Ahn Suhong PDF

Erasmus University Rotterdam
MSc in Maritime Economics and Logistics
2017/2018
Analysis of Natural Gas Supply Chain and

Optimisation of Small-Scale Liquefied Natural Gas
Supply Chain in Spain
by
Suhong Ahn
copyright © Suhong Ahn

Acknowledgements
I appreciate Dr. Zuidwijk for valuable feedback for my thesis.

Life is continuously unexpected and tough. However, I have managed it well so far
with the support from my family. My mother always believes and supports me. Without
my father, I would not be here since he introduced and helped me in the maritime
industry.
ii
Abstract
Spain has the most extensive infrastructure of small-scale LNG (SSLNG) in Europe.
This leads to the complexity of the distribution of SSLNG to the domestic destinations.
The purpose of the study is the optimisation of small-scale liquefied natural gas
(SSLNG) supply chain by truckload transportation in Spanish territory in the Iberian
Peninsula. Also, SSLNG in Spain cannot exist without large-scale LNG after trading
the resource so, the whole markets are analysed.
Seasonality of LNG causes a problem for truck companies transporting LNG since
the companies should manage the surplus from the gap between peak and off-peak
season. Another problem is a limit of LNG output truckload capacity at LNG receiving
terminals. This leads for SSLNG facility to be supplied SSLNG not always from the
nearest terminal. The last issue is derived from El Musel terminal under not operation
by Spanish government even though a tremendous amount of capital was invested
for the terminal. So, it is essential to justify the operation of the terminal to the Spanish
government not to waste the investment.
Mixed Integer Linear Programme (MILP) solves the network optimisation problems by
using software “Excel Solver”. The scope of the study for simulations is from liquefied
natural gas (LNG) import terminals to SSLNG facilities consisting of LNG fuel stations
and satellite terminals. A regression model estimates a decision variable of SSLNG
demand with seasonality for forecasting with using “Minitab” software. Three
scenarios are applied to the model based on the degree of SSLNG demand with sub-
scenario whether the operation of El Musel terminal or not.
The seasonality has effects on the supply chain by appearing or disappearing short-
term transportation of LNG. This has been observable in high demand (scenario 1).
When the high demand almost reaches the maximum LNG truckload capacity of LNG
terminals, the geographical reliance from the destinations to the terminal is weak. On
the other hand, the less demand (scenario 2) results in the high geographical reliance.
The difference in total costs comparing operation and no operation of El Musel is not
substantial. A lack of capacity in scenario 3 is not able to analysis since adjustments
of the order of terminals lead to different results while the other conditions are the
same.
Truck companies benefit from the number of trucks needed for LNG transportation
calculated from the simulations since they can mitigate the adverse effects of the
surplus of their asset in advance by setting appropriate strategy. There are three types
of SSLNG facility depends on having a single or multiple LNG import terminals
supplying SSLNG. The first and second types are not necessary to change the
number of terminals since the numbers of terminal they needed remain same
regardless of the scenario. On the other hand, the companies in the third type
experiencing the change in the number of terminals between the two scenarios should
consider making contracts to be supplied SSLNG. So, companies in the second type
should be informed to meet the demands, otherwise satisfying the demand is
impossible by losing the service terminals to other firms.
iii
Table of Contents
Acknowledgements ............................................................................................................................................... ii
Abstract ....................................................................................................................................................................... iii
Table of Contents ................................................................................................................................................... iv
List of tables ............................................................................................................................................................ vii
List of figures ......................................................................................................................................................... viii
List of Abbreviation ............................................................................................................................................... x
Chapter 1 Introduction ......................................................................................................................................... 1
1.1 The topic .......................................................................................................................................................... 1
1.2. Problem Statement ................................................................................................................................... 1
1.3 Background and Justification ............................................................................................................. 2
1.4 Deficiencies in the Evidence ................................................................................................................ 2
1.5 Audience.......................................................................................................................................................... 2
1.6 Research Objective ................................................................................................................................... 2
1.7 Research Questions ................................................................................................................................. 2
1.8 Methodology and Research Design ......................................................................................... 3
1.9 Thesis Structure .......................................................................................................................................... 4
Chapter 2 Literature Review ............................................................................................................................. 5
2.1 Seasonality..................................................................................................................................................... 5
2.2 Vertical Integration .................................................................................................................................... 6
2.3 Optimisation of Small-Scale Supply Chain .................................................................................. 7
Chapter 3 Natural Gas and Liquefied Natural Gas Supply Chain in Spain ............................. 9
3.1 Trade of Natural Gas and Liquefied Natural Gas ...................................................................... 9
3.1.1 Import of Natural Gas and Liquefied Natural Gas ......................................................... 10
3.1.2 Export of Natural Gas and Liquefied Natural Gas ........................................................ 13
3.2 Domestic Consumption of Natural Gas and Liquefied Natural Gas ............................ 16
3.3 Large-Scale Infrastructure .................................................................................................................. 17
3.3.1 Liquefied Natural Gas Import Terminals ............................................................................ 18
3.3.2 Natural Gas Pipelines ................................................................................................................... 21
iv
3.3.3 Underground Storage Tanks of Natural Gas ................................................................... 22
3.4 Small-Scale Infrastructure .................................................................................................................. 23
3.4.1 Liquefied Natural Gas Satellite Terminal ........................................................................... 23
3.4.2 Liquefied Natural Gas Fuel Loading Road Station ....................................................... 25
3.4.3 Liquefied Natural Gas Fuel Loading Ship Station ........................................................ 27
3.4.4 Liquefied Natural Gas Bunker Ships .................................................................................... 28
3.4.5 Modes of Small-Scale Liquefied Natural Gas transportation ................................. 28
3.4.5.1 Trucks ................................................................................................................................................ 28
3.4.5.2 Ships .................................................................................................................................................. 29

3.5 Vertical Integration of Natural Gas and Liquefied Natural Gas Supply Chain in
Spain ...................................................................................................................................................................... 30
3.6 Seasonality in the Natural Gas and Liquefied Natural Gas Supply Chain ............... 31
Chapter 4 Methodology and Data Collection ........................................................................................ 32
4.1 Network Optimisation Model ............................................................................................................. 34
4.1.1 Players in the Model ...................................................................................................................... 35
4.1.2 Assumption of the Model ............................................................................................................ 36
4.1.3 Scenario Description .................................................................................................................... 36
4.1.4 Description of Optimisation Model ....................................................................................... 36
4.2 Regression Model with Seasonality for Forecasting ........................................................... 38
4.2.1 Analysis of the relationships of Pearson correlation and p-value ...................... 41
4.2.2 Analysis of the regression models with P-value, R-squared .................................. 41
4.3 Autoregressive Model for Forecasting ........................................................................................ 44
4.4 Data Collection.......................................................................................................................................... 44
4.4.1 Demand for Liquefied Natural Gas Direct Consumption .......................................... 44
4.4.2 Distribution of the Demand in the Autonomous Community Scale. ................... 47

4.4.3 Distance between Liquefied Natural Gas Import Terminals and Autonomous
communities .................................................................................................................................................. 47
4.4.4 Lot Size of the Trucks ................................................................................................................... 48
4.4.5 Capacity of LNG Transportation by Trucks in the LNG terminals ....................... 48
4.4.6 Transportation Costs .................................................................................................................... 48
v
4.4.7 Design of Transportation Networks ................................................................................. 50
4.4.8 Pipeline Holding costs ................................................................................................................. 50
Chapter 5 Simulations and Results ........................................................................................................... 53
5.1 Result of Scenario 1. Full Capacity ............................................................................................... 53
5.1.1 Allocation of LNG ............................................................................................................................ 53
5.1.2 Excess capacity ............................................................................................................................... 55
5.1.3 Total costs .......................................................................................................................................... 55
5.2 Result of Scenario 2. Abundant Capacity .................................................................................. 56
5.2.1 Allocation of LNG ............................................................................................................................ 56
5.2.2 Excess capacity ............................................................................................................................... 57
5.2.3 Total costs .......................................................................................................................................... 58
5.3 Result of Scenario 3. A lack of Capacity ..................................................................................... 59
Chapter 6 Conclusion ........................................................................................................................................ 61
6.1 Summary of the problem and methods ....................................................................................... 61
6.2 Key findings ............................................................................................................................................... 61
6.3 Implication ................................................................................................................................................... 62
6.4 Limitation of the research and Suggestion of the Further Research ......................... 63
Bibliography ........................................................................................................................................................... 64
Annex A ..................................................................................................................................................................... 72
1. MILP model in Excel Solver. ................................................................................................................. 72
Annex B ..................................................................................................................................................................... 73
1.Estimated LNG demands in the autonomous community level in scenario 1. ......... 73
2.Estimated LNG demands in the autonomous community level in scenario 2 .......... 73
3.Estimated LNG demands in the autonomous community level in scenario 3. ......... 73
Annex C ..................................................................................................................................................................... 75
1.Distribution of LNG demand in Scenario 1.1. (GWh) ............................................................... 75
2. Distribution of LNG demand in Scenario 1.2. (GWh) .............................................................. 78
3. Distribution of LNG demand in Scenario 2.1. (GWh) .............................................................. 80
4.Distribution of LNG demand in Scenario 2.2. (GWh) ............................................................... 83
vi
List of tables
Table 1 Literature Matrix of SSLNG optimization. .................................................................... 8
Table 2 Pearson correlations between NG, LNG consumption and heating, cooling degree
days in Spain between 2009 and 2017. Source: Eurostat, Corporacion de Reservas
Estrategicas de Productos Petroliferos .................................................................................. 17
Table 3 List of provinces in Spain. Source: Tourism in Spain ................................................ 18
Table 4 The ownership, number of LNG tanks and storage capacity of terminals in Spain.
Source: Enagas. ..................................................................................................................... 21
Table 5 Coverage areas, length, number of sections and owners of gas pipeline in Spain.
Source: Enagas. ..................................................................................................................... 22
Table 6 Location and distribution of LNG satellite stations in each Spanish province. Source:
Gas Infrastructure Europe. ..................................................................................................... 24
Table 7 Owners/operators of LNG fuel loading ship stations in Spain. Source: Gas
Infrastructure Europe.............................................................................................................. 27
Table 8 Location and distribution of LNG fuel road and ship stations in each Spanish province.
Source: Gas Infrastructure Europe. ....................................................................................... 27
Table 9 Road tractors by type of motor energy in Spain in 2016. source: Eurostat ............... 28
Table 10 Ownership of large- and small-scale LNG facility. Source: Enagas, Gas Infrastructure
Europe .................................................................................................................................... 30
Table 11 Regression models for the consumptions by Minitab. Source: Corporacion de
Reservas Estrategicas de Productos Petroliferos .................................................................. 43
Table 12 Interpretation of the regression models of the consumptions of LNG and NG in Spain
by Minitab. Source: Corporacion de Reservas Estrategicas de Productos Petroliferos ........ 44
Table 13 Estimated demands from the destinations in Spain. Source: Enagas and Corporacion
de Reservas Estrategicas de Productos Petroliferos ............................................................. 46
Table 14 Number of trucks required for each scenario. ......................................................... 46
Table 15 Autonomous Communities having provinces with small-scale LNG facilities. Source:
Gas Infrastructure Europe ...................................................................................................... 47
Table 16 LNG truckload capacity in LNG terminals in Spain. Source: Enagas...................... 48
Table 17 Transportation costs in the scenario of the study. Source: American Transportation
Research Institute, European Commission and Dünnebeil et al. .......................................... 50
Table 18 Transit time depends on distances. Source: Reed TMS Logistics .......................... 51
Table 19 Forecasted Estimated Landed Price of LNG and Holding Costs in 2018. Source:
Waterborne Energy, Inc cited in Federal Energy Regulatory Commission ............................ 52
Table 20 Monthly allocations of LNG between LNG import terminals and destinations in
Scenario 1.1. .......................................................................................................................... 53
Scenario 1.2. .......................................................................................................................... 54
Table 22 Excess capacity with El Musel Terminal in Scenario 1.1. (GWh). ........................... 55
Table 23 Excess capacity without El Musel Terminal in Scenario 1.2. (GWh). ...................... 55
Table 24 Total costs in Scenario 1.1. and 1.2......................................................................... 56
Scenario 2.1. .......................................................................................................................... 56
Scenario 2.2. .......................................................................................................................... 57
Table 27 Excess capacity with El Musel Terminal in Scenario 2.1. (GWh). ........................... 58
Table 28 Excess capacity without El Musel Terminal in Scenario 2.2 (GWh). ....................... 58
Table 29 Total costs in Scenario 2.1. and 2.2......................................................................... 58
Table 30 Result of the distribution of LNG in December in Scenario 3 .................................. 59
Table 31 Results of excess capacity and the total costs in December in Scenario 3. ........... 60
vii
List of figures
Figure 1 Flows of Methodology and Research Design ............................................................................. 4
Figure 2 Structure of the thesis. .......................................................................................................................... 4
Figure 3 LNG supply chain in Spain. ................................................................................................................ 9
Figure 4 Total LNG and NG import in Spain. Source: Source: Corporacion de Reservas
Estrategicas de Productos Petroliferos ......................................................................................................... 10
Figure 5 Seasonal Index of Import of NG and LNG. Source: Corporacion de Reservas
Figure 6 NG import from outside of Spain. Source: Source: Corporacion de Reservas
Figure 7 NG export and import between Portugal and Spain. Source: Source: Corporacion de
Reservas Estrategicas de Productos Petroliferos .................................................................................... 12
Figure 8 NG export and import between France and Spain. Source: Source: Corporacion de
Figure 9 LNG import to Spain. Source: Corporacion de Reservas Estrategicas de Productos
Petroliferos ................................................................................................................................................................ 13
Figure 10 NG export from Spain to France and Portugal. Source. Source: Corporacion de
Figure 11 Worldwide LNG export from Spain. Source: Corporacion de Reservas Estrategicas
de Productos Petroliferos .................................................................................................................................... 14
Figure 12 LNG export to South Korea. Source: Corporacion de Reservas Estrategicas de
Productos Petroliferos .......................................................................................................................................... 15
Figure 13 Landed LNG prices in Spain and South Korea from 2013 to 2015. Source:
Waterborne Energy, Inc cited in Federal Energy Regulatory Commission .................................... 15
Figure 14 Gross inland consumption in Spain. Source: Eurostat ...................................................... 16
Figure 15 Natural gas consumption in Spain. Source: Corporacion de Reservas Estrategicas
de Productos Petroliferos .................................................................................................................................... 16
Figure 16 Network of large-scale LNG and NG supply chain in the Iberia Peninsula of Spanish
territory. Source: Enagas ..................................................................................................................................... 18
Figure 17 Monthly send out the level of regasified gas from the 6 LNG import terminal in Spain
from 2012 to 2017 Source: Gas Infrastructure Europe .......................................................................... 19
Figure 18 Seasonal Index (SI) of sending out of natural gas after regasification of LNG from 6
LNG import terminals in 12 months in Spain. Source: Gas Infrastructure Europe ..................... 20
Figure 19 Monthly level of the three gas storage tanks in Spain from 2009 to 2017 Source:
Enagas ........................................................................................................................................................................ 22
Figure 20 Seasonal Index (SI) of the level of inventory in three gas storages in 12 months in
Spain. Source: Enagas ........................................................................................................................................ 23
Figure 21 Location and distribution of LNG satellite stations in each Spanish province. Source:
Gas Infrastructure Europe. ................................................................................................................................. 24
Figure 22 Owners/operators of LNG satellite terminals in Spain. Source: Gas Infrastructure
Europe ......................................................................................................................................................................... 25
Figure 23 Owners/Operators of LNG fuel loading road stations in Spain. Source: Gas
Infrastructure Europe ............................................................................................................................................ 26
Figure 24 Locations and distribution of LNG fuel road and ship stations in each Spanish
viii
province. Source: Gas Infrastructure Europe ............................................................................................. 27
Figure 25 Numbers of vehicles fuelled by LNG in Spain from 2013 to 2016. Source: Eurostat.
........................................................................................................................................................................................ 29
Figure 26 Seasonality in NG and LNG supply chain. Source: Gas Infrastructure Europe,
Enagas and Corporacion de Reservas Estrategicas de Productos Petroliferos ......................... 31
Figure 27 Models for the research. ................................................................................................................. 32
Figure 28 Relationships between MILP model and scenarios. ........................................................... 33
Figure 29 Data collections for inputs of the MILP model. ...................................................................... 33
Figure 30 Flows of LNG and NG in Spain. Source: Enagas ................................................................ 34
Figure 31 Descriptions of the MILP model. ................................................................................................. 35
Figure 32 Players of the model. ....................................................................................................................... 35
Figure 33 Process of forecasting future demand...................................................................................... 39
Figure 34 LNG direct consumption between 2009 and 2018 in Spain. source: Corporacion de
Figure 35 NG consumption for electricity generator between 2009 and 2018 in Spain. Source:
Corporacion de Reservas Estrategicas de Productos Petroliferos ................................................... 42
Figure 36 NG consumption for conventional between 2009 and 2018 in Spain. Source:
Corporacion de Reservas Estrategicas de Productos Petroliferos ................................................... 42
Figure 37 A total NG and LNG between 2009 and 2018 in Spain. Source: Corporacion de
Figure 38 Seasonal Index (SI) of the estimated landed price of LNG in Spain in 12 months in
Spain. Source: Waterborne Energy, Inc cited in Federal Energy Regulatory Commission .... 52
Figure 39 Single or Multiple LNG import terminals to supply LNG to Autonomous Communities
in Spain....................................................................................................................................................................... 62
ix
List of Abbreviation
BOG: Boil-off gases
CNG: Compressed natural gas
Cbm: Cubic Metres
GHG: Greenhouse gas
GWh: Gigawatt hours
Km: Kilometre
LNG: Liquified natural gas
LTL: Less than truckload
MMBTU: One million British Thermal Units
R-sq: R-squared
SSLNG: small-scale liquified natural gas
TL: Truckload
US: United States
x
Chapter 1 Introduction
1.1 The topic

There is the growing interest in LNG ever since International Maritime Organization
(IMO) has imposed the regulation against pollution derived from fuels of ships
(Bittante, et al., 2017b). LNG is considered an eco-friendly resource and one of the
best alternative fuels (Bittante, et al., 2017b). Also, LNG is necessary for the isolated
area where the access of the pipeline for transportation of natural gas is not available
(Bittante, et al., 2017a).
The concept of small-scale LNG is different from the one of large-scale LNG.
Transport mode of large-scale LNG is a massive size of LNG ships and pipelines while
the one of small-scale LNG is the small size of ships, trucks and trains (Jokinen, et
al., 2015). The organisation of the mode for small-scale LNG is relatively more
complicated than the one for large-scale LNG.
1.2. Problem Statement

This difference in the level of complication is derived from the lot size and the number
of destinations. In the large-scale LNG, the LNG import terminal is the only destination
for the relatively small number of ships carrying a considerable volume of LNG.
However, In the small-scale LNG, many numbers of small-scale transportation modes
are required to handle the capacity of LNG to various destinations.
Seasonality brings the problem of surplus derived from the gap between peak and off-
peak demands (Prentice and Prokop, 2016). In SSLNG supply chain, truck companies
have issues from the surplus. In order to tackle the problems, it is necessary to set a
proper pricing strategy and allocation plan for trucks. So, the issue derived from the
surplus of trucks in the off-peak season can be solved by understanding the
seasonality of the LNG supply chain. However, the strategy of the truck companies to
handle the surplus is out of scope in the study, so it is not discussed.
Also, understanding the seasonality is vital for LNG import terminals and SSLNG
facilities to estimate demands.
Non-vertical integration of the supply chain tends to result in several issues such as
delaying of sharing information, increasing transaction costs (Williamson, 1983). The
obstacles are highly likely to discourage the optimisation of the supply chain.
It would be the best solution if there were no limit of LNG truckload transportation
output capacity at LNG import terminal. In that ideal situation, SSLNG facilities could
order SSLNG from the nearest LNG receiving terminals. However, there are limits to
the output capacity at the terminal in reality. So, the limit capacity sometimes forces
the SSLNG facilities to be provided LNG from the terminal where is not geographically
closest. This issue is relevant for LNG import terminals and SSLNG facilities since
they are decision makers to decide the distribution of the SSLNG.
EL Musel LNG import terminal was constructed entirely in 2013, but it has been
mothballed after the construction due to the regulation of the Spanish government
concerning the excessive capacity of natural gas with the expectation of less demand
of the resource (Prontera, 2017). In other words, the government has an intention of
avoiding the situation of the surplus of LNG by importing more LNG through El Musel
terminal. Although the owner of the terminal has been negotiating to utilise the
terminal in different ways such as gas storage, it would be the best way to operate
the terminal as the original purpose to recoup a considerable investment (about 380
million euros) for the terminal (El Comercio, 2017).
1
1.3 Background and Justification
In order to meet the various demands, not only the large-scale also small-scale LNG
supply chains need to be facilitated. However, the small-scale supply chain is
underestimated, unlike the large-scale one which is traditionally well developed and
optimised (Jokinen, et al., 2015). However, recently, several journals written by
Bittante, et al. (2016) Bittante, et al. (2017a), Bittante, et al. (2017b) and Jokinen, et
al. (2015) focus on SSLNG supply chain to research the area.
1.4 Deficiencies in the Evidence

Most case studies of prior researches such as Bittante, et al. (2017b) and Jokinen, et
al. (2015) regarding small-scale LNG supply chain analysed the supply chain in
Finland sometimes coupled with Sweden. So, in this paper, case studies in Spain is
conducted to fill this gap in the lack of case studies in different areas. Another reason
for opting for Spain is the size of the small-scale LNG infrastructure. According to Gas
Infrastructure Europe (2018b, Spain is the largest country regarding the number of
small-scale LNG facilities by accounting for 37% of Europe. The large numbers in
Spain must cause more issues in the distribution of small-scale than the other
countries. If the paper studied other countries’ SSLNG supply chain with a few
numbers of the small-scale facilities, the answer would be straightforward and obvious.
1.5 Audience
LNG import terminals, truck companies and small-scale LNG facilities are the
audiences of the research. LNG import terminals is the origin of the SSLNG and
distribute SSLNG by trucks companies to the destinations which are SSLNG facilities.
1.6 Research Objective

Overall objective is “To optimise SSLNG supply chain between LNG import terminals
and SSLNG facility by means of truck transportation in Spain”.
Four specific objectives are below.

1. To analysis the seasonality of NG and LNG at trade, LNG import terminals,
gas storage, consumption.
2. To measure the degree of integration among LNG import terminals, natural
gas pipeline, LNG fuel stations, LNG satellite stations.
3. To analysis the effect of the seasonality to the numbers of trucks and LNG
import terminals needed for SSLNG facilities on a monthly basis.
4. To assess the role of El Musel terminal in the SSLNG supply chain
comparing situations with and without operating the terminal.
1.7 Research Questions

There are four research questions for the objectives introduced.
1. What are the features of the seasonality of NG and LNG in import,

distribution and consumption in Spain?
2. To what extent does the level of the vertical integration exist from LNG
import terminals to SSLNG facilities in Spain?
3. To what extent do changes in SSLNG demand affect the allocation of small-
scale LNG between LNG import terminals and SSLNG facilities in Spain.
4. To what extent does El Musel LNG import terminal play a key role in different
SSLNG demands in small-scale LNG supply chain in Spain?
2
The first two questions are answered in Chapter 2. The research question 1 will be
solved by a regression model with the relevant data. Searching for the ownership of
the facilities in the supply chain solves the research question 2.
The answers for the two questions are essential to understanding the characteristics
of SSLNG industry in Spain since they are linked in the massive scale of the supply
chain. Answers for the other two are in Chapter 5. MILP model will solve the two
research questions. The different input of the SSLNG demand in the scenario will
solve the question 3. Furthermore, including or excluding El Musel terminal leads to
the answer for question 4.
1.8 Methodology and Research Design

Figure 1 shows the overview of the methodology and research design. The principal
methodology of this paper is Mixed Integer Linear Programme by Excel Solver in order
to optimise the small-scale LNG supply chain between LNG import terminals and
destinations which are LNG fuel stations and satellite terminals. Another model is the
regression model to forecast future consumption of SSLNG and price of LNG. Also,
the estimation of fuel costs is done by the Autoregressive model by Excel.
On the other hands, there are several inputs for the model. At first, SSLNG demand
in the autonomous communities’’ scale is estimated by using the regression. After
forecasting the future demand, the demand is allocated to the autonomous community
covering the SSLNG facilities based on the population ratio in the related community.
Another is the total transportation costs consisting of transportation costs and
inventory holding costs. Transportation costs consider distance, fuel efficiency and
fuel costs together to calculate the fuel costs of trucks per kilometres. Also, other
transportation costs such as expenses for insurance, repair and maintenance are
included in the costs. Pipeline holding costs are estimated by holding costs rate and
price of LNG. The estimation of the rate and price are from other journals and running
the regression model respectively.
The data of maximum SSLNG capacity for truck transportation at LNG import
terminals is acquired from the terminal’s website.
The full integration helps to improve the delay of information. Also, transaction costs
are indeed avoided by the integration. In the study, the integration is assumed in the
supply chain.
The research is designed in three scenarios having different demands of small-scale
LNG demands. Before mentioning the scenarios, the research area is limited into
Spanish Iberian Peninsula meaning that excluding Spanish Islands in order to focus
on LNG truck transportation, not other LNG transportations. The demand of LNG
almost reaching the maximum output capacity (full capacity) of LNG truckload
transportation from LNG import terminals is verified in scenario 1. For the sub-
scenarios for the main one is researching the effect of the existence (scenario 1.1.)
and non-existence (scenario 1.2.) of El Musel LNG receiving terminals. Subsequently,
scenario 2 (abundant capacity) simulates more margins of the output LNG truckload
capacity to meet the demand. The scenario two also contains the two sub-scenarios
depending on the operation of the El Musel terminal in the same way as scenario 1.
Scenario 3 examines a case of a lack of output capacity from LNG import terminals
not satisfying the demand.
3
Figure 1 Flows of Methodology and Research Design
1.9 Thesis Structure

Figure 2 illustrates the overview of this paper. At first, introductions with research
questions are stated. After that, other relevant papers about this thesis are discussed
in the literature review. Subsequently, trade of the resources and LNG supply chain in
Spain are explored. The next chapter explains methodology, and data collection for
the study. After the simulations of the model, the results are analysed. Finally, the last
chapter will conclude the study.
Figure 2 Structure of the thesis.
4
Chapter 2 Literature Review
Three topics which are seasonality, vertical integration and optimisation of SSLNG
supply chain are discussed in the literature review. Chapter 2.1 discusses seasonality
of LNG in other countries. The seasonality is the essential feature of the supply chain
that it needs to understand before simulating the network optimisation model.
Designing a robust network is vital to deal with the seasonal changes. In other words,
cooperating to design the robust network under seasonal changes is critical to
optimise the supply chain. Regardless of the seasons, the well-prepared supply chain
has a lot more efficient network available than the one does not consider the
seasonality. The flexible network is not fully demanded toward winter nor summer.
The dimension of the network, capacity, for example, is built with having quite an
efficient operation that both winter and summer.
This discussion demonstrates that the seasonality of LNG in Spain is not the
distinctive feature of Spain but occurs in other countries as well. Also, the reasons
and effect of the seasonality are discussed in other countries, and these can be
compared to the cases in Spain.
Vertical integration is discussed in Chapter 2.2. Vertical integration is necessary for
the assumption of the optimisation model in the paper. If there is no vertical integration,
it is difficult to handle the whole supply chain because every single SSLNG facility
company makes different contracts with adverse effects of non-vertical integration.
Less willingness to sacrifice for the common benefit of the entire supply chain and
more transaction costs increasing the total costs are highly likely to be observed. So,
in order to optimise the supply chain, it is assumed that the SSLNG supply chain is
vertically integrated. At first, the theory of the vertical integration is discussed in detail.
Subsequently, the theory applies to the natural gas industry to discuss the market with
the vertical integration.
Chapter 3.3 shows SSLNG optimisation model in different papers. Apparently, the
model is the major methodology so it is important to discuss. A matrix of the literature
review is provided for a clear understanding of the trends of the SSLNG study. Types
of models had been used in the majority of the papers are introduced first. The time
period of the simulation, assumptions and target areas in the relevant papers are
discussed in the order.
2.1 Seasonality
Seasonality of LNG demand in many countries is observed from various papers and
they are discussed in this section. In the US, the seasonality coupled with long-term
trend are the critical factors of domestic LNG demand (Zhu, 2008). Increases in
economy and population lead to growing demands of LNG for resident and commerce
sectors such as electric power, cooling and heating along with industry sectors in need
of the gas and storage (Zhu, 2008). Also, some other reasons such as preventing
environment by decreasing air pollutions, expending the size of houses lead to the
growing demand (Zhu, 2008).
Seasonality of the consumption of natural gas is seen in the short-term by increasing
the usage for heating purpose in cold days and cooling purpose at high-temperature
days (Zhu, 2008). This seasonality affects the price of natural gas; In the winter
season, there is relatively high price while the price is moderate in summer based on
data from The International Energy Agency (Eliston, 2009). This consumption trend is
easily seen in the residence and commercial sectors for the same purposes while
industry and transportation sectors have stable demands (Zhu, 2008).
Storing natural gas during off-peak seasons are caused by the unbalance between
the production of natural gas which is not seasonal and consumption of natural gas
5
which is seasonal (Zhu, 2008). The inventory of the natural gas enables to meet the
seasonality of the demand (Zhu, 2008). The term “peak shaving” explains the situation
above which is buying natural gas at low cost in the off-peak season and liquified it
as LNG to store (Zhu, 2008). Later, the LNG is regasified to be sold at a higher price
in the peak seasons (Zhu, 2008). Also, some inventory of LNG without regasification
are transported by trucks (Zhu, 2008).
Unlike Zhu (2008), Anne (2008) argued by using generated decorrelated time series
that there were no clear relationships between the seasonality of the price of natural
gas and demands in the US since power generations need natural gas in summer
and winter.
However, Anne (2008) mentioned that Europe market had the strong relationships
between the seasonality of price and demand for natural gas. Because most of the
natural gas was for heating in winter seasons (Anne, 2008). Eliston (2009) supported
this argument by showing the data on the LNG price seasonality in Europe from The
International Energy Agency.
In Asian countries, a strong seasonal demand was observed in South Korea (Wood,
2007). This country met the seasonal demand by buying LNG in short-term contracts
(Wood, 2007). So, it was critical for the Korean market to be affected by the fluctuation
of the LNG price in spot markets (Wood, 2007). In order to overcome the vulnerability
derived from the fluctuation, the capacity of the LNG storage tanks was expended,
and the new LNG terminal was built in Korea (Wood, 2007).
Another country in Asia is Japan also having seasonality of natural gas consumption
(Eliston, 2009). Unlike the increase in demand of natural gas for cooling purpose in
the US mentioned by Zhu (2008) and Eliston (2009), there is seasonal demand for
heating purpose in Japan which is observed by the higher price from August to
December based on the data from The International Energy Agency (Eliston, 2009).
This seasonality is similar to the one in Europe (Eliston, 2009).
2.2 Vertical Integration

Williamson (1985, p. 55) introduced the necessity of the vertical integration by
mentioning “durable investments that are undertaken in support of particular
transactions, the opportunity cost of which investments is much lower in best
alternative uses or by alternative users should the original transaction be prematurely
terminated”. Arfaa et al. (2011) stated that creating value and enhancing it are the
advantages of the vertical integration enabling a single firm to run more than one
company.
Based on the Williamson’s theory (1983, 1985), Neumann and Ruster (2006) suggest
that there are three driving factors which are asset specificity, the frequency of
transactions and uncertainty having influences on the high transaction costs in LNG
supply chain. In order to reduce the transaction costs, it is highly likely to have a
structure of the vertical integration (Neumann and Ruster, 2006).
Williamson (1983) mentioned the four types of the asset specificity: site specificity,
physical asset specificity, human specificity and dedicated asset (Williamson, 1983).
Site specificity means the proximity is required for the efficiency of processing and
reducing costs (Williamson, 1983). Also, immobility of the asset resulting in costing a
substantial capital in case of the relocation or new construction (Williamson, 1983).
The second one is physical asset specificity meaning the usage or application for
certain customers and products, not for others in general (Williamson, 1983). The third
one is human asset specificity meaning that special knowledge acquired from tasks
(Williamson, 1983). The last one is a dedicated asset about producing most of the
products to meet a specific demand (Williamson, 1983).
6
The argument of Neumann and Ruster (2006) was agreed that the more asset is
transaction-specific, the more possibility of gaining profit from opportunism (Klein,
Crawford, and Alchian, 1978, cited in Doane and Spulber, 1994). In other words, the
increase in contracting costs is more substantial than the one in vertical integration
costs (Klein et al., 1978, cited in Doane and Spulber, 1994).
Apart from the theoretical approach of the vertical integration, Neumann and von
Hirschhausen (2008) apply the vertical integration theory to large-scale LNG supply
chain along with long-term contract, not the small-scale LNG supply chain.
Three entities try to exploit benefits from the vertical integration (Neumann and Ruster,
2006). At first, companies in upstream want additional profits generated from
downstream business (Neumann and Ruster, 2006). Another entity is a transportation
firm producing margins from arbitrage between the upstream and downstream
(Neumann and Ruster, 2006). The last player is in downstream which is interested in
not only producing profits but also securing the supply of LNG from upstream to satisfy
the demand (Neumann and Ruster, 2006).
2.3 Optimisation of Small-Scale Supply Chain

A mixed integer linear programming (MILP) model was used to by a majority of
authors regarding SSLNG supply chain. Table 1 shows the matrix of literature about
SSLNG supply chain. Bittante, et al. (2018) research optimisation of the supply chain
by sea-sea shipping with MILP model. However, the authors did not address land
transports which are a part of small-scale LNG supply chain. Faaijf et al. (2017) used
the MILP model with the transports modes of the road, rail and inland short sea
shipping. Bittante, et al. (2016), Bittante, et al. (2017a), Bittante, et al. (2017b) and
Jokinen, et al. (2015) focused on small-scale LNG supply chain with truck and ship
transportation by using MILP model. Based on the trend of the using MILP models for
the optimisation of SSLNG supply chain in the other papers, the MILP model is also
used in this paper. The difference from the other journals is a selection of software
which is Excel Solver to simulate the model.
Most of the paper use 30 days as the time period of their research. For instance,
Jokinen, et al. (2015) set the time period in their paper as dividing a one year into
twelve 30-day periods. Other studies conducted by Bittante, et al. (2016), Bittante, et
al. (2017a) and Bittante, et al. (2018) also used a time horizon of 30 days. On the
other hand, single (10 days) and multi-period (30 days) cases were presented by
Bittante, et al. (2017b) which is different from the other papers. In the multi-period
results, the three factors are addressed additionally: the amount of LNG transported
by LNG by truck, the number of trucks and the related port-customers links (Bittante,
et al., 2017b). Bittante, et al. (2017b) found that less overall cost in multi-period results
which was more accurate. In a similar way of most of the journals, monthly (30 days)
time period in a year is used for the simulation of this paper.
Regarding assumptions, enough storage capacity from customers are widely
observed from the other journals (Bittante, et al., 2016, Bittante, et al., 2017a, Bittante,
et al. 2017b and Jokinen, et al., 2015). This assumption is also applied in this paper.
Several operational factors such as scheduling, supply availability linked to production
rate or inventory, boil-off loss, time windows at the terminals are not considered
(Bittante, et al., 2017b). This study also does not consider the operation factors like
the same way of Bittante, et al. (2017b).
Most case studies of prior researches such as Bittante, et al. (2017b), Bittante, et al.
(2016) regarding SSLNG supply chain analysed the supply chain in Finland coupled
with Sweden. Study of a single area of Finland was done by Jokinen, et al. (2015).
Also, Faaijf et al. (2017) focused Sweden for research area. Recently, the Caribbean
7
area was studied by Bittante, et al. (2018). On the other hand, there are still not
sufficient academic journals about SSLNG supply chain in other areas. Even though
SSLNG supply chain in Spain having a considerable amount of SSLNG facilities
according to Gas Infrastructure Europe (2018b), there are no journals covering this
country. So, this paper focuses on the supply chain in Spain to fill the gap of the other
areas.
Transportation Time Relevant

Author Year Methodology Research area
Modes horizon assumptions
-Enough storage
capacity from
Bittante, short-sea ships, 30
2016 destinations MILP Finland, Sweden
et al. trucks days
-no tank investment
costs
-Enough storage
capacity from
destinations
Bittante, short-sea ships, 30 -land transport is
2017a MILP Finland, Sweden
et al. trucks days only available on
weekdays
- no tank investment
costs
-Enough storage
capacity from
destinations
-no investment costs
10 - no consideration
Bittante, short-sea ships, days, (scheduling, supply
2017b MILP Finland, Sweden
et al. trucks 30 availability
days (production rate,
inventory),
boil-off gas, time
windows at the
terminals)
-Demand is smaller
Bittante, 30 than storage
2018 short-sea ships MILP Caribbean region
et al. days Capacity -different
LNG price
-no boiled off gas -
Faaijf et short-sea ships, Price of LNG is
2017 - MILP Sweden
al. trucks, trains assumed to be the
current value
-shipping strategy
decides the routes -
Jokinen, short-sea ships, 30
2015 Enough storage MILP Finland
et al. trucks days
capacity at
destinations
Table 1 Literature Matrix of SSLNG optimisation.
8
Chapter 3 Natural Gas and Liquefied Natural Gas Supply Chain in Spain
Figure 3 LNG supply chain in Spain.
Figure 3 illustrates the overview of Chapter 3. From chapter 3.1 to 3.3, seasonality of
NG and LNG is observable after researching the data of the NG and LNG supply
chain. Seasonality of LNG trade, large-scale LNG are shown in chapter 3.1 and 3.3
respectively. It is not available to access the direct data of flows of LNG in small-scale
LNG facilities, however, using the data of the consumption of LNG in chapter 3.2
enables to seek the seasonality in the small-scale LNG. Those data of seasonality in
the different stages in the supply chain are discussed altogether in chapter 3.6.
Chapter 3.1 and chapter 3.3 are necessary to discuss in advance of small-scale LNG
since they are in the same supply chain. Regarding a flow of LNG, small scale-LNG
in Spain would not exist unless there was the trade of LNG and large-scale LNG.
Demands for SSLNG is estimated from downstream, and the information is delivered
to upwards of the supply chain.
Vertical integration of the supply chain can reduce the obstacles derived from the
other negative factors. In other words, the full integration encourages to improve the
delay of information. In addition, transaction costs are indeed avoided by the
integration. Chapter 3.5 seeks the degree of the vertical integration in the LNG and
NG supply chain in Spain based on the information from chapter 3.3 and 3.4. So, in
order to optimise SSLNG supply chain, it is required to understand not only SSLNG
but also LNG trade and large-scale LNG since they are closely connected.
3.1 Trade of Natural Gas and Liquefied Natural Gas

The flow of the trade of NG and LNG is necessary to understand the SSLNG supply
chain since the SSLNG supply chain is closely linked to the large-scale LNG supply
derived from the trade.
The trade of the resources would not have played a key role in SSNLG if Spain
domestically produced the sufficient NG and LNG. However, the independence of the
exploits the resources is not observed in Spain. For instance, Spain merely produced
0.19% of NG out of total gas consumption in 2016 (Benito et al., 2017). In other words,
Spain heavily relies on importing NG and LNG for the domestic consumption in Spain.
The heavy reliance must affect SSLNG supply chain in Spain since the supplied LNG
is originally from the trade of LNG.
For example, in case of the issues decreasing in the production of NG and LNG in
some countries exporting the resources to Spain, players in SSLNG supply chain can
9
expect the reduction and find the solutions.
3.1.1 Import of Natural Gas and Liquefied Natural Gas

Figures 4 shows the data about LNG and NG import in Spain. At the beginning of the
given period, the amount of imported LNG approximately twice more than the one of
imported NG. However, the amount of imported NG exceeded the one of imported
LNG at most of the times since 2013. Also, the figure 4 illustrates that the seasonality
of import NG and LNG to meet the change in the demand for natural gas consumption
shown in figure 11 in chapter 2.6.
45,000 GWh
40,000 GWh
35,000 GWh
30,000 GWh
25,000 GWh
20,000 GWh
15,000 GWh
10,000 GWh
5,000 GWh
0 GWh
Jul-10
Jul-11
Jul-12
Jul-13
Jul-14
Jul-15
Jul-16
Jul-17
Jan-10
Jan-11
Jan-12
Jan-13
Jan-14
Jan-15
Jan-16
Jan-17
Total LNG import Total NG import

Total NG&LNG import
Figure 4 Total LNG and NG import in Spain. Source: Source: Corporacion de Reservas Estrategicas de
Productos Petroliferos
Benito et al. (2017) mention that most of NG are imported to Spain through Medgaz
and Maghreb pipelines (Benito et al., 2017) and figure 2 illustrates the amount of NG
imported from other countries to Spain. Figure 2 shows that except for Alegria, the
other relatively minor NG exporters to Spain are Norway, France and Portugal which
are European countries.
10
1.2 SI
1.0 SI
0.8 SI
0.6 SI
0.4 SI
0.2 SI
0.0 SI
1 2 3 4 5 6 7 8 9 10 11 12
Figure 5 Seasonal Index of Import of NG and LNG. Source: Corporacion de Reservas Estrategicas de
Figure 5 shows a seasonal index of import of the resources. The seasonal index is
relatively high in winter season during low between April and September. In other
words, Spain tends to import more NG and LNG in winter than the other seasons.
Figure 6 illustrates that the majority of NG is imported from Algeria to Spain while
Norway is the second largest country for exporting NG to Spain.
25,000 GWh
20,000 GWh
15,000 GWh
10,000 GWh
5,000 GWh
0 GWh
Feb-12
Feb-17
Jun-10
Sep-11
Sep-16
Nov-10
Jul-12
May-13
Jun-15
Aug-14
Nov-15
Jul-17
Apr-11
Apr-16
Mar-14
Jan-10
Oct-13
Dec-12
Jan-15
Dec-17
Algeria NG France NG Norway NG

Portugal NG Total NG import
Figure 6 NG import from outside of Spain. Source: Source: Corporacion de Reservas Estrategicas de
Figure 7 illustrates that even though there are both import and export of NG between
the two countries, import NG from Portugal only occurred around in 2013. It implies
that the pipelines are usually dedicated to distributing NG from Algeria to Portugal.
11
4,500 GWh
4,000 GWh
3,500 GWh
3,000 GWh
2,500 GWh
2,000 GWh
1,500 GWh
1,000 GWh
500 GWh
0 GWh Feb-12
Feb-17
Jun-10
Sep-11
Aug-14
Nov-10
Apr-11
Jul-12
May-13
Jun-15
Sep-16
Nov-15
Apr-16
Jul-17
Mar-14
Jan-10
Oct-13
Dec-12
Jan-15
Dec-17
Export NG to Portugal Import NG from Portugal
Figure 7 NG export and import between Portugal and Spain. Source: Source: Corporacion de Reservas
Estrategicas de Productos Petroliferos
The allocations of the two pipelines are through Irun and Larrau for France’s gas
market (Enagas, 2016). In a similar way of the transport of NG between Spain and
Portugal, however, figure 8 shows that Spain has been annually imported more NG
than exported NG against France through these pipelines except for 2010.
2,500 GWh
2,000 GWh
1,500 GWh
1,000 GWh
500 GWh
0 GWh
Feb-12
Feb-17
Jun-10
Nov-10
Apr-11
Sep-11
Jul-12
May-13
Jun-15
Sep-16
Aug-14
Nov-15
Jul-17
Apr-16
Mar-14
Jan-10
Oct-13
Dec-12
Jan-15
Dec-17
NG export to France NG Import from France
Figure 8 NG export and import between France and Spain. Source: Source: Corporacion de Reservas
A lack of clear explanation of NG from Norway to Spain can be explained by the

pipeline connection between France and Norway. France is one of major NG import
countries from Norway by pipeline so Norway must export NG to Spain through the
pipeline in France in a similar way to Maghreb pipeline from Algeria to Spain through
Morocco (Benito et al., 2017).
On the other hands, LNG is imported by LNG ships due to geographical restrictions
of inaccessibility of gas pipeline such as from the Middle East, Africa and America or
insufficient capacity of the gas pipeline such as Algeria. Within Europe, LNG from
12
Norway accounts for the most of LNG import (Corporacion de Reservas Estrategicas
de Productos Petroliferos, 2018c). Figure 8 shows that Nigeria, Qatar and Algeria are
the largest exporters of LNG to Spain. Figure 9 also illustrates that Norway is a
significant LNG exporter country in Europe to Spain while Peru and Trinidad & Tobago
play an essential role in importing LNG from America market.
Qatar LNG
United States LNG
Peru LNG
Libya LNG
Angola LNG
Portugal LNG
Netherlands LNG
Belgium LNG
2010 2011 2012 2013 2014 2015 2016 2017
Figure 9 LNG import to Spain. Source: Corporacion de Reservas Estrategicas de Productos Petroliferos
3.1.2 Export of Natural Gas and Liquefied Natural Gas

Figure 10 shows that Spain solely transports NG to France and Portugal regarding
the export of NG outside of Spain. It is noticeable that most of the distribution of the
NG from Algeria through Spain is on Portugal. It is explained by a lack of geological
access of Portugal than France. Portugal is more isolated from Europe since it is
located in the Iberia Peninsula. On the other hand, France has connections from many
countries in Europe.
4,500
4,000
3,500
3,000
2,500
2,000
1,500
1,000
500
May-16
May-10
Sep-10
May-11
Sep-11
May-12
Sep-12
May-13
Sep-13
May-14
Sep-14
May-15
Sep-15
Sep-16
May-17
Sep-17
Jan-11
Jan-10
Jan-12
Jan-13
Jan-14
Jan-15
Jan-16
Jan-17
France NG Portugal NG
Figure 10 NG export from Spain to France and Portugal. Source. Source: Corporacion de Reservas
13
The fact that Spain has 0.19% of NG domestic production in 2016 would make no
sense that Spain is the LNG and NG export country (Benito et al., 2017). However,
Spain was one of the European countries generating substantial profits from LNG
reloading (Timera Energy, 2013).
Reloading means the transaction of LNG from a storage tank receiving the cargo from
LNG ships to other vessels by reloading of the cargo (Zhuravleva, 2009). It is
considered as arbitrage since it is caused by price difference (Zhuravleva, 2009). In
Spain, Huelva LNG import terminal exploited the LNG reloading to generate profits
(Zhuravleva, 2009).
Even though Spain does not produce any LNG domestically, this country was the
largest exporter of LNG in 2014 because of the LNG reloading. (Enriquez and Rojo,
2016b).
70,000 GWh
60,000 GWh
50,000 GWh
40,000 GWh
30,000 GWh
20,000 GWh
10,000 GWh
0 GWh
2010 2011 2012 2013 2014 2015 2016 2017
Total Africa LNG Total America LNG Total Europe LNG

Total Middle East LNG Total Asia-Pacific LNG
Figure 11 Worldwide LNG export from Spain. Source: Corporacion de Reservas Estrategicas de
Figure 11 shows that the amount of reloading of LNG is the most significant in 2014
throughout the years. The major importers of Spanish reloading LNG are Asian
countries and America countries. Between 2011 and 2015, the amount of Spanish
reloading LNG roughly remains the same in Europe, while rapid increases of importing
of LNG are observed in America from 2012 to 2013. Also, the amount of reloading to
Asia substantially increase in the following year, so the amount is doubled from 2013
to 2014. After reaching the highest amount of the reloading in 2014, the amount
substantially decreased to approximately 15,000 GWh in 2015. Following years, the
amount of reloading barely exists.
14
2,500 GWh
2,000 GWh
1,500 GWh
1,000 GWh
500 GWh
0 GWh
May-13
Sep-11
Sep-12
Sep-13
Sep-14
Sep-15
Sep-16
May-11
May-12
May-14
May-15
May-16
Jan-12
Jan-11
Jan-13
Jan-14
Jan-15
Jan-16
Figure 12 LNG export to South Korea. Source: Corporacion de Reservas Estrategicas de Productos
Petroliferos
20 $/MMBtu
18 $/MMBtu
16 $/MMBtu
14 $/MMBtu
12 $/MMBtu
10 $/MMBtu
8 $/MMBtu
6 $/MMBtu
4 $/MMBtu
2 $/MMBtu
0 $/MMBtu
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Spain Korea
Figure 13 Landed LNG prices in Spain and South Korea from 2013 to 2015. Source: Waterborne Energy,
Inc cited in Federal Energy Regulatory Commission
Figure 12 shows there was the high frequency of reloading of LNG from Spain to
Korea between at the end of 2013 and at the beginning of 2015. The primary reason
for the reloading is explained by a vast difference in landed LNG price between the
two countries at the similar period. Figure 13 shows that there was a gap of at least 2
$/ one billion British thermal units (MMBtu) before the beginning of 2015. So, it was
possible for the Spanish market to generate profits from the reloading. However, after
January in 2015, the price difference between the two markets decrease to less than
2 $/MMBtu discouraging the reloading from Spain to Korea. In other words, the price
elasticity is relatively high for the reloading market. The price gap which is around 2
$/MMBtu between the two markets is the threshold for the decision of reloading. This
situation is also explained by Zhuravleva (2009) arguing that the price difference is
the most critical factor in the reloading among other factors such as contractual
15
limitation, technical restrictions. The price gap should cover the expense of
transaction and lead to favourable interest for aggregators (Zhuravleva, 2009).
The MMBtu is equivalent to a million of British thermal units (BTU), and BTU stands
for “the amount of heat required to increase the temperature of a pint of water (which
weighs exactly 16 ounces) by one degree Fahrenheit” (Energy Vortex, n.d., para 1.).
3.2 Domestic Consumption of Natural Gas and Liquefied Natural Gas
7,000,000 Terajoule
6,000,000 Terajoule
5,000,000 Terajoule
4,000,000 Terajoule
3,000,000 Terajoule
2,000,000 Terajoule
1,000,000 Terajoule
0 Terajoule
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016
All products Natural gas Total petroleum products

Renewable energies Nuclear heat Others
Figure 14 Gross inland consumption in Spain. Source: Eurostat
Figure 14 shows that natural gas is the second most abundant resource regarding the
gross inland consumption in Spain between 2007 and 2016.
45,000 GWh
40,000 GWh
35,000 GWh
30,000 GWh
25,000 GWh
20,000 GWh
15,000 GWh
10,000 GWh
5,000 GWh
0 GWh
Feb-12
Feb-17
Jun-10
Sep-11
Sep-16
Nov-10
Jul-12
May-13
Jun-15
Aug-14
Nov-15
Jul-17
Apr-11
Apr-16
Jan-10
Oct-13
Mar-14
Dec-12
Jan-15
Dec-17
Conventional Electricity generation

LNG for direct consumption Total
Figure 15 Natural gas consumption in Spain. Source: Corporacion de Reservas Estrategicas de

Markets of NG and LNG consumptions are shown in figure 15. According to

Corporacion de Reservas Estrategicas de Productos Petroliferos (2018), the
domestic consumptions of natural gas in Spain consists of conventional, electricity
generation and LNG for direct consumption. The conventional demand covers
16
Business, households and industries. Natural gas consumption for conventional
usage accounts for the most of the consumption during an entire of the given period
except for 2010. For instance, the usage of NG for conventional purpose accounted
for more than 75% of total consumption of natural gas in 2017. Also, seasonality of
the consumption of natural gas is also observable in figure 15. The demand for natural
gas tends to increase in the winter season while the one decrease in the summer
season. The change in demand depending on the seasons implied that the usage of
the natural gas is for heating.
NG, LNG consumption Heating degree days Cooling degree days

Pearson correlation 0.955679 -0.638376
Conventional (NG)
P-Value =< 0.0001 =< 0.0001
Pearson correlation -0.07028 0.196803
Electricity Generator (NG)
P-Value 0.4698 0.0412
LNG direct use (LNG)
P-Value 0.3544 0.6534
Total consumption
P-Value =< 0.0001 =< 0.0001
Table 2 Pearson correlations between NG, LNG consumption and heating, cooling degree days in Spain
between 2009 and 2017. Source: Eurostat, Corporacion de Reservas Estrategicas de Productos
Petroliferos
Chapter 3.1.3 explains Pearson correlations and p-value briefly. In order to check the
relevance between the consumptions and heating purpose, heating and cooling
degree days will be useful. Eurostat (n.d., para. 3) defines that “heating degree day is
a weather-based technical index designed to describe the need for the heating energy
requirements of buildings”. In addition, according to Eurostat (n.d., para. 3) “cooling
degree day is a weather-based technical index designed to describe the need for the
cooling (air-conditioning) requirements of buildings”.
Table 2 highlights the important correlation between heating days and conventional
consumption of NG. Pearson correlation close to 1 and very low p-value (less than
0.001) implies that there is overwhelming evidence supporting the positive correlation
between heating days and conventional consumption of NG (Keller, 2014). So, we
can conclude that the principal purpose of conventional consumption of NG is for
heating. Vice versa, for cooling days, the Pearson correlation is -0.64 which is
relatively strong negative figures along with a very low p-value (less than 0.0001).
This implies that the NG consumption for conventional purpose decreases during the
summer season. Other two consumptions for the electricity generators and direct LNG
use are not relevant for heating and cooling days judging from the Pearson
correlations close to 0 meaning weak linear relationships and relatively high p-value
meaningless evidence supporting the relationships (Keller, 2014).
On the other hands, since the NG consumption for conventional purpose accounts for
large parts of the total consumption of the LNG and NG, it has a significant effect on
the total consumption. This effect can be observed from a very low p-value (less than
0.001and high figures of Pearson correlations on the total consumption (Keller, 2014).
3.3 Large-Scale Infrastructure
17
Figure 16 Network of large-scale LNG and NG supply chain in the Iberia Peninsula of Spanish territory.
Source: Enagas
1 A Coruna 11 Huelva 21 Cantabria 31 Murcia 41 Navarra

2 Pontevedra 12 Sevilla 22 Burgos 32 Alicante 42 Huesca
3 Lugo 13 Cadiz 23 Segovia 33 Valencia 43 Lleida
4 Ourense 14 Malaga 24 Madrid 34 Teruel 44 Barcelona
5 Asturias 15 Cordoba 25 Guadalajara 35 Zaragoza 45 Girona
6 Leon 16 Ciudad Real 26 Cuenca 36 Soria 46 Tarragona
7 Zamora 17 Toledo 27 Albacete 37 La Rioja 47 Castellon
8 Salamanca 18 Avila 28 Jaen 38 Alava
9 Caceres 19 Valladolid 29 Granada 39 Vizcaya (Biscay)
10 Badajoz 20 Palencia 30 Almeria 40 Guipuzcoa (Basques)
Table 3 List of provinces in Spain. Source: Tourism in Spain
The blue lines in figure 16 illustrate the distribution of the natural gas pipelines. The
network of the pipeline set up on across the 47 provinces in Spain and the name of
the provinces is seen in table 3. In addition, the 7 LNG import terminals including El
Musel terminal is shown the figure 16. For the large-scale NG and LNG facilities,
LNG import terminals, gas pipelines and natural gas underground storage tanks are
discussed in the paper.
3.3.1 Liquefied Natural Gas Import Terminals

There are 7 Spanish LNG receiving terminals which are Barcelona, Cartagena,
Huelva, El Musel and Bilbao LNG terminals (Enagas, 2016). The four main functions
of the terminal are explained by the flow of LNG from arrival at LNG at the import
terminal before being distributed as a natural gas according to Enagas (2015).
18
The first function of the LNG receiving terminals is a cargo operation to import and
export LNG Enagas (2015). Unloading of LNG is operated by transporting of LNG
from LNG vessels to the LNG terminal Enagas (2015). The flow of the loading
operation of the LNG is from LNG terminals to the LNG ships Enagas (2015).
According to Enagas (2015) owning most of the LNG terminals in Spain, the largest
LNG carriers having LNG capacity of up to 266,000 m3 can enter and operate at the
majority of their LNG terminals.
Another function is storage of LNG in special tanks for cryogenic conditions with
inconsiderably more than atmospheric pressure (Enagas, 2015). Encyclopaedia
Britannica (n.d., para 1.) defines the meaning of cryogenic as “cryogenic
temperature range has been defined as from −150 °C (−238 °F) to absolute zero
(−273 °C or −460 °F), the temperature at which molecular motion comes as close as
theoretically possible to ceasing completely”.
Subsequently, LNG is regasified as natural gas by evaporators with the assistance
of seawater making the temperature of LNG to above 0 º C (Enagas, 2015).
Because of this function that LNG import terminals cover the regasification of LNG,
the LNG import terminals are also called as LNG regasification facility.
Metering and odorisation is the last stage of the flow before distributed through the
gas pipeline. These treatments are necessary for detecting natural gas in case of
leakages of it (Enagas, 2015). Apart from transporting of natural gas by the gas
pipeline, those 6 LNG receiving terminals also provide LNG truck loading service for
other destinations such as satellite terminals and LNG fuel station (Gas
Infrastructure Europe, 2018b).
Apart from El Musel, the other six terminals are under operation (Enagas, 2016).
Even though El Musel terminal was built in 2012, the terminal is not under operation
(Enagas, 2016). The underlying reason is to prevent the surplus of the LNG capacity
in Spain judging from the decrease in Spanish gas consumption for electronic
generator between 2008 and 2014 (Prontera, 2017). So Spanish government had
an intention to prevent other new LNG import terminals by enforcing Royal Decree-
Law 13/2012 (Prontera, 2017). This existence of the facility is not under operation
will be simulated in the scenarios in the research.
7,000 GWh
6,000 GWh
5,000 GWh
4,000 GWh
3,000 GWh
2,000 GWh
1,000 GWh
0 GWh
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70
Barcelona Bilbao Cartagena Huelva Mugardos Sagunto
Figure 17 Monthly send out the level of regasified gas from the 6 LNG import terminal in Spain from 2012
to 2017 Source: Gas Infrastructure Europe
19
Also, the definition of send out of regasified natural gas from LNG terminals excludes
the truckload loading service (Agency for the Cooperation of Energy Regulators,
2017). In addition, the data of sending out from Sagunto is available from 2013. Figure
17 shows that Barcelona LNG terminal tends to be the largest regarding the amount
of send out among six terminals during the given period.
1.6 SI
1.4 SI
1.2 SI
1.0 SI
0.8 SI
0.6 SI
0.4 SI
0.2 SI
0.0 SI
1 2 3 4 5 6 7 8 9 10 11 12
Barcelona Bilbao Cartagena

Huelva Mugardos Sagunto
Figure 18 Seasonal Index (SI) of sending out of natural gas after regasification of LNG from 6 LNG import
terminals in 12 months in Spain. Source: Gas Infrastructure Europe
The calculation of the seasonal index is explained in “chapter 4.2 Regression models
with seasonality for forecasting”. Figure 18 illustrates that the seasonal index of the
send out of natural gas transformed from LNG by regasification at LNG terminals is
highly likely to be low during the summer season. However, it tends to increase after
October. Subsequently, the amount of send out of natural gas tend to be peak during
winter seasons between November and February.
In terms of the owners of the terminals, there are four companies shown in table 4.
The largest company is Enagas by owning four terminals: Barcelona, Cartagena,
Huelva and El Musel LNG terminals (Gas Infrastructure Europe, 2018b). There are
no major shareholders of Enagas since 95% of the total shares is a free float (Enagas,
2017a). Other two companies are Saggas which is an owner of Sagunto LNG terminal
and Bahia Bizkaia Gas (BBG) owning Bilbao LNG terminal (Gas Infrastructure Europe,
2018b). However, Enagas is the major shareholder of BBG (50%) and Saggas
(72.5%), so Enagas wholly owns four terminals and partially owns two terminals. In
other words, Enagas engages in the ownership and operation of 6 out of 7 LNG
terminal coupled with the regasification facility in Spain (Enagas, n.d.b). The last
company is Reganosa owning Mugardos LNG terminal (Gas Infrastructure Europe,
2018b). According to Reganosa (n.d.), shareholders of the company are Reganosa
Holdco, Sojitz Corporation and Sonatrach, 75%, 15% and 10% respectively.
Based on the findings, the structure of industry about LNG import terminals in Spain
is close to monopoly by the horizontal integration of Enagas.
LNG
Gijon (El
import Barcelona Cartagena Huelva Sagunto Bilbao Mugardos
Musel)
terminal
Saggas BBG
Owner Enagas Enagas Enagas Enagas (72.5% (50% Reganosa
Enagas) Enagas)
20
Number
of LNG 6 5 5 2 4 3 2
tanks
Storage
760000 587000 619500 800000 600000 800000 300000
(m3)
Table 4 The ownership, number of LNG tanks and storage capacity of terminals in Spain. Source: Enagas.
3.3.2 Natural Gas Pipelines

There is a favourable tendency that the supply chain of natural gas by pipeline is less
greenhouse gas (GHG) and energy intensive than the one of LNG (Georgakaki et al.,
2009). The gap reduces when comparing transportation of natural gas by pipeline to
a remote area (Georgakaki et al., 2009). The leakage of the pipeline tends to occur
more often for the long distances transportation to remote destinations (Georgakaki
et al., 2009). Also, the less transport cost of natural gas through pipeline compared to
LNG transportation is another advantage of the natural gas pipeline (Georgakaki et
al., 2009).
Specially coated polyethene gas pipelines guarantee no corrosion of the steel pipeline
because of the soil (Enagas, n.d.a). The typical minimum and maximum range of the
pipelines on the ground are from 30 to 72 or 80 bar. On the other hand, pipelines
which were set undersea such as Almeria international pipelines or a connection to
Balearic Island are designed to have a pressure of 220 bar (Enagas, n.d.a).
In terms of international connections of the pipeline, there are six pipelines to transport
NG between Spain and foreign countries (Benito et al., 2017). Two of them are
Medgaz and Maghreb pipelines which are initially from Alegria; Medgaz pipeline is
connected to Almeria in Spain, and Maghreb pipeline is connected to Tarifa in Spain
through Morocco (Benito et al., 2017).
The other four pipelines are connected to France and Portugal to transport natural
gas and all of them they are belong to Enagas (Enagas, 2012). Two of them are
allocated to two cities which are Tuy and Badajoz for Portugal’s gas market (Enagas,
2012). The other two for French gas market are connected to Irun and Larrau (Enagas,
2012).
When it comes to the owners or operators of the two pipelines dedicated to importing
NG from Algeria (Maghreb and Medgaz pipelines), there are several Spanish
shareholders. At Tarifa international connection on Maghreb pipeline, an owner is
Enagas (2012). In the massive scale of the pipeline, Gas Natural Fenosa (Natrugy)
accounts for 77.2% of a total share of Europe Maghreb Pipeline Limited (EMPL) and
76.68% of a total share of Metragaz which is operating and maintaining the pipeline
for EMPL (Europe Maghreb Pipeline Limited, n.d.). Also, Gas Natural Fenosa
(Natrugy) is also a significant shareholder of Medgaz by accounting for 15% of the
total share while CEPSA which is a Spanish company has 42% shares of Medgaz
(Medgaz, n.d.).
Gas length number of
pipeline Coverage Province (km) sections Owner
890.4
Al Andalus Cordoba, Sevilla, Cadiz, Malaga, Granada, Jaen 9 6 Enagas (6), Endesa (1, co-owner)
Algete- Burgos, Palencia, Valladolid, Zamora, La Rioja,
Haro Soria, Segovia, Madrid, Guadalajara 900.2 6 Enagas (6), T.R.G (1, co-owner)
Almeria- 428.9
Chinchilla Albacete, Murcia, Almeria 9 4 Enagas (3), *missing 1
330.3
Baleares Valencia, Alicante, Ibiza, Mallorca 7 4 Enagas (3), *missing 1
Cordoba- Madrid, Toledo, Cuenca, Ciudad Real, Cordoba, 1193.
Madrid Jaen 7 12 Enagas (12)
912.3
Eje Levante Castellon, Valencia, Alicante, Tarragona, Murcia 6 9 Enagas (9), Natrugy (1, co-owner)
Eje
transversal Ciudad Real, Albacete, Valencia 264.2 4 Enagas (4), *missing 1
Extremadur
a Badajoz, Cordoba 328.8 3 Enagas (3)
21
Huelva- 550.2
Cordoba Huelva, Sevilla, Cordoba 3 4 Enagas (4), Natrugy (1, co-owner)
Noroeste- Pontevedra, A Coruna, Lugo, Asturias, Cantabria, 1005. Enagas (6), Reganosa (1, co-
Cantabrico Burgos 55 6 owner)
461.6
Pais Vasco Alava, Guipuzcoa, Vizcaya, La Rioja 3 11 Enagas (5), ETN (6)
Ruta de la Asturias, Leon, Zamora, Salamanca, Caceres, 756.4 Enagas (6), Gas Extremadura (1,
Plata Badajoz 4 6 co-owner)
Tivisa- 531.2
Barcelona Barcelona, Tarragona 8 8 Enagas (8), Natrugy (1, co-owner)
Valle del La Rioja, Navarra, Zaragoza, Huesca, Teruel, 1309. Enagas (9), Natrugy (1, co-owner
Ebro Lleida, Tarragona 37 11 with Enagas), Endesa (2)
Table 5 Coverage areas, length, number of sections and owners of gas pipelines in Spain. Source:
Enagas.
Missing 1* in table 5 means that there is no information of owner for 1 section on the
website. Table 4 shows that Enagas is the largest company owning most of the gas
pipelines within Spain. Regarding the connections of natural gas after being regasified
from the LNG import terminals, not only Enagas but also Reganosa LNG receiving
terminal owns its connection.
3.3.3 Underground Storage Tanks of Natural Gas

For the satisfaction of uncertainty and peak demands of natural gas derived from
seasonal effect, sizeable natural gas storages are required (Enagas, n.d.c).
Underground storages of the natural gas are at “old deposits, or it is injected in deep
water stratum or cavities generated in salt formations” (Enagas, n.d.c., para 2.).
The minimum pressure at the facilities is 45 bar while the maximum one is from 72 to
80 bars (Enagas, n.d.c).
Three underground storage tanks are under operation in Spain. Three of them which
are Serrablo, Gaviota and Yela storage tanks are owned by Enagas (Enagas, 2018)
30,000 GWh
25,000 GWh
20,000 GWh
Serrablo
15,000 GWh
Gaviota
Yela
10,000 GWh
5,000 GWh
0 GWh
13
19
25
31
37
43
49
55
61
67
73
79
85
91
97
103
1
7
Figure 19 Monthly level of the three gas storage tanks in Spain from 2009 to 2017 Source: Enagas
The data of Yela storage tank and Gaviota are available to access from 2013 and
2016 respectively while the data of Serrableo from 2009 to 2017 is available from
22
Enagas (2018). Figure 19 shows illustrate that Gaviota and Yela tanks accumulate
more gas inventory while the gas inventory level in Serrablo tends to remain the same
throughout the period.
1.4 SI
1.2 SI
1.0 SI
0.8 SI Serralo
0.6 SI Gaviota
Yela
0.4 SI
0.2 SI
0.0 SI
1 2 3 4 5 6 7 8 9 10 11 12
Figure 20 Seasonal Index (SI) of the level of inventory in three gas storages in 12 months in Spain.
Source: Enagas
Figure 20 shows the general trend of the seasonal index that the storages tanks tend
to be dedicated to the gas injection during the summer season and the gas withdraw
winter season (Enagas, 2018). In particular, the storage level tends to decrease from
November to April while there is a tendency that the level increase from April to
November (Enagas, 2018).
In terms of the new facility of the gas storage tank, Pinasses storage tank owned by
Gas Natural Fenosa (Natrugy) is planned to be constructed (Gas infrastructure
Europe, 2015).
3.4 Small-Scale Infrastructure
3.4.1 Liquefied Natural Gas Satellite Terminal

The first satellite terminal was built in North East Spain in December of 1970 which
means that Spain has a long experience of SSLNG (Enriquez and Rojo, 2016a).
LNG from LNG import terminal is transported by trucks or trains to satellite terminals
(JFE engineering corporation, n.d.).
The LNG satellite terminals encompass several functions. The first one is storage and
unloading of liquefied natural gas (Gu et al., 2010). Another one encompasses the
process of regasification, calorie control, blending with water gas and coal gas, and
send it to pipelines connecting to the city (Gu et al., 2010). The last one is using a
distribution station for LNG loading and transporting to LNG fuel stations (Gu et al.,
2010). Gu et al. (2010) excluded the last function for the satellite terminals but the first
and second functions. In the same way, the function as the distribution of LNG at
satellite terminals is not included in this study but the distribution centre of natural gas
after regasification of LNG.
The satellite terminals enable for the remote areas in which the access of the gas
pipeline is not possible to consume the gas (HAM, n.d.c). Because of the this, satellite
terminal is regarded as a distribution centre for isolated areas.
23
Figure 21 Location and distribution of LNG satellite stations in each Spanish province. Source: Gas
Infrastructure Europe.
Number
of LNG
Number
fuel
of Provinces
stations
provinces
in
provinces
7 1 Madrid (24)
6 2 Lleida (43), Tarragona (46)
Asturias (5), Toledo (17), Guadalajara (25), Granada
5 6
(29), Almeria (30), Murcia (31)
A Coruna (1), Pontevedra (2), Lugo (3), Alicante (32),
4 6
Zaragoza (35), Girona (45)
Badajoz (10), Cadiz (13), Valencia (33), Teruel (34),
3 5
Barcelona (44)
Leon (6), Caceres (9), Malaga (14), Ciudad Real (16),
2 7
Burgos (22), Cuenca (26), Albacete (27)
Cordoba (15), Avila (18), Cantabria (21), Sevilla (12),
1 7
Jaen (28), Navarra (41), Huesca (42)
Table 6 Location and distribution of LNG satellite stations in each Spanish province. Source: Gas
Infrastructure Europe.
Figure 21 and Table 6 illustrate that the density of the location of LNG satellite stations
is high in the centre, South and North-East of Spain. So, it is assumed that that area
has less pipeline connection that the other region in the assumption that the
distribution of natural gas is optimised and well facilitated.
24
Natrugy
HAM
Endesa
Via Augusta Gas
MONFORT
MONEGAS
AVIA
0 2 4 6 8 10 12 14 16 18
Figure 22 Owners/operators of LNG satellite terminals in Spain. Source: Gas Infrastructure Europe
On the other hand, in terms of local Spanish satellite terminals, there are six owners
or operators from figure 22. There are three major players in the satellite terminal
industry which are Natrugy, Redexis Gas and HAM. Since the title of the company “Gas
Natural Fenosa” had been changed to Natrugy (La Vanguarida, 2018), this change is also
applied in the research of the companies. Natrugy encompassing Gas natural Castilla
y Leon, Cegas and Distribution is the largest company regarding the satellite terminals
in Spain (Precio Gas, n.d.). Explore in more detail of the company, the major
shareholders of the company Criteria Caixa Holding (35.3%) and Repsol Group
(30.0%) (Precio Gas, n.d.).
Redexis Gas and HAM are the second largest companies owning 23 terminals each.
Redexis Gas Aragon, Baleares, Distribution owing some satellite terminals are part of
Redexis Gas, S.A. The major shareholders of the company are Arbejdsmarkedets
Tillægspension (33.3%), Universities Superannuation Scheme couple with Guoxin
Guotong Fund LLP (33.3%) and CNIC Corporation Limited (33.3%) (Redexis Gas,
n.d.).
Ham group does not provide a clear explanation of the shareholder structure, but it
mentions that family business is engaged in the business (HAM, n.d.a.) So, it is
possible to assume that there are no shareholders outside of the company but a family
business.
Judging from the data of the shareholder of the companies, these shareholders of the
companies have no internal relationships with the owners or operators of the LNG
import terminals. In other words, the owners or operators of the LNG import terminals
do not identify with those of the LNG satellite terminals.
There is one satellite terminal planned to be constructed at Cobisa in the province of
Toledo by Natrugy.
Apart from the satellite terminals, there are several types of SSLNG facilities which
are about fuel loading road, fuel loading ship and bunker ship explained on the
following sub-chapters.
3.4.2 Liquefied Natural Gas Fuel Loading Road Station

There are two different forms of natural gas used as a fuel for road vehicles:
compressed natural gas (CNG) and liquefied natural gas (LNG) (Arteconi et al., 2010).
25
According to Gas Infrastructure Europe (2018b), some of the fuel stations provide
service of fuelling of both CNG and LNG in Spain.
In terms of CNG fuel stations, there are two types of the stations. The first one is the
time-fill station fuelling the vehicles from a compressor delivering CNG at low pressure
to fuel tanks of the vehicles (International Energy Agency, 2017). Another one is fast-
fill stations supplying CNG without the compressor from utility line at high pressure
(about 300 bar) to the fuel tank of the vehicle in easy and quick ways (International
Energy Agency, 2017).
On the other hand, the unique requirement of the equipment such as LNG storage
tanks, safety gadget and cool down system preventing from an increase of LNG
pressure in a storage tank at the dangerous level is necessary for the LNG fuel loading
station (International Energy Agency, 2017).
Since the energy density of LNG is higher than the one of CNG, LNG is usually
consumed for the heavy-duty road vehicles (Arteconi et al., 2010). Also, LNG as a fuel
is proper for trucks in long-haul travel such as more than 100,000 km in a regular
basis because of the evaporated LNG needed to be consumed while CNG is suitable
for the small trucks less travelling and irregular operation (International Energy Agency,
2017).
AVIA MONEGAS MONFORT Via Augusta Gas

Endesa HAM Natrugy
Figure 23 Owners/Operators of LNG fuel loading road stations in Spain. Source: Gas Infrastructure
Europe
Before mentioning the structure of the LNG fuel road stations, there are two fuel
stations are cooperated by Natrugy and Repsol, and one is cooperated by Endesa
and Molgas (Gas Infrastructure Europe, 2018b). The former relation is assumed as
sole ownership by Natrugy, and the latter one is by Endesa in the study.
Figure 23 illustrates 37 LNG fuel loading road stations under operation in Spain in
2018. Natrugy and HAM are considered as large LNG fuel station companies based
on the number of the station they own. Each of the companies accounts for 40% of
the LNG fuel station market in Spain in 2018.
Three LNG fuel stations are planned to be built. The two sites of the construction are
in Madrid by HAM and Natrugy (Gas Infrastructure Europe, 2018b). The other one will
be built in Cadiz by Repsol (Gas Infrastructure Europe, 2018b).
26
3.4.3 Liquefied Natural Gas Fuel Loading Ship Station
Fuel loading ship station Owner/Operator

Port of Gibraltar Port of Gibraltar
Port Authority of
Cartagena fuel loading ship
Cartagena
Ferrol LNG loading ship (bunkering) Reganosa
Bilbao BBG
Table 7 Owners/operators of LNG fuel loading ship stations in Spain. Source: Gas Infrastructure Europe
Table 7 shows that the owners or operators of the fuel loading ships station are
identical with the port authority or owner of the LNG import terminals. In addition,
figure 24 coupled with table 8 illustrates the distribution of the SSLNG facilities in the
province scale.
Figure 24 Locations and distribution of LNG fuel road and ship stations in each Spanish province. Source:
Gas Infrastructure Europe
# of LNG
fuel
stations in # of
provinces provinces Provinces
5 1 Barcelona (44)
3 4 A Coruna (1), Valencia (33), Vizcaya (39), Guipuzcoa (40)
Cadiz (13), Burgos (22), Madrid (24), Guadalajara (25),
2
5 Cuenca (26)
Ourense (4), Badajoz (10), Sevilla (12), Malaga (14), Toledo
(17), Jaen (28), Murcia (31), Alicante (32), Zaragoza (35),
1 14
Alava (38), Lleida (43), Girona (45), Tarragona (46),
Castellon (47)
Table 8 Location and distribution of LNG fuel road and ship stations in each Spanish province. Source:
Gas Infrastructure Europe.
in terms of LNG fuel for ships, Valencia fuel loading ship station is under construction
and LNG fuel station in Algeciras is planned to be built (Gas Infrastructure Europe,
2018b).
27
3.4.4 Liquefied Natural Gas Bunker Ships
There is one LNG bunker ship “Oizmendi” (former name was “Monte Arucas”) in Bilbao
in Spain (Gas Infrastructure Europe, 2018a) (LNG World News, 2018). According to
LNG World News (2018), the first bunkering of LNG by ship-to-ship has been operated
in the Port of Bilbao in Spain on February 3, 2018. The ship has two cryogenics tanks
having a capacity of 300-cubic metres (cbm) to store LNG, and it was renovated as
the bunker barge recently (LNG World News, 2018). Basque region project covers not
only this bunker ship but also adaption of the BBG dock and regasification facility and
building and design of tugs fuelled by tug (LNG World News, 2018).
Itsas Gas is the dominant player in the project. Basque Energy Agency (EVE) is the
principal shareholder of the company by owning 49% of the shares (LNG World
News, 2018). The rest of the shares (51%) are equally owned by Remolcadores
Ibaizabal (25.5%) and Naviera Murueta (25.5%) (LNG World News, 2018).
3.4.5 Modes of Small-Scale Liquefied Natural Gas transportation
3.4.5.1 Trucks
The first transport mode is the truck equipping tanks designed for carrying LNG
(Hansson, 2008). In terms of the short distance, truck transportation of LNG is the
most suitable among other modes for minimising the total cost of the transportation
(Hansson, 2008). Approximately 300 km is the traditionally maximum distance for
LNG truckload (Eliot, cited in Hansson, 2008). However, further technical advances
enable to extend the limits of the range (Lennerås, cited in Hansson, 2008). Another
advantage is door-to-door service is available with no requirement of transfer from
picking up and delivery of the cargoes (Chopra and Meindl, 2015). The disadvantage
of the transport by trucks is not eco-friendly since most of them are powered by fossil
fuels (Hansson, 2008). In Spain, table 8 shows a majority of road tractors which is
capable of transporting LNG use diesel as fuel in Spain in 2016 (Eurostat, 2018). In
other words, the most trucks are not eco-friendly, and this supports the argument of
Hansson (2008). However, table 9 illustrates that there are a minimal number of trucks
using LNG as a fuel (Eurostat, 2018).
# of road
tractors
Total 207889
Diesel 206305
Petroleum products 1219
Liquefied natural gas (LNG) 167
Compressed natural gas
(CNG) 135
Liquefied petroleum gases
(LPG) 35
Other 24
Electricity 4
Table 9 Road tractors by type of motor energy in Spain in 2016. source: Eurostat
Figure 24 shows that there are a growing number of LNG so if the growth rate is
continuously increasing, road transportation of LNG is a more favourable option in the
future by preserving the environment.
28
180.0
160.0
140.0
120.0
100.0
80.0
60.0
40.0
20.0
0.0
Motor coaches Lorries <= 3.5 Lorries > 3.5 Road tractors
tonnes tonnes
2013 2014 2015 2016
Figure 25 Numbers of vehicles fuelled by LNG in Spain from 2013 to 2016. Source: Eurostat.
According to Eurostat, small size of vehicles such as motorcycles and passenger cars
are not equipped LNG fuelled motor engine while a large size of vehicles such as
lorries and road tractors equipped one. Figure 25 shows the significant increases in
the number of road tractors which are relatively more substantial than the other
vehicles during the given period. Also, apart from motor coaches, there are increases
in the other vehicles. So, we can conclude that there is a trend of the increase in the
number of LNG fuelled vehicles.
There are seven facilities providing LNG truck loading service in Barcelona,
Cartagena, Huelva, Bilbao, Mugardos, El Musel and Sagunto LNG terminals in Spain
(King & Spalding LLP, 2018).
The market for truck loading accounts for approximately 4% of the entire annual
conventional demand in 2015 (Natural Gas World, 2016)
3.4.5.2 Ships
The second mode is transportation by ships. However, since it is small-scale LNG
transportation, it needs to be differentiated from large-scale LNG shipping (Hansson,
2008). An example of the small-scale LNG ships under operation in Norway is 1,100
m3 while the large-scale LNG vessels are between 80,000 to 300,000 m3 (Hansson,
2008).
3.4.5.3 Trains
The final one is the railroad transportation. The loading capacity of each railway
wagon is approximately 40 tonnes (Näslund, cited in Hansson, 2008).
Fast transportation is one of the advantages of the railroad (Hansson, 2008). For
instance, a single trip by railroad across an entire of Sweden within a day is highly
likely available (Hansson, 2008). It is possible to apply in Spain since the territory of
Sweden is more extensive than Spin. Another advantage is that trains are regarded
as the eco-friendly modes if it is powered by electricity (Hansson, 2008).
On the other hand, there are some disadvantages of railroad transportation. At first, it
is not suitable for short distance transportation since it costs more than the other
alternatives (Hansson, 2008). Another issue occurs during the reloading operation.
Mostly, the delivery of LNG by trains is not door to door system, evaporating of LNG
29
is inevitable during the operation (Hansson, 2008). The term evaporating of LNG is
called as boiled off. Boiled off means the evaporation of LNG at a certain temperature
higher than its boiling point (Dobrota et al., 2013). This is caused by entering of heat
to LNG during cargo operation, transporting and storage occurs BOG (Dobrota et al.,
2013). Subsequently, it increases pressure and decreases density. So, railroad
transportation of train less favourable because of boiled off the gas (BOG) leading to
the loss of the cargo (Vikersveen cited in Hansson, 2008)
In terms of the railroad for transporting LNG in Spain, no rail loading facilities are
existing but they are planned to be constructed in Barcelona LNG terminal (Gas
Infrastructure Europe, 2018b). Based on these data, the more significant numbers of
fuel loading road facilities than those of ship facilities means the majority of
transportation of SSLNG is done by trucks. This is supported by Enriquez and Rojo
(2016a) mentioning that Spanish LNG truck loading to inside and outside of Spain
accounts for 85% of hole LNG truck loading in the EU market in 2015.
3.5 Vertical Integration of Natural Gas and Liquefied Natural Gas Supply
Chain in Spain
Large-scale facility Small-scale facility
Owner of LNG import Num Owner of Fuel Num Num
terminal bers loading station bers Owner of Satellite station bers
Enagas 6 Natrugy 16 Natrugy 53
Reganosa 1 HAM 15 HAM 23
Owner of underground
Endesa Redexis Gas
storage tanks 2 23
Enagas 3 AVIA 1 Nortegas 9

Owner of sections of Distribución y Comercializacion
MONEGAS
the gas pipeline 1 de Gas Extremadura 3
DISTRIBUIDORA REGIONAL
MONFORT
Enagas 84 1 GAS 1
ETN 6 Via Augusta Gas 1
Endesa 2
Table 10 Ownership of large- and small-scale LNG facility. Source: Enagas, Gas Infrastructure Europe
It is essential to know the degree of the vertical integration of the supply chain in the
scope of the study since one of the assumptions is the supply chain is vertically
integrated. There are co-ownership of facilities as mentioned above sub-chapters 3:
some LNG import terminals, gas pipeline and fuel stations. The co-ownerships are
supposed as a single owner who has a larger size than the other co-owner judging
from the numbers of facility they have in table 10. It is possible to conclude that the
vertical integration of the LNG supply chain between LNG receiving terminals and
LNG fuel stations coupled with LNG satellite stations does not exist from table 10.
On the other hand, regarding the large-scale facility, Enagas is the dominant company
by having six terminals out of 7 in Spain. When it comes to small-scale LNG
infrastructure, Natrugy and Ham vastly run their business in both LNG fuel loading
station and satellite station. Since the number of players of the industries is limited,
the large-scale LNG industry including the LNG import terminals, underground
storages and gas pipelines is close to Monopoly by Enagas while the small-scale LNG
industry is considered an oligopoly.
The result of the scenarios in this paper would be more feasible if there was the
vertical integration unlike this current situation between the LNG import terminals and
the destinations. Other conditions such as individual contracts or relationships
30
between companies might discourage the freedom of contracts based on only total
costs of transportation. So, without the vertical integration, the supply chain might not
be able to be fully optimised regarding the viewpoints of the costs. The full vertical
integration of the supply chain is assumed in the next chapter for simulation of the
model.
3.6 Seasonality in the Natural Gas and Liquefied Natural Gas Supply
Chain
1.4 SI
1.2 SI
1.0 SI
0.8 SI
0.6 SI
0.4 SI
0.2 SI
0.0 SI
1 2 3 4 5 6 7 8 9 10 11 12
NG and LNG import Send out at LNG terminals

Inventory of storage tanks Conventional consumption
Figure 26 Seasonality in NG and LNG supply chain. Source: Gas Infrastructure Europe, Enagas and
Corporacion de Reservas Estrategicas de Productos Petroliferos
Figure 26 illustrates the seasonality in NG and LNG supply chain in Spain. The value
of the send out at LNG terminals and inventory of storage tanks are averaged among
the related facilities introduced chapter above. Conventional consumption is chosen
since it has strong seasonality and occupies the majority of parts among the other
types of consumptions. It will represent the seasonality of SSLNG supply chain. The
trends of the seasonality of NG and LNG import, send out at LNG terminals and
conventional consumption are similar by having high value in winter while having less
value in the other season.
Direct comparison with the seasonality of the inventory of storage tanks and the other
seasonality is not available since the seasonality of the inventory storage has a
different standard. However, interpretation of the data allows to analysis with the other
values. An increase in the seasonal index at the inventory of storage means saving
natural gas while a decrease means sending out natural gas for consumption. For
example, after November, the storage tanks send out natural gas to consumers while
saving natural gas after April.
To sum up, there is a seasonality in the entire NG and LNG supply chain in Spain by
having the trend that high demand during winter and low demand apart from winter.
31
Chapter 4 Methodology and Data Collection
Figure 27 Models for the research.
Figures 27 introduce the three methodologies of the study. The different sizes of the
boxes imply the importance in the study. The primary methodology of the study is the
MILP model because it optimises the SSLNG supply chain which is the goal of the
study. Another model which is the regression model is used for estimating of SSLNG
future demand. The accurate forecasting of the demand plays a vital role in the paper
since the demand is necessary for the model. Also, the model is applied for estimation
of the price of LNG. So, instead of making a simple assumption from scratch, the
regression model is used for more reliable data.
Moreover, the other analysis in chapter 4.2.1 and 4.2.2 support the feasibility of the
demand. The last model is the autoregressive model for the estimation of the fuel
price of trucks. This model plays the least important role among the other models
since the fuel price is merely a factor of the input.
32
Figure 28 Relationships between MILP model and scenarios.
Regarding the relation between the scenarios and MILP model, figure 28 illustrates
the how the scenarios are related to the model. The different inputs of demand sharing
the same seasonal index differentiate the scenarios. This research enhances the
flexibility and efficiency of the SSLNG supply chain regardless of the different demand.
Furthermore, the supply chain becomes robust against seasonal effect which is the
characteristic of the demand by monthly simulating with the scenarios.
Figure 29 Data collections for inputs of the MILP model.
Figure 29 illustrates data collections for inputs of the MILP model. There are three
inputs are required for the simulating MILP model: SSLNG demand, Total
transportation costs and maximum capacity at LNG import terminals. Fixed costs of
33
the terminals are excluded according to the assumption of the study.
In order to acquire SSLNG demand, future consumption of SSLNG is needed by
dividing the consumption in the autonomous community scale. The autonomous
communities are selected when the locations of SSLNG facilities are covered by the
community.
Other inputs are the total transportation costs consisting of transportation costs and
pipeline holding costs. The transportation costs are calculated from the three factors:
distance, fuel efficiency and fuel& other costs per kilometres. In particular, fuel costs
are calculated by the autoregressive model. The calculation of the pipeline holding
costs is done with holding costs rate assumed and price of LNG calculated from the
regression model.
The last input is maximum capacity at LNG import terminals. This input is acquired on
the terminal companies’ website.
The detail process of the data collection is in chapter 4.4.
4.1 Network Optimisation Model
Figure 30 Flows of LNG and NG in Spain. Source: Enagas
Figure 30 shows the overview of the flows of LNG and NG. The flows of “e”, “f”, “g” by
truck transportation in Spanish territory in the Iberian Peninsula are researched. The
SSLNG transportation by ships on the flow “g” are excluded in this study. Also, the
flow of “f” is included in the flow “e” since the LNG barge initially receives the LNG fuel
from the LNG import terminals. In order to optimise the supply chain, the MILP model
is applied since all the journals about the optimisation of the SSLNG supply chain
discussed in the literature review used the model.
34
Figure 31 Descriptions of the MILP model.
Figure 31 illustrates the MILP model with inputs, constraints and decision variables
for the minimisation of the total costs. The model description is discussed in detail in
chapter 4.1.4, so this paragraph briefly shows the diagram of the simulation. There
are four inputs, but fixed costs are set to zero according to the assumption of the study,
so three inputs are actually applied. The decision variables are initially zero before
simulating the model. After simulating the model with the inputs and constraints, the
decision variables are calculated for the minimisation of the total costs in the SSLNG
supply chain.
4.1.1 Players in the Model
Figure 32 Players of the model.
Figure 32 illustrates three players in the research area: LNG import terminals, truck
companies and small-scale LNG facilities. LNG import terminals supply LNG which
does not exceed the maximum output capacity for SSLNG transportation by truck.
Truck companies provide the truck transportation service from the origins to the
destinations aligned with the demands. The small-scale LNG facilities demand the
amount of the LNG they need and received the cargo. In the aspect of the allocation
of the cargo, SSLNG facilities become a problem owner since they need to decide
which LNG import terminals to receive the cargo.
On the other hand, decision maker for the amount of the SSLNG demand transported
is SSLNG facility. The truck company is problem owner since they need to adjust the
35
number of trucks to meet the demand.
The owner of El Musel is the problem owner since the terminal is not under operation
due to the Spanish government which is a decision maker in the context. So, the
problem owner should convince the Spanish government to operate the terminal with
the positive proofs.
4.1.2 Assumption of the Model

Most of the assumptions in the papers are same as the assumptions from other
journals from “Chapter 2. Literature Review”.
At first, enough storage capacity from the destination is assumed. It means that there
would be no issues of inventory management at the SSLNG facilities.
Another is the fixed costs such as investment costs of LNG import terminals, and
SSLNG facilities are not considered. This assumption affects the MILP model by not
inputting fixed costs in the model.
Also, boiled-off gas is not considered for simple calculation. In other words, there
would be no loss of cargo during transport.
Schedule of the truck transportation is not taken into account.
Regarding the period of the research, in order to check the seasonality, monthly data
of 12 months in a year are simulated by inputting monthly changeable factors such as
different demands of SSLNG.
On the other hand, there are different assumptions from the other papers. There are
no transaction costs and delays of sharing information by assuming that there is a
vertical integration in the supply chain.
Other customers for SSLNG do not exist other than LNG satellite stations and fuel
stations.
Lot size of the truck is 0.3 GWh and truckload (TL) is assumed by having only single
origin and destination.
4.1.3 Scenario Description

There are three main scenarios with sub-scenarios. The main scenarios are set based
on the assumption of the different SSLNG demands in order to test to what extent the
supply chain does change in the different demands on a monthly basis in a year.
The scenarios simulate the supply chain in the different situations. The main scenarios
are the optimisation of SSLNG supply chain in the “full capacity”, “abundant capacity”
and “a lack of capacity”. The “full capacity” does not literally have full capacity but the
LNG transportation by trucks is close to the limits of the amount of LNG transported
from the LNG import terminals. Another one is “abundant capacity” has the relatively
lower amount of LNG is transported than the amount of the “full capacity” scenario.
The first and second scenario has two sub-scenarios to examine the effect of the
operation or no operation of El Musel LNG import terminal. The last one is “a lack of
capacity” that the demand of SSLNG exceeds the maximum output capacity of LNG
truckload transportation from LNG receiving terminal. So, the difference of allocation
of the LNG transportation between the origin and destination will be discussed in the
results. Also, the sub-scenarios are about discussing the effect of additional LNG
import terminal which is El Musel LNG terminal in Spain. Enagas (2016) mentioned
that the LNG terminal had been built however it has not been operated. So how will
the additional terminal affect the optimisation of the supply chain is studied if the
terminal would be under operation.
4.1.4 Description of Optimisation Model

The SSLNG supply chain has 7 LNG import terminals (6 terminals in sub-scenario)
36
and 14 autonomous communities in Spain. The former is the origin of the SSLNG and
the latter are the destination of the SSLNG.
The model optimises the supply chain by minimising the total costs in the supply chain.
Following inputs are required for the network optimisation model.
n = number of LNG import terminals locations

m = number of destinations for the demand (LNG fuel stations and LNG satellite
terminals in autonomous community level)
𝐷𝑗 = monthly demand from LNG fuel station and LNG satellite terminals “j” in the
autonomous community level
𝐾𝑖 = Monthly maximum capacity of the LNG terminals “I”
𝑓𝑖 = Monthly fixed costs of keeping plant “I” open
𝑐𝑖𝑗 = cost of transporting costs (fuel costs) from the LNG terminals “I” to the
destinations (LNG fuel stations and LNG satellite terminals).
Decision variables are below.

𝑦𝑖 = 1 if the location of LNG receiving terminal is at i, otherwise it is 0
𝑥𝑖𝑗 = 1 if LNG receiving terminal i supplies market j, otherwise it is 0
The formulation of the problem is below (Chopra and Meindl, 2015).
Minimise
𝑛 𝑛 𝑚
∑ 𝑓𝑖 𝑦𝑖 + ∑ ∑ 𝑐𝑖𝑗 𝑥𝑖𝑗
𝑖=1 𝑖=1 𝑗=1
Subject to
- Equation 1.1.
𝑛
∑ 𝑥𝑖𝑗 = 𝐷𝑗
𝑖=1
for j = 1,…, m
- Equation 1.2.
𝑚
∑ 𝑥𝑖𝑗 ≤ 𝐾𝑖 𝑦𝑖
𝑗=1
for i = 1, … , n
- Equation 1.3.
𝑦𝑖 ∈ {0,1}
For i= 1,…, n, 𝑥𝑖𝑗 ≥ 0
37
Excel Solver is the software to run the model in this study. Inputs of the data of
capacity of terminals, costs and demand are the first step (Chopra and Meindl, 2015).
Annex A shows the excel sheet for the simulating model by Excel Solver. The number
of origins of the SSLNG (n) is 7 or 6 (sub scenario without El Musel terminal). The
number of the destination (m) is set 14. The fixed costs (yellow area) from cell P4 to
P10 is set as zero according to the assumption of the model that there is no fixed cost
(𝑓𝑖 ). Capacity of the terminal (𝐾𝑖 ) is input from Q4 to Q10 (Green area). The grey area
(cell between B11 and O 11) illustrates SSLNG demands (𝐷𝑗 ) from the autonomous
communities in Spain. The total transportation costs (𝑐𝑖𝑗 ) are entered in the cells from
B4 to O10 (Orange area).
When it comes to decision variables, cells from B17 to O 23 (Blue area) represent the
allocation of SSLNG from origins to destinations (𝑥𝑖𝑗 ). Another decision variable (𝑦𝑖 )
about operation or non-operation of the LNG import terminals is shown between P17
and P23 (Sky blue area). The two variables are initially set as zero before running the
model (Chopra and Meindl, 2015).
Subsequently, it is necessary to set the Excel formula for the model. Excel formula “=
B11-SUM(B17:B23)” is copied from B36 to O36 (Orange area) for unmet demand
(Equation 1.1.). Equation 1.1 means that the total demand from an autonomous
community minus divided demands transported from LNG import terminals.
Equation 1.2 (Excess Capacity) is applied to Excel as “=P17*Q4-SUM(B17:O17)” by
being copied from B27 to B33 (Grey area). Equation 1.2 expresses the difference
between the maximum output capacity at terminals and the sum of SSLNG
transported from the terminals to the destinations.
The objective function is shown in the cell B39 with the formula below.
“SUMPRODUCT(B4:O10,B17:O23)+SUMPRODUCT(P4:P10,P17:P23)”
The total fixed costs (excluded from the study based on the assumption) and variable
costs are measured in the objective function (Chopra and Meindl, 2015).
The minimisation of the total cost (Cell B39) is in needs of constraints below (Chopra
and Meindl, 2015).
B17:P23 ≥ 0 {Decision variables are non-negative}
B27:B33 ≥ 0 {𝐾𝑖 𝑦𝑖 − ∑𝑚 𝑗=1 𝑥𝑖𝑗 ≥ 0 for i = 1,…,7 or 6 (without El Musel terminal)}

This formula defines that the sum of SSLNG transported from the LNG import
terminals to the destination cannot exceed the maximum SSLNG output capacity at
the terminals.
B36:O36 = 0 {𝐷𝑗 − ∑𝑛𝑖=1 𝑥𝑖𝑗 = 0 for j = 1,…,14}

This constraint is for unmet demand. It is an ideal situation that the unmet demand is
zero meaning that all the demands are satisfied.
P17:P23 = binary {Location variable 𝑦𝑖 is 0 or 1}
Finally, Simplex LP and minimization are selected in the Solver Parameters and the
model is simulated.
4.2 Regression Model with Seasonality for Forecasting

Seasonal variations mean that cycles have a short calendar repetition within a year
(Keller, 2014). It covers the patterns happens daily, weekly, monthly or even four
38
seasons (Keller, 2014). In order to estimate the seasonal effect, seasonal indexes
must be calculated (Keller, 2014). In addition, abundant time series are necessary for
the calculation of the seasonal indexes (Keller, 2014). Keller (2014) mentions that a
minimum of 4 years of time series data is needed to observe the proper seasonal
index. Keller (2014) shows the procedures for the calculation of the seasonal indexes
below.
Figure 33 Process of forecasting future demand.
Figure 33 illustrates the process of the forecasting future demands by using the
regression model. At first, data of SSLNG consumption is used for the regression
model. The data is checked among the three model which are linear, quadratic and
cubic. Based on analysis with p-value and r-squared, a suitable model for the demand
data will be selected. After the seasonal index is calculated, the time series is
deseasonalised. Finally, forecasting of the future demand is available with the
regression model, seasonal index. The process in detail is in the below paragraphs.
1. Computing regression lines. These could be linear, quadratic or cubic

regression line.
Linear:
𝑦𝑡^ = 𝑏0 + 𝑏1 𝑡
Quadratic:
𝑦𝑡^ = 𝑏0 + 𝑏1 𝑡 + 𝑏2 𝑡 2
Cubic:
𝑦𝑡^ = 𝑏0 + 𝑏1 𝑡 + 𝑏2 𝑡 2 + 𝑏3 𝑡 3
39
2. Computing the ratio for each time period. Most of the trend variations are
removed by this ration.
𝑦𝑡
𝑦𝑡^
3. Calculating the average of the ratios in step 2 for the months. The most of
the random variations are eliminated, but the seasonality is left.
4. Adjusting the averages to make an entire of the average of the seasonality 1.
Subsequently, deseasonalising the time series are necessary to acquire the

seasonally adjusted time series (Keller, 2014). The deseasonalising means that
eliminating the seasonal variation by using the seasonal index (Keller, 2014).
The procedure of the deseasonalising is computed by the formula below (Keller, 2014).
𝑇𝑖𝑚𝑒 𝑆𝑒𝑟𝑖𝑒𝑠
𝑆𝑒𝑎𝑠𝑜𝑛𝑎𝑙 𝐼𝑛𝑑𝑒𝑥
Easier comparisons of time series among months are available by deseasonalising

since there are no seasonal effects anymore (Keller, 2014).
The final step is forecasting by using the seasonal indexes, and the formula are below.
(Keller, 2014).
𝐹𝑡 = [𝑏0 + 𝑏1 𝑡] ∗ 𝑆𝐼𝑡
𝐹𝑡 = [𝑏0 + 𝑏1 𝑡 + 𝑏2 𝑡 2 ] ∗ 𝑆𝐼𝑡
𝐹𝑡 = [𝑏0 + 𝑏1 𝑡 + 𝑏2 𝑡 2 + 𝑏3 𝑡 3 ] ∗ 𝑆𝐼𝑡
where
𝐹𝑡 = 𝐹𝑜𝑟𝑒𝑐𝑎𝑠𝑡 𝑓𝑜𝑟 𝑝𝑒𝑟𝑖𝑜𝑑 𝑡
𝑏0 + 𝑏1 𝑡 = 𝑅𝑒𝑔𝑟𝑒𝑠𝑠𝑖𝑜𝑛 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (𝐿𝑖𝑛𝑒𝑎𝑟)
𝑏0 + 𝑏1 𝑡 + 𝑏2 𝑡 2 = 𝑅𝑒𝑔𝑟𝑒𝑠𝑠𝑖𝑜𝑛 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (𝑄𝑢𝑎𝑑𝑟𝑎𝑡𝑖𝑐)
𝑏0 + 𝑏1 𝑡 + 𝑏2 𝑡 2 + 𝑏3 𝑡 3 = (𝐶𝑢𝑏𝑖𝑐)
𝑆𝐿𝑡 = 𝑆𝑒𝑎𝑠𝑜𝑛𝑎𝑙 𝑖𝑛𝑑𝑒𝑥 𝑓𝑜𝑟 𝑝𝑒𝑟𝑖𝑜𝑑 𝑡
In this paper, the monthly data of LNG direct consumption, NG consumption for
conventional and NG consumption for electricity generators in Spain from 2009 to
2017 will be used for the calculation of the future consumption in Spain in 2018
(Corporacion de Reservas Estrategicas de Productos Petroliferos, 2018). In particular,
40
one of the consumption sectors which is LNG direct consumption in 20018 will be
used as a decision variable for the demand of LNG at the LNG fuel stations for trucks
and ships. So, LNG direct consumption other than these destinations is excluded from
this research. The calculation 1of the forecasting of the future consumption is done
by Minitab. In addition, in order to run the model, periods are going to be used by
setting period in ascending order such as period 1 (January 2009), period 2 (February
2009).
4.2.1 Analysis of the relationships of Pearson correlation and p-value

The value of the Pearson correlation is close to 1 meaning that it is close to a perfect
positive relationship between the two variables (Keller, 2014). In addition, the p-value
of this relation is less than 0.0001 which is very low.
4.2.2 Analysis of the regression models with P-value, R-squared

In order to interpret the regression models, it is necessary to use some figures
representing a different aspect of the models.
Because of the limitation of the coefficient of correlation only having three figures
which are -1, 0, 1, other proper measurements are required to interpret properly
(Keller, 2014). The coefficient of determination fits for the purpose (Keller, 2014). The
coefficient of determination of 0 (0%) means there are no relationships between
independent and dependent variables compared while the one of 1 (100%) expresses
that the independent variable fully explains the variation in the dependent variable
(Keller, 2014). In addition, the figures other than 0 and 1, the degree of the figures
whether close to 1 or 0 determine the relationships (Keller, 2014).
Figure 34 LNG direct consumption between 2009 and 2018 in Spain. source: Corporacion de Reservas
41
Figure 35 NG consumption for electricity generator between 2009 and 2018 in Spain. Source:
Corporacion de Reservas Estrategicas de Productos Petroliferos
Figure 36 NG consumption for conventional between 2009 and 2018 in Spain. Source: Corporacion de
Reservas Estrategicas de Productos Petroliferos
42
Figure 37 A total NG and LNG between 2009 and 2018 in Spain. Source: Corporacion de Reservas
Figure 34, 35 and 36 illustrate the three different consumption history data of NG and
LNG from 2009 to 2017 coupled with the data in the year of 2018 acquired from the
forecasting with seasonality in the regression model. Period 1 means January in 2009
and period 120 means December in 2018. The blue line is the data acquired from
Corporacion de Reservas Estrategicas de Productos Petroliferos (2018a). from the
three tables, the seasonality of the consumption is clearly observed. Deseasonality of
the consumptions is seen from the green line. The red lines show the direct results of
the regression models. The purple line is the result of multiplying of red lines in 2018
and seasonal indexes. In other words, the purple line is the consumption forecasted
in 2018. The lines of figure 37 are similar to figure 36 since conventional consumption
(figure 36) accounts for the large part of the total consumption (figure 37).
Conventional GWh (y) = 19155 + 199.3 Period − 4.327 Period2 + 0.02556 Period3
Electricity generator GWh (y) = 15294.4 − 284.77 Period + 1.9130 Period2
LNG direct GWh (y) = 665.73 + 25.829 Period − 0.51289 Period2
consumption + 0.0026954 Period3
Total consumption GWh (y) = 36995 − 261.96 Period + 1.6922 Period2
Table 11 Regression models for the consumptions by Minitab. Source: Corporacion de Reservas
Table 11 shows the regression models are used for the LNG and NG consumption
data in Spain. There are three regression models available from the Minitab which are
linear, quadratic and cubic regression models. There are two quadratic models, and
two cubic models are chosen. The choice of the proper model is based on the p-value
of the regression models and variables of the model along with R-squared (R-sq). The
figures are shown in table 12 below.
43
Conventional Electricity generator LNG direct consumption Total
P-Value P-Value P-Value P-Value
Regression 0.5103 <0.0001 <0.0001 <0.0001

Model R-sq R-sq R-sq R-sq
2.19% 74.59% 60.30% 30.56%
P-Value P-Value P-Value P-Value
Constant <0.0001 <0.0001 <0.0001 <0.0001
Period 0.1386 <0.0001 <0.0001 <0.0001
Period2 0.1306 <0.0001 <0.0001 0.0006
Period3 0.1388 <0.0001

Table 12 Interpretation of the regression models of the consumptions of LNG and NG in Spain by Minitab.
Source: Corporacion de Reservas Estrategicas de Productos Petroliferos
Table 12 shows the general overview that except for the regression model of
conventional, the others are considered as a suitable model for the data. For instance,
a p-value of the regression models is less than 0.0001 in the data given except for
conventional having a high p-value (0.5103). In a similar way, the p-value of the
constant, period, Period2 and Period3 is less than 0.001 apart from the one of Period3
in total consumption among the data apart from conventional on which p-value of the
constant is less than 0.0001. However, the figure 0.0006 is also tiny, so it has a proper
value for the p-value.
When it comes to the figures of R-squared, the data of consumption of natural gas at
electricity generator is fit to its model at a high degree (74.59%) among the 4 data.
The next one is the data on LNG direct consumption having a figure of R-squared
(60.30%) while the data of conventional consumption is poorly fit its data. The effects
of the high value of R-squared of the electronic generator and low value or R-squared
of the conventional are represented by the value of the total consumption between
them.
4.3 Autoregressive Model for Forecasting

An autoregressive model is useful for forecasting without discernible trend seasonality,
and it is believed that a correlation exists among successive residuals (Keller, 2014).
There is a formula for the autoregressive model for forecasting below.
𝑦𝑡 = 𝐵0 𝑦𝑡−1 + 𝐵1 𝑦𝑡−2 + 𝜀
This model indicates the correlation of the successive values in the time series (Keller,
2014).
This model will be used for deciding the single diesel price for trucks transporting LNG
in 2018 based on the monthly price data of diesel in Spain from 2011 to 2017. The
price is calculated by Excel Regression on data analysis.
4.4 Data Collection

This chapter explains data collection for simulating the model along with relevant
assumption for each data collection altogether.
4.4.1 Demand for Liquefied Natural Gas Direct Consumption

This data is one of the decision variables (𝑥𝑖𝑗 ) for the simulation of the study while the
44
opening or closing of the LNG import terminals (𝑦𝑖 ) is the other decision variables.
Forecasting for the demand of the LNG direct consumption is done based on the data
from Corporacion de Reservas Estrategicas de Productos Petroliferos (2018) coupled
with the regression model with seasonality.
Before calculation of the demand, it is assumed that there would be no LNG loss such
as boiled off gas during the truck transportation. So, the initial amount of LNG
transported from LNG import terminals to the fuel stations coincides with the amount
of LNG arrived at the destinations. The flowing paragraphs are related to inputs of the
research.
According to table 3 in chapter 3.2.1., the sum of the capacity of LNG truck
transportation in the 6 LNG terminals in Spain apart from El Musel which is not under
operation (Enagas, 2016) is 2160. So, it is assumed that the estimated maximum
capacity of LNG transportation by trucks to LNG satellite terminals by
𝑇𝑜𝑡𝑎𝑙 𝐿𝑁𝐺 𝑡𝑟𝑎𝑛𝑠𝑝𝑜𝑟𝑡𝑎𝑡𝑖𝑜𝑛 𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦 𝑏𝑦 𝑡𝑟𝑢𝑐𝑘𝑠 𝑖𝑛 𝐿𝑁𝐺 𝑡𝑒𝑟𝑚𝑖𝑛𝑎𝑙𝑠

− 𝐿𝑁𝐺 𝑑𝑖𝑟𝑒𝑐𝑡 𝑐𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛 𝑖𝑛 2018
The next step is the calculation of estimated LNG demand from LNG satellite terminal
without SI. This value is calculated from natural gas conventional consumption. The
reason why using natural gas consumption for LNG satellite terminal is that the input
of the satellite terminals is LNG while the output of the terminals is natural gas. In
addition, it is assumed that the usage of natural gas from LNG satellite terminals is
for the conventional purpose. So, the SI and the regression model for from
conventional usage of NG are applied for this calculation. However, the results of the
original regression model exceed the total maximum capacity of LNG truck
transportation in LNG receiving terminals; there is an adjustment of the constant in
order to limit the value below the maximum capacity. So, instead of using the model
from the original regression model of conventional usage of natural gas
𝑅𝑎𝑡𝑒 (𝑦) = 19155 + 199.3 𝑃𝑒𝑟𝑖𝑜𝑑 − 4.327 𝑃𝑒𝑟𝑖𝑜𝑑2 + 0.02556 𝑃𝑒𝑟𝑖𝑜𝑑3
, the new formula
𝑅𝑎𝑡𝑒 (𝑦) = −500 + 199.3 𝑃𝑒𝑟𝑖𝑜𝑑 − 4.327 𝑃𝑒𝑟𝑖𝑜𝑑2 + 0.02556 𝑃𝑒𝑟𝑖𝑜𝑑3
is applied by only adjusting the constant from -19155 to -500. After applying it, the
data from period 73 to 84 is used. Subsequently, SI from original data is multiplied to
this figure. This new model is for the scenario 1 which is the “full capacity”. For the
scenario 2, another model adjusted the constant to -1000 is used in the same way as
the model in scenario 1. The last scenario adjusts the constant to 200.
The formula for the scenario 2
𝑅𝑎𝑡𝑒 (𝑦) = −1000 + 199.3 𝑃𝑒𝑟𝑖𝑜𝑑 − 4.327 𝑃𝑒𝑟𝑖𝑜𝑑2 + 0.02556 𝑃𝑒𝑟𝑖𝑜𝑑3
The formula for the scenario 3
𝑅𝑎𝑡𝑒 (𝑦) = 200 + 199.3 𝑃𝑒𝑟𝑖𝑜𝑑 − 4.327 𝑃𝑒𝑟𝑖𝑜𝑑2 + 0.02556 𝑃𝑒𝑟𝑖𝑜𝑑3
The result of the calculation is shown in table 12.
45
Scenario Scenario Scenario Available
1. LNG 2. LNG 3. LNG capacity
LNG
SI of demand demand demand for
2018 Direct
Conventional from from from satellite
(GWh)
Satellite Satellite Satellite terminals
(GWh) (GWh) (GWh) (GWh)
Jan 1.32 911.6 1232.9 572.3 2156.3 1248.4
Feb 1.21 869.4 1104 497.4 1949.4 1290.6
Mar 1.16 918.1 1034.8 453.9 1845.9 1241.9
Apr 0.97 865.1 851.3 363 1527 1294.9
May 0.91 877.9 779.2 327.9 1419.9 1282.1
Jun 0.83 894.4 708.7 290.1 1286.1 1265.6
Jul 0.82 942.8 692.7 280.4 1264.4 1217.2
Aug 0.71 954.5 598.7 239.9 1091.9 1205.5
Sep 0.82 1027 694 276.7 1260.7 1133
Oct 0.9 1032.4 757.9 306.7 1386.7 1127.6
Nov 1.08 1062.5 919 375.9 1671.9 1097.5
Dec 1.25 1057.3 1075.4 449.3 1949.3 1102.7
Table 13 Estimated demands from the destinations in Spain. Source: Enagas and Corporacion de
Reservas Estrategicas de Productos Petroliferos
Also, table 13 shows the number of trucks for LNG transportation is calculated by
dividing 0.3 GWh in each scenario.
Scenario Scenario Scenario

1 2 3
Jan 7149 4947 10227
Feb 6578 4556 9396
Mar 6510 4574 9214
Apr 5722 4094 7974
May 5524 4020 7660
Jun 5344 3949 7269
Jul 5452 4078 7358
Aug 5178 3982 6822
Sep 5737 4346 7626
Oct 5968 4464 8064
Nov 6605 4795 9115
Dec 7109 5022 10022
Table 14 Number of trucks required for each scenario.
Since the figures in table 14 have the feature of the seasonality, the number of the
trucks needed for LNG transportation to align with the trend. Winter season requires
more trucks than the other season. In particular, the demand for trucks in January is
the approximately 38% higher than the one in August. In the view point of the entire
supply chain, it is necessary to deal with the surplus of the trucks in off-peak season.
However, it is not in the scope of the study so it is not discussed in the paper.
46
4.4.2 Distribution of the Demand in the Autonomous Community Scale.
Autonomous Provinces with LNG fuel
Communities Provinces with LNG satellite stations stations
Galicia A Coruna, Lugo, Pontevedra A Coruna, Ourense
Castile and Leon Burgos, Avila, Leon Burgos
Extremadura Badajoz, Caceres Badajoz
Almeria, Cadiz, Cordoba, Granada, Jaen, Malaga,
Andalusia Sevilla Cadiz, Jaen, Malaga, Sevilla
Murcia Murcia Murcia
Albacete, Ciudad Real, Cuenca, Guadalajara,
Castilla La Mancha Toledo Cuenca, Guadalajara, Toledo
Madrid Madrid Madrid
Valencia Alicante, Valencia Alicante, Castellon, Valencia
Barcelona, Girona, Lleida
Catalonia Barcelona, Girona, Lleida, Tarragona Tarragona
Aragon Huesca, Teruel, Zaragoza Zaragoza
Navarre Navarra
Asturias Asturias
Cantabria Cantabria
Basque Countries Alava, Bizkaia, Gipuzkoa
Table 15 Autonomous Communities having provinces with small-scale LNG facilities. Source: Gas
Infrastructure Europe
At first, the population data is gained from Instituto Nacional de Estadistica (2018) and
the data of the 1st of January 2018 is applied for this study. Table 15 shows that the
monthly LNG direct and NG conventional consumptions are be divided for each
autonomous community having the LNG fuel loading stations and LNG satellite
terminals respectively. For the calculation, each LNG fuel station and satellite station
are collected in a province-level at first. After that, the population in the province are
summed in the autonomous community level. Subsequently, the ratios of the
population of each autonomous community against the total population of the
autonomous community collected are calculated. This ratio will be multiplied by the
monthly estimated consumption of the LNG direct consumption and NG conventional
consumption. The figure after the calculation represents the amount of LNG demand
from the autonomous community through the truck transportation since the resource
of NG convention consumption from satellite terminal was previously LNG. So, LNG
demand from the LNG fuel stations and satellite terminals are summed up, so there
is a total demand for LNG by truck transportation in the autonomous community level.
The results are in Annex B.
4.4.3 Distance between Liquefied Natural Gas Import Terminals and

Autonomous communities
In order to calculate the distance between LNG import terminals and autonomous
community, it is necessary to collect distance data from the terminals to each LNG
fuel stations and satellite stations. The locations of the 7 import terminals are provided
by Global Energy Observatory (n.d.a) (n.d.b) (n.d.c) (n.d.d) (n.d.e) (n.d.f) (n.d.g). Also,
the locations of the LNG fuel loading stations for road vehicles and ships and the
satellite stations have been provided by Gas Infrastructure Europe (2018b) and
Dieselo Gasolina (2018) for the more accurate position of the facilities. The location
data which is not accessed by Gas Infrastructure Europe (2018b) has been searched
on the relevant website. Subsequently, based on the collected information, the
distance (kilometres) between the LNG import terminals and the destinations are
calculated on Google Maps. The shortest distance by car is chosen among the
47
suggested distances from Google Maps. After collecting all the distance data between
LNG receiving terminals and the destinations, the distance data is averaged in the
autonomous level. So, there is a single distance from the LNG receiving terminals to
the autonomous community.
Furthermore, only LNG fuel stations existing in Spanish territory in Iberia peninsula is
considered for the research. So, three LNG satellite terminals which are Cala Millor,
Can Picafort and Manacor are located in the Balearic Islands are excluded from the
research.
4.4.4 Lot Size of the Trucks

Barcel Cartag Huel El Sagu Bilb Mugar
LNG import terminal
ona ena va Musel nto ao dos
LNG truck loading bay 3 3 3 2 2 1 2*
Numbers of LNG trucks per day 50 50 50 30 40 15 35*
Daily max. capacity of LNG truckload
15 15 15 9 12 4.5 10.5
transportation (GWh)
Monthly max. capacity of LNG truckload
450 450 450 270 360 135 315
transportation (GWh)
Table 16 LNG truckload capacity in LNG terminals in Spain. Source: Enagas
Before discussing the lot size of the trucks, there is no data of LNG truck loading bay
in Mugardos, however, it can be assumed as 2* since El Musel terminals (the capacity
of 30 trucks) and Sagunto terminals (the capacity of 40 trucks) have two loading bays.
In addition, there is no data for the daily number of trucks available in Mugardos so
35* from the figure XX is calculated by
10.5 𝐺𝑊ℎ(𝑀𝑎𝑥.𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦 𝑜𝑓 𝐿𝑁𝐺 𝑡𝑟𝑎𝑛𝑠𝑝𝑜𝑟𝑡𝑒𝑑 𝑏𝑦 𝑡𝑟𝑢𝑐𝑘𝑠 𝑖𝑛 𝑀𝑢𝑔𝑎𝑟𝑑𝑜𝑠 𝑡𝑒𝑟𝑚𝑖𝑛𝑎𝑙𝑠)
0.3 𝐺𝑊ℎ(𝑙𝑜𝑡 𝑠𝑖𝑧𝑒 𝑜𝑓 𝑡𝑟𝑢𝑐𝑘𝑠)
).
Based on the data provided by the table 16, it is possible to it is possible to calculate
the maximum capacity of LNG transported by single truck although Enagas (2012)
did not mention the lot size of the trucks. After dividing from the maximum capacity of
LNG transported by trucks per day by the maximum number of trucks transporting
LNG per day, the figure of 0.3 Gigawatt hours (GWh) is calculated. So, 0.3 GWh is
considered as the maximum amount of LNG transported by a truck, and this will be
used for the research. Eurostat (2013, para 1.) explains the meaning of GWh as “GWh
is a unit of energy representing one billion-watt hour and is equivalent to one
million kilowatt hours” (Eurostat, 2013, para 1.). “Gigawatt hours are often used as a
measure of the output of large electric power stations” (Eurostat, 2013, para 1.). “A
kilowatt-hour is equivalent to a steady power of one kilowatt running for one hour and
is equivalent to 3.6 million joules or 3.6 megajoules” (Eurostat, 2013, para 1.). In
addition, the capacity of LNG in fuel station is assumed sufficient to be received the
full amount of LNG from the lot size.
4.4.5 Capacity of LNG Transportation by Trucks in the LNG terminals

Figure 18 also shows the maximum capacity of the LNG transportation by trucks in
the LNG terminals. Barcelona, Cartagena, Huelva have the largest capacities (450
GWh) while Bilbao terminal has the lowest capacity (135 GWh),
4.4.6 Transportation Costs

In this study, the total cost consists of transportation costs and inventory holding costs
during transit time. Fixed costs derived from operating the LNG receiving terminals
are excluded.
48
4.4.6.1 Fuel Price per Kilometre for the Trucks
Before deciding the fuel price for the research, it is vital to opt for the what type of the
fuel is applied. According to the table 8 in chapter 3.3.4.1., around 99 % of the road
tractor is fuelled by diesel so, all the trucks transporting LNG are considered as diesel-
fuelled vehicles.
The price data of diesel is presented on the website of European Commission (2018).
Based on the monthly data, the diesel price in Spain in 2018 will be calculated by
using the autoregression in Minitab. Since this study is about the minimisation of the
transportation costs, only one value of the fuel cost is applied in the entire of the year
to clearly observe the changes in the transportation costs regardless of the fluctuation
of the diesel price.
The Autoregressive formula of the diesel price in Spain is below.
𝑦𝑡 = 0.046517+ 1.1567𝑦𝑡−1 - 0.1957𝑦𝑡−2
After calculation, 1.13 euros are acquired for the diesel price for January 2018. This
value is considered as the fixed diesel price of a whole 2018.
In order to acquire the data of fuel consumption per kilometre, the data of the fuel
efficiency is required. In this study, the information about the fuel efficiency of the truck
transporting LNG is acquired from Dünnebeil et al. (2015). Dünnebeil et al. (2015) set
the fuel efficiency of the semi-trailer truck as 34.5L (Litre-Diesel/100km). In the
calculation, the fuel efficiency is divided by 100 to make it as litre-diesel / 1km. So,
the result is 0.38985 Euros per kilometre.
4.4.6.2 The Rest of Transportation Costs

According to Hopper and Murray (2017), many factors are consisting of the
transportation costs of the truck. Data for 2016 provided by Hopper and Murray (2017),
is applied for the simulation in 2018 apart from the fuel costs already calculated in the
sub-chapter above. The rests of transportation costs along with the calculated fuel
price are shown in the table below. In addition, since Hopper and Murray (2017)
provided the currency as dollars and the measurement of length as miles, those two
are converted to Euro and kilometres respectively: 1 dollar to 0.9 Euros and 1 mile to
1.6 kilometres.
Rate based on trucks

(Euro/km)
Fuel Costs 0.3898
5
Payments for leasing or 0.143
purchasing vehicles
Repair and Maintenance 0.093
Truck Insurance Premiums 0.042
Permits and Licenses 0.012
Tires 0.02
Tolls 0.014
Rate based on workers (Euro/km)
Driver Wages 0.294
Driver Benefits 0.087
Total (Euro/km) 1.095
49
Table 17 Transportation costs in the scenario of the study. Source: American Transportation Research
Institute, European Commission and Dünnebeil et al.
Table 17 indicates that 1.095 Euros per kilometre is the rate of the transportation costs
in this study.
4.4.7 Design of Transportation Networks

Direct shipment from a single destination is applied for the research (Chopra and
Meindl, 2015). In this assumption, the adverse effects such as complication of
coordination and operation derived from multiple destinations are removed (Chopra
and Meindl, 2015). In other words, a direct shipment with a milk run for LNG
transportation is excluded from the research, but solely direct shipping is considered
for LNG truckload transportation (Chopra and Meindl, 2015).
Transportation by means of less than truckload (LTL) has smaller lot sizes than
truckload (TL) (Chopra and Meindl, 2015). In this study, it is assumed that the amount
of shipment of LNG by the truckload is the full size of the LNG tank, so TL is applied
for the simulation. In addition, hub-and-spoke networks which are likely used for LTL
shipments are not relevant to this study (Chopra and Meindl, 2015). So, LTL is
excluded but TL transporting from a single origin to a single destination is considered
for the research.
Furthermore, it is supposed that sufficient LNG storage capacity in the LNG fuel
stations and satellite stations to satisfy the amount of LNG receiving by trucks.
Moreover, end customers directly from the LNG import terminals are excluded
4.4.8 Pipeline Holding costs

Inventory holding costs occur during the transportation of the products (Chopra and
Meindl, 2015). According to Chopra and Meindl (2015), calculation of the annual
holding costs in transit is below.
H: Annual holding costs

h: Holding cost rate
c: Price of the LNG (0.3 GWh) in the tank by truck transportation.
𝐻 =ℎ∗𝑐
𝑇𝑟𝑎𝑛𝑠𝑖𝑡 𝑑𝑎𝑦𝑠
To calculate holding costs per days 𝐻𝑑 , will be multiplied for holding costs.
365 𝑑𝑎𝑦𝑠
So, the formula for holding costs in transit days is below
𝑇𝑟𝑎𝑛𝑠𝑖𝑡 𝑑𝑎𝑦𝑠
𝐻𝑑 = ∗ℎ∗𝑐
365 𝑑𝑎𝑦𝑠
So, there are three variables are required for the calculation of the holding costs:
holding cost rate, transit time, and LNG price in truck transportation. They are
explained below paragraphs.
At first, many kinds of literature mentioned the holding cost rate around annually 25%
of the price of the good (Durlinger, 2012). The underlying reason for the figure is
argued that “It will be assumed here that a charge of ten per cent on the stock is a fair
one to cover both interest and depreciation. It is probable that double this would be
fairer in many instances” (Hariss, cited in Durlinger, 2012, p. 2). Also, 25% is also
derived from averaging of the holding costs rates in other industries between 5% and
45% (Durlinger, 2012). So, in this study, the fixed holding cost rate of 25% is used.
50
On the other hands, Yang (2014) stated that many holding costs are assumed to be
constant per unit time even though the holding cost is variable for the same products
depends on time for perishable products. In this study, the holding costs also will also
vary but the different reason which is seasonality of LNG. As mentioned before, the
seasonality of LNG price affects the price of holding costs.
Since the holding cost calculation is about annual cost, so it might not be possible that
calculating monthly holding costs with seasonality from annual holding costs. In other
words, the monthly seasonal price is only applied to a month, not a whole year.
However, in order to explore the effect of seasonality and holding costs, the monthly
holding cost is calculated from the annual holding cost.
In terms of transit time by the truck transportation, there are some factors to delay the
transportation by trucks such as traffic congestions, working hours for drivers and road
constructions (Reed TMS Logistics, n.d.). So, these factors are required to consider
for calculation of the transit time by the truckload.
The transit time is assumed based on the data of the estimated transit time for TL by
Reed TMS Logistics (n.d.). Reed TMS Logistics (n.d.) provided data of distance by
miles, so it has been converted to kilometre by multiplying 1.6 from miles. The result
is in table 18.
Distance Transit
(kilometres) time
0 ~ 639 Same day
or next day
640 ~ 960 1 day
961 ~ 1920 2 days
1921 ~ 2880 3 days
Table 18 Transit time depends on distances. Source: Reed TMS Logistics
In this study, the distance range between 0 ~ 639 km is considered as same day
service and the transit time is estimated as half of the days. So, holding cost per a
half day is calculated by dividing half from holding cost per 1 day.
Finally, the price of the LNG in 2018 is calculated based on the data from Federal
Energy Regulatory Commission (2018) by using the regression model with
seasonality for forecasting. After that, the measurement of Dollar/MMBtu is converted
to Euro/GWh. Since 1000 MMBtu is approximately equally to 0.29 GWh, it is assumed
that 1000 MMBtu is 0.3 GWh to make the GWh to the same as lot size of LNG tank
for the truck transportation (0.3 GWh) in the research. Subsequently, after converting
dollar with Euro, table 21 shows the LNG price per 0.3 GWh along with monthly
holding costs of LNG with 25% of the holding cost rate after calculations.
Annual Holding Cost

€/0.3 Holding Cost (€) per a half Holding Cost
2018 $/MMBtu GWh (€) day (€) per 1 day
Jan 9.33 8397 2099.25 2.88 5.75

Feb 8.96 8064 2016 2.76 5.52
Mar 8.23 7407 1851.75 2.54 5.07
Apr 8.34 7506 1876.5 2.57 5.14
May 8.94 8046 2011.5 2.76 5.51
June 9.68 8712 2178 2.99 5.97
July 10.03 9027 2256.75 3.09 6.18
51
Aug 10.83 9747 2436.75 3.34 6.68
1079
Sept 11.99 1 2697.75 3.7 7.39
1229
Oct 13.66 4 3073.5 4.21 8.42
1321
Nov 14.68 2 3303 4.53 9.05
1417
Dec 15.75 5 3543.75 4.86 9.71
Table 19 Forecasted Estimated Landed Price of LNG and Holding Costs in 2018. Source: Waterborne
Energy, Inc cited in Federal Energy Regulatory Commission
Table 19 shows that the holding costs gradually go up from April 2018 and in
December, the cost reaches its peak to 9.71 euros per day.
1.4 SI
1.2 SI
1.0 SI
0.8 SI
0.6 SI
0.4 SI
0.2 SI
0.0 SI
1 2 3 4 5 6 7 8 9 10 11 12
Figure 38 Seasonal Index (SI) of the estimated landed price of LNG in Spain in 12 months in Spain.
Source: Waterborne Energy, Inc cited in Federal Energy Regulatory Commission
Figure 38 illustrates that the seasonal index of the estimated landed price of LNG in
Spain acquired from on the data between 2015 and 2017 are high during winter
season whereas relatively low from March to July. The figures also related to the
seasonality of the demand for LNG and NG in Spain shown the chapters above. As
the demand for the resources goes up, the price follows the upward trend. Oppositely,
if the demand decrease, the price also decreases.
52
Chapter 5 Simulations and Results
The network optimisation model enables to test the scenarios. The significant
difference between the scenarios is the different demand of SSLNG. So, adjusting the
SSLNG demand aligning with the scenarios is done in the model by inputting different
the figures of the demands. Also, it is merely including or excluding the related cells
for El Musel LNG import terminal in the model in order to test the sub-scenarios.
Each scenario is monthly simulated to verify the monthly difference by putting the
different demands, holding inventory costs. Three significant results are acquired from
the model: Allocation of LNG, excess capacity and total costs. The three are essential
to analyse the result of running the simulations.
5.1 Result of Scenario 1. Full Capacity
5.1.1 Allocation of LNG

There are two critical factors of the distribution. The first one is that distances between
the origin and destination play a vital role in the distribution since the more distance,
the more costs occur. The amount of the LNG demand with the seasonality also
affects the allocation of LNG.
Destination of LNG (Months)

LNG Castilla Extre M Castilla Ma Val Cat Ar Basque Na
Supply Gali Asturia and madu Andalu ur La dri enc alon ag Countri var
Terminal cia s Leon Cantabria ra sia cia Mancha d ia ia on es re
Barcelona x x x x x x X x x x 12 10 X x
Cartagen
a x x x x x 2 12 4 12 x x x X x
Huelva X x x x 12 12 X x x x x x X x
El Musel X 12 5 2 x x x x 1 x x x x x
Sagunto X x x x x x x 12 7 12 2 5 x x
Bilbao X x 10 10 x x x x 7 x x 5 12 12
Mugardos 12 x x x x x x x x x x x x x
Transport Appear Disappear An
ation during during entire
pattern winter winter of year
Table 20 Monthly allocations of LNG between LNG import terminals and destinations in Scenario 1.1.
Table 20 illustrates the results of simulation 1.1 about the monthly frequency of LNG
distribution between the origins and destinations. Before analysis of the results, there
are trends of the allocation between the origin and destination. Some allocations of
relatively short months happen only during in winter season while other allocations
with a relatively more extended period occur only during the other seasons. The
amount LNG of these occurrences tend to account for the relatively low ratio against
the total LNG demand in the autonomous community level and be smaller than one
of occurring 12 months. For instance, the situation that the destinations where LNG
transportation disappears during winter can be observed at Aragon and Catalonia
supplied LNG from Barcelona terminal. Barcelona LNG import terminal is entirely
dedicated to the LNG demand from Catalonia apart from January and December
because the terminal also transports LNG to Aragon at that time.
Regarding the demand in Aragon, the amount of LNG which is supposed to transport
this autonomous community during the winter season is taken over by Catalonia since
the demand in Catalonia is higher with having lower total costs for the transportation
than the other months. Oppositely, Andalucía is supplied additional LNG from
Cartagena LNG terminal to meet the peak demand in winter. Because the capacity of
53
LNG supplied by Huelva LNG import terminal is not sufficient and there is no
alternative for Cartagena terminal.
In the viewpoints of the terminals, Mugardos terminal is the only dedicated terminal
supplying LNG to only Galicia while Bilbao terminal has six destinations which are the
most significant number of the seven terminals. In other words, Mugardos LNG
terminal and Galicia are the most isolated from the other terminals and autonomous
community. Oppositely, Bilbao terminal has a relatively close connection of the other
autonomous community than other terminals.
In the autonomous community’s point of view, Madrid is supplied LNG from 4 LNG
import terminals which are the largest among other autonomous community. The
geographical feature located in the centre of Spain is the underlying reason for it. In
other words, there are no LNG import terminals in the vicinity of the centre of Spain.
Oppositely, Galicia, Basque Countries and Navarra are only supplied LNG by one
terminal which is physically close to those autonomous communities.

LNG Ga Castilla Extre M Castilla Ma Val Cat Ar Basque Na
Supply lici and madu Andalu ur La dri enc alon ag Countri var
Terminal a Asturias Leon Cantabria ra sia cia Mancha d ia ia on es re
Barcelona x X x x x x X X x x 12 10 x x
Cartagen
a x X x x x 2 12 5 12 x x x x x
Huelva x X x x 12 12 X X x x x x x x
Sagunto x X x x x x X 12 7 12 2 5 x x
Bilbao x X 12 9 x x X X 7 x x 5 12 12
Mugardos 12 12 x 5 x x X X 2 x x x x x
Transport Appear Disappear An
ation during during entire
pattern winter winter of year
The red coloured figures in table 21 are different from the result in scenario 1.1 in
which El Musel is operated. The absence of the operation of the El Musel LNG
terminal affects the other terminals mainly to Mugardos LNG import terminals. In the
scenario 1.2, Mugardos is not only dedicated to supplying LNG to Gallica but to
additional three autonomous communities which are Asturias, Cantabria and Madrid.
In other words, some amount of LNG transported from El Musel terminal is allocated
to Mugardos terminal. Asturias is fully supplied LNG from the terminal for a whole year
while Mugardos terminal supplies LNG to Cantabria for five months during winter.
Similar to the distribution to Cantabria, Madrid is supplied LNG for two months during
winter time by the same terminal. Take a look at the distribution of LNG to Madrid in
January in both scenario 1.1 and 1.2 from annex C, the amount of LNG should have
been transported from El Musel terminal in scenario 1.1 distributed to El Musel. Also,
the 42 GWh of LNG transportation from Cartagena terminal to Madrid is taken over to
Mugardos from scenario 1.1 to 1.2. This is the notable result that El Musel terminal is
relatively located far from Cartagena terminal unlike the Bilbao and Mugardos terminal.
Other terminals affected the frequency of the monthly transportation by the absence
of the operation of the El Musel terminal are Bilbao and Cartagena terminal. In Bilbao
terminals, Castilla and Leon are supplied LNG during the entire year, and the
frequency of the monthly transportation decreased by one month. Also, the Cartagena
LNG import terminal supply LNG one more month in this scenario than the one in
scenario 1.1.
54
5.1.2 Excess capacity
Supply Sep
Region Jan Feb Mar Apr May Jun July Aug t Oct Nov Dec
Barcelona 0 0 0 28.2 41.7 53.9 45 64 23 6.7 0 0
129. 155. 177. 160. 195. 116.
Cartagena 0 0 0 9 1 6 2 5 7 85.8 0 0
113. 108. 128.
Huelva 0 24.7 31.9 85.4 100 7 3 3 92.1 76 32.1 0
224. 235. 246. 248. 250. 251. 253. 251. 249. 206.
El Musel 88.3 8 9 9 8 7 2 7 1 4 2 96.7
Sagunto 0 0 0 0 0 0 0 0 0 0 0 0
Bilbao 0 0 0 0 0 0 0 0 0 0 0 0
197. 207. 209. 223. 227. 229. 235. 226. 221. 210. 200.
Mugardos 1 1 3 2 2 231 9 3 1 8 3 6
285. 456. 477. 713. 772. 826. 794. 876. 639. 448. 297.
Total 4 6 1 6 8 9 6 8 709 7 6 3
Table 22 Excess capacity with El Musel Terminal in Scenario 1.1. (GWh).
From table 22, El Musel LNG import terminal which is currently not under operation
has the largest surplus capacity while Sagunto and Bilbao LNG receiving terminals
has no excess capacity among the 7 LNG import terminals throughout the year. Since
the demand for LNG has seasonality, the amount of the total excess capacity hare
larger during summer seasons while the less excess capacity during winter seasons.
Supply De
Region Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov c
Barcelona 0 0 0 28.2 41.7 53.9 45 64 23 6.7 0 0
129. 155. 177. 160. 195. 116.
Cartagena 0 0 0 9 1 6 2 5 7 85.8 0 0
113. 108. 128.
Huelva 0 24.7 31.9 85.4 100 7 3 3 92.1 76 32.1 0
Sagunto 0 0 0 0 0 0 0 0 0 0 0 0
Bilbao 0 0 0 0 0 0 0 0 0 0 0 0
15. 161. 175. 200. 211. 211. 207. 201. 146. 27.
Mugardos 4 9 2 1 206 7 1 219 2 2 5 3
15. 186. 207. 443. 502. 556. 524. 606. 369. 178. 27.
Total 4 6 1 6 8 9 6 8 439 7 6 3
Table 23 Excess capacity without El Musel Terminal in Scenario 1.2. (GWh).
Table 23 in scenario 1.2 shows that the absence of the El Musel LNG import terminal
does not affect the amount of excess capacity at Barcelona, Cartagena and Huelva.
The far distance between El Musel and three terminals is the underlying reason. The
effect is on the Mugardos LNG import terminals since the terminal, and El Musel
terminals are geographically closer than the other three terminals. On the other hand,
Sagunto and Bilbao LNG receiving terminals are not affected by the close of El Musel
terminal since they do not have excess capacity from the scenario 1.1.
5.1.3 Total costs

The
With El Without EL difference Ratio of the difference against the costs
Musel (€) Musel (€) (€) without El Musel (%)
Jan 421012.7 438982 17969 4.09
Feb 374939.6 379079 4139 1.09
55
Mar 369740.1 372728 2988 0.80
Apr 321065.4 322990 1924 0.60
May 309496.3 311262 1766 0.57
Jun 298968.4 300576 1608 0.53
July 305434 307000 1566 0.51
Aug 289302.9 290660.7 1358 0.47
Sept 323328.9 324903.2 1574 0.48
Oct 338150.3 339868.4 1718 0.51
Nov 380686 387001 6315 1.63
Dec 420806.8 437956.8 17150 3.92
Table 24 Total costs in Scenario 1.1 and 1.2.
The difference between the scenario 1.1 with El Musel terminal and scenario 1.2. are
significant during the winter season, and it becomes narrow during the winter season.
It is an obvious result since the more transportation of LNG to the destination to meet
the demand with the seasonality. Explore more in detail of ratio of the difference
against the total costs without El Musel in table 24, the ratio does not account for
substantial parts which are mostly below 1% apart from the one from November to
February. So, there is not an outstanding advantage of operating El Musel terminal. If
the fixed costs of operating the LNG import terminal which is excluded from the study
are considered, it would be a worse choice to run the LNG import terminal. The
positive of the operation of the terminal tend to appear if there is unmet LNG demand
or the increase in transportation costs and inventory holding costs.
5.2 Result of Scenario 2. Abundant Capacity
5.2.1 Allocation of LNG

G M Castilla Basqu
LNG ali Ast Castill Extre ur La M Val Cat Ar e Na
Supply ci uri a and mad Andal ci Manch ad en alo ag Countri var
Terminal a as Leon Cantabria ura usia a a rid cia nia on es re
Barcelon
a x x x x x X x X x x 12 12 x X
Cartagen
a x x x x x X 12 X 12 x X x x X
Huelva x x x x 12 12 x X x x X x x X
El Musel x 12 x x x X x X 1 x X x x X
Sagunto x x x x x X x 12 12 12 X x x X
Bilbao x x 12 12 x X x X 12 x X x 12 12
Mugardo
s 12 x x x x X x X x x X x x X
Ap
pe
ar
dur
ing An
Transpor wi Disappea entire
tation nte r during of
pattern r winter year
Table 25 shows that the decrease in LNG demand in scenario 2.1 has effects on the
56
distribution of LNG. At first, the seasonal distributions appearing and disappearing
during the winter season does not seem necessary anymore apart from LNG
transportation from El Musel to Madrid. The red figures represent that the difference
against the scenario 1.1. Another effect of the decrease in LNG demand is that the
autonomous communities tend to be supplied LNG from a single terminal. As a result,
some of the autonomous communities have the more frequency of receiving LNG
than the one from scenario 1.1. For instance, Castilla and Leon and Cantabria are
solely transported LNG from Bilbao terminal by decreasing in the frequency of
receiving LNG from EL Musel terminal to zero month.
On the other hand, Madrid still has 4 LNG import terminals to receive LNG compared
to the scenario 1.1. Again, the central location of Madrid substantially affects this
arrangement that having the many connections from the LNG terminals. The only
change of the distribution to this autonomous community is that the frequency of LNG
transportation increases from 7 months to 12 months from Sagunto and Bilbao
terminals.

M Castilla Basqu
LNG Castill Extr ur La M Val Cat Ar e Na
Supply Gal Asturi a and Cantabri ema Andal ci Manch ad en alo ag Countr va
Terminal icia as Leon a dura usia a a rid cia nia on ies rre
Barcelon
a x x x x x x x X x x 12 12 x x
Cartagen
a x x X x x x 12 X 12 x x x x x
Huelva x x x x 12 12 x x x x x x x x
Sagunto x x x x x x x 12 12 12 x x x x
Bilbao x x 12 12 x x x x 12 x x x 12 12
Mugardo
s 12 12 X x x x x x x x x x x x
Appe An
Transpor ar Disappea entire
tation during r during of
pattern winter winter year
Table 26 illustrates that the change in the frequency in scenario 2.2 compared to
scenario 2.1 nearly does not exists even though El Musel LNG import terminal is not
under operation. This is quite a different result from the comparison of the distribution
between scenario 1.1 and 1.2 experiencing the several changes. The only change in
the frequency is that Mugardos terminal becomes a supplier of LNG to Asturias for 12
months. This is taken from El Musel terminal and this result is the same as the one
between scenario 1.1 and 1.2.
5.2.2 Excess capacity
Supply
Region Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
81. 109. 107. 141. 146. 151. 141 147. 120. 111. 87. 72.
Barcelona 1 5 1 9 7 2 .1 4 1 7 8 1
228 283. 277. 344. 353. 361. 341 353. 300. 284. 238 208
Cartagena .9 3 7 6 7 9 .5 3 1 4 .5 .9
142 168. 169. 201. 206. 212. 213. 182. 160 144
Huelva .7 4 5 1 9 8 206 3 191 9 .8 .3
254 256. 257. 260. 261. 262. 262 263. 262. 261. 259 257
El Musel .5 5 7 1 1 1 .4 5 5 7 .8 .8
Sagunto 0 0 0 0 0 0 0 0 0 0 0 0
57
Bilbao 0 0 0 0 0 0 0 0 0 0 0 0
238 245. 246. 254. 255. 257. 252. 250. 244 240
Mugardos .9 5 1 1 8 5 256 258 5 4 .7 .3
946 1063 1058 1201 1224 1245 120 1235 1126 1091 991 923
Total .1 .2 .1 .8 .2 .5 7 .5 .2 .1 .6 .4
Table 27 Excess capacity with El Musel Terminal in Scenario 2.1. (GWh).
From table 27, the less LNG demand leads to the more surplus capacity of the LNG
transportation by trucks. There is no excess capacity in Sagunto and Bilbao LNG
import terminal which is the same as the scenario 1. However, apart from the two
terminals, the other terminals have the excess capacity the entire of the year.
Supply
Region Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
109. 107. 141. 146. 151. 141. 147. 120. 111.
Barcelona 81.1 5 1 9 7 2 1 4 1 7 87.8 72.1
228. 283. 277. 344. 353. 361. 341. 353. 300. 284. 238. 208.
Cartagena 9 3 7 6 7 9 5 3 1 4 5 9
142. 168. 169. 201. 206. 212. 213. 182. 160. 144.
Huelva 7 4 5 1 9 8 206 3 191 9 8 3
Sagunto 0 0 0 0 0 0 0 0 0 0 0 0
Bilbao 0 0 0 0 0 0 0 0 0 0 0 0
223. 233. 244. 246. 249. 248. 251. 242. 234. 228.
Mugardos 4 232 8 2 9 6 4 5 245 1 5 1
676. 793. 788. 931. 954. 975. 965. 856. 821. 721. 653.
Total 1 2 1 8 2 5 937 5 2 1 6 4
Table 28 Excess capacity without El Musel Terminal in Scenario 2.2 (GWh).
In the same way of the scenario 1.2., Mugardos terminal absorbs the LNG demand
by trucks transportation while the other terminals are not affected by the absence of
the operation of the El Musel terminal from table 28.
5.2.3 Total costs
Ratio of the difference

With El Without EL Musel Differenc against the costs without El
Musel (€) (€) e (€) Musel (%)
Jan 275428.6 276720 1291 0.47
Feb 251385.8 252510 1125 0.45
Mar 251932.2 252957 1025 0.41
Apr 223396.7 224221 825 0.37
May 219124.3 219866 741 0.34
Jun 215042 215700.1 658 0.31
July 222715.9 223349 633 0.28
Aug 217271.8 217813.3 541 0.25
Sept 239377.6 240002.4 625 0.26
Oct 247091 247783.2 692 0.28
Nov 267579.2 268428.9 850 0.32
Dec 281844.1 282860.3 1016 0.36
Table 29 Total costs in Scenario 2.1. and 2.2.
58
Table 29 shows that compared to the scenario 1, the ratio in the scenario 2 is less
than the one in the scenario 1. In other words, operating El Musel terminal become
less critical in the viewpoints of the total costs. So, it is apparent that it is a better
strategy not to run the additional terminal if the LNG demand is lower like in the
scenario 2.
5.3 Result of Scenario 3. A lack of Capacity

In this scenario, excessive LNG demand above the available capacity from the LNG
import terminals. Since the constraint which LNG demand should be less than the
available capacity is violated in this scenario, the supply chain is not able to be
optimised with a massive amount of the total costs. In other words, Excel Solver is
not able to acquire a feasible solution. In order to verify the possibility of analysis,
there are some adjustments in the model. When the orders of the terminals have been
changed, the results including the distributions of the LNG, excessive capacity and
the objective costs are different from the original model.
Suppl
y Ga Ast Castile Can Extre And M Castilla M Val Cat Ar Basque Na
Regio lici uri and tabr madu alus ur La ad enc alo ag countri var
n a as Leon ia ra ia cia Mancha rid ia nia on es re
12
Barcel 5. 324
ona 0 0 0 0 0 0 0 0 8 .2 0 0 0 0
42
Carta 3.
Ori gena 0 0 0 0 0 0 0 26.3 7 0 0 0 0 0
gin
12
al
Huelv 210. 3.
Mo
a 0 0 0 0 0 7 8 115.5 0 0 0 0 0 0
del
El
Musel 0 0 0 0 0 270 0 0 0 0 0 0 0 0
Sagun 34 53. 103.
to .7 0 62.1 30 77.0 2 0 0 0 0 0 0 0 0
13
5.
Bilbao 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Mugar 61. 628 98. 33.
dos 0 0 0 0 0 0 0 0 0 4 .2 8 70.2 2
Suppl
Regio lici uri and tabr madu alus ur La ad enc alo ag countri var
n a as Leon ia ra ia cia Mancha rid ia nia on es re
Barcel 196 628 98. 33.
ona 0 0 0 0 0 0 0 0 0 .4 .2 8 70.2 2
26
Carta 0. 189
Adj gena 0 0 0 0 0 0 0 0 8 .2 0 0 0 0
ust 12 28
ed Huelv 3. 8.
Mo a 0 0 0 0 0 30.7 8 6.8 7 0 0 0 0 0
del El
Musel 0 0 0 0 0 270 0 0 0 0 0 0 0 0
Mugar 283.
dos 0 0 0 0 31.8 2 0 0 0 0 0 0 0 0
16
Sagun 9. 53.
to 7 0 62.1 30 45.2 0 0 0 0 0 0 0 0 0
Bilbao 0 0 0 0 0 0 0 135.0 0 0 0 0 0 0
Table 30 Result of the distribution of LNG in December in Scenario 3
Supply Original Supply Adjusted

Region Model Region Model
Barcelona 0 Barcelona -576.8
59
Cartagena 0 Cartagena 0
Huelva 0 Huelva 0
El Musel 0 El Musel 0
Sagunto 0 Mugardos 0
Bilbao 0 Sagunto 0
Mugardos -576.8 Bilbao 0
Cost (€) 1802590.8 Cost (€) 1338180.7
Table 31 Results of excess capacity and the total costs in December in Scenario 3.
For instance, when the order of Mugardos terminal is located to the 5th from the 7th
line in December, the result is different from the original one. Table 30 and 31 show
the results in the adjusted model that Barcelona should have more capacity to
transport LNG to the destinations while the original results indicate that Mugardos
should have done it. So, it does not make sense to analyse the results if there is a
lack of capacity of terminals to handle the peak demand of LNG. In addition, there is
no need to compare the two situations that with and without the operation of the El
Musel LNG import terminal because the model cannot run the model in case of the
lack of capacity.
60
Chapter 6 Conclusion
6.1 Summary of the problem and methods

There are the several issues of the SSLNG supply chain in Spain. The first one is
derived from seasonality. Seasonality affects the SSLNG facilities’ choices of the LNG
import terminals to meet the demand. In addition, truck companies transporting
SSLNG are required to set out an effective strategy to handle the surplus derived from
the gap between peak and off-peak season. However, the subject is not covered by
the study so it is not studied in the paper.
Another is that not fully integrated supply chain leads to detrimental factors such the
high transaction costs and delays of the sharing information.
The third is that the limits of the SSLNG output capacity at LNG terminals deter for
SSLNG facilities to be supplied SSLNG from the nearest LNG import terminals.
In order to recoup the investment of the El Musel LNG import terminal which is in
mothballed, the owner of the terminal should know in what circumstance does the
terminal is required to operate.
MILP model by Excel Solver is used to optimise the SSLNG supply chain related to
the issues of the limit capacity and El Musel terminals. The regression model by
Minitab forecasts the input of MILP model which is the future demands with
seasonality. In addition, the regression model is used for estimated of the price of LNG
which is a factor of holding costs. The estimation of the fuel price of the trucks are
done by autoregressive model.
6.2 Key findings

With the regression model and data of the flows of the NG and LNG gas, the
seasonality of natural gas is verified. In the winter season, there is an increase in
consumption of NG and LNG for heating purpose. However, the consumption
decreases during the summer season since the primary reason for consuming the
resource is heating.
Vertical integration of small-scale LNG supply chain from LNG receiving terminals to
the destinations of the cargo which are LNG fuel stations and satellite terminal is not
observed. On the other hand, it is notable that Enagas is considered as a monopoly
player in LNG import station while oligopoly is observed in the small-scale LNG
facilities by several large companies such as Natrugy and HAM.
In the research, the network optimisation model with the three scenarios is simulated.
Throughout the entire results of the simulation, the two key factors deciding the
distribution of the small-scale is observed. The first one is that distance. Majority of
the distributions tend to occur if the origin and destinations are closer than alternatives.
Another is that the amount of LNG demand with seasonality. The demand is a root
cause for the appearance and disappearance of the frequency of the transportation
during winter time. Take a closer look at the results of each scenario. Scenario 1 is
simulated in the assumption of the full capacity. The sub-scenario 1.1 would show the
role of the El Musel LNG import terminal if the terminal was operated. Asturias is fully
supplied LNG from the terminal while the terminal provides the service of the short-
term seasonal LNG transportation for the three autonomous community: Castilla and
Leon, Cantabria and Madrid. The absence of the operation of El Musel terminal has
mostly effects on the terminals in the vicinity of the El Musel terminal. For instance, El
Musel terminal takes over the most amount of LNG transportation to Bilbao and
Mugardos terminal. However, Cartagena terminal is also affected by the no operation
of the El Musel terminal by being taken over 42 GWh of LNG which is supposed to be
transported to Madrid in January in the scenario 1.1. to Mugardos terminal.
61
The sufficient capacity is applied to scenario 2. Due to the less LNG demand, the
short-term seasonal transportations disappear. So, regarding the frequency of the
LNG transportation in scenario 2.1 compared to 2.2, the increase in the frequency
from the terminal where has relatively long-term frequency is observed by reducing
the relatively small seasonal frequency. The notable change from scenario 2.1 to 2.2
is that Mugardos terminal is taken the entire of the LNG transportation to Asturias by
the El Musel terminal.
It is impossible to optimise the model in the case of a lack of capacity in scenario 3.
Merely adjusting the order of LNG import terminals in the model with having all other
same conditions generates the different results.
The results of the simulation would be helpful to decide the distribution of LNG by
trucks depending on the number of LNG demands from the destinations and the open
of El Musel terminal.
In the scenario 1 and 2, the costs difference between operation and no operation of
El Musel LNG terminal is relatively small, and the ratio of the difference against the
total cost is also not substantial.
6.3 Implication
The three players (truck companies, SSLNG facilities, LNG import terminals) benefit
from the study. In the SSLNG supply chain, the seasonality is essential to reduce the
loss derived from surplus between peak and off-peak season. By estimating the gap
of utilising their trucks from the seasonality, truck companies well establish strategy
such as price differentiation. For instance, a higher price should be charged during
peak season which is winter to maximise total profits while lower price needs to be
charged during the off-peak season which is summer season to attract more
customers (Prentice and Prokop, 2016). Also, truck companies should consider
different types of cargo transportation other than LNG during the summer season for
more utilisation of their assets to generate profits. However, the research of the
strategies for truck companies is out of scope in the paper so it is not discussed.
Figure 39 Single or Multiple LNG import terminals to supply LNG to Autonomous Communities in Spain.
Depends on the different amount of SSLNG demand, SSLNG facility should decide
62
what terminals they need to contract to be supplied SSLNG. Figure 39 shows that
there are three types of SSLNG facilities can be discussed comparing scenario 1.2
and 2.2 which are no operation of El Musel LNG terminal. This comparison is more
feasible than the one of scenario 1.1 and 2.1 since El Musel terminal is not under
operation in reality. The first type is SSLNG facilities covering several possible LNG
import terminals supplying SSLNG in scenario 1.2 while coverage area becomes a
single terminal in scenario 2.2. SSLNG facilities in Cantabria, Andalusia, Castilla La
Macha, Catalonia and Aragon are the first type. So, SSLNG companies in the area
should set the plan to be supplied SSLNG depends on the amount of SSLNG demand.
If the demand is high, the should seek the other terminals to be supplied SSLNG to
meet the demand. If there is low demand, they can make a contract with the nearest
LNG import terminals.
The second type is the coverage of LNG import terminals are multiple regardless of
the amounts of SSLNG in scenario 1.2 and 2.2 Only Madrid belongs to the type
because of the geographical feature of Madrid. In other words, the location of Madrid
which is the centre of Spain result in the second type. So, SSLNG facilities located in
the autonomous community should consider that a single terminal cannot satisfy the
demand so always make contracts with several LNG import terminals.
The last type is SSLNG facilities having a single LNG import terminal in the vicinity of
their autonomous community regardless of the scenarios. So, it is not necessary for
the companies owning the facilities to consider what terminals they needed to be
supplied SSLNG other than one. SSLNG facilities in Galicia, Asturias, Castilla and
Leon, Extremadura, Murcia, Valencia, Basque countries, Navarre are only supplied
from only one nearest LNG import terminal.
The simulation of sub-scenarios for El Musel terminal is valuable for the owner of the
terminal. This result implies that it is a better option not to run El Musel terminal as
long as there is not higher LNG demand than the current capacity of the six terminals.
So, the owner of the terminal should claim for the necessity of operating the terminal
to the Spanish government only if there are the high NG and LNG consumption are
estimated in the future.
6.4 Limitation of the research and Suggestion of the Further Research

One of the limitations of the research is that the fixed costs of the LNG terminal are
not considered in this paper. In order to acquire a more precise result, it would be
better to consider the fixed cost of operating LNG import terminals in the future study.
Also, a limit of Excel solver which is not able to run the model with more than 200
variables discourages to simulate the scenario with the distances from the LNG
terminals to the exact location of each LNG fuel station and satellite terminals. Using
other software handling more variables would be suggested to acquire more accurate
data and more detailed analysis.
63
Bibliography
Agency for the Cooperation of Energy Regulators (2017). Frequently Asked

Questions (FAQs) on REMIT fundamental data and inside information collection.
[online] Available at: https://documents.acer-remit.eu/wp-content/uploads/FAQs-on-
REMIT-fundamental-data-and-inside-information_V4.pdf [Accessed 17 July. 2018]
Anne, N. (2008). Linking natural gas markets: is LNG doing its job?. DIW Discussion
Papers, No. 822. Berlin: Deutsches Institut für Wirtschaftsforschung (DIW).
Arfaa, N., Tordo, S. and Tracy, B. S. (2011). National Oil Companies and Value
Creation. World Bank Working Paper No. 218.
Arteconi, A., Brandoni, C., Evangelista, D. and Polonara, F. (2010). Life-cycle

greenhouse gas analysis of LNG as a heavy vehicle fuel in Europe. Applied Energy
87(6), pp. 2005-2013.
Benito, M. J. D., Couso, I. A. and Menendez, U. (2017). Oil and gas regulation in
Spain: overview. [online] Available at:
https://uk.practicallaw.thomsonreuters.com/w-011-
1357?transitionType=Default&contextData=(sc.Default)&firstPage=true&comp=pluk
&bhcp=1 [Accessed 14 July. 2018]
Best Way LNG Station (2018). The project. [online] Available at:
http://bestwaylngstations.com/the-project/ [Accessed 25 July. 2018]
Bittante, A., Jokinen, R., Krooks, J., Pettersson, F. and Saxen, H. (2016). Mixed
integer optimization of an LNG supply chain in the Baltic Sea region. Proceedings of
ECOS 2016- The 29th International Conference on Efficiency, Cost, Optimization,
Simulation and Environmental Impact of Energy Systems, June 19-23, 2016, Portorz,
Slovenia. p. 624.
Bittante, A., Jokinen, R., Krooks, J., Pettersson, F. and Saxén, H. (2017a). Optimal
design of a small-scale LNG supply chain combining sea and land transports.
Industrial & Engineering Chemistry Research. 56 (45), pp. 13434-13443.
Bittante, A., Pettersson, F. and Saxen, H. (2017b). A multi-period optimization model

for the design of new LNG supply chains. Proceedings of the 58th Conference on
Simulation and Modelling (SIMS 58) Reykjavik, Iceland, September 25th – 27th, 2017.
Linköping: Linköping University Electronic Press, pp. 332-342.
Bittante, A., Pettersson, F. and Saxen, H. (2018). Optimization of a small-scale LNG

supply chain. Energy, 148 (1), pp. 79-89.
Chopra, S. and Meindl, P. (2015). Supply chain management: Strategy, Planning, and
operation. 6th ed. London: Pearson.
Corporacion de Reservas Estrategicas de Productos Petroliferos (2018a). Natural

gas consumption. [online] Available at:
https://www.cores.es/en/estadisticas [Accessed 13 June. 2018]
Corporacion de Reservas Estrategicas de Productos Petroliferos (2018b). Natural
64
gas exports. [online] Available at:
Corporacion de Reservas Estrategicas de Productos Petroliferos (2018c). Natural gas

imports. [online] Available at:
Dieselo Gasolina (2018). Search for gas stations in Spain. [online] Available at:
https://www.dieselogasolina.com/buscador-gasolineras.html [Accessed 24 July. 2018]
Dobrota, D., Komar, I. and Lalic, B. (2013). Problem of boil - off in LNG supply chain.
Transactions on Maritime Science, 02, pp. 91-100.
Dünnebeil, F., Hausberger, S., Kies, A., Lambrecht, U., Reinhard, C. and Rexeis, M.
(2015). Future measures for fuel savings and GHG reduction of heavy-duty vehicles
[online] Available at:
https://www.umweltbundesamt.de/sites/default/files/medien/378/publikationen/texte_
32_2015_summary_future_measures_for_fuel_savings.pdf [Accessed 01 August.
2018]
Durlinger, I. P. P. J. (2012). Inventory and Holding costs. [online] Available at:

http://www.durlinger.nl/files/artikelen/Inventory-and-Holding-Costs.pdf [Accessed 01
August. 2018]
El Comercio (2017). Enagas negotiates with large traders to give activity to the El
Musel regasification plant. [online] Available at:
https://www.elcomercio.es/economia/201702/12/enagas-negocia-grandes-
comercializadores-20170212022016-v.html [Accessed 03 Sept. 2018]
Eliston, A. (2009). Convergence of Regional Liquid Natural Gas (LNG) Prices: A

review of regional LNG import prices using Engle Granger’s Cointegration approach.
University of Oslo.
Enagas (n.d.a). Gas pipelines. [online] Available at:

http://www.enagas.es/enagas/en/Transporte_de_gas/Red_de_transporte/Gasoducto
s [Accessed 09 August. 2018]
Enagas (n.d.b). Regasification plants. [online] Available at:

http://www.enagas.es/enagas/en/Transporte_de_gas/PlantasRegasificacion
[Accessed 06 July. 2018]
Enagas (n.d.c). Underground storage facilities. [online] Available at:

http://www.enagas.es/enagas/en/Transporte_de_gas/Almacenamientos_Subterrane
os [Accessed 09 August. 2018]
Enagas (2012). Transmission, storage and regasification services and infrastructure.

http://www.enagas.es/stfls/ENAGAS/Transporte%20de%20Gas/Documentos/Trans
mission,%20storage%20and%20regasification%20and%20infrastructure%20of%20
Enag%C3%A1s%20catalogue.pdf [Accessed 17 July. 2018]
65
Enagas (2015). LNG terminals. [online] Available at:
http://www.enagas.es/stfls/ENAGAS/Documentos/Folleto%20Terminales%20GNL%
20ENG_WEB.pdf [Accessed 20 July. 2018]
Enagas (2016). Leader in natural gas infrastructures. [online] Available at:

http://www.enagas.es/stfls/ENAGAS/Transporte%20de%20Gas/Documentos/CAT_E
nglish.pdf [Accessed 17 July. 2018]
Enagas (2017a). Enagas in 2017. [online] Available at:

http://www.enagas.es/WEBCORP-
static/InformeAnual2017/sites/default/files/enagas_en_2017_eng.pdf [Accessed 06
July. 2018]
Enagas (2017b). Spanish gas system. [online] Available at:

http://213.0.117.107/WEBCORP-static/instalaciones/indexE.htm [Accessed 10 July.
2018]
Enagas (2018). Underground storage facilities. [online] Available at:

http://www.enagas.es/enagas/en/Transporte_de_gas/Almacenamientos_Subterrane
os [Accessed 21 July. 2018]
Encyclopaedia Britannica (n.d.). Cryogenics. [online] Available at:

https://www.britannica.com/science/cryogenics [Accessed 08 Aug. 2018]
Energy Vortex (n.d.). British thermal unit (BTU), MBTU, MMBTU. [online] Available at:
https://www.energyvortex.com/energydictionary/british_thermal_unit_(btu)__mbtu__
mmbtu.html [Accessed 08 Aug. 2018]
Enriquez, A. and Rojo, A. (2016a). Small–mid scale LNG: “45 years of experience in
Spain”. [online] Available at: https://gastechinsights.com/article/small-mid-scale-lng-
45-years-of-experience-in-spain [Accessed 15 June. 2018]
Enriquez, A. and Rojo, A. (2016b). Spain develops small-scale LNG: Part 2. [online]
Available at: https://www.naturalgasworld.com/spain-develops-small-scale-lng-part-
2-30302 [Accessed 03 July. 2018]
European Commission (2018). All weekly oil maps since 2010. [online] Available at:
http://ec.europa.eu/energy/maps/maps_weekly_oil_bulletin/map_library_Oil_prices&
taxation.pdf [Accessed 03 August. 2018]
Europe Maghreb Pipeline Limited (n.d.). Europe Maghreb Pipeline Limited. [online]
Available at: http://www.emplpipeline.com/en/europe-maghreb-pipeline-limited/
Eurostat (2018a). Complete energy balances - annual data [nrg_110a]. [online]

Available at:
http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=nrg_110a&lang=en
Eurostat (2018b). Cooling and heating degree days by country - monthly data
[nrg_chdd_m]. [online] Available at:
66
http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=nrg_chdd_m&lang=en
Eurostat (2018c). Lorries, by type of motor energy [road_eqs_lormot]. [online]

Available at:
http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=road_eqs_lormot&lang=e
n [Accessed 05 Aug. 2018]
Eurostat (2018d). Motor coaches, buses and trolley buses, by type of motor energy
[road_eqs_busmot]. [online] Available at:
http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=road_eqs_busmot&lang=
en [Accessed 05 Aug. 2018]
Eurostat (2018e). Road tractors by type of motor energy [road_eqs_roaene]. [online]

Available at:
http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=road_eqs_roaene&lang=e
n [Accessed 25 July. 2018]
Eurostat (2013). Glossary: Gigawatt hours (GWh). [online] Available at:

http://ec.europa.eu/eurostat/statistics-
explained/index.php/Glossary:Gigawatt_hours_(GWh) [Accessed 10 Aug. 2018]
Eurostat (n.d.). Energy statistics - cooling and heating degree days (nrg_chdd).
http://ec.europa.eu/eurostat/cache/metadata/en/nrg_chdd_esms.htm#contact15106
42469560 [Accessed 27 July. 2018]
Excel (n.d.). Excel Solver [computer programme]
Federal Energy Regulatory Commission (2018). Natural Gas Markets. [online]

Available at:
https://ferc.gov/market-oversight/mkt-gas/overview/archives.asp [Accessed 17 June.
2018]
Faaijf, A., Hoefnagelsa, R., de Jong, S., Junginger, M., Petterssond, K. and
Wetterlund, E. (2017). Cost optimization of biofuel production – The impact of
scale,integration, transport and supply chain conﬁgurations. Applied Energy, 195(1),
pp. 1055-1070.
Gas Eco Suministros (2017). Estaciones de GNV 2016/2017. [online] Available at:
http://gasecosuministros.com/es/ben-gnv-menu.html [Accessed 25 July. 2018]
Gas Infrastructure Europe (2015). GSE investment database. [online] Available at:
https://www.gie.eu/index.php/gie-publications/databases/gse-investment-database
[Accessed 05 August. 2018]
Gas Infrastructure Europe (2018a). LNG data. [online] Available at:

https://alsi.gie.eu/#/ [Accessed 14 June. 2018]
Gas Infrastructure Europe (2018b). SSLNG database final out. [online] Available at:
https://www.gie.eu/maps_data/downloads/2018/20180124_SSLNG_database_FINA
67
L_out.xlsx [Accessed 02 July. 2018]
Georgakaki, A., Kavalov, B. and Petric, H. (2009). Liquefied natural gas for Europe –
Some important issues for consideration. Report EUR 23818 EN. Luxemburg: Joint
Research Centre.
Global Energy Observatory (n.d.a). Barcelona LNG Terminal Spain. [online] Available
at: http://globalenergyobservatory.org/form.php?pid=43749 [Accessed 24 July. 2018]
Global Energy Observatory (n.d.b). Bilbao (BBG) LNG Terminal Spain. [online]
Available at: http://globalenergyobservatory.org/form.php?pid=43751 [Accessed 24
July. 2018]
Global Energy Observatory (n.d.c). Cartegana LNG Terminal Spain. [online] Available
Global Energy Observatory (n.d.d). El Musel (Gijon) LNG Terminal Spain. [online]
July. 2018]
Global Energy Observatory (n.d.e). Huelva LNG Terminal Spain. [online] Available at:
http://globalenergyobservatory.org/geoid/43750 [Accessed 24 July. 2018]
Global Energy Observatory (n.d.f). Renagosa Ferrol LNG Terminal Spain. [online]
July. 2018]
Global Energy Observatory (n.d.g). Sagunto LNG Terminal Spain. [online] Available
Gu, A., Lin, W. and Zhang, N. (2010). LNG (liquefied natural gas): A necessary part in
China’s future energy infrastructure. Energy, 35(11), pp. 4383-4391.
HAM (n.d.a.). About us. [online] Available at: https://www.ham.es/en/about-us/

HAM (n.d.b.). Liquid natural gaz. [online] Available at: http://www.ham.es/liquid-

natural-gaz/ [Accessed 30 July. 2018]
HAM (n.d.c). Satellite regasification plants. [online] Available at:

https://www.ham.es/en/lng-satellite-plants/ [Accessed 13 July. 2018]
HAM (2018). Where to refuel. [online] Available at: https://www.ham.es/en/lng-and-

cng-service-stations/ [Accessed 24 July. 2018]
Hansson, J. (2008). LNG as an alternative energy supply in Sweden. Lund: Svenskt

Gastekniskt Center: Rapport SGC 197.
Hopper, A. and Murray, D. (2017). An Analysis of the Operational Costs of Trucking:

2017 Update. Arlington: American Transportation Research Institute
68
Instituto Nacional de Estadistica (2018). Resident population by date, sex and age
group. [online] Available at: http://www.ine.es/jaxiT3/Datos.htm?t=10272 [Accessed
17 July. 2018]
International Energy Agency (2017). The future of trucks implications for energy and
the environment. [online] Available at:
https://www.iea.org/publications/freepublications/publication/TheFutureofTrucksImpli
cationsforEnergyandtheEnvironment.pdf [Accessed 03 Aug. 2018]
International group of Liquefied Natural Gas importers (2018). GIIGNL 2018 annual
report. [online] Available at:
https://giignl.org/sites/default/files/PUBLIC_AREA/Publications/rapportannuel-
2018pdf.pdf [Accessed 13 July. 2018]
JFE engineering corporation (n.d.). Energy plants. [online] Available at: http://www.jfe-
eng.co.jp/en/products/energy/energy_plant/ene03.html [Accessed 13 July. 2018]
Jokinen, R., Pettersson, F. and Saxen, H. (2015). An MILP model for optimization of
a small-scale LNG supply chain along a coastline. Applied Energy, 138, pp. 423-431.
Keller, G. (2014). Statistics for management and economics. 10th ed. Stamford:
Cengage Learning.
King & Spalding LLP (2018). LNG in Europe 2018: An Overview of Import Terminals
in Europe. [online] Available at:
https://www.kslaw.com/attachments/000/006/010/original/LNG_in_Europe_2018_-
_An_Overview_of_LNG_Import_Terminals_in_Europe.pdf?1530031152 [Accessed
02 July. 2018]
La Vanguarida (2018). Gas Natural Fenosa se convierte en Naturgy. [online] Available

at: https://www.lavanguardia.com/economia/20180627/45434603305/gas-natural-
fenosa-naturgy.html [Accessed 07 Aug. 2018]
LNG blue Corridors (n.d.). GNF-Lleida. [online] Available at:

http://lngbc.eu/node/185 [Accessed 24 July. 2018]
LNG World News (2018). First STS LNG bunkering operation for Spain’s Bilbao port.
https://www.lngworldnews.com/first-sts-lng-bunkering-operation-for-spains-bilbao-
port/ [Accessed 01 August. 2018]
Medgaz (n.d.). Partners. [online] Available at:

https://www.medgaz.com/medgaz/pages/socios-eng.htm# [Accessed 02 July. 2018]
Minitab Express (n.d.). [computer programme]
Monfort Transportes (n.d.) Contact. [online] Available at:

http://www.tmonfort.es/?lang=en#contacto [Accessed 24 July. 2018]
Natural Gas World (2016). Spain develops small-scale LNG: Part 2.

[online] Available at: https://www.naturalgasworld.com/spain-develops-small-scale-
lng-part-2-30302 [Accessed 20 July. 2018]
69
Neumann, A. and von Hirschhausen, C. (2008). Long-Term Contracts and Asset
Specificity Revisited: An Empirical Analysis of Producer–Importer Relations in the
Natural Gas Industry. Review of Industrial Organization, 32(2), pp.131-143.
Neumann, A. and Ruster, S. (2006). Corporate strategies along the LNG value added
chain - An empirical analysis of the determinants of vertical integration. Globalization
of natural gas markets. Working Papers WP-GG-17, German Institute for Economic
Research & Chair of Energy Economics and Public Sector Management: Dresden
University of Technology. [online] Available at: https://tu-
dresden.de/bu/wirtschaft/ee2/ressourcen/dateien/dateien/ordner_publikationen/wp_
gg_17_ruester_neumann_LNG_vertical_integration.pdf?lang=en [Accessed 07 July.
2018]
NGV journal (2014). Spain: HAM Group opens first LNG fueling station in Galicia.
[online] Available at: http://www.ngvjournal.com/s1-news/c4-stations/spain-ham-
group-opens-first-lng-fueling-station-in-galicia/ [Accessed 24 July. 2018]
Precio Gas (n.d.). Gas Natural Fenosa. [online] Available at:

https://preciogas.com/companias/gas-natural-fenosa#comercializadoras-y-
distribuidoras [Accessed 06 July. 2018]
Prentice, B. E. and Prokop, D. (2016). Concept of transportation Economics.

Singapore: World Scientific
Prontera, A. (2017). The new politics of energy security in the European Union and
beyond: States, markets, institutions. Oxfordshire: Routledge.
Redexis Gas (n.d.). Shareholders and structure of the group. [online] Available at:
http://www.redexisgas.es/en/about-redexis-gas/shareholders-and-structure-of-the-
group/ [Accessed 06 July. 2018]
Reed TMS Logistics (n.d.). Transit Times. [online]. Available at:

http://reedtms.com/documents/TransitTimes.pdf [Accessed on 01 August. 2018]
Registro-Empresas (n.d.). Gasogas Olaberria - Gas station in Olaberria. [online]

Available at: http://olaberria-vasconia.registro-empresas.es/gas-station/gasogas-
olaberria-olaberria/ [Accessed 25 July. 2018]
Reganosa (n.d.). Shareholders. [online] Available at:

http://www.reganosa.com/en/shareholders [Accessed 09 Aug. 2018]
Timera Energy (2013). Will European LNG reloads continue? [online] Available at:
https://timera-energy.com/will-european-lng-reloads-continue/ [Accessed 03 July.
2018]
Tourism in Spain (n.d.). Autonomous regions. [online] Available at:

https://www.spain.info/en/consultas/ciudades-y-pueblos/comunidades-
autonomas.html?tv=l&rpp=20 [Accessed 03 August. 2018]
Williamson, O. E. (1983). Credible commitments: Using hostages to support
70
exchange. The American Economic Review, 73 (4), pp. 519-540.
Williamson, O. E. (1985). The Economic Institutions of Capitalism. New York: Free

Pre.
Wood, D. A. (2007). LNG trade-conclusion: Spark spread trends allow analysis of LNG
consumers. Oil and Gas Journal, 105(38), pp. 72-76.
Yang C. T. (2014). An inventory model with both stock-dependent demand rate and
stock-dependent holding cost rate. International Journal of Production Economics,
155, pp. 214-221.
Zhu, Z. (2008). Natural gas prices, LNG transport costs, and the dynamics of LNG
imports. Energy Economics, 33(2), pp. 217-226.
Zhuravleva, P. (2009). The nature of LNG arbitrage: An analysis of the main barriers
to the growth of the global LNG arbitrage market. Oxford: Oxford Institute for Energy
Studies.
71
Annex A
1. MILP model in Excel Solver.
72
Annex B
1.Estimated LNG demands in the autonomous community level in
scenario 1.
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Galicia 117.9 107.9 105.7 91.8 87.8 84 85.1 79.7 88.9 93.2 104.7 114.4
Asturias 33.5 30 28.1 23.1 21.2 19.3 18.8 16.3 18.9 20.6 25 29.2
Castile and
Leon 42 38.1 36.9 31.5 29.8 28.2 28.3 25.9 29.2 30.9 35.4 39.5
Cantabria 19 17 15.9 13.1 12 10.9 10.6 9.2 10.7 11.6 14.1 16.5
Extremadura 53.7 49.2 48.3 42 40.2 38.5 39.1 36.7 40.9 42.8 48 52.3
Andalusia 410 376.1 369.8 322.6 309.8 297.8 302.6 285 317 331.2 369.9 401.6
Murcia 89.3 82.3 81.8 72.3 70 68 69.6 66.5 73.4 76.2 83.8 89.7
Castilla La
Mancha 98.2 89.8 87.8 76 72.6 69.4 70.2 65.6 73.3 76.9 86.6 94.8
Madrid 396.2 365.4 363.2 320.8 310.9 302 308.9 295 326 338.2 372.1 398.1
235.
Valencia 280.6 259.3 258.7 229.4 223 217.4 222.8 213.7 7 243.9 267.1 284.5
372.
Catalonia 453 417.8 415.2 366.8 355.4 345.2 353.2 337.2 7 386.6 425.4 455.2
Aragon 69.6 63.9 62.9 55 52.9 50.9 51.8 48.8 54.3 56.7 63.1 68.4
Basque
countries 60.6 57.8 61 57.5 58.3 59.4 62.6 63.4 68.2 68.6 70.6 70.2
Navarre 21 18.8 17.6 14.5 13.3 12.1 11.8 10.2 11.8 12.9 15.6 18.3
2144. 1973. 1952. 1716. 1657. 1603. 1635. 1553. 172 1790. 1981. 2132.
Total 5 4 9 4 1 1 5 2 1 3 5 7

scenario 2
Galicia 117.9 107.9 105.7 91.8 87.8 84 85.1 79.7 88.9 93.2 104.7 114.4
Asturias 33.5 30 28.1 23.1 21.2 19.3 18.8 16.3 18.9 20.6 25 29.2
Castile and
Leon 42 38.1 36.9 31.5 29.8 28.2 28.3 25.9 29.2 30.9 35.4 39.5
Cantabria 19 17 15.9 13.1 12 10.9 10.6 9.2 10.7 11.6 14.1 16.5
Extremadura 53.7 49.2 48.3 42 40.2 38.5 39.1 36.7 40.9 42.8 48 52.3
Andalusia 410 376.1 369.8 322.6 309.8 297.8 302.6 285 317 331.2 369.9 401.6
Murcia 89.3 82.3 81.8 72.3 70 68 69.6 66.5 73.4 76.2 83.8 89.7
Castilla La
Mancha 98.2 89.8 87.8 76 72.6 69.4 70.2 65.6 73.3 76.9 86.6 94.8
Madrid 396.2 365.4 363.2 320.8 310.9 302 308.9 295 326 338.2 372.1 398.1
235.
Valencia 280.6 259.3 258.7 229.4 223 217.4 222.8 213.7 7 243.9 267.1 284.5
372.
Catalonia 453 417.8 415.2 366.8 355.4 345.2 353.2 337.2 7 386.6 425.4 455.2
Aragon 69.6 63.9 62.9 55 52.9 50.9 51.8 48.8 54.3 56.7 63.1 68.4
Basque
countries 60.6 57.8 61 57.5 58.3 59.4 62.6 63.4 68.2 68.6 70.6 70.2
Navarre 21 18.8 17.6 14.5 13.3 12.1 11.8 10.2 11.8 12.9 15.6 18.3
2144. 1973. 1952. 1716. 1657. 1603. 1635. 1553. 172 1790. 1981. 2132.
Total 5 4 9 4 1 1 5 2 1 3 5 7

scenario 3.
Galicia 176.4 161.4 157 134.5 128.3 120.6 121.3 110.9 124.8 133 152.3 169.7
73
Asturias 58.6 53 50.2 41.5 38.6 34.9 34.4 29.7 34.3 37.7 45.4 53
Castile and
Leon 65.9 60.1 57.9 49.1 46.4 43.1 43.1 38.8 43.9 47.3 55 62.1
Cantabria 33.1 30 28.4 23.5 21.8 19.8 19.4 16.8 19.4 21.3 25.7 30
Extremadura 79.9 73.1 71.2 61.1 58.3 54.9 55.3 50.6 56.9 60.6 69.3 77
538.
Andalusia 602.6 552.4 9 463.5 443.4 418.2 421.8 387.8 435.2 462.4 526.9 583.9
113.
Murcia 125.3 115.3 5 98.6 95 90.6 91.9 85.7 95.6 100.7 113.2 123.8
Castilla La 131.
Mancha 147.8 135.2 4 112.4 107.1 100.4 101 92.1 103.7 110.7 127.1 141.8
503.
Madrid 556.1 511.8 7 437.8 421.8 402 407.9 380.4 424.2 447.1 502.5 549.5
352.
Valencia 387.5 357.1 5 307.6 297.2 284.2 289 270.8 301.3 316.7 354.3 385.6
575.
Catalonia 635.8 585.1 8 500.5 482.3 459.6 466.4 434.9 484.9 511.1 574.5 628.2
Aragon 101.7 93.2 91.1 78.4 75.1 70.9 71.6 66 74 78.5 89.3 98.8
Basque
countries 60.6 57.8 61 57.5 58.3 59.4 62.6 63.4 68.2 68.6 70.6 70.2
Navarre 36.7 33.2 31.4 26 24.2 21.9 21.5 18.6 21.5 23.6 28.5 33.2
1483. 1366. 137 1228. 1205. 1184. 1223. 1194. 1303. 1339. 1438. 1506.
Total 9 8 2 1 8 5 2 4 7 1 4 6
74
Annex C
1.Distribution of LNG demand in Scenario 1.1. (GWh)
Suppl
Regio lici uri and tabri madu alus ur La ad enc alon ag Countri var
n a as Leon a ra ia cia Mancha rid ia ia on es re
Barcel 0. 0. 0. 450
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .0 0.0 0.0 0.0
30
Carta 0. 89 9.
J gena 0 0.0 0.0 0.0 0.0 13.7 .3 38.0 0 0.0 0.0 0.0 0.0 0.0
a
Huelv 0. 396. 0. 0.
n- a 0 0.0 0.0 0.0 53.7 3 0 0.0 0 0.0 0.0 0.0 0.0 0.0
1
8 El 0. 33. 0. 87
Musel 0 5 42.0 19.0 0.0 0.0 0 0.0 .2 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 0. 280 16.
to 0 0.0 0.0 0.0 0.0 0.0 0 60.2 0 .6 3.0 2 0.0 0.0
0. 0. 0. 53. 21.
Bilbao 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 4 60.6 0
11
Mugar 7. 0. 0.
dos 9 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 417 32.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .8 2 0.0 0.0
36
Carta 0. 82 5.
F gena 0 0.0 0.0 0.0 0.0 0.0 .3 2.3 4 0.0 0.0 0.0 0.0 0.0
e
Huelv 0. 376. 0. 0.
b-
a 0 0.0 0.0 0.0 49.2 1 0 0.0 0 0.0 0.0 0.0 0.0 0.0
1
El 0. 30. 0. 0.
8
Musel 0 0 15.2 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 0. 259 13.
to 0 0.0 0.0 0.0 0.0 0.0 0 87.5 0 .3 0.0 2 0.0 0.0
0. 0. 0. 18. 18.
Bilbao 0 0.0 22.9 17.0 0.0 0.0 0 0.0 0 0.0 0.0 5 57.8 8
10
Mugar 7. 0. 0.
dos 9 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 415 34.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 8 0.0 0.0
36
Carta 0. 81 3.
M gena 0 0.0 0.0 0.0 0.0 0.0 .8 5.0 2 0.0 0.0 0.0 0.0 0.0
ar
Huelv 0. 369. 0. 0.
-
a 0 0.0 0.0 0.0 48.3 8 0 0.0 0 0.0 0.0 0.0 0.0 0.0
1
El 0. 28. 0. 0.
8
Musel 0 1 6.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 0. 258 18.
to 0 0.0 0.0 0.0 0.0 0.0 0 82.8 0 .7 0.0 5 0.0 0.0
0. 0. 0. 17.
Bilbao 0 0.0 30.9 15.9 0.0 0.0 0 0.0 0 0.0 0.0 9.6 61.0 6
10
Mugar 5. 0. 0.
dos 7 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 366 55.
A ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .8 0 0.0 0.0
pr 24
- Carta 0. 72 7.
1 gena 0 0.0 0.0 0.0 0.0 0.0 .3 0.0 8 0.0 0.0 0.0 0.0 0.0
8 Huelv 0. 322. 0. 0.
a 0 0.0 0.0 0.0 42.0 6 0 0.0 0 0.0 0.0 0.0 0.0 0.0
El 0. 23. 0. 0.
Musel 0 1 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 54 229
to 0 0.0 0.0 0.0 0.0 0.0 0 76.0 .6 .4 0.0 0.0 0.0 0.0
75
0. 0. 18 14.
Bilbao 0 0.0 31.5 13.1 0.0 0.0 0 0.0 .4 0.0 0.0 0.0 57.5 5
Mugar 91 0. 0.
dos .8 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 355 52.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .4 9 0.0 0.0
22
M Carta 0. 70 4.
a gena 0 0.0 0.0 0.0 0.0 0.0 .0 0.0 9 0.0 0.0 0.0 0.0 0.0
y- Huelv 0. 309. 0. 0.
1 a 0 0.0 0.0 0.0 40.2 8 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8 El 0. 21. 0. 0.
Musel 0 2 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 64 223
to 0 0.0 0.0 0.0 0.0 0.0 0 72.6 .4 .0 0.0 0.0 0.0 0.0
0. 0. 21 13.
Bilbao 0 0.0 29.8 12.0 0.0 0.0 0 0.0 .6 0.0 0.0 0.0 58.3 3
Mugar 87 0. 0.
dos .8 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 345 50.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 9 0.0 0.0
20
J Carta 0. 68 4.
u gena 0 0.0 0.0 0.0 0.0 0.0 .0 0.0 4 0.0 0.0 0.0 0.0 0.0
n- Huelv 0. 297. 0. 0.
1 a 0 0.0 0.0 0.0 38.5 8 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8 El 0. 19. 0. 0.
Musel 0 3 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 73 217
to 0 0.0 0.0 0.0 0.0 0.0 0 69.4 .2 .4 0.0 0.0 0.0 0.0
0. 0. 24 12.
Bilbao 0 0.0 28.2 10.9 0.0 0.0 0 0.0 .4 0.0 0.0 0.0 59.4 1
Mugar 84 0. 0.
dos .0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 353 51.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 8 0.0 0.0
22
J Carta 0. 69 0.
ul gena 0 0.0 0.0 0.0 0.0 0.0 .6 0.0 2 0.0 0.0 0.0 0.0 0.0
- Huelv 0. 302. 0. 0.
1 a 0 0.0 0.0 0.0 39.1 6 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8 El 0. 18. 0. 0.
Musel 0 8 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 67 222
to 0 0.0 0.0 0.0 0.0 0.0 0 70.2 .0 .8 0.0 0.0 0.0 0.0
0. 0. 21 11.
Bilbao 0 0.0 28.3 10.6 0.0 0.0 0 0.0 .7 0.0 0.0 0.0 62.6 8
Mugar 85 0. 0.
dos .1 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 337 48.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 8 0.0 0.0
A 18
u Carta 0. 66 8.
g- gena 0 0.0 0.0 0.0 0.0 0.0 .5 0.0 0 0.0 0.0 0.0 0.0 0.0
1 Huelv 0. 285. 0. 0.
8 a 0 0.0 0.0 0.0 36.7 0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
El 0. 16. 0. 0.
Musel 0 3 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 80 213
to 0 0.0 0.0 0.0 0.0 0.0 0 65.6 .7 .7 0.0 0.0 0.0 0.0
0. 0. 26 10.
Bilbao 0 0.0 25.9 9.2 0.0 0.0 0 0.0 .3 0.0 0.0 0.0 63.4 2
76
Mugar 79 0. 0.
dos .7 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 372 54.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .7 3 0.0 0.0
25
S Carta 0. 73 9.
e gena 0 0.0 0.0 0.0 0.0 0.0 .4 0.0 9 0.0 0.0 0.0 0.0 0.0
p- Huelv 0. 317. 0. 0.
1 a 0 0.0 0.0 0.0 40.9 0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8 El 0. 18. 0. 0.
Musel 0 9 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 51 235
to 0 0.0 0.0 0.0 0.0 0.0 0 73.3 .0 .7 0.0 0.0 0.0 0.0
0. 0. 15 11.
Bilbao 0 0.0 29.2 10.7 0.0 0.0 0 0.0 .1 0.0 0.0 0.0 68.2 8
Mugar 88 0. 0.
dos .9 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 386 56.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .6 7 0.0 0.0
28
O Carta 0. 76 8.
ct gena 0 0.0 0.0 0.0 0.0 0.0 .2 0.0 0 0.0 0.0 0.0 0.0 0.0
- Huelv 0. 331. 0. 0.
1 a 0 0.0 0.0 0.0 42.8 2 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8 El 0. 20. 0. 0.
Musel 0 6 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 39 243
to 0 0.0 0.0 0.0 0.0 0.0 0 76.9 .2 .9 0.0 0.0 0.0 0.0
0. 0. 11 12.
Bilbao 0 0.0 30.9 11.6 0.0 0.0 0 0.0 .0 0.0 0.0 0.0 68.6 9
Mugar 93 0. 0.
dos .2 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 425 24.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .4 6 0.0 0.0
36
Carta 0. 83 6.
N gena 0 0.0 0.0 0.0 0.0 0.0 .8 0.0 2 0.0 0.0 0.0 0.0 0.0
o
Huelv 0. 369. 0. 0.
v-
a 0 0.0 0.0 0.0 48.0 9 0 0.0 0 0.0 0.0 0.0 0.0 0.0
1
El 0. 25. 0. 5.
8
Musel 0 0 32.9 0.0 0.0 0.0 0 0.0 9 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 0. 267
to 0 0.0 0.0 0.0 0.0 0.0 0 86.6 0 .1 0.0 6.3 0.0 0.0
0. 0. 0. 32. 15.
Bilbao 0 0.0 2.5 14.1 0.0 0.0 0 0.0 0 0.0 0.0 2 70.6 6
10
Mugar 4. 0. 0.
dos 7 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 450
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .0 0.0 0.0 0.0
D 31
e Carta 0. 89 0.
c- gena 0 0.0 0.0 0.0 0.0 3.9 .7 46.4 0 0.0 0.0 0.0 0.0 0.0
1 Huelv 0. 397. 0. 0.
8 a 0 0.0 0.0 0.0 52.3 7 0 0.0 0 0.0 0.0 0.0 0.0 0.0
El 0. 29. 0. 88
Musel 0 2 39.5 16.5 0.0 0.0 0 0.0 .1 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 0. 284 21.
to 0 0.0 0.0 0.0 0.0 0.0 0 48.4 0 .5 5.2 9 0.0 0.0
0. 0. 0. 46. 18.
Bilbao 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 5 70.2 3
77
11
Mugar 4. 0. 0.
dos 4 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
2. Distribution of LNG demand in Scenario 1.2. (GWh)

Suppl
Barcel 0. 0. 0. 450
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .0 0.0 0.0 0.0
26
J Carta 0. 89 7.
a gena 0 0.0 0.0 0.0 0.0 13.7 .3 80.0 0 0.0 0.0 0.0 0.0 0.0
n-
Huelv 0. 396. 0. 0.
1
a 0 0.0 0.0 0.0 53.7 3 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8
Sagun 0. 0. 0. 280 58.
to 0 0.0 0.0 0.0 0.0 0.0 0 18.2 0 .6 3.0 2 0.0 0.0
0. 0. 0. 11. 21.
Bilbao 0 0.0 42.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 4 60.6 0
11 12
Mugar 7. 33. 0. 9.
dos 9 5 0.0 19.0 0.0 0.0 0 0.0 2 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 417 32.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .8 2 0.0 0.0
36
F
Carta 0. 82 5.
e
gena 0 0.0 0.0 0.0 0.0 0.0 .3 2.3 4 0.0 0.0 0.0 0.0 0.0
b-
Huelv 0. 376. 0. 0.
1
a 0 0.0 0.0 0.0 49.2 1 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8
Sagun 0. 0. 0. 259 13.
to 0 0.0 0.0 0.0 0.0 0.0 0 87.5 0 .3 0.0 2 0.0 0.0
0. 0. 0. 18. 18.
Bilbao 0 0.0 38.1 1.8 0.0 0.0 0 0.0 0 0.0 0.0 5 57.8 8
10
Mugar 7. 30. 0. 0.
dos 9 0 0.0 15.2 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 415 34.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 8 0.0 0.0
36
M Carta 0. 81 3.
ar
gena 0 0.0 0.0 0.0 0.0 0.0 .8 5.0 2 0.0 0.0 0.0 0.0 0.0
-
Huelv 0. 369. 0. 0.
1
a 0 0.0 0.0 0.0 48.3 8 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8
Sagun 0. 0. 0. 258 18.
to 0 0.0 0.0 0.0 0.0 0.0 0 82.8 0 .7 0.0 5 0.0 0.0
0. 0. 0. 17.
Bilbao 0 0.0 36.9 9.9 0.0 0.0 0 0.0 0 0.0 0.0 9.6 61.0 6
10
Mugar 5. 28. 0. 0.
dos 7 1 0.0 6.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 366 55.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .8 0 0.0 0.0
A 24
pr Carta 0. 72 7.
- gena 0 0.0 0.0 0.0 0.0 0.0 .3 0.0 8 0.0 0.0 0.0 0.0 0.0
1 Huelv 0. 322. 0. 0.
8 a 0 0.0 0.0 0.0 42.0 6 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 54 229
to 0 0.0 0.0 0.0 0.0 0.0 0 76.0 .6 .4 0.0 0.0 0.0 0.0
0. 0. 18 14.
Bilbao 0 0.0 31.5 13.1 0.0 0.0 0 0.0 .4 0.0 0.0 0.0 57.5 5
Mugar 91 23. 0. 0.
dos .8 1 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
78
Suppl
Barcel 0. 0. 0. 355 52.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .4 9 0.0 0.0
M 22
a Carta 0. 70 4.
y- gena 0 0.0 0.0 0.0 0.0 0.0 .0 0.0 9 0.0 0.0 0.0 0.0 0.0
1 Huelv 0. 309. 0. 0.
8 a 0 0.0 0.0 0.0 40.2 8 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 64 223
to 0 0.0 0.0 0.0 0.0 0.0 0 72.6 .4 .0 0.0 0.0 0.0 0.0
0. 0. 21 13.
Bilbao 0 0.0 29.8 12.0 0.0 0.0 0 0.0 .6 0.0 0.0 0.0 58.3 3
Mugar 87 21. 0. 0.
dos .8 2 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 345 50.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 9 0.0 0.0
J 20
u Carta 0. 68 4.
n- gena 0 0.0 0.0 0.0 0.0 0.0 .0 0.0 4 0.0 0.0 0.0 0.0 0.0
1 Huelv 0. 297. 0. 0.
8 a 0 0.0 0.0 0.0 38.5 8 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 73 217
to 0 0.0 0.0 0.0 0.0 0.0 0 69.4 .2 .4 0.0 0.0 0.0 0.0
0. 0. 24 12.
Bilbao 0 0.0 28.2 10.9 0.0 0.0 0 0.0 .4 0.0 0.0 0.0 59.4 1
Mugar 84 19. 0. 0.
dos .0 3 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 353 51.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 8 0.0 0.0
J 22
ul Carta 0. 69 0.
- gena 0 0.0 0.0 0.0 0.0 0.0 .6 0.0 2 0.0 0.0 0.0 0.0 0.0
1 Huelv 0. 302. 0. 0.
8 a 0 0.0 0.0 0.0 39.1 6 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 67 222
to 0 0.0 0.0 0.0 0.0 0.0 0 70.2 .0 .8 0.0 0.0 0.0 0.0
0. 0. 21 11.
Bilbao 0 0.0 28.3 10.6 0.0 0.0 0 0.0 .7 0.0 0.0 0.0 62.6 8
Mugar 85 18. 0. 0.
dos .1 8 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 337 48.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 8 0.0 0.0
A 18
u Carta 0. 66 8.
g- gena 0 0.0 0.0 0.0 0.0 0.0 .5 0.0 0 0.0 0.0 0.0 0.0 0.0
1 Huelv 0. 285. 0. 0.
8 a 0 0.0 0.0 0.0 36.7 0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 80 213
to 0 0.0 0.0 0.0 0.0 0.0 0 65.6 .7 .7 0.0 0.0 0.0 0.0
0. 0. 26 10.
Bilbao 0 0.0 25.9 9.2 0.0 0.0 0 0.0 .3 0.0 0.0 0.0 63.4 2
Mugar 79 16. 0. 0.
dos .7 3 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
S n a as Leon a ra ia cia Mancha rid ia ia on es re
e Barcel 0. 0. 0. 372 54.
p- ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .7 3 0.0 0.0
1 25
8 Carta 0. 73 9.
gena 0 0.0 0.0 0.0 0.0 0.0 .4 0.0 9 0.0 0.0 0.0 0.0 0.0
Huelv 0. 317. 0. 0.
a 0 0.0 0.0 0.0 40.9 0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
79
Sagun 0. 0. 51 235
to 0 0.0 0.0 0.0 0.0 0.0 0 73.3 .0 .7 0.0 0.0 0.0 0.0
0. 0. 15 11.
Bilbao 0 0.0 29.2 10.7 0.0 0.0 0 0.0 .1 0.0 0.0 0.0 68.2 8
Mugar 88 18. 0. 0.
dos .9 9 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 386 56.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .6 7 0.0 0.0
O 28
ct Carta 0. 76 8.
- gena 0 0.0 0.0 0.0 0.0 0.0 .2 0.0 0 0.0 0.0 0.0 0.0 0.0
1 Huelv 0. 331. 0. 0.
8 a 0 0.0 0.0 0.0 42.8 2 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 39 243
to 0 0.0 0.0 0.0 0.0 0.0 0 76.9 .2 .9 0.0 0.0 0.0 0.0
0. 0. 11 12.
Bilbao 0 0.0 30.9 11.6 0.0 0.0 0 0.0 .0 0.0 0.0 0.0 68.6 9
Mugar 93 20. 0. 0.
dos .2 6 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 425 24.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .4 6 0.0 0.0
34
N
Carta 0. 83 7.
o gena 0 0.0 0.0 0.0 0.0 0.0 .8 18.8 4 0.0 0.0 0.0 0.0 0.0
v-
Huelv 0. 369. 0. 0.
1
a 0 0.0 0.0 0.0 48.0 9 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8
Sagun 0. 0. 0. 267 25.
to 0 0.0 0.0 0.0 0.0 0.0 0 67.8 0 .1 0.0 1 0.0 0.0
0. 0. 0. 13. 15.
Bilbao 0 0.0 35.4 0.0 0.0 0.0 0 0.0 0 0.0 0.0 4 70.6 6
10
Mugar 4. 25. 0. 24
dos 7 0 0.0 14.1 0.0 0.0 0 0.0 .7 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 450
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .0 0.0 0.0 0.0
27
D Carta 0. 89 0.
e gena 0 0.0 0.0 0.0 0.0 3.9 .7 85.9 5 0.0 0.0 0.0 0.0 0.0
c-
Huelv 0. 397. 0. 0.
1
a 0 0.0 0.0 0.0 52.3 7 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8
Sagun 0. 0. 0. 284 61.
to 0 0.0 0.0 0.0 0.0 0.0 0 8.9 0 .5 5.2 4 0.0 0.0
0. 0. 0. 18.
Bilbao 0 0.0 39.5 0.0 0.0 0.0 0 0.0 0 0.0 0.0 7.0 70.2 3
11 12
Mugar 4. 29. 0. 7.
dos 4 2 0.0 16.5 0.0 0.0 0 0.0 6 0.0 0.0 0.0 0.0 0.0
3. Distribution of LNG demand in Scenario 2.1. (GWh)

Suppl
Barcel 0. 0. 0. 322 46.
J ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 7 0.0 0.0
a 15
n- Carta 0. 63 7.
1 gena 0 0.0 0.0 0.0 0.0 0.0 .5 0.0 6 0.0 0.0 0.0 0.0 0.0
8 Huelv 0. 272. 0. 0.
a 0 0.0 0.0 0.0 35.0 3 0 0.0 0 0.0 0.0 0.0 0.0 0.0
El 0. 15. 0. 0.
Musel 0 5 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 93 204
to 0 0.0 0.0 0.0 0.0 0.0 0 62.7 .1 .2 0.0 0.0 0.0 0.0
80
0. 0. 31
Bilbao 0 0.0 24.8 8.8 0.0 0.0 0 0.0 .1 0.0 0.0 0.0 60.6 9.7
Mugar 76 0. 0.
dos .1 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 297 42.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .7 8 0.0 0.0
10
Carta 0. 58 8.
F gena 0 0.0 0.0 0.0 0.0 0.0 .7 0.0 0 0.0 0.0 0.0 0.0 0.0
e
Huelv 0. 249. 0. 0.
b-
a 0 0.0 0.0 0.0 32.0 6 0 0.0 0 0.0 0.0 0.0 0.0 0.0
1
El 0. 13. 0. 0.
8
Musel 0 5 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
11
Sagun 0. 0. 3. 189
to 0 0.0 0.0 0.0 0.0 0.0 0 57.2 7 .1 0.0 0.0 0.0 0.0
0. 0. 38
Bilbao 0 0.0 22.4 7.6 0.0 0.0 0 0.0 .7 0.0 0.0 0.0 57.8 8.5
Mugar 69 0. 0.
dos .5 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 300 42.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 7 0.0 0.0
11
Carta 0. 59 3.
M gena 0 0.0 0.0 0.0 0.0 0.0 .2 0.0 1 0.0 0.0 0.0 0.0 0.0
ar
Huelv 0. 248. 0. 0.
-
a 0 0.0 0.0 0.0 31.8 7 0 0.0 0 0.0 0.0 0.0 0.0 0.0
1
El 0. 12. 0. 0.
8
Musel 0 3 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
11
Sagun 0. 0. 2. 191
to 0 0.0 0.0 0.0 0.0 0.0 0 56.5 0 .5 0.0 0.0 0.0 0.0
0. 0. 37
Bilbao 0 0.0 21.8 7.0 0.0 0.0 0 0.0 .5 0.0 0.0 0.0 61.0 7.7
Mugar 68 0. 0.
dos .9 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 270 38.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .1 0 0.0 0.0
Carta 0. 53 52
A gena 0 0.0 0.0 0.0 0.0 0.0 .2 0.0 .2 0.0 0.0 0.0 0.0 0.0
pr Huelv 0. 220. 0. 0.
- a 0 0.0 0.0 0.0 28.1 8 0 0.0 0 0.0 0.0 0.0 0.0 0.0
1 El 0. 0. 0.
8 Musel 0 9.9 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
13
Sagun 0. 0. 7. 172
to 0 0.0 0.0 0.0 0.0 0.0 0 49.8 3 .9 0.0 0.0 0.0 0.0
0. 0. 46
Bilbao 0 0.0 18.9 5.6 0.0 0.0 0 0.0 .8 0.0 0.0 0.0 57.5 6.2
Mugar 60 0. 0.
dos .9 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 266 37.
M ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .1 2 0.0 0.0
a Carta 0. 52 43
y- gena 0 0.0 0.0 0.0 0.0 0.0 .4 0.0 .9 0.0 0.0 0.0 0.0 0.0
1 Huelv 0. 215. 0. 0.
8 a 0 0.0 0.0 0.0 27.4 7 0 0.0 0 0.0 0.0 0.0 0.0 0.0
El 0. 0. 0.
Musel 0 8.9 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
14
Sagun 0. 0. 0. 170
to 0 0.0 0.0 0.0 0.0 0.0 0 48.4 8 .8 0.0 0.0 0.0 0.0
81
0. 0. 48
Bilbao 0 0.0 18.1 5.0 0.0 0.0 0 0.0 .0 0.0 0.0 0.0 58.3 5.6
Mugar 59 0. 0.
dos .2 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 262 36.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .4 4 0.0 0.0
Carta 0. 51 36
J gena 0 0.0 0.0 0.0 0.0 0.0 .7 0.0 .4 0.0 0.0 0.0 0.0 0.0
u Huelv 0. 210. 0. 0.
n- a 0 0.0 0.0 0.0 26.7 5 0 0.0 0 0.0 0.0 0.0 0.0 0.0
1 El 0. 0. 0.
8 Musel 0 7.9 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
14
Sagun 0. 0. 4. 168
to 0 0.0 0.0 0.0 0.0 0.0 0 46.9 2 .9 0.0 0.0 0.0 0.0
0. 0. 48
Bilbao 0 0.0 17.3 4.5 0.0 0.0 0 0.0 .9 0.0 0.0 0.0 59.4 4.9
Mugar 57 0. 0.
dos .5 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 271 37.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .5 4 0.0 0.0
Carta 0. 53 55
J gena 0 0.0 0.0 0.0 0.0 0.0 .5 0.0 .0 0.0 0.0 0.0 0.0 0.0
ul Huelv 0. 216. 0. 0.
- a 0 0.0 0.0 0.0 27.4 6 0 0.0 0 0.0 0.0 0.0 0.0 0.0
1 El 0. 0. 0.
8 Musel 0 7.6 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
13
Sagun 0. 0. 6. 175
to 0 0.0 0.0 0.0 0.0 0.0 0 48.1 8 .1 0.0 0.0 0.0 0.0
0. 0. 45
Bilbao 0 0.0 17.6 4.3 0.0 0.0 0 0.0 .7 0.0 0.0 0.0 62.6 4.8
Mugar 59 0. 0.
dos .0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 266 36.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 4 0.0 0.0
Carta 0. 52 44
A gena 0 0.0 0.0 0.0 0.0 0.0 .5 0.0 .2 0.0 0.0 0.0 0.0 0.0
u Huelv 0. 210. 0. 0.
g- a 0 0.0 0.0 0.0 26.5 2 0 0.0 0 0.0 0.0 0.0 0.0 0.0
1 El 0. 0. 0.
8 Musel 0 6.5 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
14
Sagun 0. 0. 1. 172
to 0 0.0 0.0 0.0 0.0 0.0 0 46.3 5 .2 0.0 0.0 0.0 0.0
0. 0. 47
Bilbao 0 0.0 16.6 3.7 0.0 0.0 0 0.0 .2 0.0 0.0 0.0 63.4 4.1
Mugar 57 0. 0.
dos .0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 290 39.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .1 8 0.0 0.0
S Carta 0. 57 92
e gena 0 0.0 0.0 0.0 0.0 0.0 .2 0.0 .7 0.0 0.0 0.0 0.0 0.0
p- Huelv 0. 230. 0. 0.
1 a 0 0.0 0.0 0.0 29.0 0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8 El 0. 0. 0.
Musel 0 7.5 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
12
Sagun 0. 0. 1. 187
to 0 0.0 0.0 0.0 0.0 0.0 0 50.8 7 .5 0.0 0.0 0.0 0.0
0. 0. 39
Bilbao 0 0.0 18.4 4.3 0.0 0.0 0 0.0 .4 0.0 0.0 0.0 68.2 4.7
82
Mugar 62 0. 0.
dos .5 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 297 41.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .3 0 0.0 0.0
10
Carta 0. 58 7.
O gena 0 0.0 0.0 0.0 0.0 0.0 .6 0.0 0 0.0 0.0 0.0 0.0 0.0
ct
Huelv 0. 237. 0. 0.
-
a 0 0.0 0.0 0.0 30.0 1 0 0.0 0 0.0 0.0 0.0 0.0 0.0
1
El 0. 0. 0.
8
Musel 0 8.3 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
11
Sagun 0. 0. 5. 191
to 0 0.0 0.0 0.0 0.0 0.0 0 52.6 7 .7 0.0 0.0 0.0 0.0
0. 0. 37
Bilbao 0 0.0 19.2 4.7 0.0 0.0 0 0.0 .3 0.0 0.0 0.0 68.6 5.2
Mugar 64 0. 0.
dos .6 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 317 44.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .9 3 0.0 0.0
14
N Carta 0. 62 8.
o gena 0 0.0 0.0 0.0 0.0 0.0 .6 0.0 9 0.0 0.0 0.0 0.0 0.0
v- Huelv 0. 256. 0. 0.
1 a 0 0.0 0.0 0.0 32.6 6 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8 El 0. 10. 0. 0.
Musel 0 2 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 98 204
to 0 0.0 0.0 0.0 0.0 0.0 0 57.4 .3 .3 0.0 0.0 0.0 0.0
0. 0. 30
Bilbao 0 0.0 21.3 5.8 0.0 0.0 0 0.0 .9 0.0 0.0 0.0 70.6 6.4
Mugar 70 0. 0.
dos .3 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 331 46.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 7 0.0 0.0
17
D Carta 0. 65 5.
e gena 0 0.0 0.0 0.0 0.0 0.0 .3 0.0 8 0.0 0.0 0.0 0.0 0.0
c- Huelv 0. 271. 0. 0.
1 a 0 0.0 0.0 0.0 34.6 1 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8 El 0. 12. 0. 0.
Musel 0 2 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 86 212
to 0 0.0 0.0 0.0 0.0 0.0 0 61.2 .8 .0 0.0 0.0 0.0 0.0
0. 0. 27
Bilbao 0 0.0 23.2 6.9 0.0 0.0 0 0.0 .1 0.0 0.0 0.0 70.2 7.6
Mugar 74 0. 0.
dos .7 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
4.Distribution of LNG demand in Scenario 2.2. (GWh)

Suppl
J Barcel 0. 0. 0. 322 46.
a ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 7 0.0 0.0
n- 15
1 Carta 0. 63 7.
8 gena 0 0.0 0.0 0.0 0.0 0.0 .5 0.0 6 0.0 0.0 0.0 0.0 0.0
Huelv 0. 272. 0. 0.
a 0 0.0 0.0 0.0 35.0 3 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 93 204
to 0 0.0 0.0 0.0 0.0 0.0 0 62.7 .1 .2 0.0 0.0 0.0 0.0
83
0. 0. 31
Bilbao 0 0.0 24.8 8.8 0.0 0.0 0 0.0 .1 0.0 0.0 0.0 60.6 9.7
Mugar 76 15. 0. 0.
dos .1 5 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 297 42.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .7 8 0.0 0.0
10
F
Carta 0. 58 8.
e gena 0 0.0 0.0 0.0 0.0 0.0 .7 0.0 0 0.0 0.0 0.0 0.0 0.0
b-
Huelv 0. 249. 0. 0.
1
a 0 0.0 0.0 0.0 32.0 6 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8
11
Sagun 0. 0. 3. 189
to 0 0.0 0.0 0.0 0.0 0.0 0 57.2 7 .1 0.0 0.0 0.0 0.0
0. 0. 38
Bilbao 0 0.0 22.4 7.6 0.0 0.0 0 0.0 .7 0.0 0.0 0.0 57.8 8.5
Mugar 69 13. 0. 0.
dos .5 5 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 300 42.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 7 0.0 0.0
11
M
Carta 0. 59 3.
ar
gena 0 0.0 0.0 0.0 0.0 0.0 .2 0.0 1 0.0 0.0 0.0 0.0 0.0
-
Huelv 0. 248. 0. 0.
1
a 0 0.0 0.0 0.0 31.8 7 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8
11
Sagun 0. 0. 2. 191
to 0 0.0 0.0 0.0 0.0 0.0 0 56.5 0 .5 0.0 0.0 0.0 0.0
0. 0. 37
Bilbao 0 0.0 21.8 7.0 0.0 0.0 0 0.0 .5 0.0 0.0 0.0 61.0 7.7
Mugar 68 12. 0. 0.
dos .9 3 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 270 38.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .1 0 0.0 0.0
A Carta 0. 53 52
pr gena 0 0.0 0.0 0.0 0.0 0.0 .2 0.0 .2 0.0 0.0 0.0 0.0 0.0
- Huelv 0. 220. 0. 0.
1 a 0 0.0 0.0 0.0 28.1 8 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8 13
Sagun 0. 0. 7. 172
to 0 0.0 0.0 0.0 0.0 0.0 0 49.8 3 .9 0.0 0.0 0.0 0.0
0. 0. 46
Bilbao 0 0.0 18.9 5.6 0.0 0.0 0 0.0 .8 0.0 0.0 0.0 57.5 6.2
Mugar 60 0. 0.
dos .9 9.9 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 266 37.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .1 2 0.0 0.0
M Carta 0. 52 43
a gena 0 0.0 0.0 0.0 0.0 0.0 .4 0.0 .9 0.0 0.0 0.0 0.0 0.0
y- Huelv 0. 215. 0. 0.
1 a 0 0.0 0.0 0.0 27.4 7 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8 14
Sagun 0. 0. 0. 170
to 0 0.0 0.0 0.0 0.0 0.0 0 48.4 8 .8 0.0 0.0 0.0 0.0
0. 0. 48
Bilbao 0 0.0 18.1 5.0 0.0 0.0 0 0.0 .0 0.0 0.0 0.0 58.3 5.6
Mugar 59 0. 0.
dos .2 8.9 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
J
u
n-
84
1 Barcel 0. 0. 0. 262 36.
8 ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .4 4 0.0 0.0
Carta 0. 51 36
gena 0 0.0 0.0 0.0 0.0 0.0 .7 0.0 .4 0.0 0.0 0.0 0.0 0.0
Huelv 0. 210. 0. 0.
a 0 0.0 0.0 0.0 26.7 5 0 0.0 0 0.0 0.0 0.0 0.0 0.0
14
Sagun 0. 0. 4. 168
to 0 0.0 0.0 0.0 0.0 0.0 0 46.9 2 .9 0.0 0.0 0.0 0.0
0. 0. 48
Bilbao 0 0.0 17.3 4.5 0.0 0.0 0 0.0 .9 0.0 0.0 0.0 59.4 4.9
Mugar 57 0. 0.
dos .5 7.9 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 271 37.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .5 4 0.0 0.0
J Carta 0. 53 55
ul gena 0 0.0 0.0 0.0 0.0 0.0 .5 0.0 .0 0.0 0.0 0.0 0.0 0.0
- Huelv 0. 216. 0. 0.
1 a 0 0.0 0.0 0.0 27.4 6 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8 13
Sagun 0. 0. 6. 175
to 0 0.0 0.0 0.0 0.0 0.0 0 48.1 8 .1 0.0 0.0 0.0 0.0
0. 0. 45
Bilbao 0 0.0 17.6 4.3 0.0 0.0 0 0.0 .7 0.0 0.0 0.0 62.6 4.8
Mugar 59 0. 0.
dos .0 7.6 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 266 36.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 4 0.0 0.0
A Carta 0. 52 44
u gena 0 0.0 0.0 0.0 0.0 0.0 .5 0.0 .2 0.0 0.0 0.0 0.0 0.0
g- Huelv 0. 210. 0. 0.
1 a 0 0.0 0.0 0.0 26.5 2 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8 14
Sagun 0. 0. 1. 172
to 0 0.0 0.0 0.0 0.0 0.0 0 46.3 5 .2 0.0 0.0 0.0 0.0
0. 0. 47
Bilbao 0 0.0 16.6 3.7 0.0 0.0 0 0.0 .2 0.0 0.0 0.0 63.4 4.1
Mugar 57 0. 0.
dos .0 6.5 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 290 39.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .1 8 0.0 0.0
S Carta 0. 57 92
e gena 0 0.0 0.0 0.0 0.0 0.0 .2 0.0 .7 0.0 0.0 0.0 0.0 0.0
p- Huelv 0. 230. 0. 0.
1 a 0 0.0 0.0 0.0 29.0 0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
8 12
Sagun 0. 0. 1. 187
to 0 0.0 0.0 0.0 0.0 0.0 0 50.8 7 .5 0.0 0.0 0.0 0.0
0. 0. 39
Bilbao 0 0.0 18.4 4.3 0.0 0.0 0 0.0 .4 0.0 0.0 0.0 68.2 4.7
Mugar 62 0. 0.
dos .5 7.5 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 297 41.
O
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .3 0 0.0 0.0
ct
10
-
Carta 0. 58 7.
1
gena 0 0.0 0.0 0.0 0.0 0.0 .6 0.0 0 0.0 0.0 0.0 0.0 0.0
8
Huelv 0. 237. 0. 0.
a 0 0.0 0.0 0.0 30.0 1 0 0.0 0 0.0 0.0 0.0 0.0 0.0
11
Sagun 0. 0. 5. 191
to 0 0.0 0.0 0.0 0.0 0.0 0 52.6 7 .7 0.0 0.0 0.0 0.0
85
0. 0. 37
Bilbao 0 0.0 19.2 4.7 0.0 0.0 0 0.0 .3 0.0 0.0 0.0 68.6 5.2
Mugar 64 0. 0.
dos .6 8.3 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 317 44.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .9 3 0.0 0.0
N 14
o Carta 0. 62 8.
v- gena 0 0.0 0.0 0.0 0.0 0.0 .6 0.0 9 0.0 0.0 0.0 0.0 0.0
1 Huelv 0. 256. 0. 0.
8 a 0 0.0 0.0 0.0 32.6 6 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 98 204
to 0 0.0 0.0 0.0 0.0 0.0 0 57.4 .3 .3 0.0 0.0 0.0 0.0
0. 0. 30
Bilbao 0 0.0 21.3 5.8 0.0 0.0 0 0.0 .9 0.0 0.0 0.0 70.6 6.4
Mugar 70 10. 0. 0.
dos .3 2 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Suppl
Barcel 0. 0. 0. 331 46.
ona 0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 .2 7 0.0 0.0
D 17
e Carta 0. 65 5.
c- gena 0 0.0 0.0 0.0 0.0 0.0 .3 0.0 8 0.0 0.0 0.0 0.0 0.0
1 Huelv 0. 271. 0. 0.
8 a 0 0.0 0.0 0.0 34.6 1 0 0.0 0 0.0 0.0 0.0 0.0 0.0
Sagun 0. 0. 86 212
to 0 0.0 0.0 0.0 0.0 0.0 0 61.2 .8 .0 0.0 0.0 0.0 0.0
0. 0. 27
Bilbao 0 0.0 23.2 6.9 0.0 0.0 0 0.0 .1 0.0 0.0 0.0 70.2 7.6
Mugar 74 12. 0. 0.
dos .7 2 0.0 0.0 0.0 0.0 0 0.0 0 0.0 0.0 0.0 0.0 0.0
86

Ahn Suhong PDF

Uploaded by

Copyright:

Available Formats

Ahn Suhong PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ahn Suhong PDF

Uploaded by

Copyright:

Available Formats

Erasmus University Rotterdam

MSc in Maritime Economics and Logistics

Analysis of Natural Gas Supply Chain and

copyright © Suhong Ahn

I appreciate Dr. Zuidwijk for valuable feedback for my thesis.

Abstract ....................................................................................................................................................................... iii

Table of Contents ................................................................................................................................................... iv

List of tables ............................................................................................................................................................ vii

List of figures ......................................................................................................................................................... viii

List of Abbreviation ............................................................................................................................................... x

Chapter 1 Introduction ......................................................................................................................................... 1

1.1 The topic .......................................................................................................................................................... 1

1.2. Problem Statement ................................................................................................................................... 1

1.3 Background and Justification ............................................................................................................. 2

1.4 Deficiencies in the Evidence ................................................................................................................ 2

1.6 Research Objective ................................................................................................................................... 2

1.7 Research Questions ................................................................................................................................. 2

1.8 Methodology and Research Design ......................................................................................... 3

1.9 Thesis Structure .......................................................................................................................................... 4

Chapter 2 Literature Review ............................................................................................................................. 5

2.2 Vertical Integration .................................................................................................................................... 6

2.3 Optimisation of Small-Scale Supply Chain .................................................................................. 7

3.1 Trade of Natural Gas and Liquefied Natural Gas ...................................................................... 9

3.1.1 Import of Natural Gas and Liquefied Natural Gas ......................................................... 10

3.1.2 Export of Natural Gas and Liquefied Natural Gas ........................................................ 13

3.3 Large-Scale Infrastructure .................................................................................................................. 17

3.3.1 Liquefied Natural Gas Import Terminals ............................................................................ 18

3.3.2 Natural Gas Pipelines ................................................................................................................... 21

3.4 Small-Scale Infrastructure .................................................................................................................. 23

3.4.1 Liquefied Natural Gas Satellite Terminal ........................................................................... 23

3.4.2 Liquefied Natural Gas Fuel Loading Road Station ....................................................... 25

3.4.3 Liquefied Natural Gas Fuel Loading Ship Station ........................................................ 27

3.4.4 Liquefied Natural Gas Bunker Ships .................................................................................... 28

3.4.5 Modes of Small-Scale Liquefied Natural Gas transportation ................................. 28

3.4.5.1 Trucks ................................................................................................................................................ 28

3.4.5.2 Ships .................................................................................................................................................. 29

Chapter 4 Methodology and Data Collection ........................................................................................ 32

4.1 Network Optimisation Model ............................................................................................................. 34

4.1.1 Players in the Model ...................................................................................................................... 35

4.1.2 Assumption of the Model ............................................................................................................ 36

4.1.3 Scenario Description .................................................................................................................... 36

4.1.4 Description of Optimisation Model ....................................................................................... 36

4.2 Regression Model with Seasonality for Forecasting ........................................................... 38

4.2.1 Analysis of the relationships of Pearson correlation and p-value ...................... 41

4.2.2 Analysis of the regression models with P-value, R-squared .................................. 41

4.3 Autoregressive Model for Forecasting ........................................................................................ 44

4.4 Data Collection.......................................................................................................................................... 44

4.4.1 Demand for Liquefied Natural Gas Direct Consumption .......................................... 44

4.4.2 Distribution of the Demand in the Autonomous Community Scale. ................... 47

4.4.4 Lot Size of the Trucks ................................................................................................................... 48

4.4.5 Capacity of LNG Transportation by Trucks in the LNG terminals ....................... 48

4.4.6 Transportation Costs .................................................................................................................... 48

4.4.8 Pipeline Holding costs ................................................................................................................. 50

Chapter 5 Simulations and Results ........................................................................................................... 53

5.1 Result of Scenario 1. Full Capacity ............................................................................................... 53

5.1.1 Allocation of LNG ............................................................................................................................ 53

5.1.2 Excess capacity ............................................................................................................................... 55

5.1.3 Total costs .......................................................................................................................................... 55

5.2 Result of Scenario 2. Abundant Capacity .................................................................................. 56