328 UNDERSTANDING PROCESS EQUIPMENT FOR OPERATORS AND ENGINEERS

transmitter was off-zero. Most flow transmitters are off-zero. For example, let’s say that the
indicated flow is 8 B/H. The instrument is checked and found to be off-zero. The zeroed
flow is 4 B/H. Flow meters read as a function of the square root of the measured pressure
drop through the flow orifice plate. Therefore, the corrected flow (i.e., compensating for
the meter reading off-zero) is (see Eq. 41.3) as follows:

ð8Þ2  ð4Þ2 ¼ 64  16 ¼ 48 (41.4)

ð48Þ½ ¼ 7 B=H

The meter was reading 1 B/H too high. But if the control valve is shut, and the real flow
falls to zero, the operators will see a flow of 4 B/H.
Apparently, the operators in Texas City mistook a meter being off-zero for a large flow.
As they were bringing in feed continuously, and the indicated level on their panel was not
increasing (but the real level was), they blocked in the splitter bottom’s level control valve.

Loop Seal
The Blowdown Stack Loop-Seal is shown in Fig. 41.3. The liquid from the Blowdown Stack
flowed into a slightly pressurized (i.e., about 3 psig) hydrocarbon liquid collection system,
rather than into an open sewer. To prevent hydrocarbon vapors from continuously back-
ing out of this collection system, into the atmospheric blowdown stack, there was an inter-
vening loop seal. Unfortunately, as the liquid in the blow-down stack flowed upwards in
the vertical portion of this loop, the liquid lost head pressure, due to its increased eleva-
tion. As this liquid was naphtha at its saturated liquid bubble point, it partly vaporized.
The expanding vapor choked off the liquid flow. The result—“VAPOR LOCK.”
Vapor lock stopped the flow of naphtha draining from the blow-down stack. The liquid
level in the stack backed up over the inlet nozzle. The vapor flowing through the inlet noz-
zle blew the naphtha up the stack. The naphtha ignited.
Likely, a high point vent on the top of the loop seal—which is a common feature of loop
seals used in water plumbing systems—could have prevented the vapor lock.

The Ultimate Question

We can hope that BP and Marathon (the current operator of the refinery) would have
learned a lesson from this incident. But that is not the question. Would I have made
the same mistake if I had been present? Hopefully, I would have noted that the input flow
exceeded the output flow. Then, calculated that the level in the splitter must be increasing.
Had the unit’s Technical Service Engineer been involved in start-up, he or she should have
interceded with operating personnel based on the material balance.
But there is another, more fundamental action that I would certainly have taken to pre-
vent this explosion. In 1984 I was involved in an Alkylation Unit Depropanizer start-up at
 Chapter 41 • Plant Safety Fundamentals 329

the Good Hope Refinery. No one was sure of the liquid level. The level gauge glasses were
largely useless as the process fluid was water white (not yellow), and the glasses old and
scratched.
I placed a pressure gauge (Fig. 41.3) at the top vent of the level glass (P-2) and then
checked the head pressure (P-1). This is the pressure difference between the top vapor line
and top of the gauge glass. To convert this DP to feet:
Feet ¼ ðDPÞ ð2:3Þ  Specific Gravity of Butane (41.5)

At Good Hope, the measurement indicated we had 40 ft of liquid above the top of the
gauge glass. We were “Tapped-Out.” The liquid level was pumped down until the indicated
level on the panel declined from a dead-steady 80% to an erratic 70%.
In 1974 I demonstrated this method of finding levels on start-up to my operators in
Texas City. But by 2006, this simple step had been submerged in the flood of advancing
technology, to the detriment of refinery safety.
 42
Dangers of H2S, Steam, Nitrogen,
and Hot Water

Lieutenant Joe Patrocelli was the Engine Room officer in an Escort Air Craft Carrier in the
bloody World War II battle of Leyte Gulf in the Philippines. We worked together at the
Amoco Oil Refinery in Texas City on the Amine Regeneration Plant and the Sulfur
Recovery Unit.
On graveyard shift, we would exchange war stories. Joe recited naval battles against the
Japanese. I would tell tales of death and disaster that I had been associated with in process
units. “Norm, being below deck in a carrier under attack by kamikazes is hard. At least guys
topside can fight back. You know, we lost a carrier in Leyte.”
“Yeah, Joe. I read about that.”
“Of course, working on a sulfur and amine plant with this bunch ain’t all that safe either.
What with the H2S and SO2 and the pyrophoric iron. Just last week, Kenny overfilled a sul-
fur rail tank car. Could’ve killed someone.”
“Joe, refineries are really dangerous places unless you’re careful. I’ll tell you a few
stories involving fatalities I have been involved in.”
“Well, Lieberman, we got us nine hours before the day shift shows up. You talk and I’ll
listen.”

Coastal—Aruba
A maintenance supervisor (also a retired Marine) was directing a crew of pipefitters
changing a spool piece on a rich amine line on an elevated platform. It was a “Fresh
Air” job. For some reason, he removed his Scott Air Pak. The rich amine flashed. He died
of H2S inhalation.
• Remedial Action—Dealings with equipment in rich amine service is identical to H2S
service, which always requires use of “fresh air.”

American Oil—Texas City

On an Alky Unit, two pipefitters were blinding off a depropanizer reboiler. The tower was
floating on plant flare pressure. The sulfuric acid production plant (which used H2S as
feed) had an emergency shutdown, and their H2S was diverted to the flare. The H2S backed
into the Alky Depropanizer and killed both pipefitters.

Understanding Process Equipment for Operators and Engineers. https://doi.org/10.1016/B978-0-12-816161-6.00042-4 331

© 2019 Elsevier Inc. All rights reserved.
332 UNDERSTANDING PROCESS EQUIPMENT FOR OPERATORS AND ENGINEERS

• Remedial Action—Blind-off process equipment from the plant flare during a

turnaround.

Refinery “X”—Baytown, Texas

A female operator was sent to restart the water pump on the acid gas feed K.O. drum. The
drain valve on the pump suction was an ordinary ¾ inch gate valve. It should have been a
“Dead-Man’s” valve that is closed by a strong spring when not being held open by an oper-
ator. Her coworkers found her dead an hour later next to the pump.
• Remedial Action—The “dead man’s” valve is a spring-loaded valve that must be
manually held open. When released, it closes instantly and tightly by itself.

GHR Energy—Norco, Louisiana

A contractor walked into a 6 ft deep pit of hot water from a steam leak on the sulfur recov-
ery complex. He was scalded to death. I had weekend duty as plant manager during this
incident. The pit had been excavated to fix the steam leak.
• Remedial Action—Unit Supervisors must keep control of construction projects on the
process units they supervise.

American Oil—Texas City

Two maintenance men were killed when they entered a hydrocracker reactor. The reactor
had been purged with nitrogen. Entering a nitrogen atmosphere does not cause any short-
ness of breath, or any other discomfort. Only death.
• Remedial Action—Add 1% CO2 to N2. This will immediately cause shortness of breath.
Also, strictly enforce entry permit policy and always provide a “Hole-Watch.”

Coastal—Corpus Christi
A contract laborer was killed when a steam turbine, driving a hot asphalt pump (700°F)
overspeed and a small connection broke off of the pump. The asphalt sprayed onto the
laborer’s clothes and auto-ignited. The auto-ignition temperature of asphalt, or vacuum
tower bottoms, being 320°F. The turbine’s overspeed trip was “wired up” (i.e., disabled with
a piece of wire) because it kept tripping off, as a consequence of the governor speed con-
troller malfunctioning. I had noticed before the incident that the trip was disabled, but
neglected to report this to plant management.
• Remedial Action—Do not run steam turbines unless both the governor speed controller
and overspeed trip are both operational. The trip is not a back-up for the governor.
 Chapter 42 • Dangers of H2S, Steam, Nitrogen, and Hot Water 333

Without the governor functioning, the operators will sooner or later disable the
overspeed trip.

Refinery “X”—Beaumont
An operator was killed by boiling water from an eruption from the top of an open coke drum.
My investigation indicated that purge steam was accidentally opened into the bottom of the
coke drum. My client thought that a localized hot spot in the coke bed caused the eruption.
• Remedial Action—Modern coke drum unheading devices should protect against this
hazard. Existing units should be retrofitted with a “Delta Valve” on the top head.

American Oil—Texas City

An inspector was crawling across the roof of an alkylation unit’s spent sulfuric acid tank.
The roof had been reported to be corroded. It was! All they recovered was his teeth.
• Remedial Action—The inspector was trying to impress his coworker with his courage.
There are some accidents caused by people’s belief in their indestructability.

GHR Energy—Norco, Louisiana

The floor or a vertical fired heater was blown down onto two operators who were attempt-
ing to light the main burners. The pilot lights were plugged, and they were using an oil-
soaked rag as an ignition source. Both were my friends.
• Remedial Action—These pilot lights were using fuel gas rather than natural gas. They
were plugged with sulfur deposits. Use Natural Gas for pilots, not refinery fuel gas.

Amoco—Whiting, Indiana
Three operators were killed when an Asphalt Oxidizer Tank blew up. Both the steam and
nitrogen purge in the off-gas line to the oxidizer incinerator had been deliberately turned
off in an attempt to relight the oxidizer pilot light.
• Remedial Action—The process is inherently dangerous and unnecessary. Asphalt spec’s
can be met by better operation of a Vacuum Tower, without resorting to this archaic
oxidizer process.

Unocal—Chicago
A coke drum bottom head blew out onto the coke cutting ramp during coke drilling, and
killed the operator of the front-end loader removing the coke from the ramp. The bottom