Executive summary

What happened

Late evening 29 May 2021, a Leonardo Helicopters AW139, registered VH-TJK, operated by Toll Helicopters, was tasked as single-pilot helicopter emergency medical service (HEMS) flight ‘Rescue 208’. Onboard the helicopter were 2 aircrew and 2 medical crew.

During the later stages of the approach into an unimproved helicopter landing site, the aircraft’s tail rotor struck a small tree. The contact was not identified by the crew. Having assessed the landing site as unsuitable, the crew discontinued the approach and diverted to a landing site about 1 km away. After shutting down, the flight crew conducted a walkaround inspection of the helicopter and identified evidence of foliage contact on the vertical fin.

What the ATSB found

The ATSB found that while manoeuvring to land within a confined area, unintended yaw and drift of the helicopter was not identified by the crew and stopped prior to the tail rotor striking a tree.

What has been done as a result

The operator has completed the following proactive safety actions:

• amended their operational guidance on minimum clearances from terrain when operating in confined areas
• issued guidance on site selection during primary missions
• a final internal safety report was provided to the ATSB and proactively shared among the emergency helicopter network
• installed the A800 Trakkabeam high‑intensity searchlights onto the fleet of aircraft.

Safety message

This incident highlights the need for flight crew to have a heightened situational awareness when operating into a confined area and unfamiliar location in the vicinity of obstacles, as there is very little to no margin to recover from any unexpected event(s).

Crew coordination plays a vital role in HEMS operations and ensures improved situational awareness, reduced errors, and the fostering of effective teamwork. Effective coordination and communication (including of concerns) minimises the risk of misinterpretation, ensures accurate transmission of information, and reduces the likelihood of mistakes.

The investigation

The occurrence

On 29 May 2021, at about 2325 local time, a Leonardo Helicopters AW139, registered VH-TJK (callsign ‘Rescue 208’) and operated by Toll Helicopters, was tasked to conduct a helicopter emergency medical service (HEMS)[1] flight. The flight was planned under night visual flight rules (NVFR)[2] supplemented with the use of a night vision imaging system (NVIS).[3] The aircrew comprised a pilot and aircrew officer (ACO). The medical crew included a New South Wales health department (NSW Health) flight paramedic and an emergency specialist doctor.

The NSW Health aeromedical control tasked Rescue 208 to transit from Bankstown Airport to Shelly Beach (Figure 1) to assist with retrieval, stabilisation and transport of a patient to the Royal North Shore Hospital, NSW. After accepting the task, conducting pre-flight planning and briefing on the operation (including a planned winching retrieval), the pilot started the helicopter engines at about 2343.

Figure 1: Aircraft flight path

Source: Google Maps annotated by the ATSB

At about 2348, the aircrew and flight paramedic transitioned to using night vision goggles (NVG) and about a minute later, Rescue 208 departed Bankstown Airport, with the ACO in the rear cabin of the helicopter. While outbound at 1,500 ft, the paramedic was in contact the ground crew at the scene while the aircrew completed their checks. The aircrew had difficulties hearing the paramedic due to issues with the internal communication system.

The paramedic was informed by the scene commander on the ground at Shelly Beach that the plan had changed and was now for the crew to land in a playing field at Bear Cottage (Figure 2), instead of winching at the site. However, during that communication exchange, a second, closer landing option near the carpark at Shelly Beach was proposed, with the crew being advised that a HEMS helicopter had previously landed in the area (see the section titled Helicopter landing site).

Figure 2: Helicopter landing site options

Source: Google Earth annotated by the ATSB

Between 2352 and 2354, during the cruise, the doctor and flight paramedic discussed the equipment required for the rescue operation. A preparatory briefing was then conducted between the flight paramedic, the doctor and the ACO. At about 2355, the pilot acknowledged the briefing and subsequently briefed the crew on the weather, stating that there were light showers and the visibility was 5 km. Thirty seconds later, the pilot conducted the pre-landing checks, including lowering the landing gear, configuring and selecting the landing/search lighting system to ON (see the section titled Lighting and scan technique).

The aircraft arrived overhead Manly Beach at about 700 ft. The pilot then turned the helicopter south to parallel the beach and continued descent towards the proposed alternate landing area, while slowing the helicopter to an airspeed of 80 kt (Figure 3).

Figure 3: Aircraft flight path over Manly
 

Source: ATSB

At 2358, the pilot cleared the ACO to open the rear doors and continued descent while slowing to 60 kt. The ACO opened the right rear door and subsequently selected the winch power to ON, which also turned on the winch light. The pilot completed the pre-landing checks. The crew then received a radio call from the scene commander explaining the status of the patient, with a request that they land in the area near the Shelly Beach carpark if possible. The crew advised that they would assess that location and, if unable to land, they would reposition and land at Bear Cottage. About 30 seconds later, the helicopter was hovering at about 300 ft radio altitude (RA) adjacent to the proposed helicopter landing site (HLS).[4] From this position, the aircrew located the proposed landing area, primarily using the white search light rather than NVG.

At 0001, the pilot established the helicopter in a lower hover (between 100–130 ft RA) to assess the landing area. The ACO verbalised a description of the area, including the size, obstacles (advising that there were no power lines and that the trees were small), and available access. An approach plan was then discussed and the ACO then confirmed with the pilot and flight paramedic that they were happy to continue into the area. The pilot then handed over verbal control of the helicopter movement to the ACO, who then had responsibility for assisting the pilot by guiding them into the confined area HLS.[5]

The left rear cabin door was opened at about 0002 by the paramedic, who took up a position to ensure obstacle clearance to the left and rear of the helicopter. About 10 seconds later, the helicopter commenced an approach to the HLS on a heading of about 177°.

At 0003, as the helicopter descended, the ACO called ‘well clear of trees’ and guided the pilot through the movement of the tail through 20° right before stating ‘tail well clear’. The pilot then continued the descent, while undetected the helicopter yawed (see the section titled Helicopter movement) about 10° left to 145°. About 5 seconds later while at around 13 ft, the pilot asked the ACO about the slope of the site to which the ACO responded that it was ‘flat, or maybe a little bit nose to tail’. The flight paramedic responded to the comment of the slope stating it was ‘pretty heavy left’, this was not acknowledged by the other crew. During this conversation, the aircraft continued to gradually yaw further left.

Noticing the movement, the ACO told the pilot that they were drifting right and to move left. The ACO then advised that the tail was clear before again telling the pilot they were drifting to right. The ACO subsequently assessed that the helicopter was closer to the trees than they expected and called ‘climb climb climb’ and when the pilot did not respond, repeated the instruction. The pilot immediately responded to the second instruction and climbed the helicopter vertically to 100 ft (RA). Given the proximity of the trees, the ACO considered the possibility that the helicopter may have contacted them, however there had not been any indication of airframe contact (such as increased vibration).

Once out of the confined area, the pilot mentioned that they were happy to attempt another landing and the ACO agreed. The ACO did not mention the possibility the aircraft might have contacted the tree. However, the flight paramedic reiterated the increased slope on the left side of the aircraft and, after checking a second landing site in the car park, the crew diverted and landed at Bear Cottage.

After shutting down, the ACO and pilot conducted a walkaround inspection of the helicopter using hand-held torches. While inspecting the vertical fin, green material indicating contact with foliage was identified (Figure 4). The pilot then briefed the medical crew that the aircraft would be offline and notified the operator’s management of the incident.

Figure 4: Evidence of foliage contact on the vertical fin

Source: Toll Helicopters ACT/NSW
 

Context

Helicopter personnel

Pilot

The pilot had previously been employed by the operator for about 2 years from 2017–2019 and had recently returned to the role. They were cleared back to line flying on 29 April 2021. The pilot had a total of 4,594.8 flying hours, of which 1,219.7 were on the AW139. They had a total of 174.7 hours NVG flying time and since their return they had completed 16.1 NVG hours. An NVG/NVIS Capability Check Flight (CCF)[6] was completed 14 April 2021.

Aircrew officer

The ACO had 5,691 flying hours of which 2,001.1 were on the AW139. The ACO had 459 hours of NVG time at the time of the occurrence. They had completed an NVG/NVIS CCF on 19 May 2021.

Medical crew

The flight paramedic had previous experience in HEMS operations while working with another operator from mid-2015 to 2018. They started working with the operator in 2018 in road operations and in 2019 transitioned to flight paramedic. The flight paramedic was using NVG on the occurrence flight and had completed 3 hours of NVG flying within the last month.

The doctor had worked as an aeromedical doctor with another operator between August 2020 and February 2021 before joining the operator. At the time of the occurrence, they had about 10 months experience working in the HEMS operational environment. The majority of operations experienced during this time did not include night operations. The doctor was not using NVG on the occurrence flight.

Aircraft information

General

The Leonardo S.p.A AW139 is a medium-sized, multi-role helicopter, powered by two Pratt & Whitney Canada PT6C-67C turboshaft engines. The aircraft has a length of 16.66 m and a rotor diameter of 13.8 m. The AW139 aircraft is primarily used for emergency medical services (EMS) within Australia.

The tail rotor of the AW139 is a critical component of the aircraft’s flight control system, providing the necessary lateral force to counteract the torque generated by the main rotor system. Figure 5 outlines the aircraft dimensions and the area of foliage contact.

Figure 5: Area of impact with tree

Source: Leonardo Helicopters (Augusta Westland 139), annotated by the ATSB

Helicopter movement

Two important aspects of helicopter movement, especially in the context of hover operations are yaw and drift. Yaw refers to the rotation of the helicopter around its vertical axis. This is controlled by input to the tail rotor, which generates a lateral force to counteract the torque produced by engine/s driving the main rotor. The amount of yaw is adjusted by the pilot using the anti‑torque (tail rotor) pedals, which control the pitch (and therefore thrust) of the tail rotor blades. Yaw is important for controlling the heading of the aircraft and for maintaining balance in forward flight.

Drift refers to the (usually) unintended lateral movement of the aircraft as a result of wind/turbulence, flight control inputs and lateral tail rotor thrust (particularly in hovering flight). Drift can occur without the aircraft yawing and vice versa.

The Federal Aviation administration (FAA) Helicopter Flying Handbook (2019) Chapter 2 - Aerodynamics of Flight emphasises the importance of understanding and controlling both yaw and drift.

Active Vibration Control System

VH-TJK was fitted with an active vibration control system (AVCS). The system worked by sensing vibrations of the aircraft rotor system and using advanced algorithms to automatically adjust the pitch of the main rotor blades to reduce the amplitude of the vibration. The system could adjust the pitch of the blades up to 20 time per second, allowing it to quickly respond to changes in vibration levels.

External lighting

The aircraft was fitted with two pilot-steerable landing (search) lights mounted on the underside of the helicopter. Additionally, the aircraft was fitted with an ACO-steerable hoist light and each crew member had handheld torches available.

At low level, the operator required crews to make use of the external lights (white light), to assist the detection of wires and other obstacles and maintain terrain separation. This required the crew to conduct their scan with both the use of NVG and without.

Visual scan technique

Pilot

The pilot advised that during the approach their NVG goggles were in position, however they were also looking underneath and around them to use the white light to scan, and then doing a general scan through their NVG. To ensure they had adequate reference points, the pilot had one light directed to the front of the helicopter, with the second directed in the 3 o’clock[7] position.

The pilot also reported using lit houses in front and to their right, as reference points. During the descent, they used a dead tree in front of the helicopter and another identifiable tree to their right. They then transitioned to using ground references visible through the helicopter’s clear chin bubble. The pilot identified that the ground was sloping and decided they would assess the magnitude of the slope when they got to a low hover. The pilot also stated that they thought the inadvertent yaw occurred during a reduction in power associated with lowering the collective. They assessed this most likely happened when they brought their gaze down to the ground and were not using the identified trees as references. They did not believe that the unintended yaw was due to inadequate illumination of the confined area.

Aircrew officer

The ACO advised they used both NVG and the winch light to identify visual cues when descending into the HLS. They reported keeping the closest obstacle, being the contacted trees, to their right so they could observe and avoid them. They estimated the helicopter was 15‑20 ft away from the trees during the initial descent. The ACO moved from the right to left inside the helicopter and also lay on the floor to check underneath for clearance from obstacles. While monitoring the ground clearance, the ACO looked up and identified that the tail of the helicopter was significantly closer to the trees than expected, assessing that they may have misinterpreted yaw as drift when lying on the floor.

The ACO stated that they did not observe an excessive ground slope and were surprised by the paramedic’s assessment. They also advised that they were not relying on the paramedic to give advice on the helicopter clearance from objects. Finally, the ACO advised they did not consider visibility of the HLS was an issue, assessing the conditions as a relatively high visibility night.

Paramedic

The paramedic advised they were positioned on the left side of the helicopter during the approach to the confined area and, while they had NVG, they had better vision using the white light and a handheld torch. They advised that the role of the paramedic was to give negative clearances when asked, and to identify and call out potential hazards in the area. On the night, the paramedic recalled having a large clearance on the left side of the tail and they could see the ground dropping off behind the tail in the white light. They recalled that due to the references outside the aircraft they thought the aircraft had drifted.

The paramedic assessed that the underbelly lighting of the aircraft was not ideal for landing into the HLS and that with more white light the crew might have been able to identify more of the terrain.

Briefings

The crew arrived at the base at approximately 1930, for an overnight shift. They completed a shift hand-over, checked the local weather conditions and conducted a pre-flight inspection of the helicopter. The crew then went to bed at about 2200.

At 2325, the paramedic received a phone call for a priority job and woke the pilot. The pilot conducted a pre-flight operational risk assessment considering the:

• weather that included moderate to strong winds, turbulence and showers
• time of night
• moon illumination
• possibility of a winch.

The pilot determined the risk was within the appropriate levels to conduct the operation and woke the doctor and ACO.

The ACO completed an online reconnaissance of the area, to assess possible landing locations. The pilot completed the aircraft’s weight and balance and performance calculations and confirmed they allowed for an out of ground effect hover and winch, without the need for internal configuration changes.

During the transit to the scene a brief was completed, which included the crew duties, available equipment, possible landing locations, and weather.

The operator’s operations manual outlined that when a crew was operating to an unknown HLS, NVIS crews were required to conduct a thorough reconnaissance of the landing area with a white light prior to committing to an approach. This could be done either while conducting an orbit over the area or during a high hover over the site.

The crew completed 2 hovers, a high hover at approximately 300 ft and a low hover at approximately 100 ft (Figure 6).

Figure 6: Hover heights prior to entry into HLS

Source: ATSB

As a minimum, the reconnaissance and brief was required to assess the required power, wind, any obstacles and a plan for the approach and departure, typically based on the acronym PSWATP:

• P – Power available/required and therefore performance margin
• S – Size, shape, slope and surrounds
• W – Wind direction, strength and any turbulence
• A – Approach profile, departure and overshoot options
• T – Terrain, turbulence relevant to the area
• P – The plan, including crew duties, based on the reconnaissance

The pilot was to give a brief on the relevant information of the PSWATP briefing requirements and the ACO was expected to contribute to the reconnaissance.

Table 1 outlines the 6 sections of the PSWATP and at what point during the mission they were completed and by whom.

Table 1: Completion of Briefings

Section of PSWATP	Prior to departure	During flight	During reconnaissance hover
Power available/required and therefore performance margin	Pilot – completed the aircraft’s weight and balance and performance calculations. They also analysed the hover out of ground effect and confirmed it was acceptable for the mission.
Size, shape, slope and surrounds			ACO – discussed size, shape, and obstacles. Pilot and Paramedic – confirmed ACO observations.
Wind direction, strength and any turbulence	Pilot – conducted a pre-flight risk assessment based on the weather that included moderate to strong winds, turbulence and showers, time of night, moon illumination and a possible NVIS winch	Pilot – communicated information on wind and rain	Pilot – confirmed wind direction with crew
Approach profile, departure and overshoot options			Pilot and ACO – discussed initial approach into HLS
Terrain, turbulence relevant to the area	ACO – completed an online reconnaissance of the area, to assess possible landing locations		ACO – described the terrain to the crew
The plan, including crew duties, based on the reconnaissance		Paramedic & Doctor – Discussed crew duties and what equipment would be brought based on a winching	The crew – discussed the planned approach and the revised plan for each crew member after receiving an update on the patient’s condition

Helicopter landing site

The landing site was adjacent to Shelly Beach carpark and was an ‘unimproved’ or ‘non‑conforming’ HLS.[8] The open area was about 21.5 m wide and 41.5 m at its longest point (Figure 7).

The air crew and the flight paramedic recalled the area was a confined area and a ‘tight fit’, however after a low-level reconnaissance at 100 ft, they determined that although there were some obstacles (foliage and trees) in the area, it was a suitable place to land.

When manoeuvring into the landing site, the ACO did not communicate the proximity of the obstacles at the rear of the aircraft and as such both the pilot and the flight paramedic were unaware of the hazard they posed.

The HLS had a slope of 10° towards the west, which was undetected by the ACO (Figure 7). The pilot advised that slope was difficult to detect using both NVIS and white light. The operator’s operations manual indicated that the slope limitation for the helicopter was 10° in all directions. The pilot advised that they had intended to check the slope in a low hover and even if there had been no other concerns, they may not have landed in the HLS due to the slope.

The operator required a minimum safety distance of 10 ft laterally around the main and tail rotor disk, with aircraft fuselage to be maintained clear of obstacles however, 20 ft lateral separation was preferred. A minimum safety distance of 6 ft vertically below the rotor disk was recommended and landing with any obstacle under the rotor disk was to be avoided wherever practical. Under night operations, crews were asked to increase safety margins depending on the situation, aircraft configuration, operating crew and environmental conditions.

The operator stated that once the aircraft was in a low hover in a confined area, manoeuvring should be minimised as it was difficult to maintain adequate visibility and obstacle recognition in all directions, particularly rearwards.

Figure 7: HLS slope and obstacle location

Source: Toll Helicopters annotated by the ATSB

Figure 8 shows the approach and intended landing directions. The plan was that when the helicopter was established in a low hover over the landing site, the nose would be turned right so the tail would fit in the cut-out area.

The ATSB did not attend the site, however the operator’s report advised that:

At night, with an approach to the south, it would be difficult to land in the pad due to the slope at the south-eastern corner requiring the aircraft to be close to the obstacles on the non-active side of the aircraft (the side opposite to where the PIC [pilot in command] and ACO are operating). However, at the time of the incident, the crew believed that they could achieve the required minimum operating procedure of a 10-foot obstacle clearance. This assessment reduced the margins of any drift or yaw as experienced by the incident crew. Had the crew planned to achieve the preferred 20-foot obstacle clearance, it is likely that the contact with the tree may not have occurred.

The operator confirmed that the same confined area had been used for a task by day, in good conditions, on 26 January 2021, however it was approached from the reciprocal direction. The reciprocal direction allows a larger landing area with more obstacle clearance due to the reduced slope.

The pilot advised the operator that the decision to approach the HLS was influenced by the information that a company helicopter had landed at the site previously. However, they also advised the ATSB that they were not influenced by this information when assessing the HLS. The paramedic recalled that the information played a part in deciding to go to the HLS. The ACO, recalled there had been an aircraft in the HLS previously, however, they did not state if that affected their decision to whether to conduct the approach.

Figure 8: Approach and intended landing orientation

Source: Google Earth annotated by the ATSB

Weather

Prior to the flight, the pilot reviewed the weather using the Bureau of Meteorology (BOM) meteorological aerodrome report (METAR) and terminal aerodrome forecast (TAF) for Sydney Airport. At the time of the incident, the METAR showed the wind from the south-south‑west at 18 kt, 10 km visibility, showers of rain in the vicinity, few clouds[9] at 1,800 ft and broken cloud at 3,300 ft.

The astronomical conditions at the time of occurrence included moon illumination of about 87%. The moon angle was at 50.03°, which produced a shadow length of about 25 m when at 100 ft.  There was a possibility of reduced moon illumination due to the extent of the cloud and localised light rain in the area. The low angle of a rising or setting moon may reduce contrast detail and create strong shadowing effects which can mask hazards when operating under NVG. The crew reported ambient lighting generated from surrounding residential properties was present on approach to the HLS.

The operator’s operations manual states the minimum visibility for NVIS flight was 5,000 m and there should be no more than scattered cloud below 2,000 ft above ground level (AGL) within a 2 NM corridor either side of track. The weather at the time of this occurrence was above the minimum requirements for NVIS flight.

Flight data

The aircraft was fitted with a multi-purpose flight recorder (MPFR).[10] The MPFR data was found to contain the occurrence flight from Bankstown Airport to Bear Cottage, however the audio recording of the occurrence had been overwritten. Additionally, the operator provided the ATSB recorded rear cabin footage and the associated audio.

The MPFR data indicated that in the 30 seconds prior to the ACO instructing the pilot to climb, the aircraft experienced a 30° yaw to the left, a 2 ft lateral drift to the right followed by an 8 ft drift to the left, and a 3.5 ft longitudinal drift rearwards before a slight forward movement of 1 ft, as outlined in Figure 9.

Figure 9: Aircraft flight data

Source: ATSB

Training

Australian regulations required a minimum 6-month NVIS recency interval. However, the operator implemented a requirement for additional NVIS training, specifically ‘complex operations’ to confined areas, including winching in areas devoid of cultural lighting,[11] at least every 3 months. Further, 6-monthly recency flights were to be conducted with a training and checking pilot and/or ACO ‘to improve standardisation, enhance crews comfort levels and further develop the skills and knowledge required to operate on NVIS’.

The operator advised that the choice of NVIS training and checking locations was constrained by several factors, including proximity to the operator’s base (due required response time), the permission of various landowners, the presence of hazards, and to ensure sites were devoid of cultural lighting to maximise training effectiveness. As such, the operator also used simulators for recurrent pilot and ACO training and during clearance-to-line training. This included low level flying, winching, and manoeuvring into confined areas. Simulators were also used for ‘complex operations’ training for both pilots and ACOs.

Safety analysis

During the cruise, the crew were given 2 options for landing locations. The occurrence HLS was explained as a site where the operator had previously landed and had the advantage of being closer to the mission location. The advice that a company AW139 had previously landed at the site may have also influenced the crew’s decision to conduct the approach to that location.

The crew completed a reconnaissance in a hover, however they did not conduct an orbit of the site. Although not required by the operator’s procedures, an orbit would have allowed the crew to view the site from different angles and may have enabled them to identify the sloping ground, which was on the limits of the slope allowed for the helicopter type. It would also have provided an opportunity to fully assess the extent of the obstacles/available clearances. However, the crew advised that due to the prevailing weather conditions and avoiding the residential area close‑by they did not conduct an orbit.

Following the reconnaissance, the crew completed a briefing in the hover before approaching the HLS. They discussed the size and shape of the proposed HLS, they also discussed the approach and obstacles, ensuring all members of the crew were consulted. Although the non-standard HLS met the requirements of the operators 10 ft obstacle clearance, the crew were aware it was tight. As such, only small, controlled helicopter movements were allowable once established within the confined area.

During the approach, the communication between the ACO and the pilot was continual and clear. Additionally, once established inside the confines of the landing area, the ACO actively moved within the rear cabin to assess the lateral and vertical clearance to obstacles. Despite that, unintended movement of the helicopter – both yaw and drift (lateral and rearwards) was not identified by the crew and stopped prior to the tail rotor contacting foliage.

The crew did not identify the tail rotor strike until after the helicopter was shutdown at the alternate landing site. This may have been due to the relatively light foliage that was contacted, with the active vibration control system also possibly dampening any rotor vibrations from the collision.

The ACO suspected that the tail may have contacted the tree during the initial approach however, there was no physical confirmation in the aircraft airframe of this occurring. Despite this, if the ACO had advised the rest of the crew that a  strike may have occurred, it would have allowed discussion and informed decision making on subsequent actions.
 

Safety actions

Whether or not the ATSB identifies safety issues in the course of an investigation, relevant organisations may proactively initiate safety action in order to reduce their safety risk. The ATSB has been advised of the following proactive safety action in response to this occurrence.

Safety action by Toll Helicopters

The operator has completed the following proactive safety actions:

• The operator has amended their operational procedures stipulating aircrew maintain the following minimum clearances from terrain when operating in confined areas:
  • 20 ft from the main and tail rotors
  • by day only, 10 ft from main and tail rotors if operationally necessary
  • for obstacles below the main rotors, 6 ft vertically
  • 3 ft from aircraft fuselage including antennas and ancillary equipment.
• The operator issued guidance on site selection during primary missions outlining:
  • aircrew to prioritise lower risk landing and winching sites (i.e. large open areas, playing fields, parks etc.)
  • confined areas at or close to the scene should only be used if other options are not viable
  • recommended or directed sites from emergency services on scene are not mandatory and should be regarded as guidance information only
  • proximity to a scene should be regarded as secondary consideration.
• A final internal safety report was provided to the ATSB and proactively shared among the emergency helicopter network.
• Installed the A800 Trakkabeam high‑intensity searchlight onto the fleet of aircraft.

Sources and submissions

Sources of information

The sources of information during the investigation included:

• the crew of the occurrence flight
• Toll Helicopters (ACT/NSW)
• Bureau of Meteorology
• recorded data from the MPFR unit on the aircraft.
• cabin video and audio recordings

References

The Federal Aviation administration (FAA) 2019, Helicopter Flying Handbook Chapter 2: Aerodynamics of Flight

Submissions

Under section 26 of the Transport Safety Investigation Act 2003, the ATSB may provide a draft report, on a confidential basis, to any person whom the ATSB considers appropriate. That section allows a person receiving a draft report to make submissions to the ATSB about the draft report.

A draft of this report was provided to the following directly involved parties:

• Civil Aviation Safety Authority
• Toll Helicopters (ACT/NSW)
• Crew of VH-TJK
• Transportation Safety Board of Canada

No submissions were received.

Purpose of safety investigations The objective of a safety investigation is to enhance transport safety. This is done through: identifying safety issues and facilitating safety action to address those issues providing information about occurrences and their associated safety factors to facilitate learning within the transport industry. It is not a function of the ATSB to apportion blame or provide a means for determining liability. At the same time, an investigation report must include factual material of sufficient weight to support the analysis and findings. At all times the ATSB endeavours to balance the use of material that could imply adverse comment with the need to properly explain what happened, and why, in a fair and unbiased manner. The ATSB does not investigate for the purpose of taking administrative, regulatory or criminal action. Terminology An explanation of terminology used in ATSB investigation reports is available here. This includes terms such as occurrence, contributing factor, other factor that increased risk, and safety issue. Publishing information  Released in accordance with section 25 of the Transport Safety Investigation Act 2003 Published by: Australian Transport Safety Bureau © Commonwealth of Australia 2023 Ownership of intellectual property rights in this publication Unless otherwise noted, copyright (and any other intellectual property rights, if any) in this report publication is owned by the Commonwealth of Australia. Creative Commons licence With the exception of the Coat of Arms, ATSB logo, and photos and graphics in which a third party holds copyright, this publication is licensed under a Creative Commons Attribution 3.0 Australia licence. Creative Commons Attribution 3.0 Australia Licence is a standard form licence agreement that allows you to copy, distribute, transmit and adapt this publication provided that you attribute the work. The ATSB’s preference is that you attribute this publication (and any material sourced from it) using the following wording: Source: Australian Transport Safety Bureau Copyright in material obtained from other agencies, private individuals or organisations, belongs to those agencies, individuals or organisations. Where you wish to use their material, you will need to contact them directly.

[1]     Helicopter Emergency Medical Service (HEMS): the use of helicopters to provide various kinds of medical care.

[2]     Night Visual Flight Rules (VFR): a set of regulations that permit a pilot to operate an aircraft at night, in weather conditions generally clear enough to allow the pilot to see where the aircraft is going.

[3]     Night vision imaging system (NVIS): a system of internal and external lighting, combined with night vision goggles, which provides enhanced vision to crew for operation at night.

[4]     Helicopter Landing Site (HLS): an area (not located on an aerodrome) wholly or partly used for the arrival or departure of helicopter.

[5]     Confined area: an area where the operation of the helicopter is limited in some direction/dimension by terrain or the presence of obstructions or obstacles, man-made or natural.

[6]     Capability Check Flight (CCF): periodic check (6-monthly) regarding the capability of an NVIS aircrew member in accordance with Civil Aviation Order 82.6

[7]     O’clock: the clock code is used to denote the direction of an aircraft or surface feature relative to the current heading of the observer’s aircraft, expressed in terms of position on an analogue clock face. Twelve o’clock is ahead while an aircraft observed abeam to the right would be said to be at 3 o’clock.

[8]     Unimproved helicopter landing site (HLS): a place that has not been prepared or supplemented with guidance infrastructure, identified for use as a temporary landing site for the purposes of landing and take-off.

[9]     Cloud cover: in aviation, cloud cover is reported using words that denote the extent of the cover – ‘few’ indicates that cloud is covering less than a quarter of the sky, if the sky is covered between 5/8 and 7/8 with clouds, it's reported as ‘broken’.

[10]    Multi-purpose flight recorder (MPFR): includes a cockpit voice recorder (CVR) and flight data recorder (FDR)

[11]    Cultural lighting: the illumination caused by the reflection of a major metropolitan area's lighting.

Executive summary

What happened

On 23 April 2021, a Van’s Aircraft RV-7A, registered VH-XWI, was being operated on a private flight under the visual flight rules (VFR) from Winton to Bowen, Queensland. The pilot, the sole occupant and the owner of the aircraft, had been on a multi-day tour in company with 3 other pilots, each operating their own aircraft.

About 2 hours after departing Winton, and about 90 km south of Charters Towers, the aircraft sustained an in-flight break-up. The pilot was fatally injured, and the aircraft was destroyed.

What the ATSB found

The ATSB found that the pilot departed Winton with a high risk of encountering adverse weather conditions along the planned route. About 100 km into the flight, overhead Catumnal Station, the pilot most likely entered instrument meteorological conditions (IMC) and lost control of the aircraft several times, recovering control within 50 ft of the ground. The pilot then turned back towards Winton.

About 11 km into the return leg, the pilot resumed track to Bowen, climbing to above 10,000 ft. The pilot flew at multiple altitudes between 10,000 ft and 500 ft above the ground, most likely to avoid weather along the track. The pilot again likely entered weather conditions before becoming spatially disorientated, resulting in loss of aircraft control. This led to the airspeed limitations of the aircraft being exceeded before rudder flutter structurally compromised the tail group, leading to a catastrophic airframe failure and in-flight break-up.

There were no operational reasons for the pilot to continue the flight to Bowen, and the pilot probably had a self-imposed motivation or pressure to continue the flight.

Safety message

VFR into IMC accidents usually result in a fatal outcome, and they are one of the most common types of fatal accidents in general aviation. Safety lessons to minimise the risk have been published in many sources, including the ATSB publication Accidents involving Visual Flight Rules pilots in Instrument Meteorological Conditions. Some of these lessons for VFR pilots include:

• Avoid deteriorating weather by conducting thorough pre-flight planning. Ensure you have alternate plans in case of an unexpected deterioration in the weather, and make timely decisions to turn back, divert or hold in an area of good weather.
• If you encounter deteriorating weather, turn back or divert before you are caught in cloud. For a non-instrument rated pilot, even with basic attitude instrument flying proficiency, maintaining control of an aircraft in IMC by reference to the primary flight instruments alone entails a very high workload that can result in narrowing of attention and loss of situational awareness.
• Recognise that weather often does not act as the forecast predicts. You must have alternatives available and be prepared to use them—even if it means returning to your departure point.
• Use a ‘personal minimums’ checklist to help control and manage flight risks by identifying risk factors that include marginal weather conditions and only fly in environments that do not exceed your capabilities. A personal minimums checklist is an individual pilot’s own set of rules and criteria for deciding if and under what conditions to fly or to continue flying based on your knowledge, skills and experience. In other words, to be a competent pilot, you must know and fly within your own limitations.

The occurrence

Events prior to departure

On 23 April 2021, a Van’s Aircraft RV-7A, registered VH-XWI, was being operated on a private flight from Winton to Bowen, Queensland (Figure 1).

The owner/pilot was the sole occupant and had been on a multi-day tour in company with 3 other pilots, each operating their own aircraft. The 4 aircraft were all travelling to the same event, being held near Bowen.

Prior to departing from Winton, the pilots discussed the forecast weather conditions for the planned flight, including their arrival at Bowen. Due to the likelihood of adverse weather towards the coast, the other pilots elected to operate to Bowen under the instrument flight rules[1] (IFR), while the pilot of VH-XWI, being restricted to operating in visual meteorological conditions[2] (VMC), departed under a daytime visual flight rules[3] (VFR) flight plan.

Figure 1: Flight track of VH-XWI from Winton to Bowen

Source: Google Earth with Aireon data, annotated by the ATSB

Accident flight

Witnesses stated that VH-XWI was last of the 4 aircraft to depart. Recorded data showed that the aircraft departed Winton at about 0751 local time and set a north-east course to track to Bowen, climbing to 7,500 ft above mean sea level (AMSL) for the 600-km flight.

At 0826, when the aircraft was near Catumnal Station, about 100 km north-east of Winton, the aircraft descended to about 500 ft above ground level (AGL) (Figure 2). The aircraft conducted 2 right orbits about 7 km to the south-west of Catumnal Station at indicated airspeeds as low as 25 kt, before rapidly descending to less than 200 ft AGL.

Figure 2: Flight path of VH-XWI over Catumnal Station

Source: Google Earth with Aireon data, annotated by the ATSB

The aircraft then climbed back up to about 1,000 ft AGL. At about 0829, when overhead the station airstrip, the aircraft’s airspeed reduced to about 43 kt while the aircraft conducted a right turn. It then rapidly descended to less than 50 ft AGL before climbing away from the station, and turning back towards Winton (Figure 3).

Figure 3: Flight path of VH-XWI in the vicinity of Catumnal Station

Source: Google Earth with Aireon data, annotated by the ATSB

At 0838, about 11 km after turning towards Winton, the aircraft conducted a left climbing turn then commenced a track towards Bowen and climbed to about 10,500 ft AMSL.

At 0936, when about 290 km north-east of Winton, the aircraft descended to about 1,000 ft AGL. It then conducted a large left orbit before turning back towards Bowen and climbing to about 9,000 ft AMSL.

At 0950, when about 370 km north-east of Winton (and about 90 km to the south of Charters Towers), the aircraft conducted an abrupt 90° right turn, followed by a series of further turns, with airspeed and altitude fluctuations. Soon after, the aircraft was in a right descending turn. It then rolled to the left and continued until almost inverted, and then pitched towards the ground.

At the last recorded GPS position (0952:04), the aircraft was at a recorded altitude of about 6,000 ft, where it had rolled 165° to the left and pitched down 41°. It was descending at more than 10,000 ft/min at an indicated airspeed of 224 kt and was accelerating at 15 kt/s (Figure 4). This recorded airspeed was more than 20% over the never exceed speed (V_NE) for the RV-7A.

Figure 4: VH-XWI accident sequence

Source: Google Earth with Dynon data, annotated by the ATSB

Based on the wreckage distribution and the nature of the wreckage, it was evident that VH-XWI sustained an in-flight break-up before the wreckage collided with terrain. The pilot was fatally injured, and the aircraft was destroyed.

Other information

In-flight radio contact between the pilot of VH-XWI and the 3 other pilots in company was made intermittently. These pilots recalled that, during the flight, the pilot of VH-XWI indicated that they would descend from 7,500 ft to 3,500 ft due to cloud. The pilot of VH-XWI then advised the group that they were unable to ‘get through’ and decided to descend below the cloud to continue the flight to Bowen. Sometime later the pilot of VH-XWI advised that they were ‘going to turn back’.

On landing at Bowen, the other pilots checked on the progress of VH-XWI using web-based aircraft tracking software and saw that VH-XWI’s track had stopped. They then alerted the Joint Rescue Coordination Centre.

Context

Pilot information

Qualifications and experience

The pilot held a valid private pilot licence (aeroplane), initially issued in August 1975, and was qualified to fly by day under the visual flight rules (VFR). Their licence was reissued as a Civil Aviation Safety Regulation (CASR) Part 61 licence on 3 May 2021, for single-engine aeroplanes, with endorsements for manual propeller pitch change and retractable undercarriage. The pilot had a total flying experience of 276.8 hours prior to the accident flight.

After receiving their private licence in 1975, the pilot flew several times a month until about 1981 before taking a break from flying for about 3 years. In 1984 they conducted a flight review and continued flying 3-4 times a year until January 1990. Most of their total flight time (216.1 hours) was conducted prior to January 1990. The remaining 60.7 hours was conducted in VH-XWI from July 2020 to March 2021.

In July 2020, the pilot accompanied a test pilot in 17 hours of initial flight testing of VH-XWI (in a non-flying capacity). When phase 1 flight testing of VH-XWI was finalised, the pilot began to undertake flight training in the aircraft with a flying school in September 2020. The pilot flew with a number of instructors and completed 9.4 hours of dual training. Flying school records indicated that there were a number of areas that remained inconsistent when not coached by instructors and therefore the chief flying instructor believed that the pilot was not yet of a standard to be considered ready for a flight review.

In February 2021, the pilot conducted a further 3.9 hours of flight time in VH-XWI with another flying school before passing a flight review on 18 February 2021. The instructor noted that the pilot seemed careful and made good decisions, however spent considerable time focussed on instrumentation inside the cockpit to the detriment of heading and altitude. The instructor further remarked that this was not uncommon for pilots presenting for flight reviews in aircraft such as the RV-7A that were equipped with glass cockpits.[4] The instructor added that the pilot was experimenting with the glass cockpit and demonstrated the use of the autopilot during one leg of the navigation exercise.

The instructor recalled that, during initial stalling exercises, the pilot overly pitched the aircraft forward during the recovery, however subsequent stalls were conducted to an acceptable standard. During advanced stalling, with full flaps selected, the aircraft had a wing drop that the pilot recovered with aileron. After discussion and a number of other recoveries with rudder, the instructor was satisfied that the pilot had met an acceptable standard.

During the flight review, the pilot was given an engine failure and conducted a practice forced landing to an acceptable standard. Navigation diversions and instrument flying scenarios were not conducted during the flight review. The instructor stated that these were optional during a day VFR flight review.

After completing their flight review, the pilot operated VH-XWI on numerous local and cross-country flights, leading up to the planned multi-day tour with the 3 other pilots.

Medical information and recent history

The pilot was 74 years old and held a valid basic class 2 medical certificate, with the last examination conducted on 1 February 2021. This type of certificate is based on the standard required for a commercial motor vehicle driver’s licence declaration, verified by a general practitioner.

A post-mortem examination and toxicological screen were performed, with no evidence of incapacitation or effects of alcohol or other drugs.

Due to the pilot being away from home for the days preceding the accident, a detailed recent history could not be obtained. However, from the limited information available, there were no fatigue-related concerns identified.

Aircraft information

General

The Van’s Aircraft RV-7A is a low-wing, all-metal, high-performance, amateur-built aeroplane. It has a 2-seat, side-by-side configuration. It is supplied in kit form and is designed to be constructed for the education and recreation of the owner. The RV-7A has a fixed, tricycle landing gear and is suitable for cross-country flying, with speed ranges from 56–200 kt, and can be operated between +6 g and -3 g,[5] making it suitable for flying aerobatic manoeuvres.

VH-XWI was registered in the experimental category. Pilots and passengers of experimental aircraft in Australia accept the risk that the aircraft may not meet the same airworthiness safety standards as certified aircraft, and therefore operate on the basis of informed participation.[6]

Airworthiness

The Special Certificate of Airworthiness issued to VH-XWI on 26 August 2020 indicated in the annexure, among other requirements, that the aircraft be operated ‘by day and under VFR only’.

The airframe hours prior to the accident flight were unable to be established as the maintenance release was not located at the accident site. The most recent entries in the aircraft and engine maintenance logs were on 10 April 2021, about 13 days prior to the accident. They showed that the aircraft had accumulated 84.8 hours total time in service. Prior to that, the aircraft had received an oil change and a new oil filter at 25.2 hours on 10 August 2020, with no evidence of contaminants found in the filter.

Operating limitations

Aircraft operating limitations were contained within the aircraft pilot’s operating handbook (POH), and the relevant airspeed limitations are detailed in Table 1.

Table 1: Airspeed limitations

Never exceed speed

The aircraft structure was designed to support aerobatic operations with a never exceed speed (V_NE) of 200 kt. The calculation of this speed is driven by structural or aerodynamic limitations; however, control system flutter is typically one limitation that factors heavily into the calculation of V_NE.

Flutter is an oscillatory motion of a wing or flight control. Aerodynamic vibration starts when aerodynamic forces cause a structure to deflect in a way that alters the aerodynamic force being generated, resulting in a feedback loop. This oscillation may be damped, stable (non-divergent) or unstable (divergent). Undamped forms of vibration are known as flutter. Small changes to conditions can change the mode of oscillation rapidly. The onset of flutter in aircraft is usually instantaneous and normally leads to catastrophic failure.

Factors that can contribute to flutter include high speed, a reduction in structural or system stiffness, or an incorrect or a change in control surface mass distribution. The flutter speed is dependent on many individual aspects of the aircraft and the way the aircraft is being flown at any time.

The RV-7A has an open-circuit rudder control system, and cable tension in a control circuit affects the flutter speed for that circuit. Flutter speed variation can be affected by the pressure on the rudder pedals, the centre of mass of the control surface or free play in a trim tab. Therefore, a V_NE of 200 kt does not mean the aircraft will flutter immediately above this speed.

Fuel

The aircraft operated on aviation gasoline (avgas) and had a usable capacity of 159 L in 2 (80 L) wing tanks. The aircraft was refuelled prior to departure from Winton, with records indicating that about 75 L was uplifted.

Recorded data provided by the aircraft instrumentation indicated an average fuel burn across the entire flight of about 29 L/hour, with straight and level cruise sectors at about 75–80% power settings using about 35 L/hour. The data also indicated that the fuel pressure remained uniform throughout the flight, and, at the time of the accident, the aircraft had about 51 L in the left tank and about 43 L in the right.

The recorded data provided detailed fuel quantity readings that were consistent throughout the flight, leaving a high degree of certainty that 94 L of fuel remained on board the aircraft at the time of the accident. This equated to a further 2 hours flight time, which was sufficient to divert to another landing area or continue to Bowen.

Instrumentation and systems

VH-XWI was fitted with a Dynon Skyview HDX primary flight display (Figure 5). The Skyview system incorporated various communication and GPS navigation systems.

 Figure 5: Dynon SkyView HDX primary flight display

Source: Dynon Avionics, annotated by the ATSB

The aircraft was fitted with a 2-axis Dynon Skyview autopilot which had multiple modes of operation, including a level mode. When selected, the autopilot’s level mode will immediately attempt to reach zero vertical speed and zero roll angle. It will not attempt to fly the aircraft to any previous pilot-selected altitude, track or other selected inputs. When activated, level mode will cause the autopilot to engage if it was not already engaged. Level mode is engaged by pushing the LEVEL button on the autopilot panel.

The Dynon Skyview HDX included a stall warning system that incorporated an angle of attack (AoA)[7] indicator to measure and provide an audio alert if the AoA increased significantly. The Dynon Skyview HDX pilot user guide stated:

AOA Alerts: When AOA audio alerting is enabled, it enables a progressive tone that increases in intensity as AOA increases. It is similar in nature to a conventional reed-type aircraft stall warning sound, although it is much more predictable in its progression. It can be set to start its progressive tones at either the border of the yellow/green marks on the AOA bars, the middle of the yellow, or at the yellow/red border. The tones start at the level selected and get progressively quicker and closer together until a solid tone is played at critical AOA (in the red).

The test pilot confirmed the correct operation of the AoA alert system during VH-XWI’s flight testing.

Meteorological information

Bureau of Meteorology forecasts indicated there was significant cloud, reduced visibility and adverse weather predicted for the intended duration of the flight from Winton to Bowen. Although generally better weather conditions existed inland of the east coast of Queensland, these were forecast to deteriorate from greater than 10 km visibility down to about 1,000 m in isolated thunderstorms with showers, rain and broken cloud from 800 ft to above 10,000 ft. Severe icing and turbulence were expected within the proximity of isolated thunderstorms and cumulonimbus.

A witness at Catumnal Station, about 100 km north-east of Winton, heard an aircraft overhead the station for a period of time, consistent with the recorded data from VH-XWI; however, they were unable to visually identify the aircraft due to the presence of fog.

A witness at Pajingo Station, about 17 km to the north-west of the accident site, recalled hearing an aircraft operating in the vicinity at about 1000. They did not see the aircraft in flight but indicated that the weather at the station was heavily overcast and raining on and off most of the morning.

The night before the flight to Bowen, the pilot contacted the test pilot, who had conducted the initial test flying of the aircraft. In general conversation about the trip, the weather forecast for the next day’s flight was discussed. The test pilot recalled that, despite the forecast, the pilot expressed a strong desire to remain with the group, even though predicted weather may not have been conducive to visual flight.

One of the pilots travelling in company with the pilot of VH-XWI indicated that all the pilots discussed the weather for the planned flight prior to departure from Winton as a point of concern. They recalled that the weather at Winton was clear with no fog or cloud, however significant weather was expected closer to the coast. Prior to departure, an offer was made to the pilot of VH‑XWI that, should the weather be too bad for VFR flight, one of the group would return to pick them up and take them to Bowen.

Another pilot travelling in company with the pilot of VH-XWI recalled the weather as initially good at 7,000 ft with little cloud. However, as the flight progressed, the weather deteriorated into instrument meteorological conditions (IMC)[8] with cloud and rain for the last 30–40 minutes of their flight to Bowen. They recalled becoming visual on the approach to Bowen below 2,000 ft with significant wind.

The pilots of the 3 aircraft travelling in company with VH-XWI, who were conducting IFR flights, tracked about 30 km to the south of VH-XWI’s accident location.

Recorded data

The pilot utilised a tablet with a flight planning software program. Some limited GPS data was able to be extracted for the first 5 minutes of flight, before no longer recording.

VH-XWI was fitted with an Automatic Dependant Surveillance – Broadcast (ADS-B) system, which provided ground-based aircraft tracking by digital data link on average once every 6 seconds. Airservices Australia described ADS-B on its website as:

The system involves an aircraft with ADS-B determining its position using GPS. A suitable transmitter then broadcasts that position at rapid intervals, along with identity, altitude, velocity and other data. Dedicated ADS-B ground stations receive the broadcasts and relay the information to air traffic control for precise tracking of the aircraft.

The last few parameters of ADS-B data from VH-XWI were erratic, most likely due to line-of-sight limitations with the associated antennas during the in-flight break-up sequence.

The Dynon Skyview HDX was capable of internally recording various types of data through numerous sensors and inputs, such as the GPS position, accelerometers, pressure transducers and magnetometers. This enabled recording of many aircraft and in-flight parameters, some of which included:

• flight path direction, airspeed, groundspeed and altitude
• engine data: pressures, temperatures, RPM, voltage
• fuel quantity, usage and quantity remaining
• autopilot activation
• angle of attack (AoA).

The data was recorded at up to 16 times per second. The data was buffered for a short time, before being transferred the Dynon’s internal storage. Captured data over 10 minutes old was down sampled to a data point every second. 

Onboard data recovered from the Dynon consisted of 1 second data from the start of the flight up until 11 minutes before the end of the flight, and then 16 times per second up until 1 minute before the end of the flight. The last minute or so was not recorded.

Combined data from the ADS-B and Dynon instrumentation provided detail relating to VH-XWI’s airspeed and altitude before the data stopped recording. Relevant data during the flight is presented in Figure 4. In addition:

• at 0950:19, about 2 minutes before the in-flight break-up while climbing slowly through 8,700 ft at 120 kt indicated airspeed, the aircraft entered into an abrupt climb to 9,660 ft and reduced speed to 56 kt before turning right and tracking southwards in a descent with speeds increasing to 145 kt
• at 0951:04, about 1 minute before the in-flight break-up, the aircraft had descended to 8,579 ft and stabilised airspeed at about 145 kt
• at 0950:34, about 30 seconds before the in-flight break-up, the aircraft had climbed to 9,204 ft and slowed to 88 kt before entering a right descending turn with an increasing descent rate and airspeed
• at 0951:58, about 6 seconds before the in-flight break-up, the aircraft had increased speed above V_NE limitation of the aircraft and continued to accelerate
• airspeed increased to 224 kt with an increasing trend before the data stopped being recorded.

Wreckage information

The accident site was located almost directly below the last known reliable recorded data point, on partially-vegetated, open bushland with the ground significantly softened by recent rainfall. The wreckage was distributed over a distance of about 1.5 km, with the larger and heavier aircraft sections located closer to each other, and lighter, smaller sections of wreckage found to the north-west of the main wreckage, consistent with the forecast wind direction (Figure 6). All major aircraft components were accounted for, and there was no evidence of pre-impact defects.

Figure 6: Accident site scatter pattern

Source: Google Earth, annotated by the ATSB

The propeller, engine, and associated structure was located separately to the rest of the wings, cockpit, and rear tail sections.

The left wing and centre section/cockpit floor with part of the right-wing spar remained attached to the fuselage and was deformed downwards. The right wing was found nearby and showed transfer marks on its lower surface from the right main landing gear tyre.

The vertical stabiliser and the upper rudder section had separated as an assembly from the main fuselage. The rivet line at the training edge of the rudder had burst open, and the aerodynamic mass balance (normally located on the top of the rudder) was not found at the accident site (Figure 7). The lower section of the rudder had torn away immediately below its centre hinge point and was located some distance away from the vertical stabiliser and rudder assembly.

Figure 7:Rudder and vertical stabiliser

Source: ATSB

The elevators remained with the rear fuselage, attached by the elevator control horns and pushrod. Sections of the right horizontal stabiliser main spar had been torn from the horizontal stabiliser and were attached to the elevator at its hinge points. The horizontal stabilisers had separated in flight and were located some distance away from the rear fuselage in the debris field.

Appendix A provides details of 2 previous in-flight break-up accidents involving RV-7 aircraft. The damage to the rudder and vertical stabiliser exhibited similar damage and failure modes to that of VH-XWI. In both of those previous accidents, the aircraft’s speed had significantly exceeded V_NE.

Visual flight rules requirements

Civil Aviation Regulation (CAR) 172 (Flight visibility and distance from cloud) outlined that a flight under the VFR could only be conducted in VMC.[9] The Australian Aeronautical Publication (AIP) ENR 1.2-4 contained the requirements for VMC. For flight in class G (uncontrolled) airspace below 10,000 ft above mean sea level (AMSL), these were stated as:  

• a flight visibility of 5,000 m
• a minimum vertical distance of 1,000 ft and horizontal distance of 1,500 m from cloud.

In the case of aeroplane operations in Class G airspace at or below 3,000 ft AMSL or 1,000 ft above ground level (AGL) (whichever was higher), the following minimum conditions were stipulated:

• a flight visibility of 5,000 m
• the aeroplane be maintained clear of cloud and in sight of the ground or water.

In addition to minimum visibility and distance from cloud, a pilot was also required to maintain a minimum height above the ground.

CAR 157 (Low flying)[10] detailed that a pilot in command must not fly the aircraft over:

• any city, town, or populous area at a height lower than 1,000 ft; or
• any other area at a height lower than 500 ft.

This did not apply if, through stress of weather or any other unavoidable cause, when essential that a lower height be maintained.

Risks of flying in areas of reduced visual cues

When VFR pilots enter IMC (or similar environments) and lose sufficient external visual cues, this can result in various types of accidents, including a loss of control leading to a collision with terrain (most common), a loss of control leading to an in-flight break-up, or a controlled flight into terrain.

The loss of control is usually associated with spatial disorientation, which occurs when a pilot does not correctly sense the position, motion and attitude of an aircraft relative to the surface of the Earth. VFR pilots can rapidly become spatially disoriented when they cannot see the horizon and other external cues. The brain receives conflicting or ambiguous information from the sensory systems, resulting in a state of confusion that can rapidly lead to incorrect control inputs and resultant loss of aircraft control.

Spatial disorientation can either be unrecognised by the pilot (type I) or recognised (type II). When type II disorientation accidents occur, they generally involve erratic flight paths resulting from the pilot having difficulty maintaining control of the aircraft’s flight path (consistent with what was seen in this case).

The safety risk of VFR into IMC accidents are well documented. As noted by Wiggins and others (2012):

…pilots who are not authorized to operate outside [VMC] lose control of their aircraft within a relatively short period of time following entry into [IMC]… This loss of control tends to be associated with a high rate of descent and a collision with terrain, the outcome of which is typically not survivable…

For example, in the United States, there were 226 VFR into IMC accidents involving small non-commercial fixed-wing aircraft during 2010–2019, of which 194 (86%) were fatal accidents.[11] Detwiler and others (2008) found that VFR into IMC accidents were 4 times more likely to result in fatalities when compared to all other types of general aviation accidents.               

In Australia, in the 11 years from 2011 to 2021, there were 14 fatal accidents involving VFR into IMC (with 29 fatalities).[12] In comparison, there were 26 fatal VFR into IMC accidents in 1991–2000 (58 fatalities) and 10 such accidents in 2001–2010 (18 fatalities).

Of the 14 VFR into IMC fatal accidents during 2011–2021, 13 involved private flights (including 2 community service flights) and 1 involved a charter operation. The list of these accidents is provided in Appendix B. Overall, they accounted for 11% of all fatal accidents and 18% of fatal accidents involving private flights. During 2011-2021, there were also 3 other fatal accidents (with 6 fatalities) involving VFR into other adverse weather conditions (such as storms) as well as several accidents involving VFR flights into dark night conditions.

Some characteristics of VFR into IMC accidents

A significant number of studies have examined VFR into IMC accidents, and a range of factors are involved, with no one factor being involved in all accidents. These include pilots incorrectly assessing weather conditions, pilots underestimating the risk of the conditions, pilots overestimating their ability (or the ability of their aircraft) to handle the situation, and motivational (personal and social) factors (NTSB 2005, Goh and Wiegmann 2001, Detwiler and others 2008, Harris and others 2022).

Weather-related decision making is best characterised as a dynamic, ongoing process (Batt and O’Hare 2005). Wiggins and O’Hare (1995) noted that weather-related decision making in some cases can be highly complex and therefore more prone to errors:

Because of the variable nature of operations in the aviation environment, weather-related decision making is often considered a skill that cannot be prescribed during training. Rather it is expected to develop gradually through practical experience. However, in developing this type of experience, relatively inexperienced pilots may be exposed to hazardous situations with which they are ill‑equipped to cope.

Most VFR into IMC accidents happen during the cruise phase of flight. Some research has shown that the chance of a VFR into IMC encounter increases as a flight progresses, with most occurring during the second half of a planned flight (Batt and O’Hare 2005). The United States National Transportation Safety Board (NTSB, 2005) also found that the chances of a weather-related general aviation accident increased as the length of a planned flight increased.  

Various studies have shown that, compared to pilots involved in other types of accidents, pilots involved in VFR into IMC accidents are more likely to have a private pilot licence (as opposed to a higher-level licence), are less likely to have an instrument rating, and have less flight experience (Detwiler and other 2008, NTSB 1989, NTSB 2005). However, many pilots with a significant amount of experience are also involved in such accidents. Of the 14 pilots involved in VFR into IMC accidents in Australia during 2011–2021, most (12) had a private licence and 2 had a commercial licence. Only 4 pilots held a night VFR rating and none held an instrument rating (although 1 had previously held an instrument rating).

It is difficult to draw conclusions regarding research on pilot age and accident involvement, as it is often confounded with experience and other factors, and there are significant individual differences in a range of abilities (AOPA 2014). VFR into IMC accidents involve pilots from all age groups. Other information regarding pilot age includes:

• The NTSB’s 1989 study noted that pilots in VFR into IMC accidents were, on average, older than comparison groups of all active pilots and older than pilots involved in other general aviation accidents.
• The NTSB’s 2005 study compared pilots involved in weather-related general aviation accidents with a matched sample of pilots. It found that the non-accident group included a higher proportion of pilots less than 40, and the accident group included a higher proportion of pilots over 60. It also concluded that the age a pilot obtained their private licence was a better predictor of future accident involvement than age at the time of flight, with pilots obtaining their licence before 25 being less likely to be involved in a weather-related accident.
• The age of pilots involved in VFR into IMC accidents in the United States appears to have increased over time, with the proportion being over 60 being 8% in 1983–1987, 25% in 1990–2004 (Detwiler and others 2008) and 39% in 2003–2012 (Major and others 2017).[13]
• Of the 14 fatal VFR into IMC accidents in Australia during 2011–2021, at least 9 (64%) involved pilots aged over 60 (including 5 aged over 70) and at least 5 of these pilots (36%) were over 45 when they obtained their private licence (and 9 were over 35).[14]^,[15] This did not include the pilot of VH-XWI, who re-validated their licence at 74 after not flying for more than 30 years.

Safety analysis

In-flight break-up

The distribution and nature of the wreckage indicated that the aircraft experienced an in-flight break-up prior to colliding with terrain. Prior to the break-up, the pilot conducted a series of right turns with airspeed and altitude fluctuations before conducting a left turn and rolling almost inverted. The aircraft accelerated past its never exceed speed (V_NE), and it was highly likely that this initiated rudder flutter, which led to catastrophic failures of the aircraft structure and subsequent in-flight break-up.

Examination of the aircraft wreckage did not identify any pre-existing defects, however wreckage at the accident site indicated cascading failures that likely followed the rudder and vertical stabiliser separating from the aircraft. Subsequently, the horizontal stabilisers likely failed and unbalanced the aircraft’s aerodynamic forces, rapidly pitching the aircraft downward. The aerodynamic overload on the main wing structure resulted in the right-wing main spar failing in a downwards direction, and the right wing separated from the aircraft.

The nature of the in-flight break-up was similar to 2 other accidents involving the same aircraft type. In all 3 cases, the aircraft significantly exceeded V_NE while manoeuvring. In the 2 previous cases, the aircraft was involved in aerobatic activities. However, in the accident involving VH-XWI, the manoeuvres were the result of the pilot encountering adverse weather conditions.

Flight into non-visual conditions

The information available to the pilot in the evening before, and prior to departing from Winton on the morning of the accident, indicated that for at least part of that flight the conditions would not be suitable for flight under the visual flight rules (VFR).

After departure from Winton, the pilot likely entered instrument meteorological conditions (IMC) in the vicinity of Catumnal Station. This most likely led to spatial disorientation, and the subsequent stall, loss of control, and low-level recovery, as evidenced in the recorded data. On multiple other occasions during the flight, the pilot most likely attempted to avoid reduced visual conditions, as evidenced by the significant variations in altitude, from 10,000 ft to within 500 ft of the ground, orbits, and deviations from track.

When the aircraft was south of Charters Towers, it is likely that the pilot again entered into reduced visual conditions. Recorded data indicated abnormal flight manoeuvres, commencing with an abrupt turn to the south. Analysis of the recorded data, coupled with the increased likelihood of reduced visual cues, indicated that the pilot most likely experienced spatial disorientation in the period immediately prior to the in-flight break-up.

Continuing flight into IMC with no instrument rating significantly increases the risk of spatial disorientation due to the absence of visual cues and powerful and misleading orientation sensations altering the pilot’s perception of the aircraft’s position, motion and attitude. This disorientation can affect any pilot, no matter what their level of experience, and has been associated with many previous accidents.

In-flight decision making

It was very likely that after the near collision with terrain near Catumnal Station, the pilot decided to return to Winton. In most cases, return to an area of known good weather is a safe and prudent option.

However, for reasons that could not be established, about 11 km into the return journey towards Winton, the pilot elected to resume course for Bowen. This may have been prompted by airborne communications on weather and flight visibility at altitude from the pilots of the other aircraft in the group, travelling to Bowen. However, the pilot of the accident flight may not have been aware that the 3 other aircraft were tracking about 30 km to the south, and may have been less affected by the weather encroaching from the north. It is also possible that, having encountered IMC and being able to recover the aircraft, the pilot may have become more confident in their ability to operate near such weather.

Although the pilot was aware of the forecast weather, and had already encountered IMC, they still continued towards Bowen. Research has shown that many aviation accidents involve a ‘plan continuation bias’ or ‘plan continuation error’.[16] That is, pilots decide to continue with the original plan of action despite the presence of cues or information that suggests changing the course of action would be the safer option (Orasanu and others 2001, Dismukes and others 2007, Orasanu 2010). Plan continuation bias is often associated with situations involving dynamically changing risk and pilots underestimating the risk level (Orasanu and others 2001; Wiegmann and others 2002).

Research has also shown that many personal, social and organisational pressures can influence pilot decision-making, including in terms of whether to conduct or continue flights in unsuitable weather conditions (Paletz and others 2009, Michalski and Bearman 2014, Harris and others 2022). These can include observing other pilots successfully flying in the conditions and ‘reluctance to admit defeat’, as well as normalisation of deviance (or acceptance of risk with exposure).

In this case, a witness that spoke with the pilot the night before the accident about the forecast weather indicated that the pilot had a strong desire to remain with the group. Therefore, the pilot probably had self-imposed motivation or pressure to continue towards Bowen, despite the changing weather conditions. There was no operational reason for the pilot to continue the flight, given they had plenty of fuel to return to Winton or divert to other aerodromes away from the adverse weather.

Pilot experience and other characteristics

The pilot was 74 and relatively inexperienced, having only about 270 hours total flight hours, with much of this obtained over 30 years prior to the accident. The pilot had also had some difficulty regaining the level of skill required to pass a flight review. However, the extent to which flight experience, skill level, age or other factors was involved in this specific accident (as opposed to motivational factors) is difficult to determine.

All pilots of VFR flights need to understand the risks of conducting flights into deteriorating weather conditions, and a range of educational resources are available to help pilots appreciate these risks and improve weather-related decision making (see Appendix B for examples). There are also some indications that pilots who are older and obtained (or re-obtained) their private licence later in life, and/or have limited flight experience, may be at increased risk and should therefore pay particular attention to the guidance material that is available.

Findings

From the evidence available, the following findings are made with respect to the VFR into IMC and in-flight break-up involving the Van's Aircraft RV-7A, VH XWI, 90 km south of Charters Towers, Queensland, on 23 April 2021.

Contributing factors

• The pilot departed for the flight from Winton to Bowen, Queensland, knowing that the weather conditions en route were unlikely to be suitable for flight under the visual flight rules (VFR).
• The pilot likely entered weather conditions not suitable for flight in visual meteorological conditions, leading to spatial disorientation and loss of aircraft control.
• The aircraft’s airspeed exceeded V_NE (never exceed speed or maximum airframe speed), leading to rudder flutter, airframe structural failures, and subsequent in-flight break-up.

Other factors that increased risk

• Earlier in the flight, the pilot entered a degraded visual environment over Catumnal Station, most likely resulting in spatial disorientation. This resulted in a loss of control, stall, and subsequent low-level recovery.
• Although the pilot turned back to Winton after the near collision with terrain at Catumnal Station, the pilot then resumed the flight towards Bowen and degrading weather conditions.

Other findings

• There were no operational reasons for the pilot to continue the flight to Bowen, and the pilot probably had a self-imposed motivation or pressure to continue the flight.

Sources and submissions

Sources of information

The sources of information during the investigation included:

• the flight training facilities
• airborne and ground witnesses
• Van’s Aircraft
• the Civil Aviation Safety Authority
• the Bureau of Meteorology
• Airservices Australia.

References

Aircraft Owners and Pilots Association 2014, Aging and the general aviation pilot: Research and recommendations.

Batt R & O’Hare D 2005, General aviation pilot behaviours in the face of adverse weather, ATSB Aviation Research Investigation Report B2005/0127.

Detwiler C, Holcomb K, Hackworth C & Shappell S 2008, Understanding the human factors associated with visual flight rules flight into instrument meteorological conditions, Federal Aviation Administration, Washington, DC. (NTIS DOT/FAA/AM-08/12).

Goh J & Wiegmann DA 2001, ‘Visual flight rules flight into instrument meteorological conditions: An empirical investigation of the possible causes’, International Journal of Aviation Psychology, 11:359–379.

Harris MR, Fein EC & Machin MA 2022, ‘A systematic review of multilevel influenced risk-taking in heliciopter and small aeroplane normal operations’, Frontiers in Public Health, 10:1–18.

Major WL, Carney T, Keller J, Xie A, Price M, Duncan J, Brown L, Whitehurst GR, Rantz WG, & Nicolai D 2017, VFR-into-IMC accident trends: Perceptions of deficiencies in training’, Journal of Aviation Technology and Engineering, 7:50–57.

National Transportation Safety Board 1989, General aviation accidents involving visual flight rules into instrument meteorological conditions, Safety Report NTSB/SR-89/01.

National Transportation Safety Board 2005, Risk factors associated with weather-related general aviation accidents, Safety Study NTSB/SS-05/01.

Wiggins, MW, Hunter DR, O’Hare D & Martinussen M 2012, ‘Characteristics of pilots who report deliberate versus inadvertent visual flight into instrument meteorological conditions, Safety Science, 50:472–477.

Wiggins M & O’Hare D 1995, ’Expertise in aeronautical weather-related decision making: A cross-sectional analysis of general aviation pilots’, Journal of Experimental Psychology: Applied, 1:305-320.

Submissions

Under section 26 of the Transport Safety Investigation Act 2003, the ATSB may provide a draft report, on a confidential basis, to any person whom the ATSB considers appropriate. That section allows a person receiving a draft report to make submissions to the ATSB about the draft report.

A draft of this report was provided to the following directly involved parties:

• a flight crew member of an accompanying aircraft
• the Civil Aviation Safety Authority
• the Bureau of Meteorology
• Airservices Australia
• Van’s Aircraft Incorporated
• the National Transportation Safety Board.

Submissions were received from:

• the Civil Aviation Safety Authority
• Airservices Australia.

The submissions were reviewed and, where considered appropriate, the text of the report was amended accordingly.

Appendices

Appendix A – Other in-flight break-ups involving Van’s RV-7 aircraft

Two previous accidents involving Van’s RV-7/A variants involved in in-flight break-ups demonstrated similar airframe failure modes to that of VH-XWI.

New Zealand CAA investigation 17/8080

On 1 January 2018, a Van’s Aircraft RV-7, registered ZK-DVS, departed Whangarei, New Zealand, with 2 people on board. About 17 minutes after departing, the aircraft entered a high angle of bank (AoB) manoeuvre, achieving about 70° AoB. Five seconds later the AoB increased to about 130° and the aircraft began to pitch nose-down. During the resulting descent, the indicated airspeed was recorded at 244 kt, which exceeded the aircraft’s never exceed speed (V_NE). About 30 seconds after entering the high AoB manoeuvre, witnesses observed the aircraft break up in flight and then impacted terrain.

The New Zealand Civil Aviation Authority (CAA) investigation concluded:

• The aircraft entered a high-speed descent from an unusual attitude.
• The pilot did not recover the aircraft from the unusual attitude or subsequent high-speed descent, which resulted in structural failure and in-flight breakup.
• In-flight breakup occurred as a result of rudder flutter, as the aircraft airspeed exceeded the design limitations.

Figure A1 shows the damage to the aircraft’s rudder and vertical stabiliser.

Figure A1: Rudder and vertical stabiliser from ZK-DVS

Source: CAA New Zealand, annotated by the ATSB

The Civil Aviation Authority of New Zealand found that to avoid the potential of structural failure and in-flight break-up, pilots must fly within the aircraft limitations. Accidents can occur whenever the aircraft limitations and/or the pilot’s own capabilities are exceeded. It is important to understand and be familiar with the aircraft characteristics, limitations, and associated risks. These risks can be minimised by preparation, awareness, and training.

Canadian TSB aviation investigation report A10O0018

On 23 January 2010, a Van’s RV-7A, registered C-GNDY, was part of a formation of 3 aircraft that departed Lindsay, Ontario, Canada, on a flight to Smiths Falls. During the flight, 1 of the 3 aircraft diverted to Bancroft. The remaining aircraft continued with C-GNDY in tandem. The lead conducted a series of aerobatic manoeuvres, which C-GNDY was to film. While manoeuvring, the lead lost contact with the aircraft. The aircraft’s wreckage was subsequently located in a wooded area.

The Canadian Transportation Safety Board (TSB) concluded:

• After painting, the rudder was not likely balanced, nor the aircraft reweighed. As a result, the rudder was susceptible to flutter at a lower speed than designed and the aircraft was over the maximum aerobatic gross weight during the manoeuvres.
• During the manoeuvring sequence, the speed of the aircraft reached 234 knots, exceeding the 124-knot manoeuvring speed and the 200 knot never exceed speed (Vne).
• The aircraft encountered either flutter or overstress of some rudder components. Subsequently, the vertical stabilizer and parts of the rudder separated from the empennage during flight. Consequently, the aircraft became uncontrollable resulting in the impact with terrain.

Figure A2 shows the damage to the aircraft’s rudder and vertical stabiliser.

Figure A2: Rudder and vertical stabiliser from C-GNDY

Source: TSB of Canada, annotated by the ATSB

Appendix B – VFR into IMC accidents and related resources

Recent VFR into IMC accidents in Australia

During the 11 years from January 2011 to December 2021, the ATSB investigated 14 fatal accidents (with 29 fatalities) involving VFR into IMC in VH-registered aeroplanes and helicopters in Australia:

• AO-2011-085 - VFR into IMC - South Turramurra, New South Wales, 22 July 2011, VH-CIV, Bell 206L Helicopter (private flight, 2 fatalities)
• AO-2011-100 - VFR flight into dark night conditions and loss of control involving Piper PA-28-180, VH-POJ, 31 km north of Horsham Airport, Vic., 15 August 2011 (private flight – community service, 3 fatalities)
• AO-2012-076 - Visual Flight Rules in Instrument Meteorological Conditions and controlled flight into terrain involving Cessna 182Q, VH-CWQ, 15 km north of Tooraweenah, NSW, 4 June 2012 (private flight, 1 fatality) (private flight, 1 fatality)
• AO-2012-130 - VFR flight into IMC involving de Havilland DH-84 Dragon VH-UXG, 36 km SW of Gympie, Qld, 1 October 2012 (private flight, 6 fatalities)
• AO-2013-186 - Collision with terrain involving Cessna 182, VH-KKM, 19 km WSW of Mount Hotham Airport, VIC on 23 October 2013 (private flight, 1 fatality)
• AO-2015-131 Collision with terrain involving Airbus Helicopters EC135 T1, VH-GKK, 10 km NNW Cooranbong, NSW, on 7 November 2015 (private flight, 3 fatalities)
• AO-2016-006 - Loss of control and collision with water involving Piper Aircraft Corp PA-28-235, VH-PXD, 33 km SSE of Avalon Airport, Victoria on 29 January 2016 (private flight, 4 fatalities)
• AO-2017-061 - VFR into IMC and loss of control involving Cessna 172, VH-FYN, 13 km NNW of Ballina, NSW, on 16 June 2017 (private flight, 1 fatality)
• AO-2017-069 Collision with terrain involving SOCATA TB-10 Tobago, VH-YTM, near Mount Gambier Airport, South Australia, on 28 June 2017 (private flight – community service, 3 fatalities)
• AO-2018-078 - VFR into IMC and controlled flight into terrain involving Pilatus Britten-Norman BN2A, VH-OBL, 98 km west-south-west of Hobart Airport, Tasmania, on 8 December 2018 (charter flight – ferry, 1 fatality)
• AO-2019-018 - Controlled flight into terrain involving Cessna 182, VH-DJN, 14 km south-south-west of Atherton Airport, Queensland, on 8 April 2019 (private flight, 1 fatality)
• AO-2019-052 - Controlled flight into terrain involving Mooney M20J, VH-DJU, 26 km west of Coffs Harbour Airport, New South Wales, on 20 September 2019 (private flight, 2 fatalities)
• AO-2020-004 VFR into IMC and loss of control involving Wittman Tailwind, VH-TWQ, Tooloom National Park, New South Wales, on 12 January 2020 (private flight, 2 fatalities)
• AO-2021-017 - In-flight break-up involving Vans Aircraft RV-7A, registered VH-XWI, 90 km south of Charters Towers, QLD, 23 April 2021 (private flight, 1 fatality)

Multiple additional accidents that involved VFR into IMC in 2022 are still under investigation. Additional fatal VFR into IMC accidents occurred during the same period that involved non VH-registered aircraft.

During the same period (2011–2021), there were also 3 other fatal accidents (with 6 fatalities) involving VFR flights into adverse weather conditions:

• AO-2011-160 - In-flight breakup - Cessna C210, VH-WBZ, 100 km NNW Roma, Qld, 7 December 2011 (private flight, 1 fatality)
• AO-2013-063 - Loss of control and collision with water involving Cessna 210 VH-EFB, 160 km south-west of Darwin, NT, 1 April 2013 (private flight, 3 fatalities)
• AO-2017-102 - In-flight breakup involving Cessna 210, VH-HWY, 22 km E of Darwin Airport, Northern Territory, on 23 October 2017 (charter flight – freight, 2 fatalities).

In addition, during the same period, there have been 10 fatal accidents involving VFR flights in dark night conditions, which carry the same risks as VFR into IMC. Some examples of these include:

• AO-2011-102 - VFR flight into dark night involving Aérospatiale, AS355F2 (Twin Squirrel) helicopter, VH-NTV, 145 km north of Marree, SA on 18 August 2011 (charter flight – passenger, 3 fatalities)
• AO-2013-057 - VFR flight into dark night conditions and loss of control involving Cessna T210N, VH-MEQ, 2 km north-west of Roma Airport, Qld on 25 March 2013 (private flight, 2 fatalities)
• AO-2019-050 - Loss of control and collision with water involving Bell UH-1H, VH-UVC, 5 km south-west of Anna Bay, New South Wales, on 6 September 2019 (private flight, 5 fatalities).

VFR into IMC educational resources

A discussion of some VFR into IMC accidents and advice to pilots regarding how to the risk of being involved in such accidents is provided in the ATSB Avoidable Accidents publication Accidents involving Visual Flight Rules pilots in Instrument Meteorological Conditions and related communication Don't Push It - DON'T GO.

Other relevant information is provided in the ATSB reports:

• ATSB Aviation Research and Analysis Report B20070063, An overview of spatial disorientation as a factor in aviation accidents and incidents
• ATSB Aviation Research Investigation B2005/0127, General Aviation Pilot Behaviours in the Face of Adverse Weather.

A cue-based training system, called Weatherwise, was made available to pilots by the Civil Aviation Safety Authority (CASA). Additionally, CASA has produced:

• Weather to fly, an education program which focused on topics such as the importance of pre-flight preparation, making decisions early, and talking to ATC
• ‘178 seconds to live’, a campaign on highlighting the dangers of VFR flight into IMC.

Purpose of safety investigations The objective of a safety investigation is to enhance transport safety. This is done through: identifying safety issues and facilitating safety action to address those issues providing information about occurrences and their associated safety factors to facilitate learning within the transport industry. It is not a function of the ATSB to apportion blame or provide a means for determining liability. At the same time, an investigation report must include factual material of sufficient weight to support the analysis and findings. At all times the ATSB endeavours to balance the use of material that could imply adverse comment with the need to properly explain what happened, and why, in a fair and unbiased manner. The ATSB does not investigate for the purpose of taking administrative, regulatory or criminal action. Terminology An explanation of terminology used in ATSB investigation reports is available here. This includes terms such as occurrence, contributing factor, other factor that increased risk, and safety issue. Publishing information  Released in accordance with section 25 of the Transport Safety Investigation Act 2003 Published by: Australian Transport Safety Bureau © Commonwealth of Australia 2022 Ownership of intellectual property rights in this publication Unless otherwise noted, copyright (and any other intellectual property rights, if any) in this report publication is owned by the Commonwealth of Australia. Creative Commons licence With the exception of the Coat of Arms, ATSB logo, and photos and graphics in which a third party holds copyright, this publication is licensed under a Creative Commons Attribution 3.0 Australia licence. Creative Commons Attribution 3.0 Australia Licence is a standard form licence agreement that allows you to copy, distribute, transmit and adapt this publication provided that you attribute the work. The ATSB’s preference is that you attribute this publication (and any material sourced from it) using the following wording: Source: Australian Transport Safety Bureau Copyright in material obtained from other agencies, private individuals or organisations, belongs to those agencies, individuals or organisations. Where you wish to use their material, you will need to contact them directly.

Context

MPV Everest

The ship

MPV Everest is a Bahamas-registered, ice class, dynamically‑positioned, multi-purpose vessel, built by Keppel Singmarine in Singapore in 2017. The ship was owned by New Orient Marine,[9] a subsidiary of Luxembourg-based Maritime Construction Services (MCS) and, at the time of the fire, managed by Singapore-based Fox Offshore. In October 2021, about 5 months after the fire, MCS assumed direct responsibility for the safety management and operation for the ship.

MPV Everest was equipped with the necessary navigational, firefighting, lifesaving and communication equipment required by the International Convention for the Safety of Life at Sea (SOLAS)[10] for a ship of its size and operations. The ship was equipped with a Furuno VR-7000 voyage data recorder (VDR) from which information useful to the investigation was recovered.

Classification societies

During its design and construction phase, MPV Everest was ‘dual classed’[11] with Bureau Veritas (BV) and the Russian Maritime Register of Shipping (RMRS). A working agreement between BV and RMRS defined the responsibilities of the 2 classification societies with respect to various class notations required by the shipowner and specified which classification society’s rules would be applicable for their respective class notations. Among other matters, the working agreement delegated responsibility for plan approval and surveys relating to the ship’s hull and piping, fire safety and dynamic positioning (DP) to BV while RMRS was given responsibility for the ship’s ice class, electrical equipment and unattended machinery space (UMS) notations. Following construction, the ship was ‘double classed’[12] with BV and RMRS.

MPV Everest was built to the RMRS ‘Arc5’ standard, which indicated the ship’s capability for navigation in ice.[13] In accordance with the ship’s dual classification agreement, the ship was also issued with BV’s equivalent ice class notation of ‘ICE CLASS 1A SUPER’.

Polar Code compliance

The International Code for Ships Operating in Polar Waters (Polar Code) was developed to supplement existing International Maritime Organization (IMO) instruments in order to increase the safety of ships’ operations and mitigate the impact on the people and environment in remote, vulnerable and harsh polar waters. The code acknowledged that polar water operations may impose additional demands on ships, their systems and operation beyond the existing requirements of relevant binding IMO instruments.

The code is mandatory under both SOLAS and MARPOL[14] conventions. It covers a range of requirements related to design, construction, equipment, operations, training, search and rescue, and environmental protection matters relevant to ships operating in polar waters. The Polar Code entered into force on 1 January 2017.

The code identified several hazards that could lead to elevated risk for ships operating in polar waters including:

• low temperatures and darkness
• high latitudes (and their effect on navigation and communication systems)
• remoteness and possible lack of accurate and complete hydrographic data and information
• reduced availability of navigational aids and seamarks
• limited readily deployable search and rescue facilities, delays in emergency response and limited communications capability
• potential lack of ship crew experience in polar operations, with potential for human error
• potential lack of suitable emergency response equipment, with the potential for limiting the effectiveness of mitigation measures.

In order to establish procedures or operational limitations, an assessment of the ship and its equipment was required to be carried out. The Polar Code stated that the risk level within polar waters could differ depending on factors such as the geographical location and time of the year and that mitigating measures required to address identified hazards could vary and be different in Arctic and Antarctic waters. The code also stated that while Arctic and Antarctic waters have similarities, there are also significant differences and, that the legal and geographical differences between the 2 areas need to be considered.

In September 2020, in preparation for the AAD charter, MPV Everest was assessed by RMRS for compliance with the structure, equipment, fittings, radio arrangements, and material requirements of the Polar Code. The ship was subsequently issued with Polar Code certification under the authority of the Bahamas flag administration. The ship’s Polar Code certificate categorised MPV Everest as a category B ship,[15] with a maximum expected time of rescue of 5 days.

Polar water operational manual

The Polar Code required ships to carry a polar water operational manual to provide the owner, operator, master and crew with sufficient information regarding the ship’s operational capabilities and limitations in order to support their decision-making process. The manual was required to include procedures for:

• voyage planning (to avoid ice or temperatures exceeding the ship’s design capabilities)
• means of addressing any limitations of the hydrographic, meteorological and navigational information available
• operation of equipment required under the code
• implementation of special measures to maintain equipment and system functionality under low temperatures.

A copy of MPV Everest’s polar water operational manual was provided to the ATSB by the ship’s owners on request. The 113-page manual provided a detailed description of the Russian Arctic, Arctic seas and straits, including information on Arctic geography, weather patterns, ports of refuge, routing and communication, and navigation performance.

There was no information in the manual about operations in the Southern Ocean or the Antarctic.

Training

The Polar Code also introduced minimum requirements for the training and qualifications of masters and deck officers of ships operating in polar waters which became mandatory under the International Convention on Standards of Training, Certification and Watchkeeping for Seafarers (STCW Convention)[16] on 1 July 2018. This introduced a requirement for basic and/or advanced polar waters training depending on factors such as the officer’s rank, type of ship and the nature of operating conditions.

Additionally, the Nautical Institute’s[17] Ice Navigator Training Accreditation and Certification Scheme complemented training required for masters and deck officers under the Polar Code. While participation in the scheme was not mandatory, it represented industry best practice and resulted in the award of internationally recognised qualifications. The scheme built upon the basic requirements of the Polar Code and accompanying STCW amendments, with a focus on ship handling and ice operations, aimed at ensuring that ships’ officers were not only trained but also experienced in operating in ice at sea.

Compliance with the scheme involved undergoing basic polar waters training and obtaining practical or simulated experience operating in ice conditions and initially resulted in the award of an Ice Navigator Level 1 certificate. Completion of the advanced polar waters training and further practical experience at sea in ice resulted in the award of an Ice Navigator Level 2 certificate, which was the highest level of qualification available under the scheme.

There were no mandatory or other training requirements for engineers of ships operating in polar waters.

History

MPV Everest was delivered in December 2017 but remained in the shipyard for another year while commissioning works progressed. The ship was managed by its owners (New Orient Marine) until January 2018 when Fox Offshore was engaged as the ship’s technical and safety manager. In December 2018, the ship was chartered for about a month to assist with the location and attempted recovery of the cockpit voice recorder from the wreckage of Lion Air Flight 610 in Indonesian waters before returning to anchor off Singapore. From September to December 2019, the ship was engaged in sub-sea construction work in Indonesia before returning to anchor. During the periods when the ship was at anchor, it was manned by a reduced crew complement.

In March 2020, MPV Everest was chartered by the Australian Antarctic Division (AAD) to undertake 2 voyages to Antarctica over the course of the southern summer to resupply and effect personnel changes at the Australian research stations in Antarctica (Casey, Davis, and Mawson stations).

In July 2020, MPV Everest departed Singapore for Busan, Republic of Korea (South Korea) and subsequently to Kholmsk, Russia. The ship then undertook sub-sea construction work in the Kirinskoye gas condensate field, off Sakhalin Island in the Sea of Okhotsk, until 20 November when it departed for dry dock. On 24 November, MPV Everest arrived in Samkang, South Korea where it was dry docked. On 1 December 2020, following 2 weeks of quarantine, most of MPV Everest’s crew for the AAD charter joined the ship while in dry dock.

Antarctic operations

On 7 December 2020, the ship was re-floated and, the following day departed South Korea for Australia. The ship’s previous master, second mate and chief electrician sailed with the new crew to Australia for an extended hand-over en route and to assist with familiarisation of the ship.

On 28 December, MPV Everest called at Burnie, Tasmania to take bunkers (fuel) before departing for Hobart, Tasmania. On 31 December, the ship arrived in Hobart. Over the following days, more crew joined the ship, AAD personnel embarked and supplies for the AAD Antarctic research stations were loaded.

On 9 January 2021, MPV Everest departed Hobart for Casey station in Antarctica. The ship arrived off Casey on 18 January and, by 27 January, had completed re-supply and changeover of Casey station personnel and departed for Hobart. On 29 January, while en route back to Australia, a fire broke out in the ship’s battery compartment (see the section titled Shipboard incidents and risk events). The fire was extinguished using portable fire extinguishers and the ship continued its voyage, arriving in Hobart without further incident on 4 February.

On 10 February, MPV Everest departed Hobart for Davis and Mawson stations in that order. By 23 February, the ship arrived off Davis and, by 8 March, had departed for Mawson. By 12 March, MPV Everest had arrived off Mawson, but ice conditions prevented the ship from getting closer than about 80 miles from the station. Over the course of the following 2.5 weeks, re-supply operations and changeover of Mawson station personnel were conducted using its helicopters over the ice.[18] On 31 March, MPV Everest departed Antarctica for Hobart and the fire occurred 5 days later.   

Ship’s crew

MPV Everest had a multi-national crew of 37 from 12 countries, including Polish, Dutch, Russian, Finnish, Latvian, Estonian, and Indonesian nationals.

The ship’s crew complement for the Antarctic voyages exceeded the requirements in the ship’s minimum safe manning document. The deck crew complement comprised the master, a chief mate, a first mate, 2 second mates, an ice pilot, a bosun and 3 able seamen. All the ship’s deck officers joined the ship on 1 December 2020 while dry docked in South Korea, with the exception of the ice pilot who joined on 5 January 2021 in Hobart before its first voyage to Antarctica.

At the time of the fire (see the section titled Engineering challenges), the engineering crew complement comprised the chief engineer, a second engineer, 2 third engineers, a chief electrician, an electro-technical officer (ETO), and 3 engine ratings (an oiler, a motorman and an electro-technical rating). All the ship’s engineers joined the ship on 1 December 2020 in the dry dock, with the exception of the second engineer who joined the ship on 9 February 2021 in Hobart, after its first voyage to Antarctica.

Master

The master held Polish and Dutch master’s certificates of competency and the equivalent Bahamas certificate of recognition. The master had about 21 years of seagoing experience as a qualified deck officer with about 6 years’ experience as master. Their seagoing experience was generally on board vessels associated with the offshore oil and gas industry such as offshore supply vessels (OSVs), anchor handling tug supply (AHTS) vessels, drills ships, offshore installations and other specialised DP ships. This was the master’s first time working with Fox Offshore and also their first time on board MPV Everest.

The master had completed basic and advanced training for ships operating in polar waters in September 2020 (prior to joining MPV Everest). The master also held an Ice Navigator Level 2 certificate issued by the Nautical Institute (NI) although this was issued in October 2021 (after their time on board MPV Everest). The master had previous experience in ice on board research ships in the Arctic and about a month’s experience in the Antarctic on board a Polish sail training ship during a voyage to the Polish Antarctic Station Arctowski, located on King George Island.

Chief mate

The chief mate held a Russian master’s certificate of competency and the equivalent Bahamas certificate of recognition. The chief mate had about 23 years of seagoing experience with over 10 years as master on ships of various types, including roll-on/roll-off ships, offshore DP vessels and icebreakers. This was their first time working with Fox Offshore and also their first time on board MPV Everest.

The chief mate had completed basic and advanced training for ships operating in polar waters in June 2019. The chief mate also held an NI Ice Navigator Level 2 certificate and had previous ice experience in the Arctic. This was their first time working in the Antarctic.

First mate

The first mate held a Russian master’s certificate of competency and the equivalent Bahamas certificate of recognition. They had over 29 years of seagoing experience. The first mate had completed basic and advanced training for ships operating in polar waters, held an NI Ice Navigator Level 2 certificate and had previous experience working in ice in the Arctic. This was their first time working with Fox Offshore, their first time on board MPV Everest and their first time working in the Antarctic.

Ice pilot

The ice pilot held an Estonian master’s certificate of competency and the equivalent Bahamas certificate of recognition. They had completed basic and advanced training for ships operating in polar waters, held an NI Ice Navigator Level 2 certificate and had extensive experience working in ice in the Arctic. This was their first time working with Fox Offshore, their first time on board MPV Everest and their first time working in the Antarctic.

Other deck officers

The duty and off-duty second mates held Russian officer of the watch certificates of competency and the equivalent Bahamas certificates of recognition. This was their first time working with Fox Offshore, their first time on board MPV Everest and their first time working in the Antarctic.

Both second mates had completed basic training for ships operating in polar waters and had previous experience working in ice in the Arctic. The duty second mate had also completed advanced training for ships operating in polar waters and held an NI Ice Navigator Level 2 certificate.

Chief engineer

The chief engineer held a Dutch chief engineer’s certificate of competency and the equivalent Bahamas certificate of recognition. They had over 27 years of seagoing experience, with the last 15 years as chief engineer. Their seagoing experience was generally on board DP semi‑submersible drilling rigs, DP drills ships, sub-sea constructions vessels, heavy lift ships and other specialised DP ships.

This was their first time working with Fox Offshore, their first time on board MPV Everest and their first time working in polar waters and in ice.

Second engineer

The second engineer held a Latvian second engineer’s certificate of competency and the equivalent Bahamas certificate of recognition. They had over 18 years of seagoing experience with the last 8 years in the rank of second engineer. Their seagoing experience was served on board various types of merchant ships and DP vessels.

This was their second trip on board MPV Everest with Fox Offshore and their first time working in polar waters and in ice.

Other watchkeeping engineers

The third engineer on duty at the time of the fire held an Indonesian chief engineer’s certificate of competency and the equivalent Bahamas certificate of recognition. They had about 18 years of seagoing experience and this was their first time working with Fox Offshore, their first time on board MPV Everest and their first time working in polar waters and in ice.

The off-duty third engineer held an Indonesian second engineer’s certificate of competency and the equivalent Bahamas certificate of endorsement. They had about 13 years of seagoing experience and this was their second trip on board MPV Everest with Fox Offshore and their first time working in polar waters and in ice.

Chief electrician

The chief electrician held a Finnish electro-technical officer’s certificate of competency and the equivalent Bahamas certificate of recognition. They had over 44 years of seagoing experience with the last 21 years as chief electrician. The chief electrician had worked with Fox Offshore and on board MPV Everest since the ship was built and had been assigned to the ship for the previous 4 years.

The chief electrician had previous experience working in ice in the Arctic, but this was their first time in the Antarctic.

Electro-technical officer

The electro-technical officer (ETO) held a Russian electro-technical officer’s certificate of competency and the equivalent Bahamas certificate of recognition. They had about 19 years of seagoing experience and this was their first time working with Fox Offshore and their first time on board MPV Everest.

The ETO had previous experience working in ice in the Arctic, but this was their first time in the Antarctic.

Safety management system

MPV Everest was managed and operated by Singapore-based Fox Offshore from the time of the ship’s commissioning, with shore management personnel based in Singapore, the Netherlands, Indonesia, and South Korea.

The International Safety Management (ISM) Code[19] requires companies to provide for safe practices in ship operations, to assess all identified risks to ships, personnel and the environment and, to establish appropriate safeguards against these risks. The code aims to achieve this by requiring companies to develop, implement and maintain a safety management system (SMS), with instructions and procedures to ensure the safe operation of ships, to prepare for and respond to emergencies and to verify compliance with the SMS.

Fox Offshore held a valid document of compliance[20] and MPV Everest held a valid safety management certificate,[21] both issued by BV on behalf of the Bahamas’ Administration.

MPV Everest operated under the Fox Offshore SMS, titled the Vessel Management System. The SMS consisted of policies, statements and various manuals, with supporting plans, checklists, forms, documents and publications. The manuals were divided into generic manuals and ship‑specific manuals that were designed to be specific to MPV Everest. The generic manuals included a management manual, safety manual and operations manual while ship-specific manuals included an on-board emergency manual, SOLAS fire training and safety training manuals and the ship’s DP manual.

Management manual

Fox Offshore’s management manual stated that the company specialised in supporting ‘high-tech’ vessels within the offshore oil, gas and construction industry. The company’s senior management team comprised the chief executive officer, chief operating officer, technical superintendent and designated person ashore. The manual stated that the one of the principal objectives of the senior management team was to ensure that the SMS was effectively implemented, complied with, resourced and improved.

The management manual stated that all vessels were to be manned by certified, qualified and medically fit personnel and that, when transferred to a new vessel or position, crew were to be thoroughly familiarised with their new roles. It also required all new crew to be provided with an introduction to the company SMS. In particular, it stated that where the master was recruited externally, extra care was to be taken to ensure they were familiar with the SMS. Most of MPV Everest’s officers and crew who were engaged for the AAD charter, including the master and chief engineer, were new to the ship and Fox Offshore. However, there was no evidence of familiarisation with the SMS being provided before they joined the ship.

Examination of crew induction and familiarisation records indicated that all new crew had been familiarised with relevant firefighting and lifesaving information as well as their duties and work space on board. 

The management manual required the master to review the SMS annually and report any findings, defects or request for change to the company. However, inaccuracies in several ship‑specific manuals (see the section titled Emergency preparedness and fire safety) indicated that this was not being carried out effectively.

The management manual also included procedures for compliance with fatigue management regulations (see the section titled Fatigue).

Operations manual

The operations manual provided further detail and procedures relating to the safe operations of the company and ship. The manual required ships to develop ship-specific emergency prompts for various emergency situations including fire (see the section titled Emergency preparedness and fire safety).

The operations manual required the master to develop standing orders setting out the master’s expectations of what was required of watchkeeping officers at sea and in port. The standing orders were required to be submitted to the operations manager ashore for review and all watchkeeping officers were to read, understand and sign the orders on board. On board MPV Everest, standing orders signed by the ship’s deck officers dating back to December 2017, were found on the ship’s bridge. However, the orders had not been developed by the new master but appeared to be existing standing orders developed by a previous master. While the orders appeared to have been endorsed by several of the ship’s previous masters (evidenced by their signatures), the document was not signed by the ship’s new master. The standing orders also did not include any reference to polar navigation or navigation in ice, nor was there any other substantial guidance or procedures in the operations manual or other manuals relating to operations in ice.

The operations manual also required the chief engineer to develop standing instructions covering their expectations on engineering practice. The chief engineer’s standing instructions also had to be submitted to the technical superintendent ashore for review. The manual directed the chief engineer to consider several principles in drafting their standing instructions including that maintenance and repairs could be carried out by the watchkeeping engineer provided it did not interfere with normal watchkeeping and that the chief engineer was to be informed of any abnormality in machinery performance. The chief engineer’s standing instructions largely reflected these considerations and were signed by the chief engineer and all engineering staff.

There was no evidence of the master’s standing orders and the chief engineer’s standing instructions being reviewed or approved by the operations manager ashore.

The operations manual also included procedures related to the management of critical spare parts on board and the planned maintenance system (see the section titled Engineering challenges).

Safety manual

The safety manual documented requirements for the systematic reporting of shipboard safety observations, the conduct of safety meetings and guidance on the requirements for drills and exercises on board (see the section titled Emergency preparedness and fire safety).

Some relevant safety observations documented on board during the time the ship was chartered by the AAD included:

• the absence of de-icing equipment on board resulting in some life-saving and firefighting equipment being unavailable
• the forward bridge window heater not being operational and with no spares on board
• the installed search lights on the monkey island being unacceptable for ice navigation.

The safety manual required monthly safety meetings to be held and attended at a minimum by the master, the safety officer and department representatives. The safety meetings were intended to ensure safe operations across the various departments and also to provide a medium for all staff to present safety concerns.

A review of safety meeting minutes indicated that meetings were held monthly as required. Some relevant, recurring points of discussion recorded in consecutive monthly safety meeting minutes included:

• the need to improve safety practice in the conduct of man overboard drills
• the poor quality of internet on board resulting in difficulty downloading navigation chart and publication corrections and safety flashes as well as its impact on the crew’s ability to communicate with their families.

Shipboard incidents and risk events

In the time between MPV Everest’s departure from South Korea in December 2020 and the ship’s arrival at the repair yard in Singapore following the fire, the following incidents and risk events were recorded on board:

19 December 2020

On 19 December 2020, while at sea en route to Australia, the port fast rescue craft (FRC) was lowered and launched as part of a man overboard drill. Shortly after launching, the FRC crew advised that the engine had stopped, that the boat was taking on water in the engine compartment and that assistance was required. The FRC and its crew were subsequently recovered safely. The occurrence was not recorded as an incident although a drill report was filed. The occurrence was subsequently attributed to an unsatisfactory repair of a component in the engine compartment and poor maintenance.

20 December 2020

On 20 December 2020, while at sea en route to Australia, a smoke detector activated in the ship’s aft winch control room with smoke observed emanating from the compartment. The fire alarm was sounded, crew mustered, ventilation shut down and the space electrically isolated. A fire team, wearing fireman outfits and self-contained breathing apparatus, entered the space and found a smouldering circuit breaker thermal relay unit in a motor control centre (MCC) enclosure. The circuit breaker was removed and the thermal relay isolated.

The shipboard investigation found that one of the MCC’s breakers had likely come loose in the electrical enclosure as a result of a period of prolonged disuse. The investigation recommended a review and long-term isolation of all live but unused switchboards.

23 December 2020

On 23 December 2020, while at sea en route to Australia, the starboard FRC was prepared for launching as part of a man overboard drill. As the FRC davit was swung out, the FRC release hook wire was found twisted and was subsequently cut to facilitate the launching of the boat. Shortly after launching, the FRC crew advised that the engine had stopped, that the boat was taking on water in the engine compartment and that assistance was required. The FRC and its crew were subsequently recovered safely. The occurrence was subsequently attributed to an unsatisfactory repair of a component in the engine compartment and poor maintenance.

29 January 2021

On 29 January 2021, while in polar waters en route from Casey station to Hobart, a fire broke out in MPV Everest’s battery room located above the bridge deck. On activation of the battery room fire detector and the fire alarm on the bridge, the 2 second mates were sent to investigate and found smoke and flames emanating from the space. The ship’s fire alarm was sounded, public address announcement made, and all personnel (including AAD staff) mustered. While the deck fire team was readied, the second mates extinguished the fire using portable fire extinguishers.

The ATSB marine occurrence brief MB-2021-001 summarised the incident and the findings of the subsequent shipboard investigation. The investigation found that, as the ship rolled heavily in rough weather, metal straps securing several batteries failed resulting in the batteries falling to the deck. The fire was subsequently ignited by short circuits in the batteries. While most were secured by metal brackets, some batteries that had been received about 5 months earlier were secured by less robust metal straps.

4 April 2021

On the evening of 4 April 2021, while en route from Mawson station to Hobart in rough weather, a remotely operated vehicle (ROV) and its tether management system broke free of their securing in MPV Everest’s ROV hangar. In addition, 2 smaller ROVs also moved out of their stowed positions. Over the following hours, members of the ship’s crew assisted by AAD personnel, worked to restrain and re-secure the equipment. The incident resulted in damage to the ship’s railings, gratings, hydraulic oil drums and the ROV tether. The dive technician’s report on the incident noted that the equipment was not properly secured due to the absence of satisfactory securing points for the ROVs and the tether management system in the ROV hangar, and due to the lack of proper securing equipment.

There were several other reported instances of damage due to inadequately secured equipment in the accommodation (such as furniture, computers and printers) and on deck (such as the spare anchor and crane hooks). However, these were not recorded as incidents and were managed as part of the ship’s operations.

20 May 2021

On 20 May 2021 (following the fire), while at anchor off Singapore awaiting berthing at the repair yard, a failed pressure switch fitting on a seawater cooling pump in MPV Everest’s starboard pump room resulted in a pressurised jet of water being directed upwards at the deckhead. The water dispersed in all directions including downwards onto the top of a motor control centre (MCC) in the vicinity of the pump. Internal examination of the electrical enclosure found significant water ingress had drenched almost all internal circuitry and electrical components. The water ingress was found to have resulted in electrical arcing, short circuiting and a minor explosion resulting in a hole in the rear of the enclosure.

The shipboard investigation found that the enclosures were incorrectly constructed, that the top panels were not adequately sealed and that there were gaps between adjacent panels. Sealing gaskets intended to prevent the entry of fluids were located inboard of fastening holes thereby providing an unimpeded path for fluid ingress (Figure 11).

Figure 11: Top of pump room electrical enclosure

Note that only about 3 fastening holes on each side of the top panel are used to secure the panel to the enclosure leaving the remaining fastening holes empty.

Source: Fox Offshore, annotated by the ATSB

Further inspection found that other similar electrical enclosures, including those in the fire-affected port engine room, were likewise deficient. These were subsequently rectified by the manufacturer during the ship’s stay in the repair yard in Singapore.

Inspections

Flag State inspection

On 24 June 2020, the Bahamas Maritime Authority conducted an annual flag State inspection of MPV Everest. The inspection report noted that all statutory certificates were in order and that records relating to the ship’s SMS, planned maintenance system (PMS), safety meetings and drills were inspected and also found in order. The inspection included a fire drill (for a fire in the ship’s paint locker), a lifeboat drill and testing of the emergency alarms, emergency fire pump and emergency generator, all of which were recorded as being satisfactory. The report also noted that the water mist system was inspected, that it was set up to operate automatically and that the system’s nozzles were found to be in the correct positions to allow for immediate operation.

Port State inspection

On 5 February 2021, following MPV Everest’s return to Hobart after the first voyage to Antarctica, AMSA port State inspectors boarded the ship and conducted a port State inspection. The inspection included testing of various fire dampers, radio equipment, lifeboat engines, emergency steering, emergency fire pump and emergency generator. The inspection also included examination of random extracts from the ship’s emergency response manual and emergency drill records. The AMSA port State inspection did not identify any deficiencies.

The ship’s previous port State inspection was conducted by the Maritime and Port Authority of Singapore on 10 December 2018. The inspection identified 4 minor deficiencies.

Engine room and fuel oil system

Engineering redundancy

MPV Everest is a class 3, dynamically‑positioned (DP3) ship. Dynamic positioning (DP) is a vessel capability where a computer-controlled system automatically maintains a vessel’s position and heading by directing its propellers and thrusters accordingly. The ship’s DP3 equipment class required that the ship have redundancy such that no single fault in an active component or system would cause the system to fail and lose position. It also required the ship to be able to withstand fire or flooding in any one compartment without losing the ability to maintain propulsion.

MPV Everest’s DP system comprised 2 separate redundancy groups with thrusters distributed between the 2 so as to provide the necessary redundancy. Therefore, MPV Everest was designed and built with 2 separate engine rooms—port and starboard. Engines, switchboards, and fuel, lubrication, cooling, and ventilation systems were duplicated in each engine room and the engine rooms were segregated such that fire or flooding in a single engine room would only result in the loss of a single redundancy group.

Propulsion

MPV Everest’s propulsion is provided by 6 electrically driven thrusters as follows:

• 2 fixed-pitch 5,500 kW azimuth thrusters aft (port and starboard)
• a controllable-pitch 600 kW tunnel thruster aft
• a retractable 3,000 kW fixed-pitch azimuth thruster forward
• 2 controllable-pitch 2,700 kW tunnel thrusters forward.

Electrical power for the thrusters and other shipboard services was provided by:

• 4 Rolls-Royce (Bergen) B32:40V12ACD 5,530 kW diesel generators
• 2 Rolls-Royce (Bergen) C25:33L6ACD 1,825 kW diesel generators.

The 2 secondary 1,825 kW diesel generators (DG) were numbered DG1 and DG6 while the 4 main 5,530 kW generators were numbered DG2, DG3, DG4 and DG5. These were evenly split between the port and starboard engine rooms with DG1, DG2 and DG3 in the port engine room and DG4, DG5 and DG6 in the starboard engine room. The layout of the port and starboard engine rooms and their components were mirror images of each other (Figure 12).

Port engine room

In the port engine room, DG1 was located outboard at the aft end of the engine room with DG2 and DG3 located side by side, forward and inboard of DG1 (Figure 12). The engine room floor was just above tank-top level with an upper mezzanine deck known as B-Deck. The upper deckhead of the engine room was formed by the underside of A-Deck. The primary means of access to the port engine room was a watertight door at the forward end of B-Deck that led into the port medium voltage switchboard room and then to the engine control room (ECR).

At the port, forward corner of the engine room B-Deck, there were 2 electrical cabinets—a 440 V motor control centre (MCC 3A) and a 440 V power distribution board (PDB 3).

Figure 12: MPV Everest – Engine rooms

Source: Fox Offshore, annotated and modified by the ATSB

Exhaust ventilation casing

At the port forward corner of B-Deck, almost directly above the MCC and PDB, was the opening of the port engine room’s exhaust ventilation casing. The casing extended vertically upwards from the deckhead to a deck just above the upper accommodation deck level (Figure 13). The intervening decks inside the casing were either open or composed of fibre-reinforced plastic (FRP) gratings.

The port engine room was equipped with 2 supply fans which created a positive pressure in the space. The resulting airflow created a strong updraft through the exhaust ventilation casing which exited through the ventilation louvres of the casing above the upper accommodation deck. The exhaust trunkings of all 3 generators led forward to the port corner of B-Deck where they turned upwards and entered the exhaust ventilation casing. The trunkings continued up to just above the main deck level where they turned and exited the ventilation casing through its forward bulkhead to continue to the ship’s funnels.

Figure 13: Port engine room and exhaust ventilation casing (looking forward)

Source: Fox Offshore, modified and annotated by the ATSB

In addition to the exhaust trunkings of the ship’s generators, air vent pipes for several engine room fuel oil tanks and other spaces also passed through the exhaust ventilation casing. These pipes ran vertically upwards with 5 pipes terminating inside the upper level of the casing while 11 pipes continued upwards, penetrated the deckhead and terminated externally above the ventilation casing. The following air vent pipes terminated in goose neck vents inside the upper level of the port engine room exhaust ventilation casing:

• port fuel oil day tank
• port fuel oil settling tank
• number 1 fuel oil purifier air vent
• number 2 fuel oil purifier air vent
• port boiler fuel oil day tank.

The gooseneck ends of the settling tank, day tank and purifier vents were sited within a save-all with a capacity of about 150 L. The save-all had a small drainpipe that drained any accumulated fluid to the sludge tank. The boiler day tank vent was sited separately within a smaller save-all (Figure 14).

Figure 14: Arrangement of air pipes within port exhaust ventilation casing

Source: ATSB

Fuel oil tanks

MPV Everest’s fuel oil storage capacity consisted of 11 storage tanks, with a combined fuel storage capacity of about 4,448 cubic metres (m³). The ship’s DP3 redundancy arrangements meant that the port and starboard engine rooms each had their own independent, segregated fuel oil system.

From the storage tanks, fuel was transferred to the settling tanks using the diesel oil transfer pumps. The ship was equipped with 2 diesel oil transfer pumps, each with a rated capacity of 30 m³ per hour. Either pump could be used to transfer fuel to the port and starboard fuel oil settling tanks.

In the settling tanks, heavier impurities in the fuel settled to the bottom due to gravity. Fuel was then purified by 2 fuel oil purifiers in each engine room and delivered to each engine room’s day tank for supply to the generators and to the boiler day tanks for the boilers.

Port fuel oil settling tank

The port fuel oil settling tank (port settling tank) had a capacity of 122.7 m³ and was located in the port engine room extending from the main deck down to B-Deck and sited inboard of a void space on the port side (Figure 15).

Figure 15: Relative locations of port settling tank and port drain tank

Source: Fox Offshore, modified and annotated by the ATSB

The tank was fitted with level sensors and transmitters with high and low level alarms. The low level alarm sensor was set to activate when fuel levels in the tank reached 30% of tank volume and the high level alarm sensor level was set at 85% of tank volume. As a safety measure, when fuel levels in the tank reached 90% of tank volume, an outlet drained fuel via a non-return valve to the port fuel oil drain tank (Figure 16). The overflow line was also fitted with a sight‑flow glass (located in the engine room store) on a vertical section of the port settling tank’s overflow line. There was no flow alarm fitted to the overflow line, nor was one required.

An air pipe for the port settling tank ran upwards through the port engine room exhaust ventilation casing and terminated in a goose neck vent within a save-all inside the casing just above the upper accommodation deck level (see Figure 13 and Figure 17).

Figure 16: Port settling tank and drain tank schematic

Source: Fox Offshore, modified and annotated by the ATSB

Port fuel oil drain tank

The port drain tank was a double‑bottom tank located amidships to port of the centreline (Figure 15). The tank had a capacity of 17.3 m³ and a high level alarm sensor set at 50% of tank volume.

The port drain tank air pipe terminated in a vent head located externally on the main deck near the port accommodation ladder (Figure 17). The drain tank and its vent head were both respectively at a lower level than the settling tank and the settling tank’s air pipe in the ventilation casing. The vent head was fitted with an automatic closing device (float), a flame screen and an electric heating element. The vent head heating could be controlled automatically (by a thermostat located on the main deck aft) or operated manually from control cabinets. The vent head heater was most probably to have been in automatic mode at the time of the fuel transfer.

Figure 17: Relative positions of port settling tank and port drain tank air pipes

Source: Fox Offshore and ATSB

Fuel oil type

The fuel in use on board MPV Everest at the time of the fire was Special Antarctic Blend (SAB) diesel. SAB diesel is a fuel used for diesel engines operating in cold weather conditions. The SAB diesel was bunkered in Hobart and stored in the ship’s number 4 port fuel oil storage tank, with various quantities intended for delivery to the AAD’s Antarctic stations. Contractual agreements between the ship’s owners and the AAD allowed for the use of surplus SAB diesel as fuel on board the ship.

Integrated automation system

MPV Everest’s integrated automation system (IAS) was designed, manufactured and installed by ASEA Brown Boveri (ABB) in Singapore. The IAS’s functionality included:

• alarms and monitoring of tanks, pumps, piping, machinery and other ship systems
• remote operation of valves and pumps
• control of bilge, ballast and fuel systems
• automatic control of machinery including engines and drive systems.

The IAS was operated through interfaces comprising a screen, keyboard and mouse, which were located around the vessel including the ECR and bridge. There were 2 IAS interfaces in the ship’s ECR.

Fuel oil transfer

Fuel oil transfers on board MPV Everest were undertaken using the IAS. Fuel transfers to the port settling tank could be undertaken in 2 modes—automatic and manual.[22]

When the automatic transfer mode was selected, the diesel oil transfer pump started automatically when fuel levels in the port settling tank reduced to 50% of tank volume and stopped automatically when fuel levels reached 80% of tank volume. Tests conducted in the repair yard after the fire and interviews with the ship’s engineers established that the automatic transfer worked satisfactorily. The ship’s previous chief engineer also confirmed that the automatic transfer worked and that they generally conducted fuel transfers in automatic mode with a few exceptions (such as when stripping the last of the fuel from a storage tank).

When the fuel transfer was being conducted in manual mode, the pump was started manually using the IAS system. The person conducting the transfer (or another delegated person) had to monitor the tank’s level and stop the pump once the desired tank level was reached. If the pump was not stopped, the fuel transfer would continue, and tank levels would continue to rise.  

Sequence of events

On the morning of 5 April 2021, the chief engineer used one of the ship’s IAS screens in the ECR to commence a routine fuel oil transfer from the number 4 port fuel oil storage tank to the port settling tank. The exact amount of fuel in the port settling tank at the time is not known, however it was estimated to be about 85 m³.

The chief engineer elected to conduct the fuel transfer in manual mode, as was their practice on board, by using the IAS to open the necessary valves and start the number 1 diesel oil transfer pump. At interview, the chief engineer stated that they preferred conducting fuel transfers manually because, in their opinion, the IAS’s 50% low level setting (which triggered the initiation of the automatic fuel transfer to the settling tanks) was too low and did not allow sufficient time for the fuel to settle satisfactorily.[23]   

At about 0926, following the initiation of the fuel transfer, the chief engineer left the ECR with the third engineer to work in the engine room. About 90 minutes later, the chief engineer was alerted to the fire by the sound of several alarms in the ECR.

Table 1 below provides a brief sequence of events with respect to the fuel transfer based on ASTB analysis of the IAS logs, alarm data and other recorded data:

Table 1: Fuel oil transfer events and alarms

Analysis of the recorded data indicated that the port settling tank probably began to overflow just before 1053 (although it is possible that it started to overflow as early as 1051). The pump almost certainly continued to run until it was stopped by the shipboard loss of power at about 1106. Therefore, based on the pump’s rated capacity, it is likely that about 6,500 litres of diesel fuel overflowed from the port settling tank air pipe in that time.

Tank level alarms

The IAS was configured to monitor various system variables (such as tank levels, engine temperatures or oil pressures) and when the variable deviated from a pre-configured range or setpoint, the system generated an alarm.

When a variable such as a tank level sensor signal deviated from its normal range by either exceeding the high level setpoint or falling below the low level setpoint, an alarm was generated. This resulted in an audible alert being generated in the ECR, with the alarm displayed on the IAS screen and printed in the alarm log.

When the engine room was being operated in ‘attended’ mode, tank level alarms only sounded in the ECR as they were categorised as ‘non-critical’ alarms.[24]

The alarm’s audible alert was silenced by acknowledging the alarm at the IAS screen in the ECR. Acknowledged alarms remained displayed on the IAS alarm list. If the alarm was then accepted, it remained displayed while the monitored variable remained outside its normal range. When the variable returned within its normal range, the alarm disappeared from the alarm list. However, if the alarm was only acknowledged (silenced), but not accepted, the alarm remained displayed on the alarm list, even if the variable returned to within its normal range.

If an alarm that was acknowledged (but not accepted), had its associated variable return to its normal range and then deviate from it again, the alarm would sound again and need to be acknowledged again to silence it. If the variable then returned to its normal range a second time before deviating and generating an alarm for the third time (without the alarm being accepted), the alarm would sound again. However, after the variable triggered the alarm for the third time, the IAS designated the variable as ‘Abnormal’. Once a variable was assigned the ‘Abnormal’ status, the IAS stopped monitoring the variable, effectively isolating it. The variable would remain isolated with no further alarms being generated for that variable until the original alarm(s) were accepted.

IAS recorded alarm data indicated that while there was a port drain tank high level alarm in the time leading up to the fire, there was no port settling tank high level alarm generated, despite the tank’s high level sensor activating at about 1002. 

Examination of the port settling tank’s alarm history showed that the tank’s high level alarm was triggered 3 times on the evening before the fire (probably due to the ship’s rolling) and resulted in the tank’s high level variable being designated ‘Abnormal’. The variable remained in an ‘Abnormal’ state at the initiation of, and during, the fuel transfer. Consequently, there was no alarm generated when the port settling tank’s fuel level exceed 85% because the variable was effectively isolated and not being monitored by the IAS.

Table 2 below lists recorded alarms associated with the port settling tank high level variable on the day of the fire and the previous evening.

Table 2: Port settling tank high level alarms

1. ‘ACT’ indicates an alarm requiring operator attention.
2. ‘RTN’ indicates that the variable being monitored has returned to within its normal range.
3. ‘ABL’ indicates that the variable has been designated ‘Abnormal’ and is therefore not being monitored by the IAS.

There were no procedures, guidance, instruction or other administrative mechanism in place to ensure that ‘Abnormal’ variables were made active again before commencing an operation (such as a fuel transfer) where safety depended on the variable being monitored by the IAS.

Port drain tank high level alarm

The chief engineer recalled hearing several alarms sound in the ECR on the morning of the fire while they were working in the starboard engine room. They also noted that the alarms were being silenced which indicated that someone in the ECR was acknowledging the audible alarms. While there was no port settling tank high level alarm generated, a port drain tank high level alarm which had the potential to warn engineers of the situation, was generated at 1030:53 (about 22 minutes before the estimated time that the port settling tank overflowed) and was acknowledged without its significance being recognised.

The duty watchkeeping engineers (chief engineer and third engineer) were in the starboard engine room working and the off-duty second engineer and third engineer were resting. This meant that a non-watchkeeping member of the ship’s engineering team (probably the chief electrician or an engine rating) acknowledged the port drain tank alarm without investigating its significance or notifying the chief engineer.

The overflow

Shortly after the port settling tank’s high level sensor activated at 1002, fuel levels in the tank increased to 90% and began overflowing into the port drain tank. At about 1030, the port drain tank’s high level alarm activated and by 1043, the tank was likely full. Further fuel entering the tank should have overflowed through the drain tank vent head into the save-all on the main deck. Subsequently however, fuel levels began to rise past 90% in the port settling tank, and eventually overflowed through the goose neck vent into the port engine room’s exhaust ventilation casing (see Figure 16 and Figure 17).

While at the repair yard in Singapore, the ship’s owners engaged a marine surveyor to examine the drain tank, its air pipe and vent head. Examination of the drain tank vent head showed that the vent head was working normally with the wire gauze flame screen clear and unobstructed. Similarly, the drain tank air pipe was inspected by camera and checked by blowing compressed air through and found to be clear of any obstructions. The surveyor’s examination also concluded that there was no trace of fuel having ever entered the air pipe although shipboard accounts of the fire dispute this conclusion as fuel was found in the port main deck save-all after the fire.

Examination of the vent head heating system found the heater in working condition when operated in manual mode. However, it was identified that heat was not developed in the vent head when the heating system was operated in automatic mode. The investigation identified that the thermostat, which controlled the heater in automatic mode, was not working as it was incorrectly electrically terminated.

Inspection of the ship’s bridge logbook showed that air temperatures in the days leading up to the fire hovered around 0° C. Air temperatures the night before and the morning of the fire also remained just above freezing point. Weather entries also showed 6-7 m swells from the west‑north‑west (from the direction of the ship’s port beam) in the 12 hours before the fire. It is likely that the drain tank vent head on the main deck was subject to freezing spray and build-up of ice in the days and nights preceding the fire. The failure of the tank vent head heater system meant that any built-up ice on the vent head likely remained frozen, either partially or fully blocking the port drain tank air vent pipe. This likely restricted the flow of overflowing fuel out of the port drain tank air pipe resulting in fuel levels in the settling tank rising in the settling tank air pipe and overflowing within the exhaust ventilation casing. 

Fuel oil management plan

MPV Everest’s SMS required the development of a fuel oil management procedure and that a fuel oil management manual was on board. While there was no evidence of these documents, there was a fuel oil management plan on board. The plan largely comprised procedures for bunkering and for ensuring compliance with international emission control regulations. The plan also included a section covering the internal transfer of fuel with some of the relevant guidance, requiring that fuel was to:

• be tested for water and bacteria before being used on board
• be transferred to the settling tanks, then to the purifier and to the day tanks
• remain in the settling tanks for at least 12 hours.

Additionally, the fuel oil:

• settling tanks were to be sampled at the drain cock and checked for water every 12 hours and the checks logged in the engine room logbook
• day tank were to be sampled at the drain cock and checked for water every 12 hours and the checks logged in the engine room logbook.

The plan did not include any guidance on the conduct of fuel transfers either by manual or automatic means, or of the associated risks and safety measures. There was also no guidance or checklists covering any special precautions that might be necessary when conducting fuel transfers in polar waters, such as checking that the air pipe vent head heating element was operating as required or that tank vent heads were ice-free.   

There were no entries in the engine room logbook to indicate that the required fuel sampling and water checks were being conducted as required.

Air pipe design

SOLAS regulations, specifically Chapter II-2, sets out the fire protection, fire detection and fire extinction regulations for ships. The stated objectives of the regulations included the prevention of fire and explosion, reducing the risk to life caused by fire and to reduce the risk of damage caused by fire to the ship, its cargo and the environment. Regulations in Part B of Chapter II-2 were directed at preventing the ignition of combustible materials or flammable liquids by, among other means, controlling leaks of flammable liquids and preventing the accumulation of flammable vapours. These regulations included arrangements for the storage, distribution, and use, of fuel, lubricating oil and other flammable oils.

In particular, the regulations[25] stated:

SOLAS regulations also required that ships be designed, constructed and maintained in compliance with the structural, mechanical and electrical requirements of a classification society recognized by the Administration.

Classification society rules

MPV Everest’s dual classification working agreement between Bureau Veritas (BV) and the Russian Maritime Register of Shipping (RMRS) delegated responsibility for the ship’s fire safety and hull and piping arrangements to BV. Consequently, BV classification society rules applied to these arrangements and systems.

BV classification society rules were consistent with SOLAS regulations on the subject of preventing overpressure in oil tanks[26] and stated that:

On board MPV Everest, the means to prevent overpressure in the port settling tank was the overflow pipe leading to the port drain tank which was fitted with a non-return valve (Figure 16). The port drain tank in turn was fitted with an air pipe which led to the open main deck. The design of the overflow system also complied with other classification society rules relating to the capacity of the drain tank, tank level alarms and installation of a sight-flow glass in the overflow line.

The use of the non-return valve on the settling tank’s overflow line necessitated another means of supplying air to the settling tank during fuel oil transfer operations. Consequently, a pipe from the settling tank was provided, sited within a save-all located inside the engine room exhaust ventilation casing. BV stated that the open end of the pipe was located well above the drain tank air pipe, situated close to ventilation louvres and was not intended for overflow or venting purposes.

The ATSB requested clarification from BV as to whether the pipe was an ‘air pipe’ for the purpose of class rules. In response, BV advised that the definitions and meanings of pipe-systems were based on the functionality and associated risk in the specific design. BV also stated that, during the review of the ship’s plans, no risk of fuel oil overflow from the settling tank air pipe was identified with regard to the normal operations of the ship.

Nevertheless, classification society rules for the arrangement of air pipes of flammable oil tanks[27] also reflected SOLAS regulations and stated that:

RMRS rules were similarly consistent with SOLAS and BV rules for piping systems. However, RMRS advised that given the distribution of responsibilities under the dual classification working agreement, RMRS was not involved in the approval of arrangements for the fuel oil settling tank air pipe arrangement.

Class design approval processes

During the design and approval process of shipbuilding, the shipyard drafts plans for the ship and its various systems based on the shipowner’s requirements. These plans are submitted to the classification society for review and validation by surveyors at the society’s local plan approval offices (LPOs). LPO surveyors review plans against the society’s rules and, where required, challenge elements of the shipyard’s proposed design by issuing technical comments. The shipyard is then required to either justify or rectify the issues raised in the comment.

LPO surveyors are granted individual delegations under the authority of the classification society based on factors such as their theoretical and practical training, job experience and completion of qualifying examinations. There are 3 primary levels of delegations (levels 1, 2 and 3) each with different levels of autonomy with respect to the approvals of ship plans. Specifically, all LPO surveyors could carry out plan approvals however level 1 surveyors had 100% of their work supervised, level 2 surveyors had 50% of their work supervised and level 3 surveyors had 10% of their work supervised.

Approval of MPV Everest’s plans

The settling tank air pipes were originally designed by the shipyard to terminate externally above the engine room exhaust ventilation casings. However, the ship’s owners requested that the design be modified such that the pipes terminated inside the casing (to prevent icing). The shipyard modified the ship’s plans as requested by the owner and submitted the plans to BV for review and approval.

The technical comment document for the review of MPV Everest’s fuel oil transfer, overflow, purifier and auxiliary service piping system drawing showed no comments relating to the settling tank pipe and its location inside the exhaust ventilation casing.

The technical comment document for the review of the ship’s tank air vent piping system included a request for the shipyard to confirm that the location and arrangement of vent pipes for fuel oil settling tanks (among others) were such that there was no risk of ingress of seawater or rainwater in the event of a broken pipe. The yard’s response to this comment was that the settling tank vent was located inside the engine room exhaust ventilation casing and that consequently, there was no risk of seawater ingress. This justification was accepted and there were no further challenges made with regard to the siting of the settling tank air pipes.

Representatives of the ship’s builders advised the ATSB that, to the extent of their knowledge, this particular siting of the settling tank air pipes was a one-off design feature.

Engineering challenges

Hand-over in South Korea

The conduct and effectiveness of the hand-over from MPV Everest’s existing crew to the crew for the AAD charter was limited by restrictions imposed by the coronavirus disease 2019 (COVID-19) pandemic (see the section titled Influence of the COVID-19 pandemic). The chief engineer reported that there was limited face-to-face interaction during the hand-over, but that the departing chief engineer provided detailed hand-over notes.

The chief engineer reported several observations from the departing chief engineer’s hand-over notes that were indicative of the suboptimal state of the ship and its engineering systems. For example, the hand-over notes advised the relieving chief engineer that: 

• there was no permanent technical superintendent assigned to the vessel [a technical superintendent was assigned to the ship in late 2020 but it had largely operated without one until then]
• the commissioning of the vessel had left a lot of open issues
• new issues were being discovered regularly due to the poor commissioning of the vessel
• the water mist system had a major error in pump layout and connections.

Planned maintenance system

The ISM Code required companies to establish procedures to ensure the ship was maintained in accordance with the relevant regulations. To achieve this, the company was required to ensure that inspections were held at regular intervals, that corrective action was taken to rectify any non‑conformances and that records of these activities were maintained. The code also required the company to identify equipment and technical systems, the sudden operational failure of which, could result in hazardous situations and provide specific measures in the SMS aimed at promoting the reliability of these systems. These responsibilities were generally met through the implementation and use of a planned maintenance system (PMS) on board comprising a risk‑based operational maintenance routine and a database of machinery, equipment and fittings, among others.

The SMS designated the chief engineer as the responsible person for ensuring timely completion of PMS jobs. The chief engineer was also required to compile and submit a list of overdue PMS jobs to the technical superintendent ashore on a monthly basis. The technical superintendent was responsible for oversight and monitoring of the PMS and for ensuring its effectiveness. However, there was no evidence that the progress of the PMS was monitored from ashore or that overdue PMS jobs were monitored on a monthly basis.

MPV Everest utilised a PMS software system known as ‘TM Master’. Evidence indicated that there were nearly 300 overdue PMS jobs by the end of January 2021 with the number reported to be increasing daily. An examination of the PMS system by the ATSB on 15 April 2021 found that over 1,800 PMS jobs had been completed between December 2020 and April 2021. However, the ATSB also found over 160 overdue PMS jobs extending into the previous year. Similarly, an examination conducted as part of the shipowner’s investigation on 29 April 2021 found 858 jobs due for completion with at least 650 jobs overdue, extending over the previous months and into 2020. These overdue jobs were not limited to the engineering department but included deck, safety and electrical jobs.

Tank inspections

The SMS required that the ship’s tanks be inspected as scheduled in the ship’s PMS. However, there was no evidence of scheduled tank inspections in the PMS for the fuel oil settling tanks, service tanks and the emergency generator fuel tank.

Spare parts

The relieving chief engineer was advised (in hand-over notes) that the ship’s PMS was challenging, with many spare parts listed multiple times or connected to the wrong or missing components, with information such as serial numbers missing. The system’s descriptions for many routine maintenance tasks were also reported to be wrong, incomplete or ambiguous.

The SMS’s operations manual required that the PMS be used to identify, monitor and maintain a satisfactory inventory of the ship’s critical spares. However, the chief engineer reported significant challenges relating to a general shortage of spare parts and equipment on board with some examples below:

• In early March 2021, the fuel oil booster pump for DG5 in the starboard engine room was replaced with the sole available spare booster pump on board. About 2 weeks later, on 18 March, the pump failed again resulting in DG5 being rendered inoperative.
• Following the fire, DG5 (operating with a fuel oil booster pump obtained from the damaged DG4) and DG6 were the sole remaining operational generators providing power for shipboard services and propulsion. However, there were several occasions during the ship’s passage to Fremantle when DG6 had to be shut down and repaired due to water leaks from the generator’s cooling water manifold. The manifold was temporarily repaired using a clamp from DG1 in the fire-affected port engine room.
• In the port engine room, the pre-lubrication pump for DG3 was inoperative with no spare pump available on board. This meant that DG3 could not be kept in stand-by mode for DP operations and could only be started manually (the chief engineer used the stand-by lubricating oil pump to start DG3 when required).
• The fuel filters on the ship’s diesel generators required replacement relatively frequently however, by the time MPV Everest arrived in Fremantle, the ship’s stock of filters had reduced to a critically low level.

There were also several other anecdotal accounts of a shortage of spare parts for the ship’s cranes, tugger winches, and FRCs. In addition, a general shortage of lashing and securing equipment on board was reported by the ship’s crew and AAD staff.

Workload and complexity

MPV Everest was a highly complex ship with a broad range of capabilities, including the various systems, equipment and machinery that allowed the vessel to operate in ice. The chief engineer reported that the workload and engineering situation on board was challenging from the outset.

MPV Everest was built in Singapore and largely operated in the tropics or in ice-free waters off Sakhalin Island. The voyages to Antarctica were the first time the ship was operated in ice and many challenging engineering issues emerged, including issues related to cooling systems, newly installed 3-way regulating valves, water production, steam systems, and icing of sea chests. This workload was further exacerbated by issues such as manning, the design of the ship’s IAS system, the PMS and weather conditions.

Engineering complement

The planned complement of engineering watchkeepers for MPV Everest’s AAD charter called for a chief engineer, 2 second engineers and 2 third engineers. An additional second engineer was added as a contingency in the event that one of the second engineers had to be relieved at short notice in the challenging crewing environment of the COVID-19 pandemic. The non-watchkeeping members of the ship’s engineering team comprised the chief electrician, ETO and engine ratings.   

The working arrangements called for 2 second engineers to keep 12-hour watches (assisted by a third engineer and an engine rating) while the third second engineer undertook day work. The chief electrician kept the 0600-1800 watch assisted by the electro-technical rating while the ETO kept the 1800-0600 watch. The chief engineer was responsible for oversighting the engineering department, maintenance as per the ship’s PMS and procurement of spare parts among other things. During the port call at Hobart following the end of the first voyage to Antarctica, one of the second engineers disembarked on compassionate grounds and another was replaced as part of a planned crew change. The ship subsequently departed for Antarctica with 2 instead of 3 second engineers.

On 23 February, when the ship arrived at Davis station, one of the second engineers was declared unfit for engineering watchkeeping duties and relieved of their engineering duties. This left only one second engineer available for duty and, hence, the chief engineer assumed the other 12-hour watch.

From then on, the sole remaining second engineer kept the 1800-0600 watch while the chief engineer kept the 0600-1800 watch in addition to their other oversighting responsibilities.

Watchkeeping practices

MPV Everest’s SMS required the chief engineer to ensure that engine room watchkeeping arrangements were adequate to maintain a safe engineering watch and that the engineer in charge of the watch was not to be assigned, and did not undertake, any task or duty that could interfere with their watchkeeping duties.

On the morning of the fire, the chief engineer was aware that a non-watchkeeping engineer was acknowledging and accepting alarms. The chief engineer acknowledged that a non-standard practice had developed on board of allowing non-watchkeeping engine room staff, in particular the ship’s highly experienced chief electrician, to acknowledge and accept engineering alarms. The chief engineer reported that they tolerated this practice as it allowed the duty watchkeeping engineers to focus on work in the engine rooms without having to frequently attend to alarms in the ECR. The development and tolerance of this practice of convenience was attributed to several different factors including the:

• high workload and complexity associated with the ship’s engineering systems
• reduced manning of the engineering team
• IAS design which restricted the audible alert for most alarms to the ECR
• high number and frequency of engineering alarms. 

Interviews with the ship’s engineers identified that a key factor that influenced engineering practice on board was the volume of alarms generated. For example, ATSB analysis of data showed that, in the 24-hour period of 4 April 2021, with the ship out of polar waters and steaming on passage for Hobart, there were 197 engineering alarms generated (an average of 1 every 7 minutes). The categorisation of tank level alarms as non-critical alarms meant that many alarms including ballast and fuel tank level alarms only sounded in the ECR. To effectively monitor these alarms would have required a watchkeeping engineer to be permanently stationed in the ECR which in turn reduced the number of personnel available for maintenance and repairs. 

Alarm fatigue

Li and others (2017) explain that alarm fatigue refers to a distrust or neglect of triggered alarms.[28] Alarm fatigue occurs when a person is exposed to a large number of frequent alarms and becomes desensitised to them, potentially resulting in longer response times or crucial alarms being missed altogether. The research went on to explain that system design plays a large role in reducing alarm fatigue and restoring operators’ trust in the system.

In circumstances where an alarm sounds even if no actual failure conditions exists, it can result in operator mistrust in the system and consequently, disuse of the system (Wickens and Holland, 2000).[29] Incidents may then occur because a valid alarm was ignored or because an alarm drawing attention to a critical condition was never announced in the first place because the system had been turned off previously.

Of the 197 alarms recorded on 4 April 2021, about 30 resulted in the monitored variable being designated ‘Abnormal’ (including the port settling tank).

On board MPV Everest, fuel tank and ballast tank high level audible alarms were frequently generated in the ECR, particularly when the ship was rolling (due to the associated movement of fluid activating the tank’s high level sensors). The IAS’s application of the ‘Abnormal’ designation inhibited the generation of such repetitive alarms. Ship’s engineers, including the chief engineer, indicated that they were aware of the IAS’s application of the ‘Abnormal’ designation and the consequent isolation of the monitored alarm variable. Further, interview evidence indicated that frequently occurring alarms (such as tank alarms) were sometimes intentionally acknowledged (silenced) but not accepted and that this practice had developed so as to allow these alarms to be designated ‘Abnormal’ thereby reducing further incidence of such alarms.

Unattended machinery space operations

MPV Everest was approved and equipped for unattended machinery space (UMS) operations. Effective UMS operations use automation systems to allow for better fatigue management for engineers, flexibility in watchkeeping arrangements and the ability to better allocate resources to maintenance and other work. However, MPV Everest was operated exclusively in ‘attended’ mode, with engineering watchkeepers on duty at all times, including while steaming on passage. The high number and frequency of engineering alarms was one of the factors that made it impractical to operate the ship in UMS mode thereby necessitating attended engine rooms at all times.

Fatigue

The International Maritime Organization (2019)[30] defined fatigue as:

Fatigue is a hazard because it affects everyone regardless of skill, knowledge and training and may affect a seafarer's ability to do their job effectively and safely. The effects of fatigue can be particularly dangerous in the transportation sector, including the aviation and maritime industry.

Fatigue can be caused by any one or more of a range of factors including:

• lack of sleep (or inadequate restorative sleep)
• poor quality of sleep and rest
• staying awake for long periods
• stress
• excessive workload (strenuous or prolonged mental and/or physical exertion)
• inadequate food and fluid intake
• adverse environmental conditions (extremes of temperature, low light levels, vibration, rolling/pitching and confined spaces)
• periods of monotony or boredom.

Fatigue can have a range of effects on human performance, including decreased short-term memory, slowed reaction time, decreased work efficiency, reduced motivational drive, increased errors of omission, lapses in attention, increased distractibility and reduced vigilance and reaction time.

Reviews of fatigue research have made the following observations:

• a common symptom of fatigue is a change in the level of acceptable risk that a person tolerates, or a tendency to accept lower levels of performance and not correct errors
• most people need 8 hours sleep each day to achieve maximum levels of alertness and performance
• fatigue is cumulative
• there is a discrepancy between self-reports of fatigue and actual fatigue levels, with people generally underestimating their level of fatigue.

An analysis of seafarer survey data dealing with the impact of fatigue on cognitive functioning and safety (errors of attention, memory and action), contained in a Cardiff University seafarer fatigue study,[31] showed that:

Fatigue management

MPV Everest’s management and safety manuals both addressed the subject of fatigue. The safety manual described the acute and chronic consequences of fatigue and emphasised that work was to be planned to ensure that personnel were receiving adequate hours of rest. The manual also described several measures that were to be taken into consideration to reduce the potential of personnel becoming fatigued including ‘minimising the possibility of extended rotations through effective crew management’.

The management manual required the master to ensure that the fitness for duty requirements of the STCW Code were complied with by ensuring that the ship's work and watchkeeping schedule allowed for each crew member to be satisfactorily rested. The SMS reflected the rest hour requirements of the STCW Code, specifying that minimum hours of rest should not be less than 10 hours in any 24-hour period and not less than 77-hours in any 7-day period, and required the master to ensure that hours of rest were recorded for all crew. Additionally, the chief engineer’s standing instructions also stated that all staff were to record their respective hours of rest.

Recorded hours of rest

The ATSB determined that the chief engineer’s rest hour records did not accurately reflect their hours of work and rest. Examination of the chief engineer’s rest hour records for the days in April leading up to the fire consistently showed the hours between 1800-0600 logged as hours of rest. However, the records also showed the chief engineer at rest between these hours on the day of the fire and in the days following the fire, which certainly was not the case. The chief engineer’s records for the months of January, February and March were similarly inaccurate and unrealistic and showed a uniform trend of the hours between 1800-0600 marked as rest hours without allowance for other activities such as drills, port calls and emergencies.

Assessment of the rest hour records of other crew showed similar unrealistic trends of their off‑duty hours logged as rest hours regardless of participation in other necessary work activities during those periods.

For example, the rest hours records for the off-duty second engineer and second mate showed them at rest on 29 January 2021 during the time of the battery room fire and ensuing emergency muster involving all personnel. Similarly, the second engineer’s rest hour records show them resting at the time of the fire on 5 April. Analysis of records for other crew similarly show them at rest during drills and emergency musters that were recorded as having been attended by all personnel. 

Biomathematical fatigue models

Biomathematical models are tools for predicting crew fatigue levels based on a scientific understanding of the factors that contribute to fatigue.[32]

The ATSB primarily uses 2 software programs— Fatigue Avoidance Scheduling Tool (FAST®) and FAID Quantum—in seeking to quantitatively predict fatigue risk. Each tool has its own specific applications and limitations.

FAST

FAST is a software decision aid designed to assess and forecast performance changes induced by sleep restriction and time of day. No planning software, including FAST, can predict fatigue or fatigue-induced errors in all cases or for all individuals. The tool can only forecast the effects of sleep and circadian rhythms on performance and cannot account for other factors that alter performance such as training, experience, stress, illness, or other variables that are known to affect performance.

The FAST software predicted that, on the morning of the fire, the chief engineer would have started their watch with an effectiveness criterion of around 58%.[33] This improved for a brief period in the early morning, before declining to around 59% again at 0926 (the time the fuel transfer commenced). At this time, the model predicted a lapse index of 8.5 and a reaction time factor[34] of 171%.

Lapses are excessively long reaction times associated with ‘micro-sleeps’. As effectiveness decreases, the probability of lapses increases. During an average day in a well-rested person, the lapse index averages 1.0 and ranges from about 0.2 to 1.5. A lapse index of 8.5 means that lapses are 8.5 times more likely than would be expected during an average day in a well-rested person.  

At 1058 (the time the fire broke out), the FAST model predicted an effectiveness of around 57%.

The FAST analysis flagged the low level of sleep in the previous 24 hours, the chronic sleep debt[35] and the hours awake as the primary contributing factors to the calculated score.

FAID Quantum

FAID does not predict fatigue but rather predicts sleep opportunity, based on the chief engineer’s interview evidence.

A FAID score indicates different levels of fatigue exposure for different work hours and can provide an indication of the likelihood of performance impairment associated with fatigue. The higher the FAID score the higher the fatigue exposure. For example, a standard 40-hour work week (Monday to Friday 0900-1700) results in a FAID score of 41 while a 40-hour week of night shifts (Monday to Friday 2300-0700) results in a FAID score of 97. Scores between 80-100 (high fatigue likelihood) have been compared to a level of fatigue-related impairment after 21-24 hours of continuous sleep deprivation. Performance at this level has also been argued to be comparable to performance at a blood alcohol content (BAC) of 0.05 percent (Dawson and Reid, 1997).[36]

The ATSB’s FAID analysis showed that from the time the chief engineer commenced work to the time of the manual fuel transfer and subsequent fire, they were operating with a peak FAID score of 106.

Conclusion

Biomathematical modelling using FAID and FAST both predicted a reduction in the chief engineer’s alertness around the time of the manual fuel transfer and the fire.

Sleep

Inadequate quantity and quality of sleep is a primary contributor to fatigue. Most people generally require 7-8 hours of sleep to achieve a maximum amount of alertness and performance. Inadequate sleep can result in acute or chronic sleep debt, which is the difference between the amount of sleep you should be getting, and the amount acquired below that. Sleep debt can be cumulative and can result in degraded performance (Orlady & Orlady, 1999, Hawkins, 1993). While many of the effects of fatigue generally only appear after substantial levels of sleep deprivation, even the loss of sleep for one night can have negative effects on several aspects of human performance.

The chief engineer kept the 0600-1800 watch and reported that on completion of their watch, they usually went to bed between 2000-2100 and typically obtained about 6 hours of sleep. They usually woke between 0330-0400 and worked for a while in their office before having breakfast and presenting in the ECR for their watch at about 0545.

The ATSB interview of the chief engineer included obtaining a 72-hour history of their work and rest hours. The night before the fire, the chief engineer reported not getting any sleep primarily due to the ship’s rolling and, by the morning of the fire, they had probably been awake for 25-30 hours. Over the previous 2 nights, it is likely that the chief engineer had obtained 5-6 hours of medium quality sleep each night.

Workload

Workload is described by Wickens and Hollands, 2000:[37]

Humans are limited in the amount of new information their brain can process at once. Once this limit of cognitive resources has been reached their performance starts to decline with increased error rates and delayed responses, thus resulting in cognitive overload and thus mental fatigue. The relationship between workload and fatigue can be complex, predominantly as both underload and overload can contribute to fatigue (Grech, Neal, & Geo, 2009).[38] ATSB fatigue information outlines that both the cognitive and physical demands of the work, as well as pace/time pressure can all impact upon an individual’s fatigue levels.

In the 40 days before the fire, the chief engineer kept a daily 12-hour engineering watch in addition to their significant responsibilities as head of the engineering department. The chief engineer also reported that several factors contributed to an extremely challenging workload for the engineering department in general and for the chief engineer specifically. This included the:

• limitations on the hand-over in dry dock in South Korea
• fact that most crew, including the chief engineer were new to the ship and to polar operations
• reduction in manning levels following the first voyage to Antarctica
• design and set-up of the ship’s IAS and alarm systems as well as more general engineering issues.

On the morning of the fire, the damage to DG4 and the ensuing loss of power generation redundancy caused the chief engineer additional stress and increased their workload further.

Distraction

Distractions during the completion of a task increase the likelihood of error. Distractions can be related to the task or from some external, unrelated source or event. An individual, or team, can also become completely occupied (fixated) with one event or task and therefore distracted from other tasks or operations requiring attention.

On the morning of the fire, the chief engineer determined that there was significant, irreparable crankshaft damage to DG4. To restore redundancy on board, the chief engineer decided that repairing DG5 using parts from the damaged DG4 was a priority task. The chief engineer then commenced a routine fuel transfer before starting work to restore DG5 to service, which most probably distracted them from monitoring the ongoing fuel transfer.

Summary

Based on reported sleep history alone, it is likely the chief engineer was experiencing a level of fatigue known to have an effect on performance due to lack of recent sleep.

The chief engineer had been performing watchkeeping duties as well as chief engineer duties for 40 days prior to the fire and had been onboard the ship for about 4 months. It is likely that this influenced cumulative fatigue, however without accurate sleep and rest data for the weeks prior to the fire this aspect was not examined in detail. Nevertheless, it is likely that the chief engineer’s alertness was reduced due to a number of other factors including the high workload and high stress environment on board, lack of sleep and distraction by the task of repairing DG5.

Origin and cause of the fire

Examination of fire scene

Following MPV Everest’s arrival in Fremantle, ATSB investigators, including those with a background in marine operations and data recovery, attended the ship to conduct interviews and collect other evidence. ATSB investigators were also accompanied by a fire investigator from Airservices Australia (ASA)[39] engaged as a subject matter expert under the Transport Safety Investigation Act 2003. The fire investigator examined the port and starboard engine rooms, accessible areas of the port engine room exhaust ventilation casing and fire-affected areas outside the ventilation casing louvres. For reasons relating to safety and accessibility, areas of the port engine room and casing had been cleaned following the fire. This included cleaning of fuel and fuel residues and removal of some fire-affected materials, debris, and equipment. While this impacted to some extent on assessing the evidence, the following is a summary of relevant observations from these examinations.

Port engine room – Tank top

The port engine room tank top showed mild sooting increasing from aft to forward. There was increased sooting observed on the forward, upper surfaces of DG2 and DG3, with DG2 more heavily affected than DG3. Colour changes to polycarbonate light fittings on this deck indicated that temperatures had reached at least 145° C. At the forward end of the tank top, there was evidence of fuel at deck level and damage to plastic fixtures there indicated that temperatures had been higher forward than further aft.

Port engine room – B-Deck

B-Deck generally showed increasing thermal damage and soot density from aft to forward, increasing with height. Colour changes to light fittings on the centreline bulkhead (on the starboard side of the port engine room) indicated temperatures had reached at least 145° C with no evidence of thermal damage to electrical cabling and pipework in this area. In the middle of the engine room, melted light fittings indicated temperatures of up to 300° C with cabling showing signs of sooting but only minimal thermal damage. On the port side, the forward end of a light fitting was melted while the aft end was intact but deformed indicating increased temperatures forward (Figure 18). Similarly, paint on handrails on centreline of the engine room was heavily sooted but intact whereas paint on the handrails on the port side had either decomposed or blistered with the extent of the damage increasing moving forward.

At the forward end of the port engine room, light fittings located on the bulkhead displayed complete decomposition of the light fittings indicating temperatures here had exceeded 480° C. Soot density and thermal damage both increased moving from starboard to port towards the area where the exhaust ventilation casing ascended upwards from the engine room. Lagged pipework in this area displayed areas of thermal cracking increasing with height indicating rising thermal temperatures from deck to deckhead.

There was no fire or thermal damage to DG1, DG2 or DG3 although sooting density increased on DG2 and DG3 from aft to forward.

Figure 18: Light fitting on B-Deck

Source: ATSB

DG exhaust trunking

The exhaust trunkings for the 3 diesel generators comprised a metal exhaust duct surrounded by insulation material contained within an exterior galvanised sheet metal cover with brace joinery securing the trunking at various locations. The section of exhaust trunking emerging from DG3 showed sooting but no sign of oxidation while the corresponding section from DG2 showed sooting, unburnt diesel residue and some evidence of surface combustion on lagged areas. The exhaust trunking emerging from DG1 showed no thermal damage, with sooting increasing from aft to forward.

All 3 exhaust trunkings showed heavy sooting. The underneath of the trunking for all 3 generators had zinc oxide trails, particularly where the trunking bent upwards to enter the exhaust ventilation casing indicating localised surface temperatures above 450° C. The upper surfaces of DG2’s trunking showed evidence of increased oxidisation, presenting as orange oxides, indicating that the surface temperature was above 750° C, particularly around the location of the brace securing (Figure 19).

Figure 19: Upper surfaces of DG2 exhaust trunking

Source: ATSB

Electrical enclosures

The external surfaces of the power distribution board on B-Deck (PDB 3) showed thermal damage to all surfaces. Almost all indicators, lights and switches had fallen into the cabinet and the single intact switch had melted and charred indicating direct flame contact. The interior showed no signs of combustion within the cabinet, other than localised charred plastics and thermal damage to the upper door seals. There was no evidence of arcing of the lights, gauges, and switches. There was sooting present across the PDB’s bus bars but no signs of arcing. There were signs of diesel residue throughout the interior and pooled liquid diesel was present at the bottom.

The external surfaces of the 440 V motor control centre (MCC 3A) showed more thermal damage to the upper and aft-facing surfaces. The top of the enclosure showed increased thermal damage on the starboard side below the exhaust trunking of DG2 and DG3. The front of the enclosure (consisting of 5 columns of doors) showed more thermal damage to the starboard side than the port side. The port side doors showed more thermal damage to lower door surfaces, indicating combustion at deck level, while the starboard doors showed even thermal damage to both upper and lower door surfaces indicating combustion occurring above the cabinet and at deck level. The starboard end of the MCC presented with increased thermal damage radiating from lower to higher surfaces (Figure 20).

Figure 20: Electrical enclosures at forward end of B-Deck

Source: ATSB

The interior of the MCC had mild sooting throughout with intact door seals showing minor thermal damage. There was melting and charring of switches and indicator lights but no evidence of arcing. A bank of indicator lights in the third column showed deeper charring than others and was removed by the ATSB for analysis. However, further examination did not identify any obvious signs of arcing or short circuiting. There was liquid diesel at the bottom of the MCC enclosure.

Exhaust ventilation casing – A-Deck

The space within the port engine room exhaust ventilation casing at the A-Deck level generally showed sooting throughout, although bulkhead paint was intact and not blistered. The deck was constructed of FRP grating and showed increased thermal damage to its upper surfaces. Various pipes, including the port settling tank air pipe passed through this space and showed signs of sooting, but with paintwork largely intact.

The exhaust trunkings of all 3 generators passed upwards through this space and were covered in sections of synthetic fabric insulation. The insulation on DG1’s trunking showed signs of sooting but no combustion. The insulation on DG2’s trunking showed signs of sooting and combustion, with the upper section of insulation showing evidence of having been soaked in fuel. Similarly, the upper section of insulation on DG3’s exhaust trunking also showed evidence of soaked fuel. There was a gap between the upper and lower sections of insulation on DG3’s exhaust trunking, exposing a flanged joint where 2 sections of the trunking were bolted together (Figure 21). There was oxidisation present on the exposed section of the flange with the burn pattern indicative of being affected by airflow from right to left. The section of insulation located above the exposed flange showed signs of combustion with oxidisation and burn patterns radiating upwards and outwards indicative of an upwards airflow.

Figure 21: Gap in sections of insulation on exhaust trunking of DG3

Source: TMC Marine, modified and annotated by the ATSB

Exhaust ventilation casing – Main deck

At the main deck level, the exhaust trunkings of the 3 generators bent and exited the ventilation casing through the forward bulkhead thereby presenting a horizontal profile with respect to the vertical casing. The deck consisted of an FRP grating with a small section of steel deck bounding the space with evidence of liquid diesel present. The deckhead above was partially composed of FRP grating and a steel deck.

The exhaust trunking of DG1 showed mild signs of zinc oxidisation on the side facing DG2’s trunking, indicating surface temperatures were over 450° C. The exhaust trunking of DG2 and DG3 included a section of fabric insulation about midway through the space just below the bend. The insulation showed signs of combustion and evidence of being fuel soaked. The trunkings showed mild signs of zinc oxidisation below the insulation indicating surface temperatures of over 450° C. Above the insulation, the trunkings showed signs of increased oxidation evidenced by the presence of orange iron oxides indicating surface temperatures over 750° C. The increased oxidation continued along the horizontal sections of the trunking to their exit forward.

Exhaust ventilation casing – Lower forecastle deck

This section of the ventilation casing included a watertight access door which showed signs of combustion with decomposed paint, heavy sooting, and oxidisation but no fire penetration to the exterior. The bulkheads bounding the space showed a clear demarcation between lightly and heavily sooted areas indicating a boundary between areas affected by the fire plume and hot gas. Pipework in the forward end of the space were more heavily affected, with oxidisation greater near the deck increasing upwards indicating direction of the thermal source (Figure 22). Pipework located aft showed heavy sooting and accumulated unburnt diesel residue.

Figure 22: Interior of exhaust ventilation casing at lower forecastle deck level

FRP grating deck replaced on board prior to investigator access.

Source: ATSB

Cable trays located in the upper area of this space showed more combustion damage in the centre of the space (in the path of the airflow upwards) than near the bulkheads. Cables at the forward end of the space were the most fire-affected with the clean burning of the cable inner and outer sheath indicative of direct flame contact and surface temperatures exceeding 650° C.  

Exhaust ventilation casing – Forecastle, accommodation and upper accommodation decks

The section of the ventilation casing comprising the forecastle, accommodation and upper accommodation decks had no FRP grating, although the space was partially delineated by a steel deck at the accommodation deck level (Figure 23). Several of the pipes transiting these levels  bent at the accommodation deck level before continuing vertically upwards. Pipework at the forward end of the space showed even fire damage along their length with damage greater on the side facing the centre of the casing. Horizontal sections of pipe showed heavy oxidisation with minimal sooting, indicating this area was subject to direct flame and hot gases. The oxidisation was also heavier on the upper surfaces of these pipes with visible oxidised liquid trails to lower surfaces, indicative of combusting liquid fuel falling from above. Evidence of clean burn areas at the accommodation deck level and increased levels of sooting indicated areas affected by direct flame and hot gases.

Figure 23: Forecastle deck, accommodation deck and upper accommodation deck

Photograph taken looking upwards from the lower forecastle deck. The visible grating at the top of the uppermost level of the exhaust ventilation casing where the port fuel oil settling tank air pipe and save-all was located.

Source: ATSB

Exhaust ventilation casing – Upper level (Exterior)

The engine room exhaust ventilation casing terminated in a compartment on the upper accommodation deck. Access to the interior of the casing was by a manhole on the aft bulkhead. The exterior of the outboard bulkhead (port side) showed a radiating fire pattern with the highest intensity of damage at the aft end decreasing moving forward.

Airflow from the engine room exited the ventilation casing by means of 4 ventilation louvres—3 on the starboard side of the casing facing inboard and one louvre aft of the casing. The louvres could only be opened or closed manually by means of levers on the exterior of the casing. The ventilation louvres on the inboard bulkhead showed greater oxidisation and thermal damage than the aft louvre. The lower sections of the inboard louvres showed heavy sooting while the upper sections showed evidence of clean burning and complete decay of paintwork indicating temperatures over 650° C and contact with direct flame and hot gas.

The external deck which formed the upper boundary of the ventilation casing contained goose‑neck terminations of 11 pipes with 3 each in 2 save-alls and the rest terminating on the open deck (Figure 24). The goose-neck pipe terminations showed heavy sooting particularly on the side facing inboard. The deck itself showed large areas of paint damage and oxidisation. Handrails around the deck showed evidence of metal deformation due to thermal stress.

Figure 24: Top of exhaust ventilation casing

Source: ATSB

Exhaust ventilation casing – Upper level (Interior)

Inside the exhaust ventilation casing, the deck was composed of FRP grating which showed increased thermal damage to its upper surfaces. Burnt fibres from the deck grating in an area of clean burn remained in a vertical position indicating airflow from below.

The interior showed heavy sooting forward on the port bulkhead with thermal damage and a clean burn pattern evident aft. There was heavy sooting on the starboard bulkhead below the louvres with evidence of a clean burn on the louvres themselves indicating direct flame contact. Similarly, all pipework including the port settling tank vent pipe in the save-all showed oxidisation, indicating direct flame contact (Figure 25). The deckhead showed heavy sooting and increased thermal damage in the centre and aft areas.

Figure 25: Interior of exhaust ventilation casing (upper level)

Source: ATSB

Area outside exhaust ventilation casing louvres

The area outside the exhaust ventilation casing was used as a storage area for various pieces of equipment. The equipment included 2 rigid hull inflatable boats (RHIB) with outboard motors, Yokohama fenders with attached chains and shackles and several pallet cages with smaller fenders, cordage, and other equipment (Figure 26). Examination of this area showed general fire damage extending about 5 m from the ventilation louvres. The RHIBs were located directly outside the 3 ventilation louvres and generally showed greater fire damage on the side facing the louvres.

Figure 26: Boats, fenders and other equipment outside casing

Source: Fox Offshore

CCTV footage

There were several closed-circuit television (CCTV) cameras sited at various locations on board, including in the engine room, beneath the port bridge wing, looking aft at the main deck port crane and on the forecastle. At the time of the fire, the engine room CCTV cameras were not operational.

The port bridge wing camera was pointed out to sea and did not cover the exhaust ventilation casing however footage from this camera captured a plume of black smoke after the initial eyewitness sighting of flames from the ventilation casing.

Witness accounts and mobile phone footage

On the morning of the fire, personnel on the bridge included the 2 second mates and 3 AAD staff stationed on the port side of the bridge conducting seabird observations. The AAD staff were the first to sight the fire and captured mobile phone footage and photos of the flames and smoke.

VDR audio and eyewitness accounts indicated that there was an initial deflagration[40] event, with flames emanating from the inboard ventilation louvres of the casing, 3-4 seconds before recorded data showed the activation of the smoke and flame detectors in the engine room. Shortly after, witnesses reported hearing the fire alarm sounding.

Photos and video footage captured after the initial deflagration showed the RHIBs on the upper accommodation deck alight and smoke (but no flames) issuing from the louvres.

Figure 27: Photo taken just after initial deflagration

Source: Jason Mawbey

Footage captured about 35 seconds later showed a further deflagration event with large flames issuing from the ventilation louvres. Based on the above evidence, it was estimated that the initial deflagration occurred at about 1058:17 and the second deflagration occurred about 40 seconds later, at about 1058:57 (Figure 4).

Fire investigation

Based on the available evidence, the investigation identified the following 4 potential ignition scenarios:

• spontaneous ignition
• electrical arcing and spark discharge
• hot surface ignition
• electrostatic ignition.

Spontaneous ignition

Spontaneous ignition occurs when an organic material reacts with oxygen to emit heat and continues to do so until thermal runaway[41] occurs, resulting in the ignition of the material.

The tendency of a liquid fuel to ignite due to spontaneous heating depends on its flash point temperature (FP)[42] and its relationship to the fuel’s automatic ignition temperature (AIT).[43] SAB diesel has a FP of 75° C and an AIT of about 232° C.[44] Generally, liquids with a FP below 80° C evaporate readily and therefore possess little to no tendency to self-ignite. Empirical calculations also indicated that SAB diesel does not have the propensity to undergo self-heating at normal temperatures.[45] Spontaneous heating of diesel could occur under elevated temperature conditions (such as fuel soaked insulation on generator exhaust trunking). However, the time required for such heating would generally be greater than the 5.5 minutes between the tank overflowing and the activation of the first flame detector. There was also no identified physical evidence of spontaneous combustion within the fire scene.

Therefore, it was considered unlikely that spontaneous ignition was the source of ignition.

Electrical arcing and spark discharge

Electric currents have the potential to deliver a charge that, if sufficiently high, can result in either an electric arc (sustained) or spark (unsustained), which can ignite flammable products.

Examination of electrical circuits, light fixtures and other cabling in the exhaust ventilation casing and on the engine room’s B-Deck did not identify any visible signs of arcing or short circuiting. The PDB and MCC on B-Deck were examined to check for signs of arcing or short circuiting. Both electrical enclosures were fire affected externally and showed signs of sooting and pooled diesel internally. However, there were no obvious signs of arcing or short circuiting and there was no evidence of internal combustion within these electrical enclosures. Furthermore, the flame detector on B-Deck in the port engine room did not activate until about 4 seconds after the first deflagration event above at the ventilation louvres.

Therefore, it was considered unlikely that electrical arcing or spark discharge was the source of ignition.

Hot surface ignition

Hot surface ignition is a type of ignition where evaporated fuel vapours are ignited by contact with a hot surface at or above the fuel’s AIT. In addition to the fuel’s AIT, hot surface ignition is dependent on several other variables including the shape and surface area of the heated surface, quantity of fuel exposed, fuel flow, air flow and the fuel’s residence time.

SAB diesel had a lower flammable limit (LFL) of 0.7% and an upper flammable limit (UFL) of 5%.[46] Overflowing fuel would have been agitated and rapidly dispersed by the upward, positive pressure airflow resulting in the formation of a vapour cloud. In most areas of the engine room and the ventilation casing, it is likely that the LFL of the fuel was not achievable primarily due to the strong airflow upwards. However, it is likely there were areas shielded by obstructions from the airflow, where fuel splashed or settled and where the fuel/air mixture may have been optimal.

In MPV Everest’s engine room, the exhaust trunkings of DG2 and DG3 within the exhaust ventilation casing were the primary hot surfaces of concern. At the A-Deck level, there was a gap between the upper and lower sections of insulation on DG3’s exhaust trunking exposing a flanged joint. Exhaust gas temperatures for DG2 and DG3 (which were operating on the morning of the fire) were recorded in the engine room logbook as being 374°–378° C. Examination of the exposed joint found a distinct oxidisation mark on the upper flange with the section of insulation above showing charring, signs of combustion and distinct fire patterns indicating combustion leading back to the oxidised area. While the temperature at the exposed joint would have been lower than the recorded exhaust gas temperatures of about 378° C, it is possible, with sufficient time, that ignition of the fuel occurred here. Fire development would initially have been localised but then progressed up to the horizontal section of the exhaust trunking at main deck level. Calculations show that it is unlikely that a fire at this level would have been capable of igniting the fuel vapour cloud within the ventilation shaft given the sustained airflow within.

It is possible that the fire on the exhaust trunking could have ignited other fuel deposits such as on the horizontal cable tray at the lower forecastle deck level which subsequently resulted in the deflagration observed by eyewitnesses. However, the initial ignition and fire on the exhaust trunking would have produced smoke which should have been visible to the eyewitnesses on the bridge and on CCTV footage prior to the deflagration event. There were no such eyewitness reports or CCTV footage of smoke. Furthermore, the fire developing on the trunking would likely have resulted in burning droplets of fuel raining down to the engine room B-Deck and setting off the flame and smoke detectors there. However, detector activation did not occur until after the first deflagration event.

Electrostatic ignition

Liquid hydrocarbons under certain conditions can accumulate static electricity, with the potential for subsequent static discharge in a flammable fuel/air mixture. Static electricity represents electric charges that are ’static’ with the charge collected on the surface and not continually flowing in an electrical current (current electricity). There are several identified mediums of static electrification in liquid hydrocarbons. For example, electrostatic charges can be generated in hydrocarbon liquids being pumped through pipes and during splashing or spraying of liquid hydrocarbons. Intense static electric fields, in particular, can be generated where the fuel is rapidly separated (such as by the airflow in the ventilation casing) resulting in the creation of drops, aerosols, or mists. If the generated charge is significant an electrical discharge may occur.

The conductivity of a liquid determines the liquid’s capacity to generate static charges with lower conductivity improving its ability to accumulate charge. Generally, diesel is an excellent static accumulator with very low conductivity, however SAB diesel had added conductivity improvers thereby allowing charge accumulation to be minimised.[47] Nevertheless, semi‑conductive liquids can accumulate charge if charging rates are extremely high, when they are rapidly separated or where they are effectively isolated from ground.

On the morning of the fire, fuel first overflowed from the port settling tank’s air pipe into the save-all, where it splashed and was agitated, probably resulting in foaming. Foams share similar traits to aerosols, are highly flammable, oxygen enriched and capable of holding intense electric charges. Splashing, agitation and foaming probably also occurred once the save-all was full and started to overflow. As the overflowing fuel flowed downward, it was subject to various forces which probably contributed to the rapid separation and intense static electrification of the fuel including the flow of fuel over and through the FRP grating of the decks and impacting surfaces within the casing. The upward airflow from the engine room ventilation almost certainly created fuel mists and aerosols which are also capable of intense electrification. Studies have shown that fuels mists and aerosols can ignite even where temperatures are below the FP of the fuel.

It is likely that the overflow of fuel gave rise to an electrostatically charged fuel vapour cloud within the ventilation casing and also extending externally in the vicinity of the ventilation louvres. It is also likely that the fuel vapour cloud outside the ventilation louvres contained a finer mist and aerosol than the cloud inside the casing.

The area outside the ventilation louvres was used as a storage area with several isolated conductors identified that were capable of accumulating a charge from the charged fuel vapour cloud. These isolated conductors included the propeller and shroud of the RHIB (stored using rubber tyres as dunnage) and shackles on other insulated equipment. These isolated conductors were in the vicinity of grounded metal conductors such as the ship’s superstructure, storage cages and other equipment thereby providing an avenue for an electrostatic discharge and spark to occur. Calculations showed that the charged vapour cloud had sufficient charge and time to transfer charge into numerous grounded and isolated conductors that were located within the boundary areas of the fuel vapour cloud. The calculated transferred charges were also found to have sufficient charge to produce discharge energies required to ignite the surrounding flammable vapours. It is possible that static discharge from conductors occurred numerous times during the 5.5 minutes between the tank overflowing and the first deflagration, although ignition probably only occurred when a spark discharge occurred within the flammable limits of the vapour cloud.

This electrostatic discharge within the flammable vapour cloud probably resulted in ignition of the cloud. This initial deflagration event probably resulted in rapid combustion of the available fuel aerosols, mist and droplets. It probably also resulted in flaming droplets making their way down to the engine room and activating the fire detectors. There were signs of localised sustained combustion outside the ventilation casing, notably around the RHIBs and there may have been localised sustained combustion within the casing as well.

Once a secondary flammable vapour cloud developed sufficiently, a second deflagration probably occurred. It is likely that ignition of the secondary vapour cloud was due to contact with a localised fire, located either internal or external to the ventilation casing, with sustained combustion occurring thereafter.

Static electricity has been identified as a factor in numerous incidents that have resulted in deaths, injuries, and property damage across a range of industries and settings worldwide.[48] A report published by the Fire Protection Research Foundation in August 2021 identified 89 incidents since 2001 where static electricity caused an ignition or an explosion.[49] The report’s findings also indicate that the number of incidents involving static electricity is probably far greater and more common than reported or documented. Static electricity is an ever-present hazard across many industries, although it is likely that most incidents are minor and result in little damage. However, when conditions align, such as when a flammable vapour mixture in the right concentration is exposed to a static discharge of sufficient energy, severe consequences can result. 

Conclusion

It was considered unlikely that spontaneous ignition or electrical arcing (and associated spark discharge) was the source of ignition for the fire. While it was possible that the fire started as a result of hot surface ignition, as discussed above, electrostatic ignition was considered the more likely source of ignition.

Emergency preparedness and fire safety

Emergency manual

MPV Everest’s emergency response manual formed part of the ship’s SMS. The manual was intended to guide the ship’s master and crew in effectively responding to a shipboard emergency. The manual required the development and implementation of ship-specific emergency prompts for use as aide-memoires to ensure critical steps were not forgotten in emergencies. The manual also directed that these emergency prompts were to be utilised during emergency drills and training.

MPV Everest’s emergency response manual included prompts (in the form of checklists) for various emergency scenarios, including fire. While the manual’s cover page carried MPV Everest’s name, metadata from the document indicated a different ship (Jascon 25 - Emergency Response Manual).[50]

Furthermore, the emergency prompt for fire in MPV Everest’s emergency response manual contained incorrect information. For example, the prompt indicated that the ship’s engine room and paint locker were protected by a carbon dioxide (CO₂) fixed fire-extinguishing system, and directed the discharge of CO₂ in the event of fire in these spaces. However, the ship’s engine room and paint locker were protected by a water mist fixed fire-extinguishing system. Similarly, on the subject of fire drills, MPV Everest’s SMS directed that drills were to include assessment of, among other things, the ‘Use of CO₂ in the engine rooms, danger involved and precautions to be taken…’.

The emergency prompt for fire also instructed all emergency response personnel to tune their ultra-high frequency (UHF) radios to channel 1 in an emergency whereas actual shipboard practice was to use channel 5.

Drills and training

MPV Everest’s SMS stated that shipboard drills were to be used for training and to assess the ship’s emergency response plans and the ability to respond to an emergency on board. It also directed that emergency plans and checklists were to be used during drills and exercises. These plans and checklists had to be reviewed by the master and other personnel, their accuracy and effectiveness assessed, and the plans improved, if necessary. In particular, the SMS specified that drills were to be varied, realistic and, so far as practicable, conducted as if there were an actual emergency. 

Fire drills

MPV Everest’s SMS reflected SOLAS and the Bahamas Maritime Authority’s (BMA) requirements for mandatory drills. SOLAS required that every crew member participate in at least one abandon ship and one fire drill every month. SOLAS also required that fire drills be planned such that consideration was given to the various emergencies that could occur depending on the type of ship.

The Australian Maritime Safety Authority’s (AMSA) Marine Order 21[51] gave effect to SOLAS regulations relating to emergency drills and were applicable to foreign vessels in Australian waters. The marine order specified that the master was required to carry out drills for fires in the ship’s engine rooms and that drills were to include practice in shutting all ventilators for the area involved in the mock fire.

The frequency of fire, abandon ship and other emergency drills conducted on board MPV Everest were largely in compliance with the relevant requirements. However, MPV Everest’s record of drills showed that no engine room fire drills had been held in the 4 months of the ship’s AAD charter. The last drill held on board the ship involving a machinery space was on 31 January 2021 and involved the starboard boiler room. The record of the drill showed that the simulated fire was extinguished using portable extinguishers and that it did not involve familiarisation or training with the space’s water mist fixed fire-extinguishing system.

Analysis of on board training records going back to January 2020 similarly showed no fire drills involving either the port or starboard engine rooms.

Training sessions

SOLAS also required that all crew be provided with regular training and instruction in the operation and use of the ship’s fire‑extinguishing appliances at the same intervals as drills and that all the ship's lifesaving and fire-extinguishing appliances be covered within any period of 2 months. Analysis of on board training going back to January 2020 showed no training sessions involving the engine room’s water mist fixed fire-extinguishing system.

Fixed fire-extinguishing system

Design, functions and coverage

MPV Everest was equipped with a high pressure water mist fixed fire-extinguishing system designed and manufactured by Ultra Fog.

The water mist system, as designed, comprised a fresh water reservoir, a low pressure/high volume pump (600 litres/min at 3 bar) which transferred water to a feeder tank. From this tank, up to 4 high pressure pumps (with an additional spare pump) supplied water to a high pressure manifold which in turn supplied the required nozzles in the various firefighting zones (at a pressure no less than 100 bar).

In the event the fresh water reservoir ran dry, the system was capable of being supplied with sea water via the fire main. As a further contingency, the system included an accumulator unit consisting of 2 water cylinders (50 litres) and 3 nitrogen cylinders (67.5 litres) connected directly to the high pressure manifold on the pump unit. In the event power to high pressure pumps was lost or if there was an interruption of the water supply, the nitrogen propelled the water in the cylinders through to the nozzles and provided at least 1 minute of fire-extinguishing capability in the largest protected space.

SOLAS regulations required that the water mist system be capable of automatic (for ships approved for UMS operations) as well as manual release and that the water mist system be capable of rapid operation, while also reducing the risk of accidental discharge. The system was arranged to provide local application capability (to provide localised fire suppression in areas of fire hazard such as over the generators, fuel oil purifiers and boilers) as well as total flooding of protected space(s).[52]

The spaces protected by total flooding capability were: [53]

• port engine room
• starboard engine room
• aft tunnel thruster room
• port propulsion room
• starboard propulsion room
• retractable thruster room
• diving chamber complex
• diving equipment compartment
• dive emergency generator room
• number 1 bow tunnel thruster room
• number 2 bow tunnel thruster room
• emergency generator room
• port boiler room
• starboard boiler room.

The equipment and machinery protected by the water mist system’s local application capability had push button activation points located near these zones. Activation of these zones resulted in water being supplied only to the nozzles protecting the particular machinery and equipment. When total flooding was activated, all nozzles in the protected space were supplied with water. Local and total flooding could also be remotely activated from 3 control panels located on the bridge, ECR and fire control station, as well as locally from the water mist system pump starter panel (located in the engine room’s water treatment room).

In addition to the manual push-button release capability, MPV Everest’s system was also designed to activate water release automatically in the event a flame detector and a smoke detector activated in the same coverage area.

MPV Everest’s engine room fixed fire-extinguishing system also incorporated a bilge firefighting foam system which automatically activated when total flooding of the port or starboard engine rooms was activated. The bilge foam system consisted of a high pressure/low volume pump (290 litres/min at 12 bar) which supplied water through an eductor to the port and starboard engine room bilge section valves connected to the foam injector tanks. Each engine room was provided with 2 foam injector tanks, each containing 68 litres of foam concentrate.

Installation error

The engine room fixed fire-extinguishing system on board MPV Everest was incorrectly installed at the time of the ship’s commissioning and remained so at the time of the fire. The low pressure/high volume pump designed to supply the water mist system feeder tank was instead connected to the bilge foam system, while the foam system’s high pressure/low volume pump supplied the feeder tank.

Annual inspections of the system by authorised maintenance providers had failed to identify the incorrect installation, although hand-over notes to the ship’s chief engineer indicated that it was identified prior to the ship’s dry dock in South Korea in December 2020. A job to rectify the issue was included in the docking scope of work but was later cancelled for reasons that could not be determined. The ship subsequently went into service on the AAD charter with the incorrectly installed pumps, with the issue only rectified after the fire when the ship was in the repair yard in Singapore in June 2021.

Interviews with the ship’s chief engineers indicated that the error in pump installation was more likely to impact the operation of bilge foam system than the water mist system. Recorded data from the water mist system’s alarm log indicated that water mist was automatically released for the port engine room’s DG2 local application zone as a result of the activation of a flame detector and a smoke detector in the area. While the chief engineer reported manually activating other local zones within the port engine room during the response to the fire, it appears unlikely that total flooding of the space was activated. Consequently, it is unlikely that the bilge foam system was activated during the fire.

Drawings and pump technical data sheets in the water mist system installation and operation manual described the correct pump specifications and layout. However, the manual also included a specification sheet which attributed the wrong pumps to the 2 systems.

Documentation

There were 3 primary shipboard sources of documented information on the operation, coverage and functionality of MPV Everest’s water mist fixed fire-extinguishing system:

• Water mist system manual
• SOLAS fire training manual
• Ship’s fire control plan.

Examination of these documents identified inconsistencies or incomplete information relating to the system. The fire control plan omitted to indicate that a water mist system release panel was located on the bridge. Additionally, the fire control plan and SOLAS training manual indicated that the aft transformer room was protected by the water mist system while the water mist system manual made no reference to this space. There was also no mention or description of the automatic release capability of the system in the engine rooms in any of these documents.

Additionally, similar to the ship’s emergency response manual, metadata from the SOLAS fire training manual indicated a different ship (Jascon 25 – SOLAS Fire Training Manual).

Emergency response

SOLAS required that activation of a water mist fixed fire-extinguishing system should not require engine shutdown, closing of fuel oil tank outlet valves, evacuation of personnel or sealing of the space.[54]

On the day of the fire, the water mist system was triggered automatically by the activation of the port engine room B-deck flame and smoke detectors. The activation of the system and release of high-pressure water mist was also visually confirmed by the chief engineer and third engineer when initially investigating the fire alarms.

However, analysis of bridge VDR audio showed that the master consistently requested information and made decisions in the belief that the ship was equipped with a CO₂ fire‑extinguishing system. This included at least 8 references about the intention to use or release ‘CO₂’ accompanied by requests for a head count and instructions to seal the engine room. While a head count and sealing of the space on fire are routine elements of any shipboard fire response, VDR audio clearly established that the master pursued these steps in the belief that CO₂ was to be released. Early reports to AMSA’s Joint Rescue Coordination Centre (JRCC) as well as planning and briefings ashore similarly reflected the misunderstanding that the ship was equipped with a fixed CO₂ system (this information was later corrected).

The master’s first mention of a ‘water sprinkler system’ was at 1146, about 48 minutes after the fire broke out. Following a telephone conversation with the chief engineer at 1207, the master made no further mention of a CO₂ system and only referred to the ‘water mist system’. Interviews with other ship’s officers and engineers also showed varying levels of familiarity with the operation and functionality of the water mist system. For example, both the chief electrician and chief mate were unaware that the water mist system was capable of being activated automatically and believed it could only be activated manually.

Emergency muster

SOLAS regulations required ships to post a muster list detailing, among other things, the ship’s alarm signals, actions in the event of the various emergencies and emergency duties.

The muster list was also required to be revised if there was a change in the crew which necessitated an alteration. However, there was no amendment made to MPV Everest’s muster list to reflect the fact that there was only one second engineer available for duty. Consequently, the muster list in use designated the (unfit) second engineer as the commander of the engine room fire team and they accordingly mustered at their designated muster station in the ROV winch room. The engine ratings, who were also members of the engine room fire team, proceeded directly to the ECR instead of their muster station in the ROV winch room. The duty third engineer, whose muster station was also the winch room remained in the ECR while the off-duty third engineer, chief electrician and ETO reported to the ECR, which was their designated muster station. As a result, for a period of time during the initial response to the fire, almost the entire engineering complement of 9 personnel were in the ECR and separated from the fire only by the port medium voltage switchboard room.

On deck, the chief mate was the only member of the deck fire team who mustered at the designated muster station (the deck fire control station). The remaining members of the deck fire team responded variously by reporting to the bridge, the accommodation deck or by setting up hoses and commencing boundary cooling at their own initiative. The designated firefighters on the deck fire team were the bosun and an able seaman but when directed by the chief mate to don their fireman outfits and prepare to shut the ventilation casing louvres, both refused. This task was eventually carried out by the duty second mate and off-duty third engineer. However, only the 3 inboard ventilation louvres were initially closed as they were probably unaware of the existence of the aft louvre, which was not closed until much later.

Emergency generator

SOLAS regulations required that ships have an emergency source of electrical power capable of supplying electrical power, for a certain period, to services deemed essential for safety in an emergency. Where the emergency source of electrical power was an emergency generator, it had to be capable of starting automatically and supplying the required load to the emergency switchboard within 45 seconds of the failure of the main source of electrical power. The services required to be supplied by the emergency generator (for a period of at least 18 hours) included, among others, emergency lighting at key locations around the ship, a fire pump and communication systems.

MPV Everest was equipped with a Caterpillar C32 diesel engine-driven emergency generator with a rated power output of 800 kW at 1,800 rpm. The generator, its 16 m³ fuel tank and the emergency switchboard were located in a compartment on the accommodation deck (aft of, and one deck below, the upper level of the port engine room exhaust ventilation casing). The room was also fitted with an automatic polar air damper which controlled the flow of cooling air to the emergency generator.

MPV Everest’s emergency generator failed to perform as required during the response to the fire. The generator was initially isolated, and its fuel quick closing valve shut, on the master’s orders in the incorrect belief that the fire was in the emergency generator room. Following the first total loss of power on board at about 1106, fuel supply to the generator was restored and it was manually started. At about 1114, the emergency generator came on load and supplied the emergency switchboard for a short period however it failed to reach, and maintain, operating speed and subsequently shut down about 30 minutes later following several alarms.

It was subsequently found that the emergency generator’s fuel filters were clogged, probably as a result of bacterial contamination, which resulted in the generator being starved of fuel. Following the fire, the generator’s fuel tank was drained, re-filled and the emergency generator tested however the generator once again shut down following a high temperature alarm. Further investigation identified that an electrical issue with the compartment’s polar air damper restricted the flow of cooling air to the generator, resulting in high temperature alarms and the subsequent shutdown of the generator.

While records indicated that the generator was regularly tested as part of the ship’s planned maintenance and safety routines, it was generally not run on load for extended periods of time. Furthermore, there were no scheduled tasks in the ship’s PMS for the testing of the emergency generator fuel tank for the presence of bacterial contamination.

Radio communications

MPV Everest’s designated emergency communications system was channel 5 on the ship’s ultra‑high frequency (UHF) radio communication system. Shortly after the fire broke out, the chief engineer isolated all power to the port engine room, including to the port medium voltage switchboard room (adjoining the port engine room). This was immediately followed by the loss of the ship’s UHF communication system. The ship’s crew then transitioned to using the survival craft very high frequency (VHF) radios and communications were quickly re-established.

Subsequent investigation identified that the UHF system’s repeaters were supplied through the port engine room’s medium voltage switchboard room, but that this was not known to the ship’s crew at the time of the fire. The isolation of power to the port engine room and medium voltage switchboard room resulted in the loss of the UHF system, which hampered early efforts to communicate and coordinate the response to the fire. 

The master reported that the reliability of emergency communications (both UHF radio and ship’s telephones) had been a recurring concern with shortcomings identified and communicated to the ship’s managers following the aft winch control room fire on 20 December 2020. However, there had been no corrective action implemented by the time of the engine room fire on 5 April 2021.

Electrical enclosures

ATSB examination of the fire-affected port engine room found liquid diesel fuel pooled inside the MCC and PDB electrical enclosures on B-Deck.

The classification society responsible for plan approvals and surveys of the ship’s electrical equipment (including surveys of materials, equipment and components) was RMRS. Its rules required electrical equipment such as switchboards, control gear and motor starters to comply with ingress protection (IP) ratings depending on their location on board.[55] For example, the required IP rating for electrical equipment and their enclosures located in engine and boiler rooms (such as MCC 3A and PDB 3 in the port engine room) was IP 22.[56] The IP 22 rating required the equipment enclosures to be constructed such that they were protected against drops of falling liquid, which should have no harmful effect when an enclosure was inclined up to 15° from the vertical.

Similarly, the required IP rating for electrical equipment located where there was an increased danger of liquid and mechanical damage (such as in ballast pump rooms) was IP 44. The IP 44 rating required the equipment to be constructed such that liquid splashed from any direction should have no harmful effect. Furthermore, installation of such equipment in spaces where there was danger of liquid spraying, cargo dust, serious mechanical damage, aggressive fumes and/or explosive vapour or gas mixtures (such as ventilation trunks) was not advised. Similarly, these IP rating requirements for electrical equipment were also reflected in BV classification society rules.[57]

While MCC 3A and PDB 3 in the fire-affected port engine room were not installed in a ventilation trunk, they were installed directly beneath the exhaust ventilation casing, where the enclosures were exposed to diesel droplets falling from above. The presence of liquid diesel inside the 2 electrical cabinets indicated that the cabinet was not adequately protected against fluid ingress in accordance with the required IP rating. Despite that, there was no evidence that the ingress of diesel into the MCC and PDB contributed to the fire, or any other harmful consequence.

Figure 28: Fuel pooled inside electrical enclosures on B-Deck in the port engine room

Source: ATSB

The MCC and PDB enclosures on board MPV Everest were designed, manufactured and installed by Z-Power in Singapore. The fire-affected electrical enclosures in the port engine room and the seawater affected cabinet in the starboard pump room were replaced and repaired respectively in the repair yard in Singapore. Subsequently, Z-Power was engaged to carry out a detailed inspection of other electrical switchboards and enclosures on board to ensure they were properly sealed against water ingress. The inspection identified 7 other MCCs and PDBs that required remedial work to ensure they were appropriately sealed.

Australian Antarctic Division

The Australian Antarctic Program (AAP) was established in 1947 as the National Antarctic Research Expeditions. The Program comprised a range of activities, interests and capabilities including scientific research, logistics, transport and administration of the Australian Antarctic Territory. The Australian Antarctic Division (AAD) is a division of the Australian Government’s Department of Climate Change, Energy, the Environment and Water (the Department),[58] based in Kingston, Tasmania, and it leads, coordinates and delivers the AAP. 

Australia maintains 3 year-round research stations in Antarctica—Casey, Davis and Mawson—and one on sub-Antarctic Macquarie Island. The population at each station varies between 40‑100 expeditioners over summer, and between 15‑20 over the winter months. The AAD utilises air and sea transport to get expeditioners to and from the research stations. Flights[59] carrying passengers and cargo operate from Hobart, Tasmania to Wilkins Aerodrome, near Casey research station each summer and smaller fixed wing aircraft and helicopters enable travel within the continent. However, shipping is the primary means by which personnel, equipment, fuel and supplies are transported to and from Antarctica. In addition, ships are used as platforms to conduct scientific research as well as coastal and sub-sea surveys.

Australian Antarctic shipping

Since the establishment of the AAP, several ships have been chartered to support the delivery of the Program. In 1989, the Research Survey Vessel (RSV) Aurora Australis was built for the AAD at the Carrington Slipways in Newcastle, New South Wales. Aurora Australis served as the AAD’s primary supply vessel for over 3 decades, conducting over 150 scientific research and resupply voyages in that time.

In May 2017, the construction of a new Australian Antarctic icebreaker, Nuyina, commenced at the Damen Shipyards in Romania. Nuyina was to assume the role of the AAD’s primary supply vessel following the planned conclusion of Aurora Australis’ service with the AAD in early 2020. However, by November 2019, delays in Nuyina’s construction resulted in the AAD conducting contingency planning to seek alternate supply ship options for the next season, with the conclusion of Aurora Australis’s service occurring in March 2020.[60]

Request for Tender

On 26 November 2019, the Department published a request for tender (RFT) seeking a suitable ship to resupply Australia’s Antarctic stations for the 2020/2021 season. The Department’s RFT required that the:

• proposed ship be Polar Code compliant
• ship meet AMSA and international safety and environmental requirements appropriate to the work to be undertaken
• tenderer provide details of their experience providing the services to other organisations that are similar or relevant to the services being sought
• tenderer provide details of their management capability and capacity
• tenderer should provide details of any shipping accidents and serious incidents relating to the tenderer or related entities within the last 5 years
• tenderer was to ensure that the ship carried adequate equipment and spares for Antarctic service
• Department may require the master to use any special Antarctic blend fuel stored on the vessel as fuel oil for the vessel.

Regarding the operating crew, the RFT further required that:

• only crew experienced in operations in the Arctic or Antarctic would be employed, and that the tenderer would endeavour to maintain continuity of officers and crew for the duration of the charter
• a full crew list and a resume of the experience of such proposed crew would be provided to the Department in advance
• for navigation officers proposed for the vessel, the resume was to include their age, position on board, qualifications held and previous ice experience (particularly Antarctic) including the number of voyages undertaken (including the areas and the time of year).

Response to RFT

On 20 December 2019, MCS (the parent company of MPV Everest’s owners – New Orient Marine), submitted a response to the RFT proposing MPV Everest for the role of Antarctic supply ship. The MCS response to the RFT included general details about MPV Everest, the ship’s specifications and capabilities, proposed crew numbers, a summary of the company’s experience in managing offshore work and a description of 2 of MPV Everest’s previous charters. To demonstrate management capability and capacity, MCS provided an organisational chart and evidence of the existence of specific procedures related to:

• business integrity
• ethics policy
• project management
• risk assessment and management
• emergency response
• purchasing and procurement.

In response to the requirement that the tenderer should provide details of any shipping accidents and serious incidents, the MCS response noted that there was an incident investigation procedure in place.

The response to the RFT also noted that MPV Everest did not have a Polar Code certificate, but that this certification could be arranged. Additionally, in seeking to provide evidence of polar experience, MCS provided a brochure from its parent company, MRTS,[61] which had polar experience.

On 24 March 2020, the AAD and MCS signed an agreement to charter MPV Everest to undertake voyages to Antarctica at the end of 2020.

Suitability survey

In June 2020, the AAD commissioned a survey of MPV Everest to assess the ship’s suitability for its intended scope of operations. On 6 July 2020, the ship was attended by an AAD-appointed surveyor while alongside in Jurong, Singapore. The surveyor inspected the ship and examined its certificates, equipment and records. The surveyor made a few minor observations, but no deficiencies were noted and the report subsequently concluded that the ship was suitable for the intended purpose.

Planning and risk management

As part of the AAD’s planning for the use of MPV Everest, a risk assessment was conducted, which identified several potential risks. In addition to risks related to the management of COVID‑19 on board, some of the more relevant risks identified by the AAD included that:

• MPV Everest would be unable to be mobilised into the AAP
• the vessel would be unable to obtain Polar Code certification
• the vessel would become beset in ice during Antarctic operations
• the SMS used on board MPV Everest did not meet compliance
• MPV Everest’s crew was not experienced in Antarctic operations.

While each identified risk was associated with several existing and proposed risk controls, some of the more relevant measures and controls documented by the AAD, to address the identified risks included:

• the development of a project implementation plan and appointment of a project officer to manage the mobilisation of MPV Everest into the AAP
• fortnightly meetings to discuss the mobilisation of the ship into the AAP and monitor progress against relevant implementation plans
• the engagement of a surveyor to attend and undertake an inspection of the ship [with no issues subsequently identified]
• MCS development of a Polar water operational manual
• AAD provision of cold environment specialist assistance in the development of ship’s Polar water operational manual and monitoring of the development and submission of the manual to the classification society
• MCS engagement of an experienced Australian ice pilot (with AAD input) to assist in Antarctic vessel navigation
• that MCS had an SMS covering all aspects of vessel operations including vessel crewing, fatigue, maintenance, ship specific procedures, risk management processes and incident reporting
• AAD review of MCS procedures and provision of AAD standard operating procedures to MCS
• AAD standard operating procedures and the presence of an AAD voyage management team on board
• crew to be used by MCS for the AAD charter would have extensive experience in operating this ship
• MCS crew had extensive experience operating in remote and isolated Arctic regions off Russia and Europe
• the conduct of AMSA port State control inspections.

Several of the proposed risk control measures were either not implemented or could not be achieved. For example, the contents of the polar water operational manual developed for the ship was not relevant to operations in the Antarctic and the crew used for the AAD charter were almost all new to the vessel, to Fox Offshore and to the Antarctic.

The AAD charter party agreement imposed strict polar medical requirements on the ship’s crew that were in excess of the medical standards generally applicable to seafarers. This resulted in several crew who were experienced with the ship being rejected by the AAD’s Polar Medicine Unit, amplifying existing difficulties such as uncertainty associated with crew movements and the duration of crew sea service periods on board caused by the pandemic (see the section titled Influence of the COVID-19 pandemic). These led to a significantly reduced recruitment pool of crew members who were familiar with the ship and its operations.

The AAD’s project implementation plan called for MCS to provide 2 teams to crew MPV Everest. The first team was to join the ship in South Korea, sail to Australia and operate the ship until it returned to Hobart following the first voyage to Antarctica. Meanwhile, the second team was to remain in Hobart as a contingency (see the section titled Influence of the COVID-19 pandemic). Following the first voyage, the second team was to relieve the first team for the next voyage to Antarctica while the first team remained ashore as a contingency.

The progress of the plan was reviewed at fortnightly meetings commencing in July 2020. Over the following months, it became increasingly evident that travel restrictions and other measures aimed at managing the spread of COVID-19 posed a significant challenge to implementing several of the proposed risk controls, particularly those related to crewing. In November 2020, the requirement to retain a contingency crew in Hobart was removed by the AAD due to logistical and cost considerations, and the relief of crews between the first and second voyages to Antarctica did not occur. While a second engineer did travel to Australia to join the ship after the first voyage, almost all of the crew who joined in South Korea remained on board for the entire charter.

The implementation plan identified Fox Offshore as the ship’s operational managers and noted that Fox Offshore’s shipboard SMS would be used.

Influence of the COVID-19 pandemic

The first human case of a novel coronavirus was identified in December 2019 in Wuhan, China and, by 13 January 2020, the first confirmed case of the virus outside China was reported in Thailand. On 30 January, the World Health Organization (WHO) declared the outbreak a public health emergency of international concern and, on 11 March, the WHO declared COVID-19 a pandemic. As the virus spread, nations introduced strict controls on the movement of people, including measures such as mandatory quarantine, testing, isolation and lockdown requirements.

The first case of COVID-19 in Australia was reported in Victoria on 25 January 2020, followed by 3 further cases in New South Wales on the same day. On 2 March, the first confirmed case of the virus in Tasmania was reported. By 12 March, every Australian state and territory had recorded confirmed cases of the virus. On 19 March, individual states in Australia, beginning with Tasmania, began implementing strict interstate border control measures and restrictions on the movement of people. On 20 March, Australia closed its borders to all non‑residents and non‑Australian citizens. The escalating situation and the measures introduced to control it had a profound effect on the maritime industry and seafarers over the following months.

In managing the mobilisation and integration of MPV Everest into the AAP, significant effort was directed at managing the risk posed by the COVID-19 pandemic and to preventing the virus’ entry into Antarctica. The AAD’s COVID-19 plan noted that MPV Everest presented a potential single, critical point of failure in the AAD’s COVID-19 strategy. A single confirmed case of the virus on board the ship would have necessitated aborting the ship’s voyage and returning to port, thereby risking the closure and evacuation of Antarctic research stations if they could not be supplied.

The AAD COVID-19 plan included specific risk controls directed at MPV Everest including that:

• contact between outgoing and incoming crew was to be avoided in South Korea
• any confirmed cases of COVID-19 among the ship’s crew required the affected individuals to be replaced
• there was to be no shore leave during the ship’s port calls at Hobart
• there were to be no crew changes during the ship’s port call in Hobart after the first voyage to Antarctica.

On board observations

The planning, coordination and management of the AAD station re-supply operations was overseen by an AAD voyage management team, led by a voyage leader and supported by several other personnel. The voyage leader had authority over all AAD personnel on board the ship and, while the master’s authority was absolute in matters of safety, they were required to work collaboratively with the voyage leader to achieve the objectives of the voyage.

The voyage leader’s report made several recommendations in response to concerning instances, anecdotal accounts and observations related to the:

• master's knowledge of the ship's capabilities, navigation and operations in the Antarctic, and suitability for AAD operations
• conduct of musters and drills on board
• state of repair and condition of generators, availability of spare parts and their influence on redundancy
• condition of MPV Everest’s fast rescue craft
• inadequate securing of equipment on board and availability of securing and lashing gear
• familiarity of shipboard crew with AAD standard operating procedures
• sea-keeping ability and excessive rolling of MPV Everest and its suitability for operations in the Southern Ocean.

Post-accident action

Following the fire, the Department commissioned an independent investigation into the accident led by a third-party expert familiar with both charter party contract management and maritime safety. The findings of the investigation resulted in several changes to the Department’s chartering processes, including:

• the introduction of an improved process for the completion of pre-charter due diligence arrangements
• updates to the proforma contractual elements of the charter party agreement as used by the Department, with the revised agreement more directly speaking to a range of SMS requirements
• the development of a proforma bridging document that describes the functional links between the charter party’s SMS and AAD’s SMS
• the completion of season and voyage specific risk assessments
• general improvements to voyage orders as provided to the vessel’s master and the AAD voyage leader including clarification of whom reports what to whom and when.

Additionally, the AAD commissioned an independent review of the effectiveness of its procurement process and procedures for current and future chartering processes. The review identified areas of process improvement relating to:

• pre-qualification thresholds for tenderers/prime contractors, requiring demonstrable capability in ship operations
• incorporating workplace health and safety requirements across the end-to-end procurement, evaluation, implementation and operation processes
• augmenting of due diligence of prime and sub-contractor experience, capability and safety performance during the tender evaluation process
• development of standard operating procedures for implementation and mobilisation of charter vessels
• AAD emergency response procedures aboard charter vessels.

That review made recommendations with the objective of providing AAD with a framework to ensure that:

• ice-class vessels chartered from the international shipping market were fit for purpose and delivered into service with appropriately trained crews, and with robust technical support
• workplace health and safety obligations were integrated between AAD and the chartered vessel and provided AAD with visibility of safe voyage operations
• risk was assessed, controlled and transparently reported throughout the deployment.

The review also recommended that the AAD engage a maritime subject matter expert to evaluate tenderer capability and experience, in conjunction with the technical evaluation of the vessel and that the scope of experience and capability evaluations should be expanded to include organisation structure, and comparable assets under management.

Shore response

Response to MPV Everest fire

At about 1128 on 5 April,[62] following receipt of MPV Everest’s distress alert and the AAD notification of the fire, JRCC notified the Australian Defence Force’s (ADF) Headquarters Joint Operations Command (HQJOC)[63] and Maritime Border Command (MBC)[64] of the incident. JRCC requested HQJOC and MBC transmit hourly high frequency (HF) and very high frequency (VHF) digital selective calling distress relays until further notice.

At about 1139, JRCC submitted a request for ADF assistance by email to HQJOC followed by a phone call about 2 minutes later. Having considered the request, at 1156, HQJOC responded stating there were insufficient factors driving an ADF response under the circumstances and that planning activity would proceed pending receipt of further information. HQJOC planning and scoping activity eventually identified very limited feasible ADF air or maritime options for rendering assistance to MPV Everest.

As JRCC continued to work to identify suitable options to assist MPV Everest, the ship continued to provide regular SITREPs on an approximately hourly basis. At the time of the fire, MPV Everest’s location in the Southern Ocean was about 1,100 miles from Mawson station and about 1,560 miles from mainland Australia (2,060 miles from Hobart and 1,720 miles from Fremantle).

JRCC activity included gathering ship’s plans and other details from the AAD, sourcing meteorological information, identifying suitable civilian aviation and maritime assets, and contacting other regional marine rescue coordination centres (RCC) capable of assisting in the Southern Ocean. These included South Africa, France, China, Russia, New Zealand, and the United States. Subsequent advice from these RCCs as well as from other national Antarctic programs indicated that there were limited options available for immediate assistance. At about 1235, JRCC tasked a Sky Traders[65] aircraft to prepare to deploy out of Perth, Western Australia to MPV Everest’s location.

At about 1303, following discussion with JRCC, Austral Fisheries directed the Australian fishing vessel Cape Arkona, located off Heard Island and McDonald Islands, (about 925 miles from MPV Everest’s position) to proceed to MPV Everest’s location. At about 1316, the AAD confirmed that AMSA maintained operational control and responsibility for the coordination of the response and the AAD continued to provide support and information as required.

At about 1440, MPV Everest’s master advised JRCC that the fire had been extinguished and that efforts were underway to restore propulsion. JRCC subsequently stood down Cape Arkona and the Sky Traders aircraft.

Australian search and rescue arrangements

The International Civil Aviation Organization (ICAO) and the IMO coordinate member States’ efforts on a global basis to provide an effective worldwide system of search and rescue (SAR) services. JRCC Australia in Canberra has responsibility for the coordination of SAR in the Australian search and rescue region (SRR). The Australian SRR region covers the Australian continent and large areas of the Indian, Pacific and Southern Oceans as well as the Australian Antarctic Territories, an area of about 53 million square kilometres encompassing one tenth of the earth’s surface.

In Australia, the national SAR system involves Commonwealth, State and Territory authorities and organisations. At the Commonwealth level, AMSA (the Australian SAR authority) and the ADF are the relevant SAR organisations, and the national search and rescue manual is the standard reference document for use by all Australian SAR authorities and agencies.  

SAR in the Antarctic and Southern Oceans

A memorandum of understanding (MOU) between AMSA and the AAD set out the division of responsibilities for SAR activities in the Antarctic and Southern Oceans. The MOU acknowledged that the size of Australia’s Antarctic region, its remoteness from major centres, sparseness of available resources and extremes of weather meant that response options were generally limited and that SAR operations in these areas may be challenging and protracted.

The MOU essentially delegated responsibility for overall coordination of maritime SAR activities in the Antarctic and Southern Ocean areas of the Australian SRR to AMSA. While the MOU delegated responsibility to the AAD for maritime SAR activities related to AAD-sponsored operations, it also allowed for this responsibility to be transferred to AMSA in the event coordination of the response was beyond AAD capabilities (as occurred during this incident). Under the MOU, the AAD could also request ADF operational support for AAD-led SAR and emergency response operations.[66] For AMSA-led SAR operations, ADF support requests went through existing AMSA-ADF arrangements (by liaison between JRCC and the HQJOC).

ADF Southern Ocean SAR plan

The ADF Southern Ocean SAR Plan covered ADF responsibilities, capabilities and other factors that influence ADF’s ability to provide SAR response and support in Antarctic and Southern Ocean waters. Under the plan, the ADF is responsible for the provision of SAR response for all ADF and visiting foreign military forces in the Australian SRR as well as SAR for deployed ADF personnel and assets outside of the Australian SRR (subject to the approval of the relevant nations).

ADF assets may also be requested to assist in Australian-led civilian SAR operations. However, the ADF is not obliged to automatically respond to a request for assistance from JRCC for a civilian SAR incident that is not in the immediate response range of an ADF unit.

The plan noted that for the greater part of these waters, there were no dedicated SAR assets and in general, they would need to be deployed across long distances requiring up to 5-6 days’ sailing to reach some areas of the Southern Ocean depending on weather, sea and ice conditions. The plan also acknowledged that, in the Australian SRR, factors such as the particular circumstances of the Southern Ocean, Antarctica and its meteorological and ice conditions, the scarcity of support, and the limited infrastructure meant that there was added complexity to the deployment of SAR assets.

ADF aviation assets capable of responding to a Southern Ocean SAR incident were exclusively fixed wing aircraft, capable of equipment drop and search and observation functions but no 'rescue' capability.

Naval surface assets also face significant challenges when it comes to operating in and responding to SAR incidents in the Southern Ocean due to limitations on the availability of ice‑strengthened vessels, the need for winterisation of available vessels and the distances and duration involved in the response. Additionally, factors such as sea temperatures, sea state, and ice all place stresses on vessel hulls and engines and may significantly reduce the serviceable lifetime if operating in Antarctic or sub-Antarctic waters, even for short periods.

JRCC/ADF interaction

Generally, requests for military assistance in support of civilian SAR operations are to occur via the JRCC liaison with HQJOC.

In the event of a civil SAR response where JRCC considers that the ADF can assist, the practice was for JRCC to call the HQJOC joint operations room (JOR) watch commander to advise of the situation and the expected tasking. The watch commander was responsible for objectively analysing the situation and balancing any request for assistance against impacts to ADF operational capability. The JOR then scoped the circumstances of the SAR incident including location and expected tasking, contacted the service representatives to determine the disposition of appropriate forces and operational impacts and notified the chain of command as appropriate. The JOR watch commander decided if ADF assistance was feasible in the circumstances and advised JRCC of the outcome following which JRCC decides on whether to issue a formal request for ADF assistance.

In the MPV Everest accident, JRCC submitted a request for assistance to ADF followed by a phone call to HQJOC. The request resulted in the ADF undertaking planning and scoping activity to identify potential aviation and naval assets capable of assisting MPV Everest.

ADF planning subsequently identified that available aviation assets were unable to meet the scale of survival equipment that would have been required for MPV Everest’s complement of 109 personnel. While there was scoping activity conducted for an aircraft with basic capabilities deploying from Perth, Western Australia, ultimately very limited options were identified owing primarily to factors such as distance and aircraft capability.

ADF planning activity for a maritime response was similarly informed by the scope of potential response capabilities such as the possibility that all 109 personnel would need to be taken onboard, the need to tow MPV Everest and the need for medical assistance. However, it was subsequently concluded that a surface response was not feasible based on the available vessels and the transit time to MPV Everest.

Similar occurrences

Over the years, the ATSB has investigated 3 fires involving Antarctic re-supply ships. While the specific factors that contributed to these fires were different to those in this investigation, all 3 fires involved elements of the ship’s engines or fuel system.

Aurora Australis

22 July 1998

On 22 July 1998, a fire broke out in the engine room of the Australian-registered Antarctic research and supply vessel Aurora Australis. At the time, the ship was at the Antarctic ice edge about 1,300 miles south of Tasmania with 54 expeditioners and 25 crew on board. The fire, located at the forward end of the port main engine, around the turbocharger, was attacked by engineers using portable extinguishers and apparently extinguished while the engine was stopped, and the fire alarms sounded. However, a few minutes later, a fireball erupted, and the engineers were forced to evacuate the engine room.

The expeditioners and crew were mustered, fire teams deployed, the engine room’s Halon fixed fire-extinguishing system activated, and a distress message was broadcast. About 13 hours after the fire broke out, the engine room was ventilated, and inspected. Over the next 3 days, the ship’s crew carried out repairs to restore propulsive power to the starboard engine and electrical power to ancillary equipment. Aurora Australis arrived back in Hobart on 31 July, under its own power.

The ATSB transport safety investigation report number 135 found that the fire was caused by diesel fuel from a split in the flexible fuel hose in the spill line from the main engine coming into contact with a component of the port engine turbochargers. The investigation found that the failure of the hose was due to age and ‘wear and tear’, and there was no record of approval from the manufacturer, the classification society or AMSA regarding the fitting of the hoses. The investigation also identified that the Halon system had only partially discharged due to a poor electrical design and that a high turnover of engineering staff had resulted in lack of continuity in maintenance of the Halon system and a loss of knowledge relating to the fitting of the hoses.

14 January 1999

On 14 January 1999, a fire broke out in the engine room of Aurora Australis shortly after departing Fremantle for Antarctica with 16 expeditioners and 24 crew on board. The ship’s engines were stopped, alarm sounded, crew mustered, fuel quick closing valves shut, and vents and fans stopped. The engine room’s Halon fixed fire-extinguishing system was activated, and an urgency message was broadcast. Entry was made into the engine room about 3 hours later and several hours were spent checking the space and cooling any hot spots that were found with fire hoses. The port engine was eventually re-started, and the ship made its way back to Fremantle escorted by a tug, arriving without further incident the next day.

The ATSB transport safety investigation report number 143 found that the fire was caused by the failure of a flanged joint on the starboard main engine fuel supply pipework with the resulting spray of diesel fuel igniting on the turbochargers or exhaust pipework. The investigation identified that the flanged joint failed after 2 of its 4 screws failed in fatigue, and the other 2 had worked loose.

L’Astrolabe

On 11 November 2006, a fire broke out in the engine room of the French-registered Antarctic supply ship L’Astrolabe while en route from the French Antarctic base at Dumont d’Urville to Hobart. The ship’s crew operated the remotely operated fuel quick closing valves, which shut down the main and auxiliary engines, isolated all ventilation to the engine room, and then activated the engine room’s halocarbon gas fixed fire-extinguishing system. The fire was declared extinguished about an hour after it began and a few hours later, propulsion was restored and the ship resumed passage for Hobart, arriving without further incident 2 days later.

The ATSB transport safety investigation report number 234 found that the fire was the result of a spray of diesel fuel from a leak in the starboard main engine fuel system that had been ignited by contact with the hot surfaces on the engine. The investigation also found that the ship’s procedures for re-entry into the engine room after the operation of the fixed fire-extinguishing system did not adequately consider the time required to cool the fire scene and did not provide the master with adequate guidance about when to safely re-enter the space.
 

Safety analysis

On 5 April 2021, a fire broke out in the port engine room of the Bahamas‑registered, 145 m multi‑purpose vessel MPV Everest while en route from Mawson Station, Antarctica to Hobart, Tasmania with 37 crew and 72 Australian Antarctic Division staff on board. The ship’s crew responded, and the fire was contained and eventually extinguished using the engine room’s water mist fixed fire-extinguishing system about 2.5 hours later. The port engine room sustained substantial damage with most of the power generation equipment and machinery located within rendered inoperable. There were no reported injuries or pollution of the sea as a result of the fire. Power and propulsion were subsequently restored using the starboard engine room’s machinery and the ship diverted to Fremantle, Western Australia, where it arrived without further incident about a week later.

This analysis will consider the factors and circumstances leading up to the uncontrolled overflow of fuel oil in the ship’s engine room, the fire mechanism, the effectiveness of the shipboard response and general emergency preparedness on board. The analysis will also consider elements of the ship’s design, the effectiveness of MPV Everest’s safety management system (SMS), the extent to which the ship and its crew were prepared for Antarctic operations and the effectiveness of the Australian Antarctic Division’s (AAD) pre-charter due diligence and risk management arrangements. 

Finally, in considering the circumstances of this accident, it is important to recognise that the impact of the COVID-19 pandemic and associated restrictions on the maritime industry was significant. Seafarers shouldered a heavy burden during the pandemic in maintaining global trade and sustaining operations that required maritime capabilities. Disruptions to travel and quarantine measures adversely affected the ability to relieve crews and effect crew changes in a timely manner leading to uncertainty and extended time served on board, while restrictions on shore leave forced seafarers to remain on board in port. These factors, along with the general operational hazards of seafaring, particularly in the Southern Ocean and Antarctica, all played a part in creating a stressful environment and increasing the risk of fatigue for all those on board MPV Everest.

The fire

Fuel oil overflow

On the morning of the fire, the chief engineer manually initiated a routine fuel oil transfer from a storage tank to the port settling tank using the integrated automation system (IAS) in the ship’s engine control room (ECR) to start the diesel transfer pump. The chief engineer then left the ECR to carry out repairs to a diesel generator in the starboard engine room. 

The diesel transfer pump continued to run until it was eventually stopped by the loss of shipboard power about 1.5 hours after it was started. Recorded data from the IAS showed fuel levels in the port settling tank increased to 85% of the tank’s capacity, which activated the tank’s high level sensor signal (but not the associated high level alarm). The fuel level in the settling tank continued to increase until it reached 90% of tank capacity and overflowed to the port drain tank. IAS data showed the port drain tank level increasing, followed by the activation of the port drain tank high level alarm. Fuel levels continued to increase in the port drain tank until the tank was full and the settling tank level began to increase once again (see the section below titled Port fuel oil drain tank air vent).

Shortly after, personnel on the ship’s navigation bridge observed flames and smoke from the ventilation louvres of the port engine room’s exhaust ventilation casing. A few seconds later, the IAS recorded the activation of smoke and flame detectors, as well as the automatic release of the water mist system, on B-Deck in the port engine room in the vicinity of the exhaust ventilation casing.

After the fire was extinguished, inspections of the fire-affected engine room found fuel dripping down into the space from within the port engine room’s exhaust ventilation casing. Further investigation identified that the port settling tank air pipe terminated within the upper most level of the casing and that the save-all within which the pipe was located contained fuel. The upper level of the ventilation casing, where the settling tank air pipe termination was located, connected directly with the engine room several decks below. Fuel and fuel residue was also identified on all intervening decks and structures located within the casing leading down to the engine room below where fuel was also found.

The unchecked transfer of fuel to the port settling tank resulted in the overflow of fuel from the tank’s air pipe in the port engine room exhaust ventilation casing and provided the fuel for the fire.

Manual fuel transfer

Fuel oil transfers on board MPV Everest could be conducted manually or automatically. In the automated transfer mode, the IAS would initiate the fuel oil transfer when fuel levels in the settling tank dropped to 50% of tank capacity or below. The system would then automatically stop the transfer pump thereby ceasing the filling of the tank when fuel levels reached 80% of tank capacity. The automated transfer and its associated safety mechanisms were verified to be operating correctly and fuel transfers had been conducted safely in the past using the automatic transfer mode.

The chief engineer’s decision to conduct the fuel transfer manually was based on their personal and professional preference and was standard practice during their tenure on board. However, to be conducted safely, manual transfers required close supervision as the pump would continue to run and tank(s) would continue filling until the pump was stopped. The ship’s SMS and fuel oil management plan offered no specific guidance on the risks, benefits or safety measures associated with manual or automatic fuel transfers.

In this case, after starting the pump and initiating the transfer, the chief engineer left the transfer unmonitored to carry out other tasks. The chief engineer also did not delegate another engineer to monitor the operation, nor was anyone else informed of the transfer in progress. There were other safety measures in place to alert personnel that a tank was being filled past safe levels, such as tank high level alarms and the overflow drain tank. However, in this instance, these measures failed, and the transfer continued unchecked until the settling tank overflowed.

Design of port fuel oil settling tank air pipe

The main objective of the International Convention for the Safety of Life at Sea (SOLAS) is to specify minimum standards for the construction, equipment and operation of ships to ensure an adequate level of safety. In meeting these objectives, SOLAS included regulations aimed at controlling the leaks of flammable liquids, limiting the accumulation of flammable vapours and preventing the ignition of combustible materials or flammable liquids.

Flag States are responsible for ensuring that ships under their flag comply with the requirements of the Convention. In practice, this is achieved by designing, constructing and maintaining ships in compliance with the requirements of a classification society recognized by the flag State. MPV Everest was classed with Bureau Veritas (BV) and the Russian Maritime Register of Shipping (RMRS) under a working arrangement, with BV responsible for design plan approval and survey at the shipyard of the ship’s hull and piping systems, among others.

SOLAS regulations required that oil tanks be provided with a means of preventing overpressure. They also required that air pipes discharge to a position where there was no risk of fire or explosion from the emergence of oils and vapours, and that it should not lead to machinery spaces. BV rules were consistent with the SOLAS regulations and required that air pipes discharge to a safe position on the open deck where no danger would be incurred from issuing oil or gases.

MPV Everest’s fuel oil settling tanks were fitted with an overflow pipe (to prevent over‑pressurisation) which led to a drain tank fitted with an air pipe which terminated in a save-all on the open main deck. However, the settling tanks were also fitted with pipes whose purpose was to allow air into the settling tanks during fuel transfer operations. The settling tank air pipes were sited inside the engine room exhaust ventilation casing which were contiguous with the port and starboard machinery spaces several decks below. While the air pipes were situated in save‑alls, the capacity of the ship’s transfer pumps and limited volume of the save-all meant that, in the event fuel issued from the air pipe, it would inevitably overflow into the ship’s machinery spaces, as occurred during this fire.

According to BV, settling tank air pipes were not intended for the egress of fuel, nor was such a risk identified or foreseen in the course of the normal operation of the ship. Nevertheless, had the air pipes been sited to discharge to a safe position on the open deck, in compliance with the SOLAS regulations and class rules, it is likely that the engine room fire would have been prevented.

Port fuel oil drain tank air pipe

The port fuel oil drain tank, and its air pipe on the main deck, were mandated measures aimed at mitigating the risk of overfilling and over pressurisation of the port settling tank. However, in this case, the drain tank filled with fuel overflowing from the settling tank, followed by the fuel level in the settling tank increasing past the tank’s maximum capacity and overflowing into the ventilation casing and engine room.

Subsequent investigation identified that the thermostat for the port drain tank air pipe vent head’s heating element (installed to prevent the icing over of the vent head’s wire gauze flame screen) was not operational. Furthermore, weather and sea conditions in the hours and days before the fire were conducive to the port drain tank air vent head being subject to spray and icing. The drain tank vent head heating was believed to be in automatic mode which relied on the operation of the thermostat to initiate the generation of heat and melting of any ice blocking the vent head’s flame screen. In the absence of vent head heating, it is likely that any icing of the flame screen either partially, fully or temporarily obstructed the port drain tank vent head resulting in the overflow of fuel from the settling tank air pipe.

Shipboard accounts indicated that some fuel was found in the main deck save-all where the port drain tank air pipe vent was located. This allowed for the possibility that the overflow of the settling tank could not be exclusively attributed to the icing over the vent head alone but possibly to some other hydrostatic, environmental or mechanical factors, or combination of factors, either separately or in conjunction with the vent head icing. Nevertheless, the majority of evidence, including the inoperative thermostat, prevalent weather and IAS data leading up to the overflow support the likelihood that the port drain tank air pipe vent head was probably obstructed, by icing over of the vent head’s flame screen. This probably hindered, at least temporarily, the unobstructed flow of air and fuel through the port drain tank air pipe vent head and contributed to the overflow of fuel through the settling tank air pipe.

Ignition

The first indications of a fire on board MPV Everest were the sighting of flames issuing from the ventilation louvres of the port engine room’s exhaust ventilation casing. Following this initial deflagration, eyewitness accounts and mobile phone footage showed a brief lull during which there was smoke issuing from the louvres while watercraft and other equipment located just outside the louvres burned. This was followed by the activation of fire detectors on B-Deck in the port engine room and a further deflagration event with flames once again emanating from the louvres.

It is estimated that up to 6,500 litresof fuel probably overflowed from the port settling tank air pipe into its save-all. This fuel flowed out of the save-all, down into the casing and into the port engine room below. As the fuel made its way down, it settled on and in various structures within the casing, including on an exposed section of DG3’s exhaust trunking, on light fittings, cable trays and electrical enclosures in the port engine room below. The casing was also subject to a strong updraft of ventilation air which would have resulted in the rapid separation and static electrification of the fuel as well as the formation of fuel mists and aerosols which would have been expelled through the ventilation louvres.

Based on the available circumstances evidence, the investigation identified 4 potential ignition scenarios as follows:

• spontaneous ignition
• electrical arcing and spark discharge
• hot surface ignition
• electrostatic ignition.

The investigation considered it unlikely that either spontaneous ignition or electrical arcing (and associated spark discharge) was the source of ignition for the fire. There was no identified physical evidence of spontaneous combustion within the fire scene, nor did the investigation identify any obvious signs of arcing, short circuiting or internal combustion associated with electrical equipment in the affected spaces.

It was however considered possible that the fire was ignited as a result of hot surface ignition from exposure of fuel to the exposed surface of DG3’s exhaust trunking. Evidence of oxidation at the exposed section of exhaust trunking indicated that combustion had occurred at this location, and it was possible that the fire developed further and ignited a flammable vapour cloud in the casing. However, evidence associated with such a scenario, including early development of smoke and the activation of engine room fire detectors before the conflagration event was not supported by accounts and recorded data.

The investigation found that the available evidence supported the conclusion that electrostatic ignition was the probable source of ignition of the fire on board MPV Everest. The overflowing fuel subject to agitation and misting due to the strong airflow in the casing probably resulted in the formation of an electrostatically‑charged flammable vapour cloud. The area outside the ventilation louvres (where the vapour cloud was being expelled) was used as a storage area with several isolated and grounded metal conductors located in close proximity to each other. Calculations showed that the charged vapour cloud had sufficient charge and time to transfer charge into several grounded and isolated conductors located within the boundary areas of the fuel vapour cloud. An electrostatic discharge probably occurred with a grounded object, which ignited the vapour cloud and resulted in the fire.

Fatigue

In the time leading up to the fire, the chief engineer was experiencing a level of fatigue known to have an adverse effect on performance due to a lack of recent sleep.

The chief engineer reported that they obtained no sleep the night before the fire due to the ship’s rolling. By the time of the fire, they had probably been awake for about 25-30 hours and it is very likely the chief engineer was experiencing an acute sleep debt. Biomathematical modelling confirmed this, showing a predicted reduction in their alertness around the time of the fuel transfer and fire. This would almost certainly have impacted the chief engineer’s performance and would have contributed to fatigue‑related errors affecting memory, particularly short term memory (of the fuel transfer in progress) and attention (distraction by another task). In this case, the chief engineer probably suffered a lapse in attention and/or was distracted from the task of conducting the manual fuel transfer and forgot to monitor the progress of the manual fuel transfer.

It is also possible that the chief engineer was experiencing a cumulative sleep debt after performing watchkeeping duties as well as chief engineer duties for 40 days prior to the fire as well as having worked on board the ship in challenging conditions for 4 months. Additionally, other factors including high mental and physical workload, stress and distraction probably contributed to the chief engineer experiencing a reduced level of alertness.

Shipboard emergency response

Fire drills, training and emergency preparedness

In addition to the various factors that contributed to the fuel oil overflow and ignition of the fire, there were elements of the ship’s emergency preparedness and fire safety that made the response to the fire more challenging and less effective, and increased safety risk in general.

SOLAS and the ship’s flag State regulations required regular fire drills to be held with consideration given to the various potential emergencies on board. The Australian Maritime Safety Authority (AMSA) provided further clarity by specifying that shipboard fire drills were to include drills for fires in the engine room. The regulations were also consistent in requiring drills to be realistic and conducted as if there were an actual emergency on board. While MPV Everest’s SMS did not specifically mandate engine room fire drills, it was largely consistent with SOLAS regulations with regard to drills and instruction and reflected the general requirement that drills be realistic and varied. The regulations also required instruction in the use of the ship’s firefighting and life-saving appliances be provided at the same intervals as drills.

In the 4 months leading up to the fire, there had been no fire drills conducted in the ship’s engine rooms, nor was there evidence of training sessions covering the engine room’s water mist fixed fire‑extinguishing system. With a few exceptions, all the officers and crew who joined the ship for the duration of the AAD charter, were new to the ship. Consequently, when the fire broke out, most of the ship’s officers and engineers, including the master and chief engineer, had not participated in an engine room fire drill on board the ship.

The conduct of a realistic engine room fire drill would have offered an opportunity for the ship’s crew to practice the response to such an emergency and build confidence in their abilities to respond to a fire in that space. It would also have allowed the crew to identify any challenges associated with the response to a fire in the engine room.

For example, during the response, the designated firefighters in the deck fire team (bosun and able seaman) refused to carry out their duties. This resulted in the second mate and third engineer assuming their roles and proceeding to close 3 of 4 ventilation louvres on the casing. However, they omitted to close the fourth louvre on the aft side of the casing most probably due to not being familiar with the vents and their location.

Analysis of the ship’s voyage data recorder (VDR) audio, recorded during the emergency response, identified that the master consistently used the term ‘CO₂’ (carbon dioxide) to refer to the engine room’s water mist fixed fire-extinguishing system. In addition to the demanding responsibilities of command, the master was subject to similar stress and fatigue risk factors as many of the crew, particularly during the emergency response. Although it is possible that, under these circumstances, inadvertent reference to ‘CO₂’ could be made, VDR audio and interview evidence certainly indicated that this was not the case and the master’s decision-making and planning, at least initially, was based in the incorrect belief that the engine room was equipped with a CO₂ system. The automatic activation of the engine room water mist system meant that the master’s misunderstanding had no impact on the overall response. However, it increased the risk of flawed decision-making resulting in inappropriate actions and response. It also initially resulted in incorrect information being provided to shore authorities and increased the risk of adversely influencing their planning and response.

The emergency prompt for engine room fires required by the SMS also incorrectly indicated that the ship’s engine room was equipped with a CO₂ fixed fire-extinguishing system. Similarly, there were inconsistencies, relating to the coverage, operation and activation of, the water mist system, across the water mist system manual, SOLAS fire training manual and ship’s fire control plan. While it is possible the master’s misunderstanding was influenced by the incorrect emergency prompt, a realistic engine room fire drill would have allowed the crew to identify the error and to review the accuracy and effectiveness of the ship’s emergency prompts. It would also have allowed the master, other officers and crew the opportunity to familiarise themselves with the operation, coverage, capabilities and limitations of the water mist system and identify inaccuracies in supporting documentation.

Another example of a situation where a realistic fire drill would have allowed an opportunity for improvement was the failure of the ship’s ultra-high frequency (UHF) radio communication system. During the early stages of the response to fire, power to the engine room and adjacent medium voltage switchboard room was isolated. This inadvertently and unexpectedly resulted in power loss to the UHF system’s repeaters, which effectively disabled the system. While the crew re‑established communications quickly and effectively using survival craft radios, a drill involving simulated isolation of power to the affected spaces and subsequent analysis of affected services, would have allowed the crew to better prepare by designating a back-up communication system or by building additional redundancies into the system.

Water mist system

MPV Everest’s engine room water mist fixed fire-extinguishing system was incorrectly installed during the ship’s commissioning and remained so at the time of the AAD charter and fire. The error resulted in the bilge foam flooding system’s eductor pump being installed to supply the reservoir from which the water mist system’s high pressure pumps drew water and vice versa. This issue with the pump layout was known before the ship was dry docked in South Korea. However, the issue was not rectified, and the ship subsequently sailed to Antarctica with the incorrectly arranged pumps.

The general consensus among the ship’s engineers was that the incorrect arrangement was unlikely to have had a significant adverse effect on the operation of the water mist system. However, the potential effect of the incorrect pump arrangement on the performance and effectiveness of the bilge foam system was unknown. While this error did not adversely influence the water mist system’s performance in this fire, it increased the risk of an ineffective firefighting response in the event of a fuel oil fire in the engine room bilges.

Emergency generator

MPV Everest’s emergency generator failed to operate consistently for the required duration. While the generator was started, it ran for a short time before shutting down probably due to contamination in the generator’s fuel tank and an electrical issue with its automatic damper resulting in a restricted flow of cooling air. While the ship’s engineers were able to start a generator in the starboard engine room and supply the emergency switchboard, the failure of the emergency generator left the ship without power and without the means to fight the fire for a brief period of time.

Electrical enclosures

Electrical enclosures including motor control centres and power distribution boards in MPV Everest’s engine room were found to be inadequately sealed against ingress of fluids. During examination of the fire-affected port engine room, evidence of fuel ingress and pooled diesel fuel was found inside electrical enclosures in the ship’s port engine room. There was no evidence to indicate that the fuel inside was ignited by electrical arcing or spark discharge within the cabinets. However, a separate occurrence a few weeks after the fire resulted in seawater entering an electrical enclosure in the ship’s pump room causing a minor explosion, short circuiting and arcing. This incident highlighted the increased risk posed by electrical enclosures that were inadequately sealed against fluid ingress.

Engineering watchkeeping

In the time leading up to the fire, as fuel levels in the port settling tank rose, no high level alarm was generated. This was because the port settling tank level variable had been designated ‘Abnormal’ the previous evening, effectively isolating it. This feature of the IAS was known to ship’s engineers and was arguably a reasonable control measure intended to address alarm fatigue. It was also likely that practice on board had developed to allow the IAS to designate the alarm variables as ‘Abnormal’ thereby serving to reduce the incidence of alarms. However, there was no mechanism or process in place to ensure that ‘Abnormal’ variables were reinstated and monitored by the IAS before commencing an operation (such as a fuel transfer), the safety of which depended on IAS alarms. Consequently, the port settling tank high level variable remained isolated during the fuel transfer and an alarm that had the potential to alert the watchkeeping engineer to an imminent tank overflow was not generated.

On board MPV Everest, audible alerts for most alarms, including tank level alarms were only generated in the ship’s ECR. This design characteristic, combined with the high volume and frequency of engineering alarms, necessitated having an engineering watchkeeper in the ECR at all times to acknowledge, accept and manage alarms. However, the ship’s complement of available watchkeeping engineers was restricted to the chief engineer, second engineer and 2 third engineers. Further, the workload on board was generally high (and particularly so on the morning of the fire) and dedicating an engineer to managing alarms in the ECR would have reduced the resources available for planned maintenance and necessary work to keep machinery and systems operational. Consequently, it had become accepted practice on board for non‑watchkeeping engineers, such as the chief electrician, to acknowledge and accept alarms.

About 22 minutes before the port settling tank overflowed, a port drain tank high level alarm was generated. This would have triggered an audible alert in the ECR and a visual alert on the IAS screen. The chief engineer who was in the starboard engine room reported hearing audible alarms being generated in the ECR and subsequently being acknowledged (silenced). This almost certainly included the port drain tank high level alarm, although the ATSB could not conclusively establish who silenced the alarm. The significance of this alarm should have been immediately apparent to a watchkeeping engineer. However, the alarm was silenced, most probably by a non‑watchkeeping member of the ship’s engineering team, without consideration of its significance, and without further investigation.

Search and rescue in the Southern Ocean

The Southern Ocean is known for its strong winds, cold temperatures, intense storms and rough seas. Most of these waters are remote and offer limited access to safe harbours, ports or sheltered waters with very low volumes of other shipping traffic. All these factors contribute to the significant challenge of mounting effective search and rescue (SAR) operations for ships in distress in these waters.

Following notification of the fire, Australia’s Joint Rescue Coordination Centre (JRCC) assumed overall coordination of, and responsibility for, the shore response. Although the fire was eventually extinguished and propulsion restored, the initial scope of JRCC planning necessitated consideration of the need for towage assistance to recover the ship and its complement, and for SAR in the event the ship had to be abandoned.  

JRCC efforts to seek and identify suitable surface and/or aviation assets in a position to render assistance to MPV Everest were timely, appropriate and expansive. However, despite these efforts, only very limited practical options could be identified. Coordination with the rescue coordination centres of other countries bordering Australia’s SAR region and with other nations’ Antarctic programs did not identify any merchant ships or other vessels in a position to assist. Similarly, planning by the Australian Defence Force (ADF) identified very limited aviation or maritime options owing primarily to factors such as distance, associated transit time to MPV Everest, and the availability and capabilities of aircraft and naval vessels.

Ultimately, a vessel of opportunity, the fishing vessel Cape Arkona which was about 3-4 days steaming from MPV Everest’s position, was the most suitable surface asset identified and was diverted to assist MPV Everest. While an aircraft was also readied, its utility would have been limited to observation, communication and possibly providing material aid. At the time, MPV Everest, with 109 persons on board, was over 1,500 miles from mainland Australia, without power or propulsion, with a major engine room fire and very rough weather forecast. In the event the fire could not be brought under control and/or MPV Everest had to be abandoned, any ensuing SAR and recovery operations would have been extremely challenging, with a successful outcome far from assured.