Fire in a Steam-Methane Reformer reactor of a refinery

The incident happened in a steam methane reformer unit of a refinery, constructed ten years before. A pigtail ruptured in the furnace of the unit and caused a major fire, which destroyed most of the tubes (54) in one section of the furnace.

No personnel were hurt.

Operators heard a noise in the furnace area and went to investigate. Looking in the firebox, they saw one tube swinging visibly with fire at its base. They returned to the control room and initiated shutdown procedures. About 15 minutes later, a much louder noise was heard, and black smoke was seen to be coming from the furnace stack.
When the fire was finally extinguished about two hours later, examination of the furnace showed that 54 tubes had failed. Almost all failures were at the top tube weld some 11 feet (3.35 m) below the firebox roof. Pieces of the broken tubes, catalyst, and damaged refractory littered the bottom of the narrow firebox.

The INITIATING CAUSE was the failure of aged pigtails at the root of the weld, due to thermal expansion and tube bowing.
Main technical causes were bad weld quality and shortening of pigtails over the life of the plant, which significantly increased the cyclic stress at the root of the weld from thermal expansion and tube bowing. After the failure of one pigtail, the downstream hydrogen flew back into the reformer and ignited together with the hydrocarbons present there. Together with the original fire from one ruptured tube, this additional massive quantity of flammable material contributed to the escalation, with the rupture of the other tubes.
A shortcoming was also found in the shut-down procedure. It was optimised to protect the catalysts during transient, but it resulted far too slow such a tube rupture.
The ROOT CAUSE could be attributed to shortcoming in design and manufacturing.

steam reformer, pigtails

DESCRIPTION OF THE SMR FURNACE
The furnace was the primary reformer in hydrogen production unit supplying approximately 97% purity hydrogen to the hydrocracking reactors of the refinery. The unit had a capacity of 75 MMSCFD (approximately 0.18 tons / day).
The reformer was built as a steam methane reformer and is still known by this name today (SMR), although it has since been modified to accept butane or naphtha feed. The maximum operating temperature of the tube walls was 1720°F (955OC). Although the source does not report it, typical pressure values were in the range of 15 to 30 bar.
The reformer was side-fired and had 256 catalyst tubes grouped in two cells.
The outlet “pigtails” were short sections of straight Incoloy 800H pipe connecting the catalyst tubes to the sub-headers.

McCoy et al. (see references) reported the following lessons learnt from the event:
1. Small design modifications executed to solve an immediate problem, can compromise the basic safety design of a whole process equipment. Sometimes, as in this case, several such changes have an additive effect.
2. Operating procedures must aim in first instance at protecting people and equipment,
and only secondarily at protecting the catalysts.

McCoy concluded on a pssimistinc note: "There are two general lessons to be learnt from this incident. These are lessons that have been learnt before and unfortunately, will probably have to be learnt again "

At many years from this event, a more generic historical view can be concluded. Pigtails are a weak components of a reformer, and have initiated repeatedly similar accidents. However, improvement based on their whole lifetime, from material selections and metallurgy (micro-structure) to cycling behaviour up to inspections, have reduced significantly their negative effects on the reformer operation, as reported by the second article of 2014 among the references.

The technical causes of the incident were corrected by lengthening the pigtails approximately two inches and rewelding them with better weld quality and profile, using higher creep strength Inconel 625 material.

Operating procedures were revised to allow a faster shutdown in emergency. Also, new purge facilities of far greater capacity than before were provided to sweep hydrocarbon from the feed desulphurisation section through the catalyst tubes before steam flow is stopped.