Supply chain design and planning for LNG as a transportation fuel
Lopez Alvarez, Jose
DOI:
10.33612/diss.131459842
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2020
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Lopez Alvarez, J. (2020). Supply chain design and planning for LNG as a transportation fuel. University of Groningen, SOM research school. https://doi.org/10.33612/diss.131459842
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Chapter 6
Discussion and future research
In this section, we present a discussion of the implications and limitations of the research projects developed in this thesis, together with possible directions for future research. This discussion is presented separately for each of the challenges studied in this thesis. In Section 6.1, we discuss the infrastructure development of the LNG distribution network, and in Section 6.2, the management of LNG inventories taking account of quality. In Section 6.3, we present a discussion related to the interrelation between these challenges.6.1
Infrastructure development of the LNG
distribution network
In the research projects developed in Post et al. (2018) and Chapter 2, we study some of the strategic decisions related to the infrastructure development of the LNG distri-bution network. In practice, infrastructure development for LNG as a fuel is particu-larly capital-intensive owing to the highly specialized and novel technology required to keep LNG in its liquid state. Accordingly, it is crucial to identify the infrastruc-ture required to enable the adoption of LNG as a fuel. With this aim, in Post et al. (2018), we developed an approach to identify the minimum number of refueling stations (for road transport) required to render the adoption of LNG trucks in the
Netherlands economically and environmentally beneficial. In Chapter 2, we studied the design of the distribution network for LNG as a transportation fuel, in which we analyzed the minimum infrastructure (i.e. satellite facilities and transportation vehicles) required to fulfill the demand for LNG in some of the major corridors of the TENT-T network.
In discussing these projects, it is important to note that they are not aimed at addressing the question of whether investing in LNG and its infrastructure is an effective direction to tackle the environmental and social challenges that stimulated the development of alternative (cleaner) fuels in the first place. We have seen in the last decades that several investment decisions regarding alternative fuels have been reconsidered owing to a mismatch between the ultimate goal of alternative fuels and the actual results of their adoption. For instance, various agricultural projects were executed in order to generate biomass to feed the production of first-generation biofuels such as ethanol and biodiesel, which led to concerns regarding the increase in food prices and deforestation, among other things. Hence, it is of vital importance to study and anticipate the effects of adoption of alternative fuels before investments are made.
One major implication of the adoption of LNG as a transportation fuel is that the high investments in infrastructure can only be recuperated if LNG is used for a long period of time, essentially creating a “lock-in effect” for the use of natural gas, which, despite its environmental benefits compared with fuel oils, is nonetheless a fossil fuel. In principle, this lock-in effect could be mitigated as Bio-LNG absorbs the demand for fossil LNG as a transportation fuel. This “LNG as a transition fuel” argument is frequently used by advocates for rapid LNG infrastructure investment. Naturally, this statement is only valid if the Bio-LNG is produced in a environmental, economically and socially sustainable manner. This could be done, for instance, by liquefying synthetic natural gas produced from power-to-gas systems, where the surplus of energy from wind turbines and solar panels is used to generate synthetic natural gas from water and CO2. Currently, existing processes to produce Bio-LNG
in a sustainable manner are significantly limited by the lack of availability of bio-mass, the efficiency of the processes and/or their costs. Hence, there is a great deal of uncertainty as to whether Bio-LNG would be able to absorb a large part of the demand for LNG in the future.
cur-Discussion and future research 107 rent distribution network for LNG needs to be integrated with the local distribu-tion of Bio-LNG. Typically, bio/synthetic gas is produced in various small-scale farms/facilities and then delivered to a small-scale liquefaction plant where the Bio-LNG is finally produced. Accordingly, in an integrated network, liquefaction plants for bio/synthetic gas would also serve as supply sources of the fuel. The integra-tion of the networks for fossil LNG and Bio-LNG can have a great influence on the network design of the distribution network as a whole. In a typical distribution net-work, such as that presented in the case study in Chapter 2, the import terminal is the point of origin of all the LNG distributed to the market. When considering the suppliers of Bio-LNG in the network design problem, the existence of geographic-ally spread Bio-LNG suppliers can reduce the need for satellite facilities, reduce the amount of capacity required in those facilities and/or influence their location.
Based on these considerations, a possible direction for future research is to study the decision-making problems arising in the development of a local Bio-LNG net-work. A critical decision to be considered is the location and capacity of the lique-faction plants for Bio-LNG. Typically, the sources of bio/synthetic gas are geograph-ically scattered and have an uncertain output of bio gas, which means that the costs and the output of the liquefaction plants for Bio-LNG would also be highly uncer-tain. Additionally, it is also important to study network design problems related to the integration of the fossil LNG and Bio-LNG networks. The design of an integrated network must take into account that multiple suppliers are available, all of which can supply the fuel to both customers and potential satellite facilities. One key element here is that fossil LNG suppliers can guarantee the supply of fuel regardless of the amount of LNG required, whereas Bio-LNG suppliers have limited and uncertain capacity to fulfill demand.
6.2
Management of LNG inventory considering
quality
This thesis is the first to consider quality aspects in an LNG inventory management context. Chapters 3–5 provide strong indications that quality considerations affect inventory management decisions, where replenishment orders are done in response not only to potential (future) quantity shortages, but also to quality issues. In
study-ing the role of quality in effectively managstudy-ing LNG inventories, several modelstudy-ing choices have been made—particularly with regard to which LNG quality aspects need to be considered and how they are included in the proposed inventory man-agement models. Below, we first discuss these modeling considerations and then continue with a broader discussion.
6.2.1
Modeling considerations
Quantity decayIn practice, the boil-off rate of LNG is a complex process that strongly
depends on the properties of the storage tank (Dobrota et al., 2013). Therefore, as in previous studies (e.g. Ghiami et al. (2019)), this thesis relies on estimates of the rate at which the LNG in stock boils off. The estimated amount of boil-off can differ significantly depending on whether the boil-off rate is assumed to be constant or dependent on the inventory level (Głomski and Michalski, 2011). In the context of above-ground storage tanks, which is the setting of our studies, it was shown that the amount of boil-off generated is constant and independent of the inventory levels at higher LNG quality levels (Migliore et al., 2015). Only when the LNG composi-tion changes considerably (e.g. due to long-term deterioracomposi-tion) does the amount of boil-off generated start to decrease. Indeed, the LNG compositions common in the distribution network for LNG are well within the “constant range” of boil-off, i.e. the range where a constant amount of LNG boils off per unit time. We have thus modeled boil-off accordingly. Specifically, in Chapters 3 and 4, where we study in-ventory systems with deterministic demand, we assume that the boil-off is part of the demand itself. In Chapter 5, where demand is random, we explicitly model the quantity decay in the problem.
Quality deterioration. In contrast to the quantity decay, which has been
con-sidered in previous studies on the design and management of the LNG supply chain, this thesis is the first to consider quality aspects. Quality deterioration of LNG oc-curs because its lighter components vaporize preferentially with respect to the heav-ier components, leading to a relative increase in heavy hydrocarbons in the LNG and hence to a lower methane number. Determining or predicting the variations in quality over time involves complex equations that are dependent on several chem-ical and physchem-ical factors. Currently, there is no standard procedure to predict the quality deterioration of the fuel; however, various approaches have been proposed in the scientific literature (e.g. Migliore et al. (2015)). In a project with DNV GL, we
Discussion and future research 109 estimated the quality deterioration of the fuel through a distribution network using a simulation model. When fitting the deterioration curves to polynomial functions, the resulting functions were found to be approximately linear when the quality range of the LNG was within the ranges typically seen in the distribution network for LNG as a fuel. Accordingly, in this thesis, we make the simplifying assumption that the methane number of the LNG deteriorates at a fixed rate per unit time.
LNG mixture. Determining the methane number when mixing two loads of LNG
involves solving various complex chemical equations that are dependent on the spe-cific composition and the quantity of the LNG loads involved. Since the scope of our thesis is on the inventory control implications of the physicochemical properties of LNG—rather than on those properties themselves—we have made the simpli-fying assumption that the quality of a mixture of LNG is the weighted average of both loads involved in the mixture. Albeit simplifying, this assumption proved to be reasonably accurate after testing some practical examples using W¨artsil¨a software (W¨artsil¨a, 2018) to estimate the actual methane number when mixing different loads of LNG, as shown below. Table 6.1 presents the different LNG compositions con-sidered in our example. LNG of Type A closely resembles the composition of LNG produced in Nigeria, which has a quality that is common for the LNG imported to the European market. All the other compositions in Table 6.1 are deteriorated ver-sions of Type A LNG; that is, with smaller amounts of light components compared with heavy components. In our examples, the different types of deteriorated LNG are upgraded using the Type A LNG.
Table 6.1: Composition (mol%) and methane number of different types of LNG
Type Nitrogen Methane Ethane Propane n-Butane Isobutane Methane number
A 0.10 91.30 4.60 2.60 0.80 0.60 71
B 0.00 90.29 5.19 2.93 0.90 0.68 69
C 0.00 89.06 5.85 3.31 1.02 0.76 67
D 0.00 87.46 6.71 3.79 1.17 0.87 65
Table 6.2 shows a comparison between the actual methane number (i.e. that de-termined using W¨artsil¨a software) and the methane number based on the weighted average (i.e. the assumption made in the thesis) when loads of LNG of different compositions and quantities are mixed. The results indicate that the weighted av-erage is a good approximation to the actual methane number for the loads of LNG considered in this study. Indeed, it is fully accurate for five of these mixtures. In the
Table 6.2: Estimated and actual methane numbers for different mixtures
Types of LNG mixed Volume Methane number Type 1 Type 2 Type 1 Type 2 Actual Weighted average
B A 1 1 70.0 70.0 C A 1 1 69.0 69.0 D A 1 1 68.0 68.0 B A 1 2 70.0 70.3 C A 1 2 70.0 69.7 D A 1 2 69.0 69.0 B A 1 3 70.0 70.5 C A 1 3 70.0 70.0 D A 1 3 70.0 69.5
other four mixtures, the error between the actual and the estimated methane number is reasonably small.
6.2.2
Discussion and future research
In Chapters 3–5, we studied various inventory replenishment problems inspired by the distribution network of LNG as a transportation fuel. In these studies, we provided various insights into how to manage LNG inventories while taking into consideration the quality aspects of the fuel. In practice, however, quality is rarely taken into account when managing inventories of LNG. This can be partly explained by the fact that there are no regulations governing the quality of LNG as a transport-ation fuel. However, as engines become more energy-efficient, fuel quality require-ments become more stringent, since the engines become more prone to problems arising from quality levels different from those specified. It is therefore to be ex-pected that quality management issues will be discussed more actively in the near future.
Currently, the management of LNG inventories in storage and refueling facil-ities is driven mainly by economic and safety considerations. First, the slot costs associated with loading a tanker truck or a bunker barge with LNG at import ter-minals are very high. As a consequence, replenishment policies for LNG storage and refueling facilities are typically aimed at maximizing the capacity utilization of the transportation vehicles. Second, in most cases, replenishment orders are placed with the nearest import terminal in order to reduce transportation costs (Bio-LNG is rarely used, because its current price is high and its availability is very limited). Third, in practice, safety considerations also play a role in replenishment decisions.
Discussion and future research 111 Specifically, boil-off of the LNG increases the pressure in the storage tank, which raises safety concerns. One way to cope with this is to replenish the storage tank of a facility with cold LNG coming from the replenishment vehicle.
Consideration of the main drivers of replenishment decisions in practice leads to inventory control policies where most of the replenishment orders, which are placed with a single supplier, exploit all the capacity of the transportation vehicle while, occasionally, some smaller replenishment orders are placed to reduce the pressure in the tank. However, in Chapter 5 of this thesis, we learned that if the quality critic-ality of the LNG is high (i.e. if the qucritic-ality offered by the import terminal is close to the minimum quality requirement), optimal policies make use of frequent replenish-ment of LNG, especially when the fuel is not being removed from the storage tank as a means to control quality. In the current state of practice, it can therefore be ex-pected that in such settings, end-users sometimes refuel with LNG below the quality requirement. Naturally, this risk also depends on the throughput time of the LNG in the distribution network. When the LNG remains for long periods in the network before it reaches the end-user, the quality deterioration can be significant enough to bring the quality of the fuel below the minimum requirement.
In some regions, the quality of the LNG supplied by the import terminals is by default below the minimum requirement. Although currently this may not be the case in most European nations, this issue will become more prominent when tech-nological advances in the efficiency and power of LNG engines are coupled with more stringent quality specifications for the engines. When that becomes the case, mixing LNG will be the most viable alternative to upgrade the quality of the LNG. Interestingly, the use of mixing to upgrade quality can promote the growth of Bio-LNG production. Bio-Bio-LNG is a good alternative to upgrade the quality of fossil Bio-LNG because its quality is very high, it can be produced locally, it is compatible with the equipment/infrastructure used for fossil LNG and it can be mixed with fossil LNG.
We see ample opportunities for future research. First of all, this thesis is the first study to use operations research techniques to support replenishment decisions considering quality aspects in LNG inventory management; to do so, we simplified the physicochemical properties of LNG. Therefore, an interesting way to continue this line of research would be to integrate the chemical and physical models that characterize the properties of LNG with operations research models. One of the main challenges in doing this, however, is that the methane number of LNG and
its deterioration process depend on various physical and chemical variables such as composition, temperature and pressure. Consideration of all these variables and their interactions would compromise the tractability of the solution techniques to find replenishment policies for an LNG inventory system. Accordingly, in order to pursue this line of research, a thorough analysis of how to model the physical and chemical variables of the problem needs to be done.
Other opportunities for future research involve study of the inventory routing problems for multiple LNG facilities. In practice, there are companies that man-age several LNG refueling facilities within the same region. These companies must design replenishment plans to fulfill the demand in all their facilities while ensur-ing that the quality of the fuel always complies with the minimum requirement. In this setting, optimization opportunities arise from the economies of scale that can be achieved by replenishment of multiple facilities with the same replenishment vehicle. This is particularly interesting with regard to the replenishment orders placed with the Bio-LNG supplier, since a single refueling station would typically require much less than a truckload of Bio-LNG to upgrade the quality of its invent-ories.
6.3
Integration of network design and inventory
management of LNG inventories considering
quality
This thesis first addressed new decision problems related to infrastructure develop-ment for LNG as fuel, followed by inventory replenishdevelop-ment problems considering the quality of LNG. When looking at the thesis as a whole, it becomes clear that both challenges are in fact interrelated. First, the throughput time of the LNG in the dis-tribution network determines the extent to which the quality of the fuel deteriorates. Hence, network design decisions can be affected when quality is taken into account, and those decisions affect the management of LNG inventories and their quality. For example, when considering quality, large tank sizes for LNG storage/refueling facilities might not be suitable, because the quality deterioration of the fuel discour-ages the use of large replenishment orders. Second, when LNG is transferred from
Discussion and future research 113 a storage tank to transportation vehicles (and vice versa), the boil-off rate is higher, as a consequence of which the deterioration of the fuel is also higher. Hence, the network design of an LNG network must take this aspect into consideration. Third, technology can be used to mitigate the quality deterioration of the fuel. Accordingly, decisions regarding specialized equipment such as storage tanks with different insu-lation capabilities and other refrigeration equipment can be incorporated in the LNG network design problems.
Simultaneously considering network design and quality management for LNG as a fuel leads to interesting problems for future research. One possible direction is to study the network design problem for an LNG distribution network, consider-ing the deterioration of the fuel. In that problem, various features encountered in practice could be incorporated, such as the possibility of purchasing storage tanks with different insulation capabilities. Another interesting project would be to integ-rate capacity selection decisions with the inventory control problems presented in Chapters 3–5. In such a setting, it could be interesting to explore the possibility of having multiple storage tanks in a single facility, each of which could store LNG with different qualities.
