Accident Analysis Modeling and Case Study of Hydrogen Refueling Station Using Root Cause Analysis (RCA)

Abstract

As the global transition to carbon neutrality accelerates, hydrogen energy has emerged as a key alternative to fossil fuels due to its potential to reduce carbon emissions. Many countries, including Korea, are constructing hydrogen refueling stations; however, safety concerns persist due to accidents caused by equipment failures and human errors. While various accident analysis models exist, the application of the root cause analysis (RCA) technique to hydrogen refueling station accidents remains largely unexplored. This study develops an RCA modeling map specifically for hydrogen refueling stations to identify not only direct and indirect causes of accidents, but also root causes, and applies it to actual accident cases to provide basic data for identifying the root causes of future hydrogen refueling station accidents. The RCA modeling map developed in this study uses accident cause investigation data from accident investigation reports over the past five years, which include information on the organizational structure and operational status of hydrogen refueling stations, as well as the RCA handbook. The primary defect sources identified were equipment defect, personal defect, and other defects. The problem categories, which were the substructures of the primary defect source “equipment defect,” consisted of four categories: the equipment design problem, the equipment installation/fabrication problem, the equipment reliability program problem, and the equipment misuse problem. Additionally, the problem categories, which were the substructures of the primary defect source “personal defect,” consisted of two categories: the company employee problem and the contract employee problem. The problem categories, which were the substructures of the primary defect source “other defects,” consisted of three categories: sabotage/horseplay, natural phenomena, and other. Compared to existing accident investigation reports, which identified only three primary causes, the RCA modeling map revealed nine distinct causes, demonstrating its superior analytical capability. In conclusion, the proposed RCA modeling map provides a more systematic and comprehensive approach for investigating accident causes at hydrogen refueling stations, which could significantly improve safety practices and assist in quickly identifying root causes more efficiently in future incidents.

1. Introduction

Globally, approximately 137 countries have pledged carbon neutrality by 2050, and major countries including Korea, Japan, and China, as well as the European Union, which announced the Green Deal containing carbon emission neutrality goals, have declared carbon neutrality [1]. In Korea, to participate in the international community’s efforts and create a healthy future, the ‘2050 Carbon Neutrality Declaration’ and ‘2050 Carbon Neutrality Vision’ were announced in 2020. Hydrogen’s role as an alternative energy to fossil fuels is significant in minimizing carbon emissions and achieving a carbon-neutral era. Hydrogen energy, as an environmentally friendly energy source, is an attractive energy resource that can simultaneously solve many problems such as global environmental pollution, global warming, and energy depletion [2]. Therefore, many countries consider infrastructure construction for hydrogen economy activation as an important development area, and hydrogen refueling stations play a major role in this infrastructure construction. In Korea’s case, as of 2024, approximately 220 hydrogen refueling stations have been built, and approximately 90% of hydrogen refueling stations supply hydrogen using tube trailers [3]. While Korea continues to build hydrogen refueling stations and is dedicated to establishing infrastructure for hydrogen economy activation, research related to accident and safety management in hydrogen refueling stations is significantly lacking.

Considering the hydrogen-related accident statistics released by the U.S. Department of Energy, among the 120 hydrogen accidents that occurred during 1999–2019, laboratory accidents accounted for the highest proportion at 38.3%, while hydrogen fuel supply stations (hydrogen refueling stations) and hydrogen-related commercial facilities accounted for 10.6% and 9.0%, respectively [4]. After examining the causes of these accidents, human error and equipment failure emerged as the main causes; in particular, equipment-related factors such as pipe/fitting/valve failures, storage device failures, and fuel cell component failures were reported to have high accident potential. In Korea’s case, from 2017 to June 2022, a total of 14 hydrogen-related accidents occurred [5]. Notably, the hydrogen gas-related accidents reported in the first half of 2022 were all incidents that occurred at hydrogen refueling stations, where charging hoses connected to ruptured hydrogen tube trailers or high-pressure hose connections were damaged, resulting in accidents where leaked gas from damaged pipes and accessories connected to hydrogen storage containers ignited.

The necessity for hydrogen refueling stations continues to emerge, and accordingly, the construction of hydrogen refueling stations is expected to increase rapidly. However, issues regarding the risks of hydrogen refueling stations and accident occurrences continue to arise. Therefore, the importance of preliminary risk assessment and safety design for the safe operation of hydrogen refueling stations is emerging. In the study of Park et al., after conducting a risk assessment assuming leakage from major facilities of hydrogen refueling stations built in urban areas, they evaluated that chargers for charging hydrogen to tube trailers and vehicles had the highest risk; additionally, tube trailers and refueling station operating devices (priority panel) were analyzed to have relatively higher risks than other facilities [6]. Further, various studies have conducted risk assessments for the safe operation of hydrogen refueling stations. Certain representative risk assessment studies have presented comprehensive risk assessment methodologies for hydrogen refueling stations, derived risks of on-site hydrogen production refueling stations, and examined safety distance calculation methodologies for hydrogen refueling stations [7,8,9,10,11,12]. These studies derived strategies for hazard identification, safety measures, and operational safety based on methodologies such as scenario-based risk assessment, CFD simulation studies, and probabilistic risk assessment.

In the field of system safety, various accident analysis models are utilized to clearly identify the causes that lead to accidents. Representative accident analysis models can be categorized into linear causation models, models reflecting human factors, and systemic accident analysis models [10]. Linear causation models view accident causes as results from a sequence of individual events occurring in a specific chronological order. Most traditional accident models such as domino theory, CCA, FTA, and ETA fall into this category. In models reflecting human factors, accidents are considered to occur through the combination of latent factors such as management functions and organizational culture with active failures. In these models, human error is the first factor causing accidents, directly interacting with process and technological regulations. Representative methodologies include bow-tie, threat and error management, human reliability analysis (HRA), and root cause analysis (RCA). Systemic accident analysis models analyze accidents by examining systems from a holistic (integrated) perspective beyond a partial understanding of systems, considering complex internal relationships, and explain accident causes in complex socio-technical systems as interactions between system components. Representative methodologies include AcciMap, system-theoretic accident model and process (STAMP), and functional resonance analysis method (FRAM) [13].

Among the aforementioned accident analysis models, the RCA technique is a systematic approach to identify the core of problems or inefficiencies and find the optimal way to solve them. As a series of working methods, various studies utilized it as an analysis method designed to organize identified facts and data to reach logical discoveries and conclusions through understanding the direct/root causes of actual or potential occurrences of events and accidents, and their precursors, situations, and conditions [14,15,16]. The RCA method can prevent or reduce the possibility of future recurrence by deriving effective corrective actions and countermeasures, and correct core process and system issues in a way that prevents future problems. However, few studies have applied the RCA technique as an accident analysis model for accidents at hydrogen refueling stations, and investigations into the causes of gas accidents in Korea primarily focus on identifying immediate causes rather than root causes. Therefore, this study aimed to develop an accident cause analysis model for automotive hydrogen refueling stations using the RCA technique, as one of the methods that could derive not only direct and indirect causes, but also the root causes of accidents when they occurred; compare the causes identified in existing accident investigation reports of actual accident cases with the causes derived using the RCA technique; and propose it for identifying root causes of accidents at hydrogen refueling stations in the future by applying it to actual accident cases.

2. Methods

2.1. Analysis Target: Element of Hydrogen Refueling Station

Hydrogen refueling stations are largely divided into on-site stations with hydrogen production facilities within the station and off-site stations that receive hydrogen from external sources. In Korea’s case, most hydrogen refueling stations, accounting for 90%, are off-site stations that receive hydrogen from external sources via tube trailers [3]. Therefore, we selected off-site stations as the analysis target. Figure 1 shows the overall equipment configuration diagram of an off-site station.

An off-site station primarily consists of hydrogen compression, hydrogen storage, hydrogen cooling, and hydrogen charging equipment. The hydrogen compression equipment is designed to compress hydrogen gas transferred from hydrogen cartridges at 3–20 MPa to 50–87 MPa, store it in medium- and high-pressure vessels, and then supply it to hydrogen vehicles at 70 MPa through dispensers. For the hydrogen storage equipment, the tube trailer cartridge serves as the primary storage facility with a maximum charging pressure of approximately 20 MPa; subsequently, the hydrogen gas compressed to 50–87 MPa through the hydrogen compressor is stored in medium pressure (Mid Bank, Type 1) and high-pressure vessels (High Bank, Type 1). Hydrogen vessels of type 1, 2, 3, and 4 are used in Korea, and most of the hydrogen vessels used in hydrogen tube trailers in Korea are predominantly type 1. The cascade control method is used to store hydrogen gas discharged from the compressor in medium- and high-pressure vessels. The hydrogen cooling equipment is designed for pre-cooling to reduce compression heat generated during the process of compressing and charging hydrogen into hydrogen vehicle containers (Type 1). The hydrogen charging equipment (Dispenser) is the facility for charging hydrogen into hydrogen vehicles. The interface between the hydrogen vehicle and the dispenser is critical to charge hydrogen into vehicles. For safe hydrogen charging, after determining whether hydrogen charging is possible through infrared communication modules and hydrogen charging protocols between the dispenser and hydrogen vehicle, the charging procedure begins [17].

2.2. Development of RCA Model in Hydrogen Refueling Station

This study first developed an RCA modeling map for hydrogen refueling stations, and then applied actual hydrogen refueling station accident cases to the map to validate the root cause model and identify the root causes. Subsequently, a comparative analysis was conducted between the main causes identified in existing accident investigation reports and the causes derived from the RCA modeling map (Figure 2).

2.2.1. RCA Steps

The RCA consists of four main steps, as shown in Figure 3. Step 1 is the data collection phase, which involves collecting data related to the event; this is the basic material for identifying causal relationships and root causes of the event. Step 2 is the causal factor charting phase, where the causal factor chart can systematically analyze collected information as a logical structure that explains the events leading to an incident and the conditions surrounding these events. Step 3 is the root cause identification phase, where after identifying all factors, the fundamental reasons for each causal element are identified using the RCA map. Step 4 is the recommendation generation and implementation phase, where after identifying the root causes for specific factors, recommendations are prepared to prevent recurrence.

2.2.2. Development and Validation of Hydrogen Refueling Station RCA Modeling Map

In this study, we aimed to develop a hydrogen refueling station RCA modeling map by referencing accident cause investigation data from accident investigation reports over the past 5 years related to the organizational structure and operational status of hydrogen refueling stations, and the RCA handbook. Based on that, we validated the root cause model of hydrogen refueling station accidents by applying hydrogen refueling station accident cases (Figure 4) [18].

The accident case was a gas leak accident at a hydrogen refueling station that occurred in Korea at approximately 08:17 on 28 January 2022, where at hydrogen refueling station A, a nut for connecting a high-pressure hose to the trailer outlet was damaged, resulting in a fire due to gas leakage that caused injuries to two people. Figure 3 is an illustration explaining the overall process of the accident.

3. Results

3.1. Development of Hydrogen Refueling Station RCA Modeling Map

Hydrogen refueling stations typically operate with small organizational structures of 3–4 people, focusing on hydrogen refueling and routine equipment inspection. When equipment failures occur, facility maintenance is managed by relying on external partners such as technical personnel from equipment manufacturers. Additionally, considering the causes of accidents at hydrogen refueling stations over the past 5 years, out of five accidents, three were due to product defects and second were due to worker errors. Although product defects occurring in the product design and manufacturing stages of hydrogen refueling stations have no direct causal relationship with station operations, these defects were incorporated into the RCA modeling map as they occur frequently [8]. The RCA modeling map of hydrogen refueling stations developed by integrating these factors is shown in Figure 5.

As shown in Figure 5, in the RCA modeling map for hydrogen refueling stations, the primary defect sources consist of three categories: equipment defects, personal defects, and other defects. The problem categories, which are the substructures of the primary defect source “equipment defects,” consist of four categories: the equipment design problem, the equipment installation/fabrication problem, the equipment reliability program problem, and the equipment misuse problem. Additionally, the problem categories, which are the substructures of the primary defect source “personal defect,” consist of two categories: the company employee problem and the contract employee problem. The problem categories, which are the substructures of the primary defect source “other defects,” consist of three categories: sabotage/horseplay, natural phenomena, and other.

By observing the second substructure, i.e., the root cause category, among the first substructure problem categories, the root cause categories that are substructures of equipment design problem consist of two categories, design specification and design review; the root cause categories that are substructures of equipment installation/fabrication problem, equipment reliability program problem, equipment misuse problem, company employee problem, and contract employee problem are related to and consist of equipment records, equipment reliability program design less than adequate, equipment reliability program implementation, administrative/management systems, procedures, human factors engineering, training, immediate supervision, communications, and personal performance. As a substructure of the root cause category, the near root cause exists; as a substructure of the near root cause, the final substructure exists, i.e., the root cause.

3.2. Hydrogen Refueling Station RCA Modeling Map Verification

3.2.1. Accident Cause Review

Here, a review of the accident causes from the accident investigation report is presented; the actual accident-related parts and tools are shown in Figure 6.

(1)

Based on CCTV footage from the accident scene, the statements indicated that the fire occurred the moment the container valve was opened after connecting the high-pressure hose. Judging from the worker’s position on CCTV, we presumed that the accident occurred during trailer replacement work.

(2)

The hydrogen charging line connection pipe nut in the accident fractured from the boundary between the groove and neck toward the nut neck direction. Based on fracture tests showing similar breakage patterns when tightening nuts with a wrench, and the discovery of deformation and cracks in 4 mm neck thickness nuts during site inspections, we presumed that the nut fractured because the neck thickness was too thin to withstand the tightening force. Additionally, a structural analysis showed that the stress applied to the nut owing to the difference in diameter between the outlet and nipple had less impact compared to the tightening force, suggesting that this was not a major cause of the nut fracture.

(3)

We confirmed that the tightening force was transmitted to the connection pipe nut’s neck, stress was concentrated at the boundary between the groove and the connection pipe nut’s neck, and the nut fracture also began at the boundary between the groove and neck.

(4)

Given that the color of the nut fracture surface varied in different areas and traces of shifted nipple contact positions were identifiable on the nut and nipple contact surface, we presumed that the connection pipe nut’s fracture occurred as deformation increased and cracks grew during repeated tightening processes.

(5)

While various factors such as friction heat, sparks, and static electricity could serve as ignition sources during high-pressure hydrogen discharge, the specific cause of ignition could not be determined.

(6)

The cause of manufacturing the 4 mm neck thickness nut could not be confirmed because statements from the connection pipe nut in accident manufacturer and the contractor did not match.

3.2.2. Cause of Accident

According to the accident investigation report, the accident was presumably caused by the nut being manufactured too thin, leading to deformation and cracking as it could not withstand the forces applied during each tightening. When these deformations and cracks reached their limit, the nut fractured due to the high pressure acting on the high-pressure hose at the moment the gas valve was opened, resulting in gas leakage. Furthermore, based on the discoloration traces found on the accident nut’s fracture surface and the pattern of shifted nipple contact surface in the nut neck area, we presumed that the nut’s deformation increased during repeated tightening processes.

The following problems were identified from the accident investigation results:

(1)

Nuts that were too thin could fracture as they could not withstand the tightening force. However, site inspections revealed that nuts were not standardized, with different thicknesses installed at different refueling stations.

(2)

There was a risk of nut deformation and fracture if workers tightened nuts with excessive force.

(3)

Without an emergency shutdown device installed on the tube trailer, there was a possibility of the incident developing into a major accident.

3.2.3. Application of RCA Modeling Map

Causal Factor Charting Analysis

The causal factor charting analysis results for the presented accident case are shown in Figure 7, and the corresponding content is shown in

Table 1. Causal factor charting results for accident cases.

No.	Contents	Shape	Elements
1	Hydrogen tube trailer supplier provides 10 mm neck-thick sample nuts to piping construction company (‘21.11)	□	Event
2	Piping construction company provides one 10 mm sample nut to the precision machinery company and requests the production of the same product (‘21.11)	□	Event
3	Manufacturing nuts by precision machinery company and delivering them to piping construction company (‘21.11)	□	Event
4	When the piping construction company constructs the hydrogen refueling station where the accident occurred, it is constructed using one nut supplied by a precision machinery company (‘21.11)	□	Event
5	High-pressure hose of the hydrogen chargingstation is fastened to the hydrogen tube trailer for hydrogen introduction into the hydrogen refueling station (‘22.01.28)	□	Event
6	After opening the inlet valve of the hydrogen refueling station, gas leakage and fire occur upon opening the container manual valve of the tube trailer. (2 people injured in fire.) (‘22.01.28)	◊	Accident
7	To be used as a connector nut for a high-pressure connection pipe to receive hydrogen from a hydrogen tube trailer to a hydrogen refueling station	○	Condition
8	Specifications of the high-pressure hose connecting to the tube trailer vary from station to station	○	Condition
9	High-pressure hose nut connecting to tube trailer is not legally standardized	⬡	Cause factor (root cause)
10	The precision machine company does not confirm that the sample nut requested to be manufactured by the piping construction company is suitable for the required specification	○	Condition
11	Failure to confirm that the requested nut specification and the delivered nut specification are different	○	Condition
12	Insufficient quality management system of the precision machine company	⬡	Cause factor (root cause)
13	The piping construction company does not conduct quality inspection on nuts supplied by the precision machine company	○	Condition
14	Failure to confirm that the requested nut specification and the delivered nut specification are different	○	Condition
15	Insufficient quality management system of the piping construction company	⬡	Cause factor (root cause)
16	The statement of the precision machinery company that manufactured the nuts and the piping construction company that received the nuts did not match, so it was impossible to confirm the 10 mm nuts’ delivery	○	Condition
17	As a result of the investigation, we confirmed that all nuts supplied by the precision machinery company to three other stations were approximately 4 mm thick, and all of them were cracked	○	Condition
18	During the construction of piping at the charging station, we did not confirm that the nut was a 10 mm product	○	Condition
19	No torque wrench was used when nutting the high-pressure hose to the tube trailer (engaged using a regular spanner)	○	Condition
20	The tools required for fastening work are purchased and used by tube trailer drivers, so safety such as explosion proof tools is not considered	○	Condition
21	No standards or procedures for fastening high-pressure hoses	⬡	Cause factor (root cause)
22	Workers lack awareness of the use of explosion-proof tools and work safety to prevent hydrogen explosion	○	Condition
23	Insufficient hydrogen safety and safety work training	⬡	Cause factor (root cause)
24	Ignition source was not confirmed	○	Condition
25	Locked using a manual valve attached to the tube trailer when gas leaks or fires occur due to nut breakage	○	Condition
26	The hydrogen tube trailer does not have an emergency shut-off valve	○	Condition
27	No legal regulations for installing emergency shut-off valves on hydrogen tube trailer	⬡	Cause factor (root cause)
28	In the event of a gas leak or fire, inaccessible for manual valve use depending on the situation	○	Condition

.

Causal factor charting displays events and causal factors chronologically at the bottom, while related conditions and information are listed and shown at the top. Among the analysis results, the parts marked in red text as cause factors correspond to root causes, and the direct contributing causes to the accident are marked with red border lines.

Root Cause Identification

According to the root cause analysis steps, the analysis results of root causes for the presented accident case are described as follows, referencing the causal factor charting analysis results in Figure 6 and Table 1, and the RCA modeling map shown in Figure 4.

(1)

The nuts of high-pressure hoses connecting hydrogen tube trailers to refueling stations were not legally standardized. Upon checking the nut specifications of high-pressure hoses installed at 90 hydrogen refueling stations nationwide, 19 stations had nuts with neck thicknesses of 4–8 mm, while 71 stations had nuts installed with thicknesses over 8 mm. Considering this, the lack of legal standardization for the nut neck thickness of high-pressure hose has resulted in different nut neck thicknesses being used at different stations.

(2)

The quality control system for purchases at piping construction company A was inadequate or non-functional. The piping construction company claimed that they received sample nuts with a neck thickness of 10 mm from the hydrogen tube trailer supplier for connecting to the tube trailer’s high-pressure hose, and requested precision machinery manufacturing company A to manufacture one with the same specifications. However, the precision machine manufacturing company A claimed that they received nuts with a neck thickness of 4 mm from piping construction company B and manufactured and delivered them according to those specifications. Given these conflicting claims between the two parties, there were evidently issues with the quality management system, such as piping construction company B’s failure to properly manage the product purchase process.

(3)

Precision machinery manufacturing company A’s quality control of raw materials and manufactured products was insufficient. Given that both parties’ claims about nut specifications differed after the accident, indicating that the precision machinery company failed to properly verify desired specifications when receiving the manufacturing order from the piping construction company, there were clearly problems with precision machinery manufacturing company A’s quality management system.

(4)

Hydrogen refueling station A and hydrogen tube trailer operators lacked standards or work procedures for connecting high-pressure hoses. When connecting high-pressure hose nuts to hydrogen tube trailers for hydrogen delivery to a hydrogen refueling station, tube trailer operators disregarded safety by arbitrarily purchasing and using non-explosion spanner tools and not using torque wrenches, indicating a lack of established standards or work procedures for hose connection work.

(5)

Hydrogen safety and safety work training was insufficient. Given that non-explosion tools were used and torque wrenches were not employed when handling hydrogen (i.e., a high-pressure flammable gas), we determined that safety training for hydrogen tube trailer operators and refueling station workers regarding hydrogen safety and safe work practices was inadequate.

(6)

Legal regulations for installing emergency shutdown devices on hydrogen tube trailers were not in place. While emergency shutdown devices should be installed to quickly shut off hydrogen gas in case of gas leaks or fire risks during loading and unloading operations for tube trailers, currently only manual operation valves are attached to hydrogen tube trailers. With manual operation valves, depending on the situation during an accident, access to the valve might be difficult, potentially preventing quick shutdown and leading to accident escalation.

The analysis revealed the following direct contributing causes to the accident:

(1)

Failure to verify the difference between the requested nut specifications and the manufactured and delivered nut product specifications. Both piping construction company A, which ordered the nuts, and precision machinery manufacturing company A, which manufactured and delivered them, overlooked the verification of nut specification accuracy during the ordering and product delivery processes.

(2)

Failure to verify different nut product specifications during piping construction at the hydrogen refueling station. During piping construction, there was a failure to reconfirm and construct according to the specifications for the relevant nuts.

(3)

Work was performed without using a torque wrench when connecting high-pressure hydrogen hose nuts to the hydrogen tube trailer. While a torque wrench should have been used to perform torque work according to the set torque for each nut specification during nut connection work, the nuts were damaged due to excessive tightening using regular spanner tools instead of torque wrenches.

4. Discussion

In this study, we developed an RCA modeling map to analyze the root causes of accidents at hydrogen refueling stations. Based on that, we derived improvement recommendations to prevent similar accidents in the future through data collection, causal factor charting, and root cause identification stages. Furthermore, by analyzing accident cases in domestic hydrogen refueling stations using the developed RCA modeling, we compared the accident cause investigation results specified in the accident investigation report with those applying RCA modeling.

The existing accident investigation reports identified three main causes and issues: (1) as nuts were not standardized, nuts with different thicknesses (4–10 mm) were installed at each hydrogen refueling station, and thin nuts with a neck thickness of 4 mm were damaged by wrapping them to withstand the fastening force; (2) in the fastening operation, a normal spanner for non-explosion was used without a torque wrench; and (3) because there was no emergency shut-off device installed, a major accident could possibly occur in the event of a fire in the hydrogen tube trailer.

Then, the RCA methodology identified nine causes and issues: (1) the high-pressure hose nut connected to the tube trailer was not legally standardized; (2) the quality management system of the piping construction company was insufficient; (3) the quality management system of the precision machine company was insufficient; (4) standards or procedures for fastening high-pressure hoses did not exist; (5) the hydrogen safety and safety work training for workers was insufficient; (6) legal regulations for installing emergency shut-off valves on the hydrogen tube trailer did not exist; (7) the requested and delivered nut specifications were not confirmed to be different; (8) during the construction of piping at the charging station, the nut was not confirmed to be 10 mm; and (9) no torque wrench was used when nutting the high-pressure hose to the tube trailer (engaged using a regular spanner).

The accident cause investigation results identified in the existing accident investigation report only included legal aspects, but the RCA modeling map analyzed all three aspects—equipment, personnel, and others—identifying not only legal factors, but also management aspects as root causes. After comparing these results, we concluded that when analyzing accident causes at the time of occurrence, the RCA methodology was a more systematic and effective approach to identifying the root causes of accidents.

Nevertheless, this study had certain limitations in identifying root causes, as we could not directly participate in accident investigations and relied solely on accident investigation reports. Additionally, common failure modes such as pressure rises and mechanical failure were not analyzed. Future research should integrate various accident cause analysis tools applicable to domestic gas facilities and systematically assess potential failure mechanisms to enhance the safety analysis of hydrogen refueling stations.

Additionally, to verify the practical applicability of the proposed RCA model, we consider it necessary to analyze it through the developed RCA model at the initial accident investigation stage rather than verifying it based on accident case reports. Further research focusing on the replicability of accident analysis methods is necessary to establish clearer standards for assessing consistency across different analysts.

In addition to improving safety management, revitalizing the hydrogen economy requires technological advancements and policy support to enhance commercialization potential and ensure economic viability [19,20]. Optimizing the hydrogen supply chain requires an integrated consideration of production, storage, and transportation technologies, as these technologies encompass diverse characteristics and cost factors. In this context, both government and industry must not only invest in technology development and infrastructure, but also conduct economic analyses and provide policy support to establish an optimal hydrogen supply chain [21,22,23,24,25].

5. Conclusions

In this study, an analysis was conducted using the RCA modeling map, which has never been applied to accidents occurring at hydrogen refueling stations. The analysis demonstrated that the RCA modeling map is effective in identifying the root causes of accidents. This finding provides valuable insights for safety management at hydrogen refueling stations and can be used as foundational data to prevent similar accidents in the future.

These findings emphasize the potential of the RCA modeling map as a systematic tool for accident analysis, paving the way for improved risk management and preventive measures at hydrogen refueling stations. By establishing a foundation for identifying root causes, this study contributes to enhancing the overall safety and reliability of hydrogen infrastructure.