The United States Coast Guard’s 47-page report on the Carnival Splendor fire on November 8, 2010, released in full on July 15, 2013, pointed to a variety of factors that led to the fire and subsequent propulsion and power loss onboard the ship, leaving it drifting off the West Coast.
The Coast Guard report talked about everything from heat exchange/air cooler issues to significant human error that started with a fire alarm panel being turned off on the bridge, in addition to a CO2 system which did not work properly.
Cruise Industry News’ editorial team dissected the report:
The relatively new vessel in the Carnival Splendor suffered what the Coast Guard said was a “major mechanical failure” in its fifth diesel generator – which was in one of two separated engine rooms.
The engine fire eventually ignited cable runs overhead, which then proceeded to burn and rendered the 2007-built ship without power.
“It took fire teams approximately two hours to locate the fire in the cable runs. Once located, the fire teams attempted to extinguish it with CO2 and dry powder portable extinguishers. However, the fire was not fully controlled by these agents due to a lack of cooling of the cable conductors which held heat and caused the cable insulation and jacket materials to continue to burn,” said the report.
Five hours later, the Captain evacuated the engine room to use the CO2 system, which did not work on either attempt for the 113,323-ton, 3,006-passenger vessel.
The Coast Guard went on to say in its executive summary that the local Hi-Fog system was activated 15 minutes after the fire started, due to a bridge watchstander resetting the fire alarm panel on the bridge multiple times.
“This was a critical error which allowed the fire to spread to the overhead cables and eventually caused the loss of power,” added the report, which also went on to note the failure of the CO2 system and “poor execution of the firefighting plan to the ineffectiveness of the crew’s firefighting effort.”
The Coast Guard investigation went minute by minute, analyzing the fire and ship’s response.
At 5:58 am on the morning of November 8, Diesel Generator 5 experienced a torsional vibration alarm (indicative of an imbalance in an operating engine and may be caused by a variety of conditions), which engineers went to investigate. They immediately heard an explosion and saw smoke, thus retreating to the engine control room, which also began to fill with smoke.
Seeing flames on their camera displays, the Chief Engineer was notified, as the engineers started emergency procedures for an engine room fire.
At 6:01, the desk officer announced the fire to the bridge, and called the fire team. At the same time, smoke detectors were activated above the generators, and a bridge officer reset them.
“As a result, the Hi-Fog system for local protection was not automatically triggered by the fire detection system,” said the Coast Guard explanation.
The Captain was on the bridge by 6:04, and the Quick Response Team staged at 6:06 with three fire teams at the engine room entrance. The fire teams entered the engine room at 6:09 and rotated firefighting duties until 10:54.
A minute after the fire teams went in, the vessel lost all primary power, and the Emergency Diesel Generator kicked in.
That lasted for a minute, and then the generator stopped, meaning the battery system kept emergency lighting and other critical systems running.
Things got worse at 6:11, when the Hi-Fog system pumps experienced a fault condition and nitrogen back-up cylinders came in – with the Hi-Fog system activating a few minutes later and providing much-needed water mist.
Passengers first heard of the fire from the Cruise Director at 6:14, with a general alarm following some 17 minutes later.
The Emergency Generator came back at 6:36, providing key power for the Hi-Fog pumps to get back up to working order.
The engine room fire was extinguished just before 9:00, but started up again roughly an hour later, before being put out and starting up yet again.
As the fire re-flashed at 10:54, per the Coast Guard report, citing the Chief Engineer on the Splendor, the Captain decided to use the CO2 system, verifying the engine room was evacuated.
The CO2 system activation proved unsuccessful from a remote station near the engine control room.
The Emergency Generator meanwhile quit again (it would start soon after, however), and the Captain opted to use the back-up controls to activate the CO2 system from Deck 11. It failed, with crew noting multiple gas leaks.
Smoke was still in the engine room come early afternoon, and the temperature inside was some 165°F, the fire was gone and a fan brought the temperatures down. A little while later, at 15:11, the crew extinguished a fire in the cabling and the ship set up regular fire patrols.
Due to the extensive damage, the ship did not have power and was towed to port.
An inspection of the Diesel Generator involved a failure primarily involving components for one set of cylinders. The Coast Guard stated that the failures of two connecting rods and piston assemblies destroyed the engine.
“The heaviest component ejected was the crankshaft counterweight which landed on the starboard side of the engine between DG (Diesel Generator) 5 and DG 6 in the primary burn area. The free end side of the counterweight of crank number one had been ejected from the crank case damaging the engine block on its way. The lower halves of each connecting rod were seized on the crank pins (crank shaft journals). The lower parts of the cylinder liners were each destroyed and the engine block severely damaged.”
Consulting engineers retained by Carnival Cruise Lines noted a connecting rod that fractured with a fatigue crack that most likely started as a result of increased surface stresses.
The engineers said that for this to happen, it would take between three and 324 operating hours at a constant speed of 514 RPM – an estimate Carnival said was conservative.
As for the engine? That had been started a day prior, and was running for eight hours and 24 minutes when it failed.
Engine operating records were requested from Carnival, but of the six main engines on the Splendor, only the main bearing temperature of one engine was recorded and maintained for historical purposes, according to the Coast Guard, adding that Carnival kept no historical records in regards to engine sensor data.
The Coast Guard explained that the ship had the technology and software to collect data, but those systems were not set up to do it. Engineers did not manually record and store engine data either, and were not required to review engine operating parameters
“Further investigation revealed that this was common practice on many Carnival vessels,” noted the report.
Work was recorded on the engine that failed on November 4, with air cooler problems experienced the prior summer. Records of this work (referred to as screen prints internally) were not provided.
Interestingly, additional data such as start and stop data along with electrical load information and “performance and efficiency measurements” were provided through a special software system back to the offices in Miami.
Torsional Vibration Alarms
Working with engine manufacturer Wartsila, the Coast Guard investigated the alarm that went off moments before the engine went – a torsional vibration alarm, which is indicative of an imbalance in an operating engine and may be caused by a variety of conditions.
A review of available data showed the alarm not to be a once-in-a-lifetime event, with noted alarms attributed to electrical overload, high temperature cooling, and other incidents on October 21 and November 5 that resulted in head gasket replacements.
Following the November 5 head gasket job, there were no torsional vibration alarms recorded from the Wartsila powerplant.
One issue with Carnival Dream-class ships is documented air cooler receiver problems. Carnival, Wartsila and Fincantieri were aware of the issue which mainly regarded cooling air going through the engine’s turbocharger.
The three companies had what amounted to a technical service bulletin on the issue calling for individual drain piping lines from each engine to a storage tank, using an air-gap to a funnel to confirm drainage was happening, and increasing the drainage pipe diameter size.
Other Carnival vessels had instances where hydrolock was a concern due to poor heat exchanger drainage, according to the Coast Guard, which cited emails from Wartsila and Carnival Dream (technical staff) having excessive accumulations of water within receiver and charger air cooler of the engine noted only after an incident with an engine.
On July 12, 2010, records shows the B bank of the failed engine was severely corroded, destroyed and leaking in regard to the air cooler heat exchanger.
After the incident, check valves were installed in the piping for three engines. “During the investigation, a number of Carnival employees stated that fluids and moisture from operating engines would back flow into the air coolers of engines that were secured due to the design of the piping system and absence of check valves. The back flow was due to the fact that they shared a common line leading to the tank and lacked check valves.”
The engines onboard the Splendor were not fitted with pressure gauges locally that measured the pressure drop across the heat exchangers, according to the Coast Guard’s findings, but the ship did have a hand-held device that could perform the task.
Closer inspection of the air coolers on the failed engine did show significant quantities of ferrous oxide and rust in the air cooler base area and outlet areas, but the heat exchangers were tested and found not to be leaking.
The report goes on: “Considering the heat exchangers were replaced in July 2010, the deterioration of the (engines) air coolers was alarming. As such, other factors may have contributed to the deterioration of the heat exchangers such as a lack of coatings, the use of dissimilar metals, and insufficient electrical grounding.”
And there is more: the report mentioned possible hydrolock conditions in addition to a "slow turn" system for the Wartsila 46C engine, enabling it to start easily.
"According to Wartsila and Carnival, the slow turn feature does not generate sufficient force to damage an engine during the slow turn ... As noted previously, a review of all available alarm data showed no slow turn mechanical failures associated with DG5 except during July of 2008."
While the first smoke/fire alarm came on at 6:00:20 and another alarm followed 13 seconds later “the Hi-Fog system was programmed with a 40 second time delay before the zone valve(s) for the aft engine room are opened.”
Moments later, a bridge watch office reset the alarm panel on the bridge, a key part of the Splendor story, and as three fire/smoke alarms went off again, the watch officer reset the alarms a second time.
The sensors alerting the bridge of the fire were destroyed, putting them in fault status by 6:03 and thus rendering them “not active.”
More alarms went off and the system was finally used for the aft engine room at 6:15.
“As a result of the intervention by the bridge watch officer, the activation of the Hi-Fog local protection system for DG 5 and 6 was delayed by approximately 15 minutes. By the time the Hi- Fog system was activated, the fire had spread to the cable runs over DG 5 and 6. Since the Hi- Fog system was designed for local protection, the nozzles were positioned below the cables runs and the Hi-Fog system was not effective in suppressing or extinguishing the fire in the cable runs.”
CO2 System Fails
The CO2 system on modern cruise ships is designed to put out a fire near-instantly, killing off the oxygen supply (assuming the area is evacuated). On the Splendor, two attempts to use this system resulted in two failures.
The Coast Guard cited leaking connectors, and control valve handles “incorrectly fitted,” as well as “a number of the fasteners securing the actuating arms to different control valves were discovered loose.”
An analysis of the system by Lucius Pitkin, Inc. found quite a few problem spots, starting with actuating levers not attached correctly to valve stems, water present in the system (corrosion), numerous leaks, and incorrect installation with pieces of hemp used to seal screwed connections which had been carried through the system and became lodged in the non-return valves preventing them from seating.
Perhaps even worse: the levers on the “ball valves in one control cabinet had been incorrectly fitted, such that the valves were open when the lever positions showed the valves to be closed.”
Sister company Princess Cruises had reported problems to Carnival Corporation with its CO2 system aboard the Crown Princess in 2009, using similar hemp thread sealant. An advisory notice followed that there was a potential risk for accidental release of CO2. While the ship had been approved by class society RINA and Carnival, Wilhelmsen found issues with the vessel’s CO2 system in 2009.
These problems included remote release stations on Deck 0 in non-functional status and challenges with the non-return valves on Deck 11. Furthermore, the pneumatic actuating system for the directional valves for the aft engine room and the forward engine room were installed backwards, among other items.
“As a result of these issues, Carnival took aggressive action to correct the deficiencies and ensure all sister vessels based on the original Carnival Destiny design were checked by their operators. Additionally, Carnival reviewed inspection reports and conducted additional inspections for the remaining Carnival vessels built in the other yards of the (Fincantieri) group.”
It does not stop there, however, as the firefighting manual aboard the vessel referred to a different CO2 system with photos and more being inaccurate, said the Coast Guard.
“Another document that appears to be part of a RINA approval letter dated December 28, 2008, describes the operational procedure in exact reverse order.”
While the onboard fire team’s response time was quite good, the Coast Guard said the team lacked familiarly with the engine room, had bad smoke boundaries, and used poor choice in extinguishing equipment.
In addition, hampering efforts was the “the Captain’s decision to ventilate the aft engine room before the fire was fully extinguished.”
The Coast Guard pointed to issues with training as well – with very little documentation on fire drills, several drills lasting under 30 minutes, and fire teams conducting drills in simulated spaces, that being, not the actual space.
The report continues that it “did not appear the Captain and bridge crew participated in the fire drills,” and that from May to November, 2010, there were no fire drills run in the engine room.
“While it appears that Carnival had met the SOLAS and STCW requirements for the training of crewmembers, the performance of the crew in this instance indicates that the training and fire drills did not sufficiently prepare the crew for this type of contingency.”
The Coast Guard has concluded the failure was started by a hydrolock event in the engine, which went undetected and led to the fatigue fraction in the connecting rod.
The poor condition of the air cooler also contributed, as did excessive rust particles.
“The poor design of the air cooler system contributed to both the deterioration of the system and the excessive accumulation of fluid in the B side air cooler for the engine,” the report said.
“While Carnival had recognized the problem with the air coolers on similar vessels, replaced the air cooler on the Carnival Splendor in July 2010 and took steps to prevent and manage the accumulation of fluid in the air coolers, the lack of formal procedures to document these measures and the poor condition of the air coolers (as observed post-casualty) suggests that these procedures were not routinely followed. Furthermore, the solutions devised after the July 2010 air cooler replacement were short-term fixes which relied on human intervention to manage the accumulation of water in the air cooler versus long-term engineering fixes to prevent the accumulation of fluid in the air coolers.”
Furthermore, the fire would have been manageable, said the Coast Guard, if the Hi-Fog system stepped in quickly and it could have even burned out on its own.
But prior to being put out, the fire ignited cables causing a secondary fire that ignited for hours after the initial issue.
“As such, the lack of manual activation of the Hi-Fog system by the engineering watchstanders, as well as the resetting of the fire alarm panel by the bridge watchstanders were critical mistakes which allowed the initial fire to burn without impediment for several minutes and propagate to the overhead cable runs.”
The Coast Guard pointed to the multitude of issues that made things less than ideal for the ship: ranging from firefighting problems to a non-working CO2 system, which all added up and caused more damage than necessary.
Recommendations included ditching the 40-second delay on the Hi-Fog system, improving the air cooler system, better training, and looking into "the susceptibility of the Carnival Splendor and all Dream class vessels to a complete loss of power due to the design and arrangement of electrical system and components."
Not one to sit on its hands, Carnival has launched its own investigation into the incident and is committed to continuous improvement, according to a released statement.
“We agree with the U.S. Coast Guard’s conclusions surrounding fire detection and firefighting processes and took numerous actions throughout our fleet as a result, including the creation of a Fire Safety Task Force,” said the Miami-based cruise operator. “Those efforts resulted in the implementation and enhancements of a number of processes, training and equipment, as well as the formation of a new Marine Safety department.”
“These actions directly contributed to the rapid detection and suppression of the fire on Carnival Triumph in February of 2013. Additionally, on April 17 of this year, the company announced a $300 million fleet-wide enhancement program that is currently underway to significantly enhance emergency power capabilities, further strengthen existing fire safety systems by installing the most advanced technology available, and improve the level of operating redundancies across our entire 24-ship fleet.”
In addition, and very important, there were no injuries on the Splendor or Triumph as a direct result of the engine room fire.
Earlier this year, Carnival appointed Richard J. O'Hanlon, a former rear admiral in the U.S. Navy, as vice president of nautical and safety operations for the line. He has an impressive marine background, including serving as commanding officer of a nuclear-powered aircraft carrier.
Part of the April announcement was the deployment of a company Safety & Reliability Review Board, created to “provide independent third-party perspective and to drive continuous improvement across the line’s fleet. The core objectives of the board are to review Carnival Cruise Lines’ current policies, practices and performance and to guide the company in incorporating best practices from other relevant fields and industries.”
Carnival named four new members to the Review Board on July 15, they are: Rear Admiral Mark H. Buzby, Rear Admiral Joseph F. Campbell, Ray Valeika and Dr. John K. Lauber.
Update: Carnival released this information on July 16, pertaining to specific steps the cruise line has taken in regards to the Coast Guard's report.
Photos: United States Coast Guard
Text: Cruise Industry News Editorial Staff
Original Report: Click Here