The causes and responses to cargo hold fire accidents in RoRo ships using AcciMap

ABSTRACT The recent string of cargo hold fire incidents on RoRo ships represents one of the most challenging types of ship fire accidents. The densely stowed vehicle cargo loading structure and large volume of RoRo ship cargo holds are factors that complicate firefighting efforts. In addition, vehicle cargoes have a higher likelihood of self-ignition compared to other types of cargo and produce dense smoke when a fire occurs. Combustible materials inside vehicles can also contribute to the spread of fires after they start. Particularly, the recent increase in demand for electric vehicles calls for changes in how we respond to potential cargo hold fires on RoRo ships and heightens the risk of such incidents. In this study, we analyzed major domestic and international RoRo cargo hold fire incidents using AcciMap, a system analysis method. Through this analysis, we identified needs for improved management of vehicle cargo conditions, responses to fixed firefighting systems and system inspections, as well as improvements to fire detection systems. By using a systems analysis approach, we were able to identify areas for improvement needed by ships, ports, governments, shippers, and shipping companies.


Introduction
In March 2022, the car carrier ship Felicity Ace, en route from Germany to Rhode Island, USA, with approximately 4,000 vehicles on board, experienced a fire believed to have originated in its cargo hold.Following the incident, the vessel lost stability and eventually sank in the Atlantic Ocean while being salvaged by the M/V Bear.The property damage resulting from the sinking is significant, with the estimated value of the cargo of vehicles on board being roughly 400 to 500 million dollars (Allianz, 2022).
Recently, there have been consecutive cargo hold fire incidents on RoRo vessels similar to the above case, and the extent of the damage is significant.The characteristics of RoRo vessels and cargo make it even more challenging to extinguish cargo hold fires, thereby amplifying the scale of the damage.Let's examine some of the contributing factors to this escalation.
Firstly, the cargo holds of RoRo vessels have densely packed cargo configurations, making it difficult to extinguish fires when they occur and allowing for rapid fire spread.Secondly, the cargo holds, which occupy the largest volume on the ship, pose challenges for firefighting due to their size in the event of fire spreading.Their complex structure also increases the potential for casualties during firefighting efforts.Thirdly, vehicle cargo has a high propensity for self-ignition, and after a fire outbreak, the presence of flammable materials within vehicle fuel and general vehicle products generates smoke and further spreads the fire.
In addition, the increasing use of electric vehicles equipped with lithium-ion batteries, known for their relative difficulty in fire suppression, and hydrant vehicles powered by fuel cells, further complicates firefighting on RoRo vessels.Electric vehicles have seen a rapid increase in popularity, with sales expected to reach 1.45 billion units by 2030, accounting for 7% of all road vehicles (IEA, 2021).Despite these environmental changes in RoRo vessels, research on the causes and responses to cargo hold fires remains insufficient.Therefore, in this study, we aim to analyze major Korean and international cargo hold fire incidents on RoRo vessels over the past 10 years using the AcciMap system accident analysis model to investigate the causes and responses to these incidents.
In this study, we can systematically identify the main causes of RoRo vessel cargo hold fire incidents and their corresponding responses.This will enable us to contribute to enhancing laws, regulations, technologies, and systems related to cargo hold fire accidents on Ro-Ro vessels in the future.

Pctc(pure car and truck carrier) case
A RoRo vessel, abbreviated from Roll-on/Roll-off, refers to a type of cargo ship designed to transport wheeled cargo, and it is contrasted with LoLo (Lifton/Lift-off) vessels like container ships.The typical representative vessels of RoRo ships are commonly known as PCC (Pure Car Carrier) or PCTC (Pure Car and Truck Carrier), which are cargo ships specialized in loading bulk quantities of vehicle-shaped cargo.PCTCs have cargo hold structures focused on vehicle cargo, with onboard accommodations densely located on the uppermost deck of the ship.The primary fire suppression equipment for cargo holds is fixed CO2-based fire suppression systems.For case analysis, references were made to domestic records from the Korea Maritime Safety Tribunal's investigation reports and accident investigation reports, as well as international sources such as IMO GISIS, NTSB, other authorities' accident investigation reports, and relevant articles.The representative cargo hold fire cases identified through this research over the past approximately 10 years are listed in Tables 1 and 2.

Ropax(roll-on/Roll-off passenger ship) case
ROPAX, short for Roll-on/Passenger, is a type of ship commonly referred to as a passenger ship that can simultaneously load both passengers and vehicle cargo.Unlike PCTCs, when passengers board, most of the space on ROPAX vessels, except for the lower vehicle loading area, is designed as residential areas.The fire suppression equipment in ROPAX cargo holds has been used with water-based systems with sprinklers.

Analysis of causes of fire in cargo hold of RoRo ship
Based on the causes and responses to RoRo vessel cargo hold incidents, we conducted an analysis of the major factors contributing to large-scale fires.The analysis method employed the AcciMap, which is one of the systematic accident analysis methodologies.
AcciMap is a systematic accident analysis model introduced in 1997 (Rasmussen, 1997).In addition to AcciMap, there are other systematic accident analysis models such as FRAM (Hollnagel, 2004) and STAMP (Leveson, 2004).AcciMap classifies the causes of the resulting events into three main layers: the External level, which explains government, regulatory, and societal aspects; the Organizational level, which deals with the organizational systems behind the outcomes; and the Physical/Actor, Events, Processes, and Conditions level, which analyzes physical matters and actors  directly related to the accident (Branford et al., 2009).AcciMap arranges and connects various causes across these layers to identify the factors contributing to the occurrence of accidents.This allows us to go beyond the scene of the accident and identify the fundamental societal and technological systems that led to the accident.AcciMap has been used to analyze various incidents, including water contamination incidents (Vicente & Christoffersen, 2006), outbreaks of acute respiratory syndrome (Piché and Vicente, 2005), gas plant explosions (Hopkins, 2000), and the Glenbrook train collision (Hopkins, 2005).In the context of maritime transportation, AcciMap has also been applied to analyze the Sewol ferry disaster (Lee et al., 2017).
In this study, we conducted an analysis by integrating relatively common findings identified based on several similar incidents, rather than applying AcciMap to a single specific event.This approach allows us to increase the generality for preventing similar incidents in the future, rather than limiting the analysis to events unique to a specific incident.The sequence of AcciMap analysis is shown in Figure 1  ( Branford et al., 2009), and the results of the accident analysis are presented in Figure 2. Based on the analyzed accident results, the major causes of accidents and response issues are as follows.

Improper loading vehicle condition
The primary cause of cargo hold fires on RoRo vessels, especially in the case of PCTCs, often originated from used vehicles during loading.These incidents predominantly occurred during port operations or within several days after departure.Such characteristics can be attributed to defects in the used vehicles themselves, such as aging, but it is also suspected that prolonged stacking of vehicles for loading, followed by their mobilization for cargo handling, led to electrical shocks and overheating on board the vessel.Most of the fire origins in accident investigation reports were also presumed to have originated from cables and batteries within the used vehicles.Incidents of vehicle fires due to these causes have been quite common, and recent statistics on Korean vehicle fire accidents over the past five years also indicate that the majority of causes for used vehicle fire incidents, as shown in Table 3, were related to electrical and mechanical faults.Furthermore, fires occurring when vehicles were not in operation were also typical, with approximately 30% of vehicle fire incident locations being analyzed as empty or parking areas where vehicles were not in operation, as shown in Table 4.
In the case of ROPAX vessels, in addition to the mentioned cases, there were numerous instances where incidents were presumed to have occurred in equipment connected to special vehicles (such as refrigerated and frozen vehicles and vehicles used for transporting live fish).
Generally, vehicle cargo often contains various combustible materials, leading to its classification as hazardous cargo, which requires proper segregation.Reflecting this, internal combustion engine or batteryequipped vehicles are also classified as dangerous goods (UN No. 3166, 3171) under the IMDG Code (International Maritime Dangerous Goods Code).However, there are exemptions under this regulation for cases when these vehicles are loaded onto RoRo vessels, and there is no leakage within the vehicles, and precautions are taken to prevent short-circuits involving batteries, among other conditions.Nevertheless, in practical terms, checking these exemption criteria in the field is sometimes overlooked due to limitations in time and personnel, even at the organizational level.For instance, the Auto Banner incident in Incheon Port, South Korea, revealed the operational limitations when approximately 26% of unaffected vehicles were deemed at risk of fire following a fire incident, leading to decisions to abandon loading due to fire concerns.

Lack of understanding and structural limitations of fixed fire extinguishing systems
The primary fire suppression systems installed for cargo holds on RoRo vessels are designed to use either CO2 or hydrant systems.Fixed CO2 fire suppression systems work by increasing the concentration of CO2 within the enclosed space of the cargo hold, effectively lowering the oxygen concentration and achieving a smothering effect.However, in many cases of largescale fires in the cargo holds of RoRo vessels, even when these suppression systems were utilized, they often did not effectively suppress the fire.There are three reasons for this firstly, an adequate level of sealing to achieve effective smothering was not ensured.To enhance the smothering effect of fixed fire suppression systems, sufficient sealing is necessary.However, in many instances, there was a failure to allow sufficient time for sealing by focusing on the initial firefighting efforts before using the fixed fire suppression systems.
Secondly, there was a lack of understanding among the crew about the fixed CO2 extinguish systems.The application of fixed CO2 extinguish systems for cargo holds differs somewhat from their use in engine machinery spaces.To activate the fixed fire suppression system in a compartmentalized cargo hold fire, a lineup operation is required to direct the CO2 or hydrant to the relevant cargo compartment.However, it was identified that even managementlevel crew members had a limited understanding of this lineup procedure and device operation.Thirdly, there were many instances where the appropriate timing for the use of fixed fire suppression systems was missed.In particular, there was often ambiguity in the clear decision criteria for activating fixed CO2 fire suppression systems.In most cases, the use of fixed CO2 extinguish systems for fire suppression was only considered after initial firefighting efforts failed due to factors like thick smoke in cargo hold.This often resulted in missing the appropriate timing to shut ventilation for CO2 extinguish system.

Failure to detect fire in a timely manner
Initial firefighting is a crucial element in fire suppression.However, in many cases of cargo hold fires on RoRo vessels, the fire was not initially discovered by the FAS(Fixed Alarm System, thereafter FAS) but rather by crew members or land-based personnel.This signifies that the FAS is activated after the fire has already progressed to some extent, triggering an alarm.However, in the case of vehicle fires, when a fire that originated inside the vehicle has already spread outside, it usually means that the fire has already escalated significantly.Vehicle fires contain a greater amount of combustible material inside compared to outside, making firefighting considerably challenging once flames or flames are visible externally.Especially, recent incidents of thermal runaway in electric vehicle batteries highlight the urgency of timely fire detection.

Inadequate fire patrols
Connected to the issue of fire detection mentioned above, if early detection is deemed important while there are no mechanical improvements to the FAS, then fire patrols other than FAS may be crucial.However, RoRo vessels, due to their faster cargo handling capabilities, often face time constraints within the port, and the nature of unpackaged cargo demands more manpower for cargo verification tasks.This increased demand for manpower can lead to a workload burden after departure, making fire patrols more challenging and potentially disrupting early fire detection efforts.

RoRo ship cargo hold fire prevention and countermeasures
Based on the causes of cargo hold fires analyzed using AcciMap, cargo hold fire prevention and response measures are as follows.

Random checks of the condition of vehicles scheduled for shipment
The primary cause of most RoRo vessel fires is traced back to the vehicles being transported, particularly used cars.Vehicle cargo is classified as hazardous material, but it may not be considered as such under IMDG Special Provision 961.These exemption criteria primarily apply to cargo areas designated under SOLAS Chapter 2-2, Rule 20, which include vehicle spaces, special category spaces, Ro-Ro spaces, and areas specifically designed for the carriage of vehicles on the open deck of RoRo vessels.These criteria require that there should be no evidence of leakage, such as from batteries, engines, fuel cells, compressed gas cylinders, or fuel tanks, and that measures should be in place to prevent battery short-circuits.
However, when used cars are stored in open yards for extended periods before loading, there is often a high likelihood that these exemption criteria are not adequately verified.Nonetheless, inspecting all vehicle cargo before loading is economically impractical.Therefore, loading ports and shippers may conduct random inspections at an appropriate level (typically 5-10% of the total cargo volume) and consider expanded inspections based on identified findings.Such inspections can be conducted based on factors like the location where the vehicles were stored, the time period, vehicle model years, etc.Government authorities may also establish relevant regulations for these inspections.

Improvement of fixed fire extinguishing system
PCTC (Pure Car and Truck Carrier) ships have often failed in extinguishing fires in cargo holds compared to ROPAX (Roll-On/Roll-Off Passenger) ships.There are several reasons, but generally, the fixed fire suppression system of a PCTC requires an environment where the air inside the cargo hold can be completely sealed.For extinguishing fires in the cargo hold areas of RoRo ships, at least 45% of the maximum sealable volume must be sufficiently supplied with gas, and more than 2/3 of the required gas must be discharged within 10 minutes.Vehicle loading cargo spaces require over 50% additional gas compared to a typical cargo hold's required volume of 30%, indicating that more meticulous sealing operations are necessary.However, it can be observed that such ventilation operations are not sufficient in most cases of large-scale cargo hold fires.To improve some issues including these issues, the following countermeasures are proposed.
First, it is essential to secure a document distribution system that provides information about the location of vent closing devices indicated in the emergency muster list.Typically, the muster list onboard only displays tasks related to ventilation and sealing operations, without further education or training.Consequently, responsible personnel may not know the location of the sealing devices they are responsible for.This reliance on documented systems rather than hands-on training often places a significant burden on the competence of Chief Officers and Masters to plan and command fire extinguishing procedures.Therefore, it should be considered that documentation detailing the location of sealing devices to be sealed along with their operating procedures be stipulated in the safety system of shipping companies and legislated by international agreements and government regulations.
Second, enhancing initial firefighting capabilities is crucial.After discovering a fire, attempts are typically made to extinguish it initially using nearby fire extinguishers, followed by wearing breathing apparatus and re-entering for further firefighting efforts.However, re-entry often fails due to thick smoke and other hazards.Moreover, adequate sealing work is often neglected during re-entry.Therefore, it is necessary to strengthen emergency response capabilities to ensure that sufficient sealing work is carried out during re-entry.Given that most vehicle fires generate a large amount of smoke, immediate firefighting upon discovery becomes more critical.Therefore, increasing the number of portable fire extinguishers in vehicle spaces and prioritizing initial firefighting is effective.After the initial firefighting, re-entry should be conducted with the use of fixed fire extinguishing systems in mind.
Third, there should be an emphasis on education and simplifying diagrams to enhance understanding of fixed fire extinguishing systems.Unlike machinery spaces, additional line-up work is required for fixed fire extinguishing systems in cargo hold areas.However, crew members responsible for this task, even if they hold management positions, often lack sufficient understanding, leading to errors such as discharging agents into the wrong areas or insufficient discharge.To prevent these issues, individuals responsible for discharging agents should undergo additional training and evaluations to ensure they can respond to various fire scenarios.Additionally, safety systems should be reinforced to support these efforts.In particular, the labeling and diagram display on control panels of fixed fire extinguishing systems should be made more straightforward and intuitive.Utilizing labels with high visibility colors and numbers and adopting diagram formats that are simplified and immediately comprehensible can significantly improve the clarity of valve positions and operations based on the cargo area.These changes should be based on regulatory developments and corporate guidelines.

Improved fire detection systems and fire patrols
As previously mentioned, early detection of vehicle fires is crucial.However, in some incidents, it has been identified that the Fire Alarm System (FAS) did not activate until after the crew identified the fire.Typically, one of the most immediate ways to detect a fire when it occurs in a vehicle is by monitoring changes in temperature.Flames inside the vehicle are often not immediately detectable, and smoke detection can be delayed due to air circulation within the cargo hold.However, heat detectors installed on ships must receive heat to activate, which means that the fire needs to have progressed significantly for these systems to operate.Therefore, most cargo hold areas are equipped with smoke detectors, which have a relatively high detection rate.Consequently, detecting fires originating inside the vehicle itself remains somewhat challenging at the current level of system.To address this issue, the following methods may be required: First, to enhance the speed of vehicle fire detection, the installation of thermal imaging monitoring systems using infrared technology may be necessary.These systems can measure temperature changes, allowing for the identification of fires even when they occur at a distance.This enables the detection of fires before they spread significantly.
Second, increased fire surveillance is required in areas where the frequency of fire incidents is high, such as within the port or during the 1-2 days after a vessel departs from the port.If immediate installation of thermal imaging monitoring is difficult, portable monitoring equipment can be used for enhanced fire patrols during periods of heightened fire risk, using personnel to carry out these patrols.

Conclusion
RoRo vessel cargo hold fire incidents are distinct from typical maritime accidents as they result in losses not only to the ship but also to all the loaded vehicle cargo.Despite the presence of areas where adequate response is possible, many incidents escalate into major accidents, highlighting the need for improved identification of causes and responses.Accordingly, this study analyzes major cargo hold fire incidents in RoRo vessels over the past 10 years using the AcciMap method of system accident analysis and proposes the following causes and improvement measures: Firstly, a significant portion of accident causes can be attributed to inadequate loading of vehicles.To prevent this and reduce excessive labor and time consumption, we propose random detailed inspections of scheduled loading vehicles.
Secondly, incorrect responses through fixed fire extinguishing systems after an accident made firefighting difficult.To address this, we recommend providing additional information to crew members responsible for ventilation and sealing operations, considering ventilation and sealing from the initial firefighting, ensuring an effective training system for personnel responsible for the lineup of fixed fire extinguishing systems, and simplifying relevant diagrams for better comprehensibility by all.
Thirdly, since fire detection through fire detectors is often delayed, we propose the use of thermal imaging monitoring systems using infrared technology to address this issue.Additionally, deploying additional fire monitoring personnel during periods with a high incidence of accidents is suggested.
Recognizing that accidents are not attributed to a single problem, this study used the AcciMap method to analyze accidents from multiple perspectives.The proposed improvements encompass not only shiprelated enhancements but also improvements involving shipping companies, cargo owners, ports, governments, and more.For this study, we used data from mostly credible institutions, but there were cases where information was limited or omitted.Therefore, there may be limitations in that unidentified facilities and systems may exist.Nevertheless, presenting actionable improvement suggestions based on identified findings, this study contributes insights into preventing cargo hold fire incidents and can aid future legal and systemic improvements.

Table 1 .
The status of the fire accident in the cargo hold of PCTC vessel.

Table 2 .
The status of the fire accident in the cargo hold of ROPAX vessel.