There were 2,135 passengers and 911 crew aboard at the time. All of the passengers were evacuated without injury. The explosion not only damaged the boiler room, but also damaged bulkheads, doors and doorframes three decks above.
Six of the eight killed were members of a six-person watch that managed the boilers. According to a spokesman for the Miami-Dade medical examiner’s office, the six men died of thermal injuries caused by exposure to superheated steam in excess of 900º F. The steam also killed the seventh, a second steward. It is unclear what he was doing in the vicinity of the boiler at the time of the blast. The eighth person to die was a cook who had been visiting in the engine room.The other injured crewmembers all suffered varying degrees of steam-related burns.
The explosion occurred in boiler No. 23, one of the ship’s four main propulsion boilers, about one hour after power was reduced in the system. At the time, three of the ship’s boilers were in normal in-port operating status supplying steam to electrical generators and auxiliary equipment while the ship was docked.
Boiler No. 21, located aft of boiler No. 23, was secured and not in operation when the explosion occurred. These types of boilers typically operate at 900 pounds per square inch and produce superheated steam with temperatures exceeding 950º F.
Investigators from the National Transportation Safety Board reported that the damage to boiler No. 23 from the explosion was extensive.
“The outer casing (of boiler No. 23) is ruptured, allowing the interior of the boiler to be visible on the starboard aft section,” the NTSB said. Investigators also found that the water wall header of the boiler had significant material failure.
Headers are devices that facilitate the distribution of water or steam in a boiler system. The water wall header introduces cold water into the system via a series of pipes attached to a reservoir, and steam headers distribute steam produced by the system.
Investigators have yet to determine which element failed first, the outer casing or the header.
Kevin Brandt, professor of engineering at Maritime College of the State University of New York at Fort Schuyler, said, “I’ve heard of tube and valve failures, but you hardly ever hear of a rupture and the actual boiler letting go. There are all sorts of safety devices built into these systems.”
Brandt suspects that the accident may have been the result of human error, but said that material failure and operational failure of a valve or other control were also possible.
“When problems or failures occur in these systems, it is usually when the boiler system is in a transition period, such as when ‘lighting off,’ when there is a change in temperature or pressure, never in a steady state like when the vessel is underway,” he explained.
Brandt said that if he were investigating the accident, there would be five questions that he’d need to answer:
• Were the flame scanners operable? If not, excess fuel could be delivered to the system, setting up an explosive situation.
• Was the low-water cutoff functioning? A low water level could cause the boiler to overheat.
• Was the main fuel solenoid shutoff functioning? Again, this would prevent excess fuel from being delivered to the burners.
• Were all safety valves operational? If not, pressure could build to dangerous levels.
• Were the fuel-oil combustion controls functioning? This system senses the steam pressure and assures that fuel delivery is appropriate.
The explosion came just 10 days after Norway passed a U.S. Coast Guard safety inspection in the Bahamas. Because of the Bahamian registry of the ship, inspection of the boiler system by the Coast Guard is not required. The boiler system of the 42-year-old ship was built in France by Chantiers de l’Atlantique and later rebuilt in Germany in 1999. The French company Bureau Veritas last inspected it in July 2002. According to Carol Carmody of the NTSB, there was a routine cleaning of the boilers on May 15, after 3,000 hours of service.
Harold Reemsnyder, a retired Bethlehem Steel engineer and a metal fatigue and fracture specialist, said that metal fatigue in the boiler or the header might have been a factor in the explosion. Boilers and components such as headers are subject to metal fatigue and corrosion even if they are cleaned and their water treated with anti-corrosion compounds. Poor maintenance and incorrect water chemistry can lead to premature corrosion of the water wall header or the boiler. Metal fatigue can also be a result of improper operation, such as heating up the boiler too fast during startup procedures commonly referred to as lighting off.
Corrosion and fatigue weaken the boiler walls, and only specialized testing during a survey can determine the degree of damage.
NTSB spokesman Keith Holloway said it was unclear whether or not such testing was done by Bureau Veritas during the July 2002 inspection. He said the NTSB was in the process of determining what went into the inspection and what role the header may have played in the accident. Both the NTSB and local police officials have ruled out terrorism as a cause.