This fatal Hotwork Incident is a reminder of the dangers of welding inside Confined Spaces, ESPECIALLY those that contained a flammable liquid or gas! The USCG did an excellent job with this investigation and report, laying out the timeline and facts leading up to this incident that caused this incident to occur. It is a MUST read for those who permit HW activities within confined spaces.
On 9 October 1993 at approximately 20:00, a 632 foot long U.S. tankship suffered a major explosion and fire in the No. 5 port cargo tank while anchored. The fire burned for approximately 5 hours. There were 3 fatalities and numerous other injuries. The damage loss resulting from this accident was approximately $15,000,000.00. The explosion peeled open the main deck in way of the No. 4 port and No. 5 port wing cargo tanks, laid it forward and above the main deck in way of the No. 3 port cargo tank. The explosion caused the bulkhead between No. 5 port cargo tank and the port saddle (aft ballast) tank to be blown aft into the port saddle tank. The bulkhead between the port saddle tank and engine room was breached by the force from the explosion, causing flooding of the engine room. The explosion also caused the side shell in way of the Nos. 5 port and 4 port cargo tanks to be blown away. Following the explosion the vessel was engulfed in a major conflagration in way of the No. 5 port and No. 4 port cargo tanks and also in way of the port cargo manifold. The vessel took on a list to port due to flooding and settled on the bottom of the 32 foot deep anchorage, leaving her decks above the waterline and her port stem quarter awash. Fearing subsequent explosions from the fire on deck, the master ordered his crew, except five officers and a deck cadet, to abandon ship. The pumpman was found on deck, fatally injured, and was taken off the ship with the crew. The crew was able to step off the fantail to an waiting crewboat which had come to their rescue. The master and six officers searched unsuccessfully for the missing welder and his fire watchman before they themselves abandoned the vessel and boarded the pilot boat GALTEX, which came to their aid. The Coast Guard and several other vessels searched unsuccessfully for the missing welder and his fire watchman, whose bodies were later located by divers in the port saddle tank. The vessel burned on deck at the port cargo manifold, and in the Nos. 4 port, 5 port, 5 center and 5 starboard cargo tanks. The vessel suffered a subsequent explosion at approximately 2110 in way of No. 5 starboard cargo tank. The fire was extinguished at about 0120 on 10 October 1993. The vessel was declared a total constructive loss.
The hot work was performed to repair three separate leaks. Two of the leaks were between the SC cargo tank and the pumproom, the third between the SP cargo tank and the port saddle tank. The leaks between SC and the pumproom were repaired by hot work performed in the SC. The third leak, SP, was to be repaired by hot work in the port saddle tank. For comparison, the three leaks could be considered to be in only two locations: SC and the port saddle tank.
The company policy requires spaces where hot work is to be performed be adequately cleaned, gas freed and inspected, prior to the commencement of hot work. Spaces adjacent to hot work are also required to be properly prepared prior to start of hot work.
The master followed the company’s policy as written in the Fleet Standing Orders, Section S.6.1, on cleaning and gas freeing of adjacent tanks prior to hot work which allowed for completely filling the adjacent tank with ballast water. NFPA 306, paragraph 2-3.7(b) also allows for flooding with water. However, the master deviated from company policy when he allowed the ballast water to be pumped down without properly ensuring the gas free condition in the tank using the guidance provided in section 5.4.2, which requires sampling at several openings and various levels to ensure full representation of the tank atmosphere is obtained. Furthermore, NFPA 306, paragraph 2-3.?(b) requires that the hot work be performed at least 3 ft below the level of the water and that the gas content of the atmosphere above the water does not exceed 10 percent of the lower explosive limit.
CONCLUSIONS
- The apparent cause of this casualty was the ignition of explosive gasoline vapors in the No. 5 port cargo tank. The source of ignition was a welder’s arc in the adjacent port saddle ballast tank which undercut into the common transverse bulkhead.
- Contributing to the cause was the failure to properly gas free the No. 5 port cargo tank.
- Contributing to the cause was the failure of shipboard personnel to detect the presence of explosive vapors in the No. 5 port cargo tank.
- Contributing to the cause was complacency and failure to assess risk on the part of the master and chief mate, in that they improperly followed established procedures for gas freeing and tank testing.
- Contributing to the cause was the inappropriate use of a “landing plate” or doubler to repair the port side shell longitudinal, which originally passed through the forward bulkhead of the port saddle tank. The use of doublers or “landing plates” on cargo tank bulkheads is not appropriate nor consistent with good marine practice.(NVIC 7-68, 130)
- Also contributing to the cause was the decision of the master, chief mate, and owner’s representative to stop a leak of questionable origin, by fillet welding around the improperly installed doubler plate.
- It is a conclusion of this investigation that the oxygen detector was properly calibrated and operating satisfactorily when used to test No. 5 port cargo tank and the port saddle tank on the afternoon of the day of the explosion.
- It is a conclusion of this investigation that an explosive atmosphere existed in the No. 5 port cargo tank, resulting from gasoline cargo residue and clingage.
- It is also a conclusion that a similar atmosphere existed in the No. 5 starboard cargo tank.
- This casualty has shown that “overflowing” did not adequately displace the cargo residues from the No. 5 port and starboard cargo tanks and is not good marine practice. Additionally, deballasting without inerting the No. 5 cargo tanks prior to the welding negated any benefit that may have resulted from “overflowing” those tanks with water.
- The use of a certified marine chemist, required by regulation because of the vessel’s location, would have been consistent with good marine practice. A certified marine chemist may have detected the explosive atmosphere in the No. 5 port cargo tank.
- Use of inert gas in the vapor space or forced ventilation of the No. 5 port cargo tank may have prevented the explosion by replacing the explosive atmosphere.
- The guidance provided in the OMI Standing Orders for gas freeing, was misinterpreted by the master and chief mate to allow 11overflowing.”
- It is a conclusion of this investigation that certain areas of the bulkhead at frame 49 were significantly wasted, particularly in way of and near the location of the doubler. Although averaged readings appear to be within published guidelines, individual readings are significantly outside those parameters.
- All deaths and injuries aboard the OMI CHARGER on the date of the incident were a result of the explosion.
CLICK HERE for the official USCG Report (pdf)
