sating photoelectric setting to reg-
If such a thing does happen, NO
ister minor extinction differences.
For the future, the possibility is
being studied of removing the NO
and, if possible, also the unsaturated
hydrocarbons and N,O from the gas
at low temperatures by means of an
adsorbent. Molecular sieves and acti-
vated carbon are being considered
for this purpose.
Behavior of NO resins
It is possible to make a comparison
of the behavior of NO resins in coke-
oven gas splitters and in units for
washing converted gas with N, since
we have operating experience with
both types. The conditions in the plant
are shown in Table 2.
It is apparent from this table that
the gas for the coke-oven gas splitters
contained more NO and much more
unsaturated hydrocarbons than the
converted gas. Nevertheless, there
were no particular difEculties under
these conditions in the coke-oven gas
units, whereas in the nitrogen wash
unit, a spontaneous explosion occurred
during normal operation.
Without going into details on the
many differences between the two
units, it can be stated that one differ-
ence must be considered fundamental,
namely that in the coke-oven gas
splitters any NO resins formed will
always be amply moistened by the
large excess of condensing hydrocar-
bons. In view of the uncommonly low
vapor pressure of the hydrocarbons in
the unit for converted gas, the resins
formed there will actually remain dry.
We now think it probable that this
difference is the cause of the greater
unstability of the resins in our nitro-
gen-wash unit.
For this reason we believe that for
gases with a low hydrocarbon content
still more stringent requirements
should be made as regards NO content
and accumulation than in the case of
coke-oven gas units. # 1960, approximately 70 ft. of a sched-
LITERATURE CITED
1. Wyier, Delwiche. Plant and Soil, V, NO.2.
Questions and answers
KERRY-American Air Liquide, N. Y.,
N. Y.: The problem of contaminants
in N2 scrubbing units dates back over
thirty years. Although considered only
recently in this country, a great deal
of published information is available
in Europe where intensive study has
been done on the reactions of NO or
NTO with dienes. The way they solve
them in coke oven gases is to actually
induce the formation or the chemical
reaction of NO or NTO with the
dienes in empty reactors before the ing.
coke oven gas plant itself (i.e., the obviously has to be removed, and I
exchanger system). They do that by think low temperature absorption may
permitting the feed gas to go through be practical.
an empty tower right after the com- In Europe most of the hydrocarbon
pressor. After a period of time the separation units use the front exchang-
empty reactor is washed down with ers to remove water, in other words,
caustic, and these gums are removed. as driers in terms of sequence or
The percentage of NO encountered switch changes. We prefer to use driers
in coke oven gas operation is extreme- ahead of them. While driers are cer-
ly low (0.1 is very high, while 0.05 tainly no cure-all insofar as absorption
is more acceptable). Obviously you of contaminants, we feel that they
can not afford to shut a plant down may absorb some of the dienes that
every several months just to wash it cause trouble. This is no solution if
down. With respect to treatment of the NO goes beyond 0.01%.
other hydrocarbon gases, engineers Acetylene is another problem that
want to know the quantities of NO, has been looked at for a good many
heavy hydrocarbons, and C,H,. In years. If the C,H, content is serious
most cases, the NO, very fortunately, enough, we always recommend that
is extremely low. As far as we are the feed gas be hydrogenated unless,
concerned, we can only afford to ig- of course, there is sufficient C,H, in
nore NO if it's magnitude is about the feed gas. In that case, we normal-
0.01%. With concentrations beyond ly put in a special exchanger which
that, steps have to be taken to treat is nothing more than an C,H, reflw
NO and to remove it. NO can come condenser. We use C,H, to wash out
into a feed gas which has been pre- C,H, since the two form an azeo-
pared by partial oxidation not neces- trope. However, you must have a
sarily through combustion, but through quantity of C,H,, in the order of 20-
external means such as water wash- 25 times the C,H, content.
J. H. CONKLIN
DuPont Co.
Nitrogen wash incident
Process safety modifications include replacement
of carbon steel pipe subjected to low temperatures,
better instrumentation and control systems.
O N THE AFTERNOON OF MARCH 12, After the cold warm-ends had re-
covered, feed gas was reintroduced
ule 40 carbon steel line containing and shortly thereafter trouble began.
H, at a pressure of 325 lb. was sud- The first symptoms of trouble were
denly subjected to liquid N, ternpera- cycling temperatures and gas composi-
tures causing the line to rupture with tions from the wash column. As had
explosive violence. Despite the fact
that fragments were thrown in all di-
been done many times in the past,
control instruments were placed in
rections with sufficient force to embed manual in an effort to break this
themselves into the lagging of adjacent vicious temperature and composition
equipment, no one was injured and cycle that was obviously becoming
little damage was done to nearby lines more extreme as time went on. Final-
or equipment. ly, the temperature alarm in the
Just prior to the line rupture, the carbon steel piping sounded, which
operation of the N, wash system had indicated that the temperature had
not been stable or satisfactory. In fallen to -20°C.
fact, only 4 hr. earlier it had been The operator pressed the toggle
necessary to back feed gas out of the switch in the 60 point potentiometer
boxes because of cold warm-ends on to read the temperature, and found it
the heat exchangers. to be -120°C and falling rapidly. He
CHEMICAL ENGINEERING PROGRESS, (Vol. 57, No. 4)
 ' LOW TEMPERATURE '
box and entering the carbon steel
lines.
An instant before the line ruptured,
a sudden pressure rise was noted in
the carbon steel line. This pressure
rise was caused by liquid N, vapori-
zing rapidly in the relatively warm
piping. Thus, the carbon steel was
simultaneously subjected to a rapidly
increasing pressure and a very heavy
thermal shock. It is not at all sur-
prising that the rupture was a violent
one.
It was the conclusion of all investi-
gating parties that the operator took
the proper action; but once liquid N,
had entered the carbon steel lines, all
control was out of his hands and
there was nothing he could do to pre-
vent the inexorable progression of
events.
Process modifications
The plant was shut down for a
number of weeks while the damage
was cleaned up and extensive modi-
fications made to the warm-end piping
as well as the instrumentation associ-
ated with the cold boxes.
All of the carbon steel piping that
Upper level operating platform after the line rupture at DuPont's nitrogen could be subjected to low tempera-
wash unit. tures, that is, the purified H,, the
recycle N,, and the waste gas lines
immediately took emergency correc- umn was at a level greater than 100%. leaving the boxes were replaced with
tive action by diverting the flow During the 2 to 2% hours in which type 304 stainless steel piping. The
through another parallel heat exchang- feed gas was not in the boxes, the carbon steel lines leading to the
er, but it was too late. The tempera- liquid N, reflux to the scrubbing col- compressors joined the stainless steel
ture continued to fall and in 15 to umn was not reduced since the level lines 6 ft. downstream of a stainless
20 sec, had reached -140°C at which did not appear to rise abnormally. steel automatic block valve. A stain-
time the line ruptured. The field Undoubtedly, the column was nearly less steel back-pressure control valve
operators responded well to the erner- full- of liquid when feed gas was rein- was installed some distance upstream
- of each block valve to maintain the
gency and had the boxes valved off troduced.
with purge N, flowing into the feed- The initial warm feed gas that desired pressure on the system and
gas lines before the head count was entered the base of the column caused vent the excess gas to a stainless steel
completed. heavy boiling and entrainment of stack. Each of the stainless steel auto-
liquid in the overheads line. The first matic block valves was tied to the
What happened and why? low temperature alarm that the opera- signal of the proper thermocouple in
The rather exhaustive investigation tor observed was caused by the cold the warm ends of the final exchangers
on the part of DuPont and Air Prod- gas preceding the liquid. By the in the boxes.
ucts revealed that the trouble prob- time he had selected the proper tog- The instrumentation was so ar-
ably started at the time the feed gas gle switch and read out the tempera- ranged that at -lO°C a panel alarm
was first withdrawn from the boxes. ture, liquid was already leaving the would sound to indicate not only that
At that time, the N, scrubbing column a low temperature was imminent at
( a sieve-plate column) dumped, and the exit of a heat exchanger, but also
a 100%liquid level existed in the base in which exchanger the trouble ex-
of the column. Shortly after the col- J. H. Conklin has isted. If the temperature continued
umn had dumped, the level fell to been with the Ex- to fall, a low temperature trip set at
60%and the operator assumed that the plosives D e p a r t - -20°C would automaticallv dose the
automatic level control valve had dis- ment of the DuPont proper block valve to prevent cold
Co. since receiving gases from passing from the exterior
posed of the excess liquid. a B.S. in Ch.E.
However, what actually happened from Drexel Insti- stainless steel piping into the carbon
when the level exceeded 100% was tute of Technology. steel piping leading to the compres-
that liquid N, backed up into the He is presently sors. Of course, when this automatic
Superintendent of block valve closed, the stainless steel
upper impulse line to the level con- the anhvdrous am-
troller. This impulse line was not back monia plant at the ~ e ~ a hWorks, o back-pressure control valve would
bubbled with purge N2. The liquid Gibbstown, N.J., which is the Explo- then vent the cold gases up the proper
in the impulse line gave the false level sives Department's largest installation. stainless steel stack to the atmosphere.
signal that indicated a somewhat Other job assignments ranged from
manufacture of explosives to work in A so called "panic" button also was
normal level when actually the col- the Research Department. installed in the control room so that

CHEMICAL ENGINEERING PROGRESS, (Vol. 57, No. 4)

April 1961 17
the operator would not have to rely
upon the automatic interlocks if he
saw the temperatures getting out of
hand.
In an effort to obtain earlier know-
ledge of impending temperature up-
sets on the warm ends, alarms were
installed on the thermocouples in the
lines leading to the final heat ex-
changers within the cold boxes. Also,
all of the points on the 60-point
potentiometer were duplicated with
new recording thermocouples. Thus,
the operator would have before him
a visual record of all critical AT
values within the cold equipment.
Likewise, the five new thermocouples
that were installed on the manifolds
of the five warm ends leaving the
boxes were fed to a high.speed re-
corder which monitored and printed
out each temperature every 5 sec.
The liquid level impulse lines on
the scrub column and the liquid N,
phase separator, as well as the column
pressure tap, were back bubbled with
purge N,. As an additional means of
detecting column flooding, a properly
back-bubbled recording differential
pressure instrument was installed
across the wash column.
Modifications satisfactory
All of these modifications have
proved very satisfactory during sub-
sequent operation. As a matter of
fact, this automatic diversion system
has literally saved our equipment a
number of times. The ability to take
low temperatures exterior of the boxes
and still keep iunning has enabled us
to uncover a number of shortcomings
and make the proper modifications to
the cold equipment.
I seriously doubt that by now, even
at the expense of more ruptured lines,
we would be operating the cold boxes
safely and in a controlled manner if
we did not have this ability to work
through the temperature upsets, de-
velop new operating techniques to pull
ourselves out of the difficulty, and
then keep going regardless of exterior
cold lines. #

uestions and answers

WALTON-Rohm & Haas Co., Pasa- your procedure for bringing these ex- which has a poor thermal conductiv-
dena, Texas: After the dump valve changers back in service after a nor- ity. We figure we can get quite a AT
to the flare opens, what means do you mal defrost. across the thickness of the material
use to stop the compressor? Is a com- SVOBODA, Dow Chemical Co., Plaq- and then contractual forces occurring
pressor shutdown device tied in with uemine, La.: Did you purge your dif- in the length of the line. Several com-
this system, or does the compressor ferential pressure or delta p indicator panies mentioned they have dumped
kick off on low suction pressure? on the N, scrubbing tower with N, liquid N, into a brass line and into a
CONKLIN: No, actually our compres- also? stainless steel tank. How often have
sors then go on automatic recycle ANONYMOUS: Yes, we did. I, for they done this and what was their
from, for instance, a 4th stage back one, am in Mr. Conklin's category of experience with thermal fatigue in
to 3rd, 2nd back to 1st. Should these those who have experienced ruptures such a line?
recycle valves be slow, or any other of carbon steel lines at low tempera- KRYZER-Standard Oil (Ind.), Whit-
thing happen, we do have low pres- tures on both the air separation plants ing, Ind.: The line that we use is gen-
sure suction trips on the machines. and the N, wash box. We have had erally cooled all the time. Because it
ANDRIES-Spencer Chemical Co., numerous occasions where the defrost- also carries the continuous purge from
Vicksburg, Miss.: I was wondering if ing lines or safety valve lines cracked, our auxiliary vaporizer, we do not
you have ever had any trouble with usually at welds, when they got ex- shock it with a large dump. It's chilled
channeling in your exchanger which tremely cold because of process upsets. continuously.
would result in the same condition? LAWRENCE, U. S. Industrial Chem- WALTON-SunOlin Chem. Co., Clay-
CONKLIN: No, we have had no indi- ical Co., Tuscola, Ill.: I wonder what mont, Del.: We have two blow-down
cation that we have had channeling in your basis was for the statement that systems, one is a schedule 10 alumi-
any of our heat exchangers. 6 degrees below was dangerous. It num 2-inch line which runs about a
ANONYMOUS: What type of level actually is, of course, about 14 de- thousand feet to the aluminum vapor-
indicating instrument or level sensing grees warmer than the code for car- izer. We have used this many times,
instrument was on this tower? bon steel, and most low pressure am- dropping liquid N, out there and rich
CONKLIN: The one that was on it monia systems operate lower than that. liquid bottoms from the air plant.
was a DP cell. ELLIS-Du Pont: Du Pont has never There have been no failures in this
HEPP-Sun Oil Co., Marcus Hook, agreed with this minus 20' because of line. It does contain expansion loops.
Pa.: We cracked one low pressure line some experiences that we've had on We also have a blow-down system for
and although the crack was minor the carbon steel equipment. We generally the hydrogen wash box made of a
manner in which we cracked it might divide the materials into carbon steel stainless steel type 304. It goes to a
be of some help for everyone. We be- down to 32°F. When we go from tank (about 2%-ft. in diameter by
lieve this was cracked when the opera- 32°F to O°F, we consider all the car- about 4-ft, high) made of stainless
tor was putting the unit back on bon steels with respect to their use steel, and is vented to a stainless steel
stream, after regenerating the ex- and the consequence of their use in stack. We have dumped cold liquid in
changer. Our procedure is as follows. this temperature range. Below 0°F we here many times. The coldest material
To put an exchanger back on, you put go into specified materials with im- has been the bottoms of the nitrogen
a crack of cold product gas through pact tests, special heat treatment, wash tower which is about -310°F.
without any exchange until it cools special fabrication considerations with We have had no sign of failure in this
down to a point where a thermal the equipment. system for over seven years.
bump is avoided in the process when ANONYMOUS: You mentioned that LAWRENCE-U. S. I.: On your specs
vou put it in service. The operator in you dumped the liquid N, into the on carbon steel, do you also disagree
this case cracked this valve and, we stainless steel line several times. We with the code on shapely tested steel
believe, allowed it to go too long. Of have always had a great concern for going down to minus 30 or minus 50
course, in time, the outlet line will go the thermal fatigue that might occur if it's A-300 or special type steels?
on down to the inlet temperature. because of the rapid application of ELLIS-Du Pont: We do carry the
Thus, you might wish to review the cold gases to the stainless steel A-300 material down to minus 50.

CHEMICAL ENGINEERING PROGRESS, (Vol. 57, No. 4)