Analyzing risk factors in ship-to-ship liquefied natural gas bunkering operations

Choi, Young-Seo; Listan Bernal, Maria; Krivoshapkina, Margarita; Yeo, Gi-Tae Article Analyzing risk factors in ship-to-ship liquefied natural gas bunkering operations Asian Journal of Shipping and Logistics (AJSL) Provided in Cooperation with: Korean Association of Shipping and Logistics, Seoul Suggested Citation: Choi, Young-Seo; Listan Bernal, Maria; Krivoshapkina, Margarita; Yeo, Gi-Tae (2025) : Analyzing risk factors in ship-to-ship liquefied natural gas bunkering operations, Asian Journal of Shipping and Logistics (AJSL), ISSN 2352-4871, Elsevier, Amsterdam, Vol. 41, Iss. 1, pp. 52-60, https://doi.org/10.1016/j.ajsl.2025.01.003 This Version is available at: https://hdl.handle.net/10419/329756 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/4.0/ Analyzing risk factors in ship-to-ship liquefied natural gas bunkering operations Young-Seo Choi, Maria Listan Bernal, Margarita Krivoshapkina , Gi-Tae Yeo * Graduate School of Logistics, Incheon National University, Incheon 22012, South Korea ARTICLE INFO Keywords: Risk factors Ship-to-Ship LNG Bunkering Operation CFPR ABSTRACT This study evaluates the risk factors during ship-to-ship liquefied natural gas bunkering in South Korea using the consistent fuzzy preference relations method. The study showed that the detailed factor “The operator’s inability to recognize the operational condition, system status, or measuring instrument status” was ranked first with 0.112, alluding to the human error principal factor. “Control equipment malfunction (0.073)” and “Operation of the Emergency Shut-Down (ESD) System (0.071)” ranked in second and third place, respectively. Based on this study’s industrial implications, expert opinions were integrated to prioritize risk factors, enabling industrial managers to proactively prepare. 1. Introduction In 2018, the International Maritime Organization (IMO) set a landmark goal to reduce greenhouse gas (GHG) emissions from international shipping. This objective, aimed at reducing emissions by at least 50 % by 2050 (compared with 2008 levels), represents a significant milestone in global climate change governance (Garcia et al., 2021). Achieving the objectives of the Paris Agreement, aimed to stabilize global temperatures at “well below 2◦C, and toward 1.5◦C,” requires reaching net-zero CO 2 emissions by 2050–2070 or earlier and possibly transitioning to net-negative emissions thereafter. These targets apply to all sectors of the economy (Bataille, 2020). Selecting an optimal solution to comply with regulations and address economic pressures poses a significant challenge for shipowners and decision-makers. Choosing options requires considering environmental, technological, and economic factors, often involving conflicting priorities and tradeoffs. In the context of shipping, vessels can choose between integrating emission abatement technologies (such as scrubbers) using more expensive low-sulfur fuels, such as marine gas oil (MGO) or marine diesel oil (MDO), and adopting liquefied natural gas (LNG) (Dalaklis et al., 2017). Despite its fossil origin, LNG is widely recognized as more environmentally sustainable, particularly when compared to coal and oil. Consequently, it has become a substantial component of the energy mix in numerous developed economies. It demonstrates value as a bridge fuel, facilitating a shift from fossil fuels to renewable energy sources (Srinivasan et al., 2024). Among the LNG bunkering options, Ship-to-ship (STS) is a type of direct transshipment in which goods are transferred directly from the mother vessel to the feeder vessel while the ships are moored next to one another in the open sea (Al Samrout et al., 2024). Therefore, STS bunkering offers advantages, such as reduced fueling supply time, increased speed, and larger capacity for loading and unloading, making this compendium of benefits a significant contribution to operational cost savings (Liu et al., 2024). Additionally, when constructing fixed installations at port locations has restrictions, STS bunkering serves as a practical and efficient alternative, especially for vessels with brief port stays (Aneziris et al., 2022). Thus, identifying risk factors ensures safe STS bunkering practices. Analyzing real STS bunkering case scenarios underscores the critical importance of leaking accidents (Nubli et al., 2022), the designation of the pertinent safety zones (Park et al., 2018a; Aneziris, 2022; Zhang et al., 2024; Duong et al., 2023), ship collisions (Arici et al., 2020; Sokukcu & Sakar, 2022), human error (Uflaz et al., 2022a), and other factors. Additionally, uncertainties and incidents arising from regulations and guidelines loopholes (Aneziris et al., 2022) should be considered simultaneously. Despite these studies, research on the risk factors associated with STS operations in South Korea is limited. This study identifies risk factors using the consistent fuzzy preference relation (CFPR) methodology and weighs and ranks a series of risk factors obtained from previous * Corresponding author. E-mail addresses: [email protected] (Y.-S. Choi), [email protected] (M. Listan Bernal), [email protected] (M. Krivoshapkina), ktyeo@ incheon.ac.kr (G.-T. Yeo). Contents lists available at ScienceDirect The Asian Journal of Shipping and Logistics journal homepage: www.elsevier.com/locate/ajsl https://doi.org/10.1016/j.ajsl.2025.01.003 Received 17 November 2024; Accepted 27 January 2025 The Asian Journal of Shipping and Logistics 41 (2025) 52–60 Available online 30 January 2025 2092-5212/© 2025 The Authors. Published by Elsevier B.V. on behalf of The Korean Association of Shipping and Logistics, Inc. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ ). literature and experts’ opinions. 2. Current status of LNG STS bunkering 2.1. Carbon emission regulations In July 2023, the International Maritime Organization (IMO) adopted the “2023 Strategy for the Reduction of Greenhouse Gas Emissions from Ships” during the 80th session of the Marine Environment Protection Committee (MEPC), amending the target timeline for achieving net-zero greenhouse gas emissions in international shipping. The target year for net-zero emissions was brought forward from 2100 to 2050. Accordingly, the previous reduction goal of a 50 % cut by 2050 was revised to reducing carbon emissions by 30 % by 2030 and by 70 % by 2040. By 2050, as implied by the term “Net-zero,” carbon emissions must converge to zero, as shown in Table 1. 2.2. Response strategies to carbon emission regulations Gradual response strategies have been established to decommission existing ships and introduce new ships. Table 2 shows a proposed threestep countermeasure strategy, with the first step already under implementation. LNG and methanol are considered bridge fuels in the second step of the response strategy. In particular, LNG is frequently classified as a fossil fuel owing to its attributed CO 2 emission levels of 2.7 tons per ton of fuel. To align with the IMO’s carbon regulations, changes to ship fuel, the main source of carbon dioxide emissions, are essential. Currently, LNG is prioritized as a bridge fuel to achieve the IMO’s short-term target of reducing emissions by at least 30 % by 2030. LNG emits 25 % less carbon dioxide compared with bunker oil and is more economically advantageous than fuel oil. Despite being a fossil fuel with inherent limitations, LNG has been chosen as the primary bridge fuel due to factors involving ship engine compatibility, supply chain, infrastructure, and price competitiveness. Therefore, simultaneous LNG bunkering operations, involving LNG bunkering and cargo handling at the same time, are being implemented for more efficient port operations. Initially, when supplying LNG to ships, all other tasks had to be halted as a precaution against potential accidents. However, because large-scale bunkering operations can take two to three days, and shipping companies prefer early departure, simultaneous LNG bunkering operations have begun to be introduced. 2.3. Current Status of LNG Bunkering Against this backdrop, the expansion of bunkering infrastructure is necessary for the increased use of LNG fuel. As of March 2022, 15 % of the total number of ship orders and 33 % of the total tonnage were for ships using LNG fuel propulsion. According to Clarkson Research Table 1 IMO Carbon dioxide (CO 2 ) emissions reduction target for 2050. Period Carbon dioxide (CO 2 ) emissions reduction target for 2050 (Relative to the carbon emissions of 2008) Year 2015–2019 10 % Year 2020–2024 20 % Year 2025–2030 30 % Year 2030–2050 70 % Year 2050 100 % Source: Compiled by the author based on data retrieved from the International Maritime Organization (n.d.). Table 2 Response Strategies to carbon emission regulations. Step 1 Step 2 Step 3 Fuel type Conventional fuel Bridge fuel Alternative fuel Scope of the measures Operational Technical Transition to eco-friendly (Use of low-carbon and carbon-free fuels) Target (ships) Decommissioning of the existing fleet before 2030 - Decommissioning of the existing fleet before 2040 - Introduction of new fleet before 2030 Introduction of new fleet after 2030 Response strategy Decelerated operation (optimal operation speed), optimization of shipping lanes (shipment), and application of smart ship solution platform Introduction of LNG propulsion ships, use of biofuels, and installation of energy-saving devices (e.g., introduction of wind power devices to help propel ships) Ammonia or hydrogen propulsion line (bridge fuel not used) Source: POSCO flow Table 3 In-depth interview organizations. Organization Name Work Related Korea LNG bunkering LNG bunkering ship operator H Line Shipping Company LNG bulk ship operator Korea Research Institute of Ships & Ocean Engineering LNG bunkering ship development organization Korean Register of Shipping LNG bunkering vessel approval organization Ulsan Port Authority Port authority that operate LNG bunkering facilities POSCO Flow Supervisor for verification work of the STS method’s simultaneous LNG bunkering and unloading operation Table 4 Factors extracted. Principal factor Detailed factor Component error Ship pump malfunction Pressure or level sensor malfunction in functionality, audibility, or visibility Detector failure Control equipment malfunction Organizational error Lack of a well-documented operational procedure Insufficient preparation before the operation Lack of communication between interdisciplinary teams (including lack of communication between multinational sailors) Lack of existence/implementation of accident prevention strategies (high wave, wind, ignition, dropped object) Human error The operator’s inability to recognize the operational condition, system status, or measuring instrument status Operator’s inadequate performance due to high workload or work dissatisfaction System error of LNG supply & receiving vessel Electricity blackout, communication error Pipeline leakage Overpressure/overheating of equipment Operation of the ESD System Work process error Inadequate purging of the loading arm Over-fueling (only for fueled ships) BOG (Boil-off gas) removal malfunction Flame ignition Poor ballast or mooring line failure Y.-S. Choi et al. The Asian Journal of Shipping and Logistics 41 (2025) 52–60 53 (2023), the major types of vessels are shifting to LNG propulsion, and over the next 10 years, the average conversion rate to LNG-powered of newly built ships is projected to be 31 %. Furthermore, by 2043, global LNG bunkering demand is expected to increase to 53 million tons. In Asia, countries such as Japan and China are developing LNGpowered ships. Singapore developed its first bunkering vessel and declared its position as an LNG hub port. Additionally, in 2021, the first simultaneous ship-to-ship (STS) LNG bunkering operation was successfully conducted at the Port of Singapore. In 2020, South Korea’s Ministry of Oceans and Fisheries enacted the “Act on the Promotion of the Development and Distribution of Environmentally Friendly Ships,” known as the “Eco-Friendly Ship Act,” and established a basic plan set to continue until 2030. This legislation includes the mandatory use of eco-friendly fuel-powered ships for newly constructed government public vessels, and support for R&D related to eco-friendly ships. Additionally, from September 2020, the five major Table 5 Summary of expert’s questionnaire. Interview Date February 2024–April 2024 Experts area Shipping company Research Institute LNG bunkering ship approval authority Port Authority STS-based LNG bunkering and unloading simultaneous work demonstration Academic Experts experience (years) Less than 10 years 0 0 0 1 0 0 10–15 years 2 1 1 0 0 2 16–20 years 0 2 0 0 1 0 More than 21 years 0 0 1 0 0 0 Total experts 11 2 3 2 1 1 2 Fig. 1. Principal factor result. Fig. 2. Component error detail factor result. Y.-S. Choi et al. The Asian Journal of Shipping and Logistics 41 (2025) 52–60 54 domestic ports—Busan, Incheon, Yeosu-Gwangyang, Ulsan, and Pyeongtaek-Dangjin—were designated as Emission Control Areas (ECA), where the sulfur content of ship fuel is regulated and maintained at or below 0.1 %. In 2015, truck-to-ship bunkering was primarily conducted for small LNG-powered vessels. However, through a national R&D project in 2022, South Korea is developing its first LNG bunkering vessel and operational technologies for LNG ship bunkering. Additionally, under the joint support of the government, research institutions, and the industry, including the Ministry of Trade, Industry, and Energy, the Ministry of Oceans and Fisheries, KOGAS, POSCO, and H-line LNG, a bunkering demonstration project was conducted. The demonstration operation took place on October 28, 2023, at Gwangyang Port, where 1000 tons of LNG were simultaneously bunkered on a vessel engaged in cargo handling. Ulsan Port is home to the country’s first and largest LNG bunkeringexclusive terminal, offering numerous expected benefits in line with ecofriendly port policies. Additionally, in July 2023, Ulsan Port successfully became the first in the world to bunker green methanol using the pipeto-ship (PTS) method. 3. Literature review The STS bunkering process for LNG is a relatively novel concept in the maritime industry, emerging alongside the global increase in the utilization of LNG (Uflaz et al., 2022b). Aneziris et al. (2022) argued that STS bunkering is an advantageous refueling method for ports accommodating vessels of varying sizes —from small to very large capacity- —especially those with brief port stays. They emphasized that STS bunkering is a logical substitute for ports where fixed installations are either prohibited or unfavored. Depending on the policies of the port authority, STS bunkering can occur either at the pier or anchorage in open seas. Although various studies underscore the benefits of STS bunkering, this practice introduces potential risks. Aneziris et al. (2022) identified Fig. 3. Organizational error detail factor result. Fig. 4. Human error detail factor result. Y.-S. Choi et al. The Asian Journal of Shipping and Logistics 41 (2025) 52–60 55 several possible risks associated with LNG bunkering. The accuracy of the bunkering equipment, as highlighted by Tam (2022), also plays a crucial role in mitigating risks. Moreover, the lack of data, limited understanding of safe storage and bunkering of LNG, and a scarcity of comprehensive risk assessment studies contribute to the overall uncertainties in LNG bunkering (Gerbec & Aneziris, 2022). Iannacconea et al. (2018) and Aneziris et al. (2022) specify potential scenarios that may lead to LNG release, fires, or explosions during bunkering. These include malfunctions in boil-off removal, overfilling of ship tanks, inadvertent valve closures, external fires, and additional loads caused by events such as collisions, grounding, poor ballast, and mooring line issues. Furthermore, natural events such as earthquakes, tsunamis, or high winds can impose extra loads on hoses, thereby increasing the associated risks. System failures, encompassing component malfunctions and human errors, are also considered potential hazards (Aneziris et al., 2022). To investigate the probability of human error, Fan et al. (2022) investigated the role of human factors in LNG bunkering safety. Human errors arise from a combination of onsite conditions and personal psychological factors. The concept of safety philosophical factors highlights the influence of psychological factors on human performance. A robust safety culture has been proposed to foster a positive mental state among workers during LNG bunkering tasks, thereby reducing the likelihood of human error (Fan et al., 2022). Uflaz et al. (2022a) explained that enhancing shipboard safety significantly relies on human reliability. As stringent environmental regulations are enforced, ship owners are exploring the adoption of cleaner alternative fuels to minimize ship emissions and ensure compliance with prescribed limits. Uflaz et al. (2022b) also emphasized that bunkering operations experience undesirable incidents, including overflow, leakage, and sea pollution. The challenges of the overall LNG bunkering were discussed by Aymelek et al. (2014), who divided them into five categories: systemic, operational, technical, safety, and conjectural. Macro-level challenges present significant obstacles for business entrepreneurs aiming to achieve micro-level objectives. Political instabilities, the risk of war, financial crises, fluctuating natural gas prices, and potential additional regulatory measures for environmental concerns have emerged as significant systemic challenges confronting LNG Fig. 5. System error of LNG supply & receiving vessel detail factor result. Fig. 6. Work process error detail factor result. Y.-S. Choi et al. The Asian Journal of Shipping and Logistics 41 (2025) 52–60 56 bunkering (Aymelek et al., 2014). The ability to handle cold materials within pertinent ship structures and the potential asphyxiation risk for individuals engaged in bunkering, contingent upon circumstances during an LNG spill, pool fire, vapor cloud fire, explosions, and rapid phase transition, pose significant safety challenges. These challenges necessitate preventive measures through enhanced technological advancements and training programs for both the crew and bunkering personnel (Aymelek et al., 2014). Park et al. (2018b) highlighted the lack of guidelines for establishing safety zones, pointing to their crucial role in mitigating hazards associated with LNG bunkering. In the event of gas leakage, wind speed and direction, surrounding conditions, ship draft, and cargo loading planning were identified as critical in determining the need for safe zones. Among the studies addressing the simultaneous operations performed during the bunkering process, Fan et al. (2021) reviewed the literature on their role in safety and risk assessment, namely, people movements, cargo loading/unloading, supply/crew transfer, crane maneuvering, port operations, and bunkering. The authors concluded that the main limitations corresponded to the lack of consensus in safety regulation guidelines, the difference in the understanding of safety philosophy concerning LNG bunkering, and the absence of a standardized risk analysis methodology. As Surinov (2023) stated, the safety zone for STS LNG bunkering must be mutually determined by the two involved vessels, and their compatibility must be verified and confirmed before bunkering activities begin. The risks associated with STS bunkering include mooring failure, cargo transfer hose failure, fatigue, availability of personnel, and concurrent operations. Sultana et al. (2019) identified several categories of STS bunkering hazard issues, including component error (issues related to pump malfunction, control equipment, pressure or level malfunction), organizational error (lack of well-trained human resources, lack of inter-team communication, operational procedure documentation), human error (lack of condition awareness, missing or wrong operational decision), software error (software bug, lagging, or sabotage), system or design error (or electrical blackout, equipment overpressure, or overheating pipeline leaks), and external events (wind, waves) Wu et al. (2021) used a Bayesian network to evaluate risks associated with dangerous accidents during LNG bunkering. Additionally, the EPE, an energy-based model, was also introduced to derive the causal structure. The Bayesian network-based risk assessment model proposed through the analysis identified risks and evaluated the evolutionary process from the cause to the effect of LNG accidents. Furthermore, a sensitivity analysis was conducted to quantify the correlation between each factor considered in the LNG accidents and identify the principal risks. The proposed risk assessment framework can be used to evaluate accident safety and prevent losses during LNG bunkering in ports. To evaluate the potential risks arising during ship operation in bunkering work, Goksu and Arslan (2023) proposed a quantitative marine safety analysis based on the fuzzy failure mode and effect analysis methodology. As a result of the analysis, the high-risk factors were derived as fatigue/individual errors, very strong winds, excessive heat, low tide, and increased/decreased ship speed. Additionally, the control of the preliminary failure mode was necessary for the ship’s safety. Their study revealed that it was easy for marine safety inspectors, safety researchers, and health, safety, environmental, and quality managers to identify potential risks, effects, and results during anchoring/coastal work. LNG bunkering methods include truck-to-ship (TTS), STS, and pipeline-to-ship (PTS). The types of problems that arise may vary depending on the working method. For example, Duong et al. (2023) studied the outflow of LNG and NH3 generated using the PTS method. A mathematical model was presented to investigate the LNG/NH3 leakage dispersion characteristics and determine the safety distance during bunkering between ships. A numerical study was conducted for 26 scenarios considering various operating and environmental conditions, leakage hole diameters, leakage rates, wind speeds, and wind directions. According to the results, NH3 exhibited a larger dispersion range than LNG under the same operating conditions. Therefore, NH3 requires a wider safe zone distance, and its dispersion time is longer than that of LNG. The weather conditions (wind speed and wind direction) and leakage characteristics (leak rate and leakage period) were identified as important parameters. Jeong et al. (2018) introduced a practical new method for establishing safety exclusion zones during LNG bunkering. Their results show that probabilistic risk assessment, which focuses only on population-independent analyses, is somewhat inappropriate when determining safety exclusion zones. The established hypothesis showed that the range of safety exclusion zones tends to be unrealistic. The approach proposed in this study proved useful for determining the zones more realistically. Several studies have been conducted to establish safety-related standards and regulations for LNG bunkering. This study evaluates the risk factors associated with STS operations involving LNG bunkering methods. Exploring research topics not examined in existing domestic studies is significant, providing valuable reference materials for STS Table 6 Global Importance Result. Principal Factors Detailed Factors Global Importance Rank Component error (0.232) Ship pump malfunction (0.226) 0.053 8 Pressure or level sensor malfunction in functionality, audibility, or visibility (0.239) 0.056 6 Detector failure (0.221) 0.051 9 Control equipment malfunction (0.313) 0.073 2 Organizational error (0.158) Lack of a well-documented operational procedure (0.241) 0.038 16 Insufficient preparation before the operation (0.249) 0.039 13 Lack of communication between interdisciplinary teams (including lack of communication between multinational sailors) (0.242) 0.038 15 Lack of existence/ implementation of accident prevention strategies (high wave, wind, ignition, dropped object) (0.269) 0.043 12 Human error (0.172) The operator’s inability to recognize the operational condition, system status, or measuring instrument status (0.651) 0.112 1 Operator’s inadequate performance due to high workload or work dissatisfaction (0.349) 0.060 5 System error of LNG supply & receiving vessel (0.239) Electricity blackout, communication error (0.228) 0.055 7 Pipeline leakage (0.272) 0.065 4 Overpressure/overheating of equipment (0.204) 0.049 10 Operation of the ESD System (0.296) 0.071 3 Work process error (0.198) Inadequate purging of the loading arm (0.192) 0.038 17 Over-fueling (only for fueled ships) (0.195) 0.039 14 BOG (Boil-off gas) removal malfunction (0.187) 0.037 18 Flame ignition (0.245) 0.048 11 Poor ballast or mooring line failure (0.180) 0.036 19 Y.-S. Choi et al. The Asian Journal of Shipping and Logistics 41 (2025) 52–60 57 workers, port research officials, and government authorities. 4. Research methodology This study implemented the CFPR method, based on Herrera-Viedma et al. (2004), to identify the risks produced in STS bunkering during LNG operations in Korea. Herrera-Viedma et al. (2004) proposed the CFPR method to structure complex decisions by drawing on the relative importance of factors and their priorities at each hierarchical level per pairwise comparison preference matrix. According to Chen and Chao (2012), pairwise comparison procedures -such as AHPpresent (n ×(n −1))/2 survey comparisons for a group (n) criterion. This can lead to inconsistencies and expert confusion due to the extensive number of survey items. Therefore, the CFPR method seeks to uphold logical consistency in decision-making processes by implementing a reduced number of simple survey items, which shrinks (n ×(n −1))/2 comparisons to only an (n −1) comparison for group (n) criteria (Chen & Chao, 2012). Chao and Chen (2009) and Wang and Lin (2009) further explored the foundational concepts in the CFPR methodology. In these studies, the CFPR implementation of multiplicative preference relations is the ratio shown to represent the relative importance of different criteria in decision-making. In contrast, fuzzy preference relations enable flexibility in expressing preferences, thereby reducing uncertainty. In multiplicative preference relation (1), a set of experts (E = {e1, e2,…,em,m≥2}) must state their choice over a finite array of possibilities (X). (Х= {x1,x2,…,xn,n≥2}), indicated by a preference relation matrix А⊂X×X,А=(aij),∀i,j∈ {1,…,n}, aij ∈[1 5,5]that highlights the preference ratio of the possibility xi to the possibility xj. In the equation, aij =1 indicates xi and xj equivalency and aij =5 denotes that xi is the best decision compared with xj. Thus, preference relation (A) is a multiplicative reciprocal aij ×aji =1∀i.j∈ {1,…,n}(1) Concerning fuzzy preference relations (2), in a decision-making context, the expert criteria over a finite array of possibilities establish the preference ratio of the possibility xi over xj; X is denoted by a positive relation matrix P⊂X×X with membership function μ p(xi,xj)= pij. Additionally, pij =1 2 indicates that the experts are unbiased, showing no reason to prefer xi to xj(xi∼xj). This is referred to as indiscrimination. Conversely, pij =0 denotes a preference for xi over xj xi>xj, which is referred to as precedence. The preference relation (P) is an additive reciprocal because the sum of their relative preferences always equals one. pij +pji =1∀i.j∈ {1,…,n}(2) The decision matrixes in CFPR stem from the three main propositions: Number 1. A set of alternatives about a multiplicative preference relation is adopted, Х= {x1,x2,…,xn,), A= (aji), with aij ∈[1 5,5]. The reciprocal additive preference is expressed as Pij =g(aij)=1 2(1+log5aij)(3) In the preceding context, g represents a transformation function, and log5aij is related to the implementation of aij ∈[1 5,5]. Number 2. In the context of this reciprocal fuzzy preference relationship, P =g(A), where P =pij corresponds to Pij +Pjk +Pki =3 2∀i,j,k(4) Pij +Pjk +Pki =3 2∀i<j<k(5) Number 3. The reciprocal additive fuzzy relation, P = (pij), is expressed as: Pij +Pjk +Pki =3 2∀i<j<k Pi(i+1)+Pi(i+1)(i+2)+…+Pj(i−1)+Pji =j−i+1 2∀i<j(6) In support of the original guidelines set by Herrera-Viedma et al. (2004), Chen and Chao (2012) highlighted the necessity for a transform function to normalize values in a decision matrix that falls outside the interval [0,1]. This step is crucial to maintaining reciprocity and additive consistency. Utilizing a linear transformation, the values in the decision matrix will be contained within a broader interval [ − a,1+a]. Therefore, f :[ − a,1+a]→[0,1]allows for the following transformation function: f(Pk ij)=(Pk ij +a 1+2a)(7) 5. Empirical analysis 5.1. Selection of factors This study utilized four principal factors and 27 detailed factors identified by Sultana et al. (2019) to extract the risk factors associated with STS LNG bunkering operations. Additionally, ten detailed factors were added under the principal factor “Work process error” presented in Aneziris et al. (2022). These newly added factors were incorporated into an in-depth interview to select the most appropriate factors for the final questionnaire composition. In-depth interviews were conducted between January 4 and January 19, 2024, to determine the risk factors during STS LNG bunkering in South Korea. A risk factor composition table consisting of five principal factors and 37 detailed factors was delivered to the interviewees in advance. The interviewees were asked to identify items deemed important as risk factors. Additionally, interviewees were asked to freely suggest any factors they considered missing from the listed principal and detailed factors. Six experts from six STS LNG bunkering work-related organizations participated in the in-depth interviews. The expert profiles and organizational affiliations are listed in Table 3. Eighteen factors were confirmed as important by at least three of the six experts during the in-depth interviews. Each item was selected and revised to determine the final factors. Following these results, the detailed factor “Operation of the emergency shut-down (ESD) system” was added to the detailed factors during the in-depth interviews. Furthermore, reflecting the experts’ opinions, the name of the principal factor, “System error,” was revised to “System error of LNG supply & receiving vessel.” The detailed factor “Lack of communication between interdisciplinary teams” was also revised into “Lack of communication between interdisciplinary teams (including lack of communication between multinational sailors).” The risk factors determined through indepth interviews were five principal factors and 19 detailed factors, as shown in Table 4. 5.2. Results The factors derived from the literature and in-depth interviews were classified into five main factors and 19 sub-factors and used for the analysis. The CFPR questionnaire was developed to compare the main factors and sub-factors in the pairs. Table 5 summarizes the eleven LNG bunkering and logistics experts who participated in the survey. Y.-S. Choi et al. The Asian Journal of Shipping and Logistics 41 (2025) 52–60 58 Focusing on the principal factor rankings, the results are as follows: “System error of LNG supply & receiving vessel” (0.239), “Component error” (0.232), “Work process error” (0.198), “Human error” (0.172), and “Organizational error” (0.158). The highest ranked factors, “System error of LNG supply & receiving vessel” and “Component error,” do not refer to human defects, but those defects in the system directly related to bunkering work and risk; it is considered a high-risk factor. Contrastingly, “Human error” and “Organizational error” are direct humancaused problems that can be improved through safety education. There is no significant difference in the weight values, as shown in Fig. 1. Therefore, all elements of the principal factors were considered equally important. To fully comprehend the connotations of the analysis, it was necessary to revise the results for the detailed factors per principal factor. Regarding the “Component error principal factor,” the detailed factors were ranked as follows: “Control equipment malfunction” (0.313), “Pressure or level sensor malfunction in functionality, audibility, or visibility” (0.239), “Ship pump malfunction” (0.226), and “Detector failure” (0.221). Notably, both “Control equipment malfunction” and “Pressure or level sensor malfunction in functionality, audibility, or visibility” refer to risk factors that can lead to hazardous accidents due to the volatile LNG fuel characteristics, especially when related to malfunction. The results of the principal factors of organizational error are closely aligned in value, with “Lack of existence/implementation of accident prevention strategies (high wave, wind, ignition, dropped object)’ (0.269), “Insufficient preparation before the operation” (0.249), and “Lack of communication between interdisciplinary teams (including lack of communication between multinational sailors)” (0.242), respectively. The “Lack of existence/implementation of accident prevention strategies (high wave, wind, ignition, dropped object)” factor’s high-ranking value may be related to the effectiveness of response flexibility in emergencies when guidelines on how to prevent natural disasters and unexpected accidents are provided. There were no significant differences in the weight values. All the factors were deemed to be equally important. For the “Human error principal factor,” the results showed “The operator’s inability to recognize the operational condition, system status, or measuring instrument status” (0.651), and the “Operator’s inadequate performance due to high workload or work dissatisfaction” (0.349). “The operator’s inability to recognize the operational condition, system status, or measuring instrument status” weight of the factor is nearly twice as high as the second-place factor. Therefore, errors in system-wide situational judgments are directly related to risk. Attending to “System error of LNG supply & receiving principal factor results,” “Operation of the ESD System” (0.296), “Pipeline leakage” (0.272), “Electricity blackout, communication error” (0.228), and “Overpressure/overheating of equipment” (0.204) ranked closely. Furthermore, the factor weight values from 1st to 4th place showed no significant differences. As for the detailed factors corresponding to the “System error of LNG supply & receiving vessel error,” the overall risk is similar to the other detailed factors of this category. Lastly, the results related to the principal factor “Work process error” were in the following order: “Flame ignition” (0.245), “Over-fueling (only for fueled ships)” (0.195), and “Inadequate purging of the loading arm” (0192). Except for the highest-ranked factor, “Flame ignition,” the weight values of the remaining factors are similar. This factor’s standout might be attributed to LNG’s inherent risk of fire owing to the nature of gas fuels. Table 6 presents the results for global importance, highlighting the overall analysis value of the factors extracted in this study. Global importance can be derived by multiplying the importance value of the corresponding principal factor by each detailed factor and the importance value of the detailed factor to examine the comprehensive result. For example, the global importance value of the first detailed factor, the “Ship pump malfunction,” is 0.226, and the principal factor to which the factor belongs is the component error, whose value is 0.232. Therefore, the principal factor*detailed factor (0.226 *0.232) generates a global importance of 0.053 for the “Ship pump malfunction” detailed factor. Global importance is regarded as the most important result of a CFPR analysis because it identifies the overall ranking of all factors. Regarding the global importance analysis results, “The operator’s inability to recognize the operational condition, system status, or measuring instrument status” was ranked first at 0.112. The detailed factor falls under the principal factor of human error, a topic extensively covered in logistics risk-related studies. Despite the rise of Logistics 4.0 and the activation of automation systems in the logistics sector, certain aspects still require human intervention (Cimini et al., 2021). Accordingly, the occurrence of human error in the LNG bunkering area needs to be addressed. Giusti et al. (2019) conducted a study to reduce human errors when handling air cargo. Stefanova (2021) investigated factors concerning the human error that occurred in the logistics industry and studied the reason for the occurrence. Thus, despite advancements and technological changes, we found that “human error” remains significant in logistics, highlighting the necessity to enhance detailed logistics services in a constantly evolving external environment. The first and fifth factors are human-related risk factors. Although improvement methods have been applied, the recurrence of this factor indicates its ineffectiveness as a solution; therefore, it appears to be a high-ranking risk factor. In second place is “Control equipment malfunction” (0.073), which belongs to the “Component error.” This means that the second-place factor indicators that control equipment operations occur principally. Kader et al. (2015) mentioned that it is necessary to set elements for a safety model during LNG operation by setting the components required for LNG unloading. Pio and Salzano (2018) calculated the efficiency by deriving components to consider flammability, which is a characteristic of LNG. Therefore, setting the risk zone during LNG bunkering operations and controlling the LNG gas outflow is important. Malfunctioning of control equipment may have arisen as an important priority because of its direct relationship with accidents. In third place is the “Operation of the ESD System” (0.071). The ESD represents the automatic shutdown of the system in the case of an emergency during LNG operations. As LNG is a combustible material, there is always the possibility of fire and gas leakage (Gopalaswami et al., 2017). This detailed factor belongs to the “System error of LNG supply & receiving vessel” (0.239) and is an important risk factor that should be prioritized for LNG bunkering work. 6. Conclusion Following the IMO and EU environmental regulations, efforts in the shipping sector continue to move toward achieving net zero (zero carbon emissions) by 2050. Accordingly, various policies to reduce carbon emissions have been implemented, and further measures are being discussed. Some of these alternatives include slowing down the operation of ships, coating the hull, and improving equipment efficiency. However, to achieve a comprehensive reduction in carbon emissions, it is imperative to address ship fueling due to its significant contribution to overall carbon emissions. Consequently, LNG’s role as a bridge fuel has gained prominence, given its effectiveness in aiding the transition toward net-zero carbon emissions. Therefore, considering LNG’s potential as a bridge fuel, international studies on this topic are flourishing. However, this topic remains relatively unexplored in Korean domestic studies. Hence, this study identifies and prioritizes the risk factors associated with STS activities during LNG bunkering in Korea. Building on previous research and expert surveys, this study gathered 11 expert surveys. These surveys were based on five principal factors and 19 detailed factors, which were then analyzed using the CFPR methodology. Following the results of global importance, “The operator’s inability Y.-S. Choi et al. The Asian Journal of Shipping and Logistics 41 (2025) 52–60 59