The recent CSB report involving a catastrophic failure of a pressure vessel is causing quite a stir among engineering, safety, and process safety circles. I have received several e-mails and a couple of phone calls claiming the CSB incident is “very rare” and that the vast majority of pressure vessels are managed properly. Much like their criticism of me for stating things “you have no proof of”, I ask them to provide me the data to demonstrate their statement that the “vast number of pressure vessels are managed properly”. They always go back to the LACK OF ACCIDENTS as their proof. I don’t think I need to point out to the safety community the flaws in using the “lack of accidents” as a very poor indicator of safety performance. Fact of the matter is, no one really knows the “failure rate” of PV’s as no one knows the number of PVs in use on the globe or even in the USA. No one has a definitive number of how many are actually in service, how many failed, etc. But we can look back at previous incidents (excluding fired pressure vessels such as boilers) and see that pressure vessel management is not what most think it is! Just a quick look at CSB accidents and we find another incident involving a pressure vessel that was IMPROPERLY welded on, weakening the vessel. As with most accidents, there were multiple failures involved. In this 2003 accident CSB found that the facility had no documentation on this vessel or ones similar to it. No 1UA sheets, no name plates, no operation history, no inspections, etc. One highlight of this report is:
The CSB found that Marcus Oil had altered Tanks 5 – 8: each had a 24-inch diameter temporary opening cut into one end to install a steam pipe that heated the wax above the melting temperature. Following the pipe installation, the opening was welded closed.
…
The weld used to reclose the temporary opening on Tank 7 failed during the incident because the repair weld (Figure 5) did not meet generally accepted industry quality standards for pressure vessel fabrication. The original, flame-cut surface was not ground off the plate edges before rewelding the joint. The weld did not penetrate the full thickness of the vessel head. Furthermore, the welds contained excessive porosity (holes from gas bubbles in the weld). These defects significantly degraded the strength of the weld.
The CSB identified the following most likely failure scenario:
- Operators pressurized Tanks 6 and 7 with nitrogen gas containing 18 percent oxygen instead of the intended concentration of not more than 8 percent oxygen.
- The internal pressure of Tank 7 (67 psig) likely exceeded the strength of the defective weld on the patch plate. The weld completely failed, severing the plate from the tank.
- The tank rapidly depressurized through the 24-inch hole. Hydrocarbon vapor, compressed air, and hot liquid wax were ejected through the hole.
- Sparks were generated when the patch plate struck the concrete pad and likely ignited the wax and hydrocarbon vapors.
- The oxygen concentration in Tank 7 permitted the flame to flash back into the vessel. An internal deflagration blew the vessel head into multiple fragments (Figures 1, 6, and 7).
- The 50,000 lb vessel was propelled into surplus equipment stored nearby and came to rest against a warehouse on an adjacent property more than 150 feet away (see cover photograph).
- Burning polyethylene wax was blown against the warehouse and other equipment and ignited combustible materials. The resulting fire burned for nearly seven hours
Improper alteration of Tank 7 led to its failure: the closure weld used to reinstall the patch plate was fused less than 25% through the thickness of the plate and contained numerous flaws.
The effect of the inadequate closure weld was two-fold. First, it decreased the strength of this part of the tank by more than 75 percent.
Second, the lack of weld fusion and poor quality likely caused pressure-cycle-induced fatigue cracks to develop, further reducing
the weld strength.
The internal pressure in Tank 7 was most likely 67 psig, based on the nitrogen regulator test (Section 3.3). Operators stated that the regulator was normally set at about 45 psig but that they increased it when necessary to maintain wax flow to the processing unit. The higher pressure in Tank 7 at the time of the incident likely strained the poorly made weld to the failure point.
So for all those calling on me to quit “crying wolf” about pressure vessel safety I ask you this… how much evidence do we need before we pull our proverbial heads out of the sand and begin to acknowledge the level of risks associated with these vessels. How many vessels have to fail before we begin to realize the danger potential associated with having these vessels around workers? I am in no way trying to inflate the numbers or drive fears and I do NOT provide pressure vessel services so I am not trying to drum up business. I am merely trying to put credible facts and evidence in front of those who may not be aware of the mistakes that can be made regarding pressure vessel management needs. Failure to learn from others mistakes almost dooms us to suffer the same failures!
CLICK HERE to download the CSB report on the accident mentioned in this post.