oil freight flows to 2050: Delphi perceptions of maritime specialists. Energy Policy
2013;63:553-561.
[5] Lin DY, Tsai YY. The ship routing and freight assignment problem for daily frequency operation of maritime liner shipping. Transportation
Research Part E: Logistics and Transportation Review 2014;67:52-70.
[6] Halvorsen-Weare EE, Fagerholt K. Routing and scheduling in a liquefied natural gas shipping problem with inventory and berth constraints.
Annals of Operations Research 2013;203(1):167-186.
[7] Maxwell D, Zhu Z. Natural gas prices, LNG transport costs, and the dynamics of LNG imports. Energy Economics 2011;33(2):217-226.
[8] Biresselioglu ME, Demir MH, Kandemir C. Modeling Turkey’s future LNG supply security strategy. Energy Policy 2012;46:144-152.