OTC 19738

Distributed Compressed Natural Gas Sea Transport

Michael Nikolaou and Michael J. Economides, SPE, University of Houston; Xiuli Wang, SPE, XGas; and
Matteo Marongiu-Porcu, SPE, Texas A&M University
Copyright 2009, Offshore Technology Conference

This paper was prepared for presentation at the 2009 Offshore Technology Conference held in Houston, Texas, USA, 4–7 May 2009.

This paper was selected for presentation by an OTC program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Offshore Technology Conference and are subject to correction by the author(s). The material does not necessarily reflect any position of the Offshore Technology Conference, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Offshore Technology Conference is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of OTC copyright.

Abstract
While liquefied natural gas (LNG) is the dominant technology for marine transport of natural gas, compressed natural gas
(CNG) is an economically attractive alternative for transportation of relatively small amounts of natural gas over short
distances (up to 6 Bcm/year transported over 2000 km, respectively). Because the main capital expenditure in a CNG project
is on marine transport vessels, careful design of CNG transport fleets and compatible distribution schedules is important.
Design through solution of an all-encompassing optimization problem would be desirable but complicated and intuitive
unappealing. In this work, a structured optimization framework is applied to potential marine CNG transport from a source
in the Trinidad/Venezuela area to island countries in the Caribbean. Two general patterns emerge in the results: A "hub-and-
spoke" pattern for servicing of the largest four consumption markets (Puerto Rico, Dominican Republic, Jamaica, and The
Bahamas) and a cyclical "milk-run" pattern for the remaining (smaller) markets.

Introduction
The role of natural gas as a fuel is slated to increase dramatically in the coming decades, owing to the efficiency and
environmental friendliness of natural gas in comparison to other fuels. However, transportation of natural gas from points of
production to consumption markets remains a challenge. Two well established gas transportation technologies currently
dominate the market: Pipeline and liquefied natural gas (LNG) accounting for 70 and 30 percent of transported gas,
respectively. Pipelines generally offer the most economically attractive solution and are the preferred choice whenever
feasible, such as for land transport. Subsea pipelines are restricted by the distance and terrain they can traverse. As a result,
transport of natural gas by sea vessels is favored for transport over long distances spanning sea water. While LNG is
currently the dominant technology used for sea transport of natural gas, a number of recent studies have shown that
compressed natural gas (CNG) is economically more attractive than LNG for sea transport of relatively small volumes of
natural gas over small distances (Marongiu-Porcu, et al., 2008, Wang and Marongiu-Porcu, 2008). In comparison to LNG,
which requires costly liquefaction and regasification facilities at the shipping and receiving sites, respectively and is energy-
intensive, CNG requires minimal investment in facilities at the shipping and receiving sites and uses less energy. The main
capital cost for CNG is incurred for transportation vessels. Although the cost for transportation vessels is higher for CNG
than for LNG (stemming from corresponding gas compression ratios of 300:1 vs. 600:1, respectively), overall economics
favor CNG for short distances and small loads, as summarized in Figure 1. Finally, because the demands for the entire chain
of operations to be well coordinated and in place are easier to meet for CNG (compression / transportation / distribution) than
for LNG (liquefaction/transportation/regasification), the former provides an additional degree of flexibility that is welcome in
situations of changing market conditions.
Even though CNG is used widely in land applications (buses, passenger cars) and a proof-of-concept for sea transport of
CNG has existed since the 1960s (Broeker, 1969), CNG has yet to make inroads into the transportation market for natural
gas. Three factors have contributed towards that. First, investment emphasis internationally has been primarily on LNG, for
understandable reasons (see Figure 1). Second, CNG vessel designs and projects have been envisioned to eke a bite out of
the LNG pie, which is not necessarily a good approach. Third, innovative CNG vessel designs for low cost and high
efficiency have become available only in recent years. There are certainly several regions worldwide whose energy needs,
geography, and access to natural gas sources would make them good candidates for application of CNG sea transportation
(Figure 2). For each of these regions there exist multiple scenarios for CNG distribution, in terms of number of vessels,
vessel capacities, and itineraries. Identification of promising scenarios is necessary to determine project economics and
possibly guide future technological developments.
2 OTC 19738

Figure 1. Ranges of economically preferred options for transportation of stranded gas to markets (Wood, et al., 2008)

Figure 2. Regions actively investigating CNG projects (Dunlop and White, 2003)

In this work, we present the results of a study on computing the optimal number of vessels and corresponding vessel
capacities as well as related itineraries for CNG distribution from a hub to multiple destinations in the Caribbean Sea. The
effect of geography and size of markets on optimal solutions is studied. The results of our study demonstrate that through the
proposed approach natural gas can be brought to previously inaccessible markets, thus replacing conventional fuels such as
diesel and resulting in sizeable economic and environmental benefits.

Distributed Sea Transport of CNG

Sea transport vessels for CNG

There are new technologies that are becoming available for commercial use in CNG transport. In general, they rely on a
combination of compression and mild refrigeration (but far from liquefaction) e.g. gas compression to 200-250 atm (3000-
3700 psi) (Sea NG Corporation, 2008) or about 130 atm (about 2000 psi) and refrigeration down to -40°C (-40°F) (Enersea
Transport LLC, 2008). Creative solutions have been proposed for the choice of materials (e.g., steel, composites),
configuration of gas containers (e.g., vertical or horizontal cylinders, coiled pipe), and loading or offloading techniques.
There is also flexibility in the choice of transport vessels, which can be ships or barges, depending on a number of factors, as
shown in Table 1. For the analysis presented here, it is assumed that ships are suitable for the weather conditions prevailing
over sea transportation routes. A typical itinerary for a CNG vessel involves a cycle consisting of the following steps: gas
loading at the source, transportation to delivery site(s), offloading, and return to the source. The number and capacity of
these ships as well as related itineraries will be the focus of the following analysis. Some economic issues will be discussed
as well.
OTC 19738 3

Table 1. CNG sea transport vessels (Dunlop, 2008)

Articulated Tug Barge Ship
Volume 0.7-2 MMcm (25-75 MMscf) 8-29 MMcm (300-1000 MMscf)
Loading/unloading rates 0.3-2 MMcm/day (10-75 MMscf/day) 2-14 MMcm/day (75-500 MMscf/day)
Distance 100-1000 km (50-500 nautical miles) 250-5000 km (135-2700 nautical miles)
Speed <25 km/hr (<14 knots) <33 km/hr (<18 knots)
Estimated cost $15-35 million $150-350 million

CNG transport: Vessel capacity and schedules

Before any economic optimization is attempted for CNG transport, it is worth exploring the physical requirements of a
related venture, in terms of number of vessels required, capacity of each vessel, and itineraries followed. This is because the
main capital expenditure (roughly of the order of 80%) for CNG projects is for transportation vessels. In the following
analysis, we make the simplifying assumption that natural gas has to be delivered to each receiving site at a constant rate
year-round, without taking seasonal variation into account. An annual average rate (Figure 5) is used for each receiving site,
although the same analysis could be easily repeated for peak rates as well.
As explained in the following sections, the preferred path for CNG transportation vessels may follow "hub-and-spoke" or
cyclical "milk-run" patterns (Figure 3) depending on consumption rates at receiving sites. The hub could be at the production
site of the natural gas or a trans-shipment point, possibly with storage. In the latter case, gas is delivered to the trans-
shipment point by large vessels, and then transferred to other vessels for final distribution to the ends of the spokes.
Destinations with high enough consumption rates to be serviceable by large enough vessels would be best serviced in a hub-
and-spoke pattern, namely each destination would receive gas through a chain of CNG vessels directly starting at the hub.
By contrast, destinations with low consumption rates that would be ideally serviceable by vessels smaller than a minimum
available size should avoid hub-and-spoke service by vessels of that minimum available size, because such vessels would
waste capacity. Rather, destinations with low enough consumption rates should be visited by vessels following a cyclical
milk-run type of path originating at the hub, delivering gas to some destinations, and returning to the hub. For this type of
service, local storage at each gas destination would be necessary for gas supply to be adequate for consumption during the
period between successive visits to that site by a CNG vessel.

Figure 3. "Hub-and-Spoke" (left) and cyclical "Milk-Run" paths for CNG distribution to N receiving sites (terminals T1,…, TN).

Case Study: CNG transport in the Caribbean

Caribbean geography and electricity consumption patterns

A map of Caribbean island countries with corresponding annual electricity consumption rates is shown in Figure 4. The
islands of the Caribbean basin are predominantly net energy importers, with the exception of Trinidad and Tobago, which is a
significant LNG exporter (Energy Information Administration, 2008). Most electricity generation stations on the islands
currently use diesel fuel or fuel oil and could easily be converted to use natural gas. Some Caribbean islands are already
seriously considering natural gas as an alternative to diesel and fuel oil. Two gas-fired power plants are in operation in
Dominican Republic and one in Puerto Rico, and after initially considering LNG, Jamaica is now considering CNG as an
4 OTC 19738

alternative (Energy Information Administration, 2008). The corresponding amounts of natural gas needed for power
generation to cover electricity production in Caribbean islands are shown in Figure 4 and Figure 5. The amounts clearly
correspond to the low end of the range shown in Figure 1 to favor CNG over other means of natural gas transport. In
addition, the transport distances would in the range of 2,000 km or less for gas from the offshore natural gas deposits between
the Venezuelan mainland and the islands of Trinidad and Tobago. Consequently, it would be interesting to examine in more
detail the transport vessel, storage, and itinerary possibilities for CNG transport from the area to Caribbean islands. It is
conceivable that an all-encompassing optimization problem could be formulated, whose solution would entail the optimal
number of vessels, vessel capacities, and itineraries. Such an optimization problem would be complicated, and its numerical
solution alone would provide little insight into the structure of the problem at hand. To generate insight as well as physically
rather than economically optimal solutions, we present next an analysis of the overall problem of CNG delivery to the
Caribbean island countries.

Figure 4. Amount of natural gas required annually for power generation to cover electricity consumption in Caribbean island
countries (CIA World Factbook, 2007); data for Guadeloupe are from http://www.freegk.com/worldatlas/guadeloupe.php; and for
Martinique and Anguilla from (UN Data, 2005). Conversion from GWh of electricity to required amount of natural gas is based on the
assumption of 45% efficiency of gas turbines used for electricity generation.
OTC 19738 5

Individual HbarsL and cumulative HdotsL CNG requirement @BcmêyearD

P uerto Rico
Dominican Republic
Jamaica
Trinidad and Tobago
Bahamas
Guadeloupe
Netherlands A ntilles
United States V irgin Islands
Barbados
Martinique
A ruba
Bermuda
Haiti
C ayman Islands
Saint Lucia
Grenada
Saint Kitts and Nevis
Saint V incent and the Grenadines
A ntigua and Barbuda
Dominica
A nguilla
British V irgin Islands
Turks and C aicos Islands
Montserrat
0 2 4 6 8 10 12

Figure 5. Distribution and cumulative distribution of natural gas required annually for power generation to cover electricity
consumption in Caribbean island countries. Trinidad and Tobago is the only significant exporter of natural gas.

Distributing CNG to Caribbean islands: The Big Four and a "Hub-and-Spoke" delivery scheme
Given the natural gas amounts required to accommodate the power consumption patterns shown in Figure 5, it follows
(Nikolaou, 2009) that four island countries (Puerto Rico, Dominican Republic, Jamaica, and The Bahamas, Figure 6) have
high enough consumption rates to justify CNG delivery using ships of reasonable capacity (i.e. between 8 and 29 MMcm, cf.
Table 1) as shown in Figure 7. What Figure 7 shows is individual vessel (ship) or total fleet capacities that would
accommodate CNG delivery rates exactly. While larger capacities would be feasible, they would obviously be wasteful. For
ship capacities constrained as in Table 1, Figure 7 suggests that possible fleets could consist of 3-8, 3-6, 3-4, and 2 ships for
Puerto Rico, Dominican Republic, Jamaica, and The Bahamas, respectively. Note that it is infeasible to service Puerto Rico
with fewer than three ships, because of the large consumption rate. The total capacity of a fleet would be minimum for the
largest possible value of ships in the fleet, namely 8, 6, 4, and 2 ships for each of the above island countries, respectively.
While keeping the total capacity of a fleet at a minimum would also tend to decrease capital costs, increasing the number of
ships in a fleet while decreasing the capacity of each individual ship would increase operating costs (since travel time per
ship would be the same) and might reach a point where economies of scale must be taken into account. Therefore, an
economic optimization should follow the physical optimization performed in this work. Such optimization will be performed
elsewhere.

Distributing CNG to Caribbean islands: Small Consumers and "Milk-Run" CNG delivery
Starting again with the natural gas amounts required to accommodate the power consumption patterns shown in Figure 5,
it follows that only the preceding four states (Puerto Rico, Dominican Republic, Jamaica, and The Bahamas) can be serviced
by ships of reasonable capacities (8-29 MMcm, Table 1). The remaining island countries do not have high enough
consumption rates to justify CNG delivery using reasonably sized ships (or even barges for some countries) as shown in
Figure 8. Of course, using ships or barges of capacity larger than necessary would be feasible, but it would clearly waste
resources. Therefore, a different CNG distribution strategy will be used. Its main ingredients are the following:
- Rather than servicing a single receiving site individually, each CNG transportation ship will service multiple
receiving sites during a cyclical trip from the gas source to delivery points and back.
- Each receiving site will have some capacity for local CNG storage, thus not requiring the continuous presence of a
CNG vessel from which to offload gas. Rather, gas will either be offloaded from a CNG transportation vessel,
when present, or will be delivered from local storage while the CNG vessel completes its cycle. The same
technology used to contain CNG on transportation vessels can be easily used for local storage, either in moored
barges or in land-based storage (Enersea Transport LLC, 2008).
- An optimal schedule following this strategy will be determined by first determining a minimum cyclical route
(starting and finishing in Trinidad) that includes all island countries shown in Figure 8, and then determining the
vessel size and number required to service all receiving sites.
6 OTC 19738

Figure 6. Hub-and-spoke plan for CNG delivery to four island countries of the Caribbean with the highest electricity consumption.
OTC 19738 7

Puerto Rico
Time
@days D

Of f loading time
8
Travel and loading time
6

Ò of ships, n
3 4 5 6 7 8 9 10 11 12 13

Dominican Republic
Time
@days D

15 Of f loading t ime
Travel and loading time
10

0 Ò of ships, n
2 3 4 5 6 7 8 9 10 11 12

Jamaica
Time
@days D

8
Of f loading time
Travel and loading time
6

0 Ò of ships, n
2 3 4 5 6 7 8 9 10 11 12

Bahamas
Time
@days D

8
Of f loading time
Travel and loading time
6

0 Ò of ships, n
2 3 4 5 6 7 8 9 10 11 12

Figure 7. Left: Possibilities for individual ship and total fleet capacities suitable or CNG transportation from Trinidad to Puerto
Rico, Dominican Republic, Jamaica, and The Bahamas ( tconnect = 1 h, v = 25 km/h, f = 5% , qload = 11.4 MMcm/day). The shaded band
indicates feasible values for individual ship size (Table 1). Right: Corresponding time taken by each ship to offload CNG at market
consumption rate, and to complete the travel-to-source / load / travel-to-sink cycle.
8 OTC 19738

Figure 8. Maximum capacity, Gn,max, of a CNG vessel that would be sufficient (without excess capacity) for gas distribution to each
individual Caribbean island country shown in the above plot.

Analysis for the island countries of Figure 8, produces the following results (Nikolaou, 2009).
Figure 9 shows the minimum cyclical route for vessels to visit all islands that receive CNG. At this point, it is assumed
that each island will have an individual receiving site for CNG. Clearly, for islands that are close to each other, subsea
transmission lines could be used to transfer electricity from a single station on one island to multiple neighboring islands.
Nevertheless, the amount of CNG to be delivered would be the same, and the following analysis would change little.
The capacity of each individual vessel in a fleet of n similar vessels is a function of the loading / offloading rates, as
shown in Figure 10. The total capacity of the fleet is shown in Figure 11. The local CNG storage capacity required at each
delivery point is shown in Figure 12. As in the Hub-and-Spoke case, increasing the number of vessels results in closer
approximation of a pipeline, and results in total capacity reduction for both vessels and storage. However, increasing the
number of vessels would increase the operating cost, and eventually the fixed cost (due to reverse economies of scale).
Therefore, an economic optimization will have to be performed to determine the economically optimal solution.
It is worth noticing that the results shown above are sensitive to inclusion of distant small markets in the distribution path.
For example, exclusion of Bermuda and Cayman Islands (Figure 13) results in the ship and fleet capacities shown in Figure
14 and Figure 15, respectively, and local storage capacities shown in Figure 16. Even though the reduction in delivered
natural gas resulting from exclusion of Bermuda and Cayman Islands is about 22%, the resulting reduction in ship capacity is
close to 60%.

Conclusions
Transportation of natural gas as CNG is attractive for relatively small volumes and short distances. As a result, the design
of a network for CNG distribution to a collection of neighboring small markets poses interesting challenges. In this work, we
present an approach for physical optimization of a network for distribution of CNG from a single source to a number of
markets. Rather than formulating the problem as an all-encompassing optimization, we decomposed it to sub-problems that
provide intuition about the structure of the solution. The case of the Caribbean island countries was studied, to demonstrate
the applicability of the approach. The preceding analysis indicates that either hub-and-spoke or cyclical-path itineraries may
be preferred, depending on market sizes and distances from the source. An economic assessment of proposed physically
feasible solutions will follow in the future.
OTC 19738 9

Figure 9. Optimal path for CNG delivery to island countries of the Caribbean with relatively low electricity consumption

Figure 10. Effect of loading/offloading rate, qload, on capacity, Gn, required for each ship in a fleet of n ships following the cyclical
path shown in Figure 9. The shaded area refers to feasible ship capacities, as shown in Table 1.
10 OTC 19738

180

160 n =4

Gtotal @MMcm D
140
n =5

n =6
120
n =7
n =8
n =9
n = 10
100

4 6 8 10 12 14
qload @MMcmêdayD

Figure 11. Total capacity required for a fleet of n ships following the cyclical path shown in Figure 9.

Figure 12. Local storage capacity required at each CNG delivery point for a fleet of n ships following the cyclical path shown in
Figure 9.
OTC 19738 11

Figure 13. Optimal path for CNG delivery to island countries of the Caribbean with relatively low electricity consumption.
12 OTC 19738

Figure 14. Capacity, Gn required for each ship in a fleet of n ships following the cyclical path shown in Figure 13. The shaded area
refers to feasible ship capacities, as shown in Table 1.

80

n =4
70
Gtotal @MMcm D

n =5
60

n =6

n =7
n =8
50 n =9
n = 10

4 6 8 10 12
qload @MMcmêdayD

Figure 15. Total capacity required for a fleet of n ships following the cyclical path shown in Figure 13.

Figure 16. Local storage capacity required at each CNG delivery point for a fleet of n ships following the cyclical path shown in
Figure 13.
OTC 19738 13

Nomenclature
Gn = natural gas capacity of a vessel in a fleet of n vessels
tconnect = time needed to connect or disconnect a vessel to a supply (source) or distribution line
L = distance from natural gas source to receiving site
L jk = distance from site j to site k
v = vessel velocity of sea travel
n = number of vessels in a CNG fleet
f = fraction of gas load used as fuel for transportation
qc = gas consumption rate
qload = gas loading rate
N = number of natural gas receiving sites (terminals T1 ,..., TN )
Gload,k = natural gas load delivered by a vessel to site k
Gstorage,k = local natural gas storage capacity at site k
ttravel = time needed for a vessel to complete a gas distribution cycle from source to receiving sites and back
t = time between arrivals of two successive CNG ships at a receiving site
MMcm = million standard cubic meters
Bcm = billion standard cubic meters

References
1. Marongiu-Porcu, M., Wang, X., and Economides, M.J., 2008. The Economics of Compressed Natural Gas Sea
Transport. in Russian Oil & Gas Technical Conference and Exhibition. Moscow, Russia SPE 115310.
2. Wang, X. and Marongiu-Porcu, M., 2008. The Potential of Compressed Natural Gas Transport in Asia. in International
Petroleum Technology Conference (IPTC). Kuala-Lumpur, Malaysia IPTC 12078.
3. Broeker, R.J., 1969. CNG & MLG-New Natural Gas Transportation Processes. American Gas Journal: 138-140.
4. Wood, D., Mokhatab, S., and Economides, M.J., 2008. Technology Options for Securing Markets for Remote Gas. in
87th Annual Convention, GPA. Grapevine, TX
5. Dunlop, J.P. and White, C.N., 2003. CNG Transport Technology is Delivering on Promises SPE 84254.
6. Sea NG Corporation. 2008; Available from: www.coselle.com.
7. Enersea Transport LLC. 2008; Available from: www.enersea.com.
8. Dunlop, J.P., 2008. April 22, Presentation at XGas.
9. Energy Information Administration. Country Analysis Briefs - Caribbean. 2008; Available from:
http://www.eia.doe.gov/emeu/cabs/Caribbean/pdf.pdf.
10. CIA World Factbook. Energy Statistics > Electricity > Consumption. 2007 [cited 2008]; Available from:
http://www.nationmaster.com/graph/ene_ele_con-energy-electricity-consumption.
11. UN Data. Total Electricity. 2005 [cited 2008]; Available from: http://www.freegk.com/worldatlas/guadeloupe.php.
12. Nikolaou, M., 2009. Distributed Compressed Natural Gas Sea Transport. XGas - Internal Report: Houston, TX.