Catalyst Catastrophes II

This Catalyst Catastrophes paper is the sequel to the one presented at the 48th AIChE Meeting, Orlando ,Florida
in 2003. It covers some of the more frequent, but still costly incidents relating to catalysts with the objective of
pulling together learning to prevent reoccurrence. It suggests some pre-thinking that plants might consider as
part of preparing for emergency response and reflects on the value of staff training in the prevention and
handling of incidents especially at start-ups and shutdowns which are becoming less frequent events as the
reliability of all systems improves.

John Brightling and Dr Mike Roberts

Johnson Matthey Catalysts

Introduction of the dangers and lessons we at Johnson Matthey Cata-

lysts have learned over the last half a century with the
As was shown in the first paper [1] for every cata- objective that we avoid relearning them the hard way.
lytic step in the ammonia process there have been ex-
amples of significant incidents which have either re- Catalysts are wonderful materials that enable us to
sulted in risk to personnel, major plant equipment perform the chemistry of our processes at ever improv-
damage or irrecoverable catalyst damage resulting in a ing energy efficiencies while also reducing the envi-
plant outage. ronmental impact. For the most part they do their work
in a quiet sustained way and perform their tasks as we
The cumulative costs of preventable incidents in would like. This, however, can beguile us into a false
the industry is not known, but we can assume that it sense of security, and we forget about the other reac-
amounts to many US$10’s billions, which underlines tions that they can perform if we give them the chance.
the benefits from us all sharing our experiences as these
Safety Symposiums have aimed to do over the last 50 Also, our plants, equipment, and catalysts deliver
years. incredible reliability today. Unlike in the past, we are
now seeing much longer periods of time between tran-
Catalysts normally quietly perform the reactions we sient operations such as start-ups, shut downs and the
want. However, if we give them the opportunity they occurrence of other serious problems. This gives us all
will also perform other reactions, many of which can the problem of how to ensure we retain corporate
generate large amounts of heat, and others produce knowledge that has been built up over the many years
toxic materials or other dangers to life, equipment or
the environment. The target here is to highlight some

AMMONIA TECHNICAL MANUAL 136 2005

we have operated plants. We already have two full not removing heat. Also the flow through the tubes is
generations of experience we need to keep alive. low, so the heat transfer coefficient and hence rate are
much lower than during normal operation. This also
Our attempt here is to capture some of the painful means that the temperature readings (e.g., tube exit and
experience we at Johnson Matthey Catalysts have headers) that can be relied upon during normal opera-
gained either as owner-operators of ammonia plants or tion will not provide a representative reading and there-
through our extensive interactions with our ammonia fore can not be relied upon.
catalyst customers. All of the incidents described here
are based on real events. We hope that these reminders The previous lessons are not repeated fully here.
will help us avoid relearning them the hard way. Most The key points at start-up being (a) there is no substi-
of the incidents are from recent experience in the indus- tute for visually monitoring the tube appearance to
try and for reference where previous AIChE papers ex- check tubes are not being overheated and (b) proce-
ist detailing similar incidents, these reference cited dures to ensure that the correct sequence for staged
lighting of burners must be adhered to at each stage of
We have structured our paper by reactor duty and start-up to keep the pattern/heat generation even, Figure
then included some generic considerations including 1.
those related to loading.

The ‘lessons’ listed in each area are intended to be

prompts for consideration of what might be appropriate
and shall not to be taken as advice. Each plant will
have individual circumstances and these should be re-
viewed in the light of these ‘lessons’.

Primary Reforming
Figure 1 : Typical burner lighting order at start-up
Tube Failure on start-up revisited.
Lessons:
In 2003 [1] we said that annually somewhere in the
world a reformer loses a significant number….if not all • Light burners in a staged start-up pattern
their tubes. The trend continues and we are aware of a • Need to see/monitor what is happening
number of incidents which have occurred since 2003 • Do not deviate from standard operating
involving either total or partial reformer burndowns. In procedures with out careful consideration
the last twelve months, Johnson Matthey is aware of at of the implications
least 7 such incidents including one virtually complete
reformer burndown. Carbon Formation

As is invariably so, in most cases the damage was It is well known carbon formation can lead to catas-
caused during a reformer transient, either at start-up or trophic catalyst failure. There are a number of causes
most commonly following a short trip [2]. With the of carbon formation including catalyst poisoning or op-
importance of maintaining high on-stream factors, there eration at too low a steam to carbon ratio [3].
is always pressure to bring the plant back on line as
soon as possible; in these circumstances, some opera- Depending on the cause, different carbon species
tors deviate from the normal start-up procedures. can be formed on the catalyst surface. Two of the most
common species are pyrolytic and whisker carbon.
The heating duty for a primary reformer on modern
ammonia plants is approximately in the range 100-150 Pyrolytic carbon is typically formed at high tem-
MW (340-510 mmBTU/hr) this heat input is supplied peratures (>600ºC/>1,112°F) from thermal cracking of
by burners each liberating 1-2 MW (3.4-6.8 hydrocarbons leading to the deposition of carbon pre-
mmBTU/hr). At start up the endothermic reforming re- cursors on the catalyst particle, as Figure 2 illustrates.
action which normally absorbs 50-60% of the heat is

2005 137 AMMONIA TECHNICAL MANUAL

Such carbon formation results in catalyst deactivation Case 1
and increasing pressure drop.
In a naphtha feed plant during the introduction of
nitrogen during start-up, the reformer pressure drop in-
creased from 1.4 to 7 bar (20-101 psi) in approximately
Pellet surface 2 minutes.

Investigations showed that the nitrogen line had

been filled with naphtha due to inadequate isolation
blinding (slip plating). The feed line was 100mm (4”)
Carbon diameter and around 1 km (3,300’) long so capable of
holding 10 metric ton (11 st) of naphtha.

The plant operating procedures stated that a drain

Figure 2: Pyrolytic Carbon on catalyst pellet valve shall always opened on the nitrogen line but on
this occasion this valve was not opened.
At even lower temperatures (>450ºC/842°F) it is
well known that reactions with hydrocarbons or carbon Case 2
monoxide over transition metals can lead to the forma-
tion of carbon filaments. The whiskers have a tubular A less extreme situation, which can cause a slug of
appearance and most have a Ni-particle at one end, as trapped feed to be deposited on the catalyst is shown in
highlighted in Figure 3. As the Nickel –crystallites are Figure 4. This is a problem that particularly affects
not obscured or covered during this carbon forming plants that utilise a liquid feed or where a large propor-
process, in a steam free atmosphere the rate of carbon tion of the feed can condense out under shut down con-
formation can be enormous. ditions. Here it can be seen that the liquid feed is
trapped between the trip valve and feed flow control
valve.
Hollow
carbon fibre

Nickel
0.0001mm (1/250 thou) crystallite

Figure 3: Whisker Carbon on Ni-reforming

catalyst

The whiskers are very strong and as they are Figure 4 : Example of trapped feed during trip
formed within the catalyst pores can cause significant
damage to the pellet structure; this is especially the case Case 3
when an attempt to remove the carbon with steam is
made as gasification can cause additional stresses deep A plant was the sole user of a gas supplied by a
within the catalyst pellet. Let us review some case stud- pipe line that passed underneath a large river
ies where carbon formation has caused major reformer
catastrophes. After an extended shutdown, at the point in the
start up, when the burners were being lit, an operator
saw a burning liquid flowing from a number of the
burners onto the tubes. While the operator was in the
process of leaving the vicinity of the reformer in order

AMMONIA TECHNICAL MANUAL 138 2005

to radio the control room to shutoff the fuel, a tube After around 20 start-ups from cold conditions,
burst leading to significant damage to the furnace re- there was significant damage to the catalyst and this led
fractory and adjacent tubes. to poor gas distribution within the reformer and hot
The root cause of this problem was that during the spots developed which forced the operator into a cata-
plant shut down a significant amount of hydrocarbons lyst change out.
had condensed within the cool section of pipe under the
river. Figure 5 shows the agglomerated nature of the cata-
lyst which had fused together and had to be mechani-
Lessons: cally broken up before it could be vacuumed; these are
both symptoms of a wetted catalyst that has been dried
• Inadequate purging during trip can lead to too quickly.
feed being stored in the HDS system or feed
pipework – on plant restart N2 purge pushes
this hydrocarbon into the reformer.
• Heavier feeds (including natural gases with a
large concentration of higher hydrocarbon,
LPG’s or Naphtha’s) have a high risk of feed
condensing and sitting in dead legs.
• Gas flowmeters largely ignore the presence of
condensed higher hydrocarbons and during
start-up flowmeters will read in error.
• Flowmeters that are not compensated for tem-
perature and pressure will read in error.
• Operating procedures shall always be fol-
lowed. Figure 5: Fused catalyst from primary reformer
subjected to wetting
Condensation on Catalyst
Lessons:
Earlier papers [1,3] explained how wetting the cata- • Rapid drying generates high steam pres-
lyst is unusual but does not always have catastrophic sures within pellets than can even break
effects. However, the actions that are taken once you reforming catalysts.
have a wet catalyst can have serious consequences. • Careful drying will minimise the damage
to the catalyst.
In a recently commissioned hydrogen plant, insuffi-
cient nitrogen purge was available due to compressor Secondary Reforming
limitations for the initial phase of start-up. Utilising ni-
trogen, the inlet pipe work to the reformer could only Perhaps more than any other reactor within the
be heated to about 80ºC (176°F). The remainder of the ammonia plant, there have been a number of incidents
warm-up was carried out with 420ºC (788°F) steam in- related to mal-operation of the secondary reformer.
troduced to a system pressure with a dew point of
115ºC (239°F). Therefore during each start-up, conden- Virtually all are related to air addition in respect of
sation occurred in the ‘cold’ and rather long pipe work either inadequate isolation, control of the air flowrate or
between the feed and steam mixing point and reformer the performance of the burner and the impact on the
tube inlet. process that occur within the mixing zone.

Furthermore with steam as the heating medium, the Inadequate isolation of Air
heat-up rates at the reformer inlet were between 170-
200ºC/hr (306-360°F/hr) which far exceeds the normal Adding air into the process in the secondary re-
heat-up rate for dry catalyst 50ºC/hr (90°F/hr) let alone former without combustion is a major hazard as there is
that of wetted material. potential to form an explosive air/fuel mixture in the
plant or generate extremely high oxidation exotherms

2005 139 AMMONIA TECHNICAL MANUAL

on the HTS/LTS catalysts that can easily cause vessel tion of the secondary reformer showed an approxi-
damage. mately 500mm diameter x 2000mm ( 20” by 80”) deep
hole had been burned into the bed, Figure 7.
Figure 6 shows an example of a fused HTS catalyst
resulting from and incident in which the process air Catalyst samples removed from the edge of the hole
compressor was running to cool coils in the convection showed that the temperature had been high enough to
section, but the vent to atmosphere was not fully open fully fuse and melt the ceramic catalyst support, indi-
and the process air valves to the secondary reformer cating that the temperature had exceeded 2,100ºC
were passing. (3,812°F). The nickel level in the fused catalyst had
been reduced from around 10% to less than 1% causing
the high alumina support balls to turn blue due to the
formation of nickel spinel.

Figure 6 : Fused HTS catalyst

Other examples of this issue can be found where

both the HTS catalyst and vessel were severely dam-
aged [4,5] and in another case [6] where the tempera-
ture generated by the HTS bed exotherm caused the
downstream process line to reach approximately 760ºC Support balls turned blue!
(1,400°F) before it failed. Figure 7: Schematic of secondary bed damage
showing fully fused ceramic catalyst
Inadequate control of air
Although this is a serious incident in terms of plant
In the secondary reformer the control of the air to availability, the flame remained central and did not im-
process feed rate is critical to achieve the desired hy- pinge onto the refractory.
drogen to nitrogen ratio to suit the type of process; The extensive damage to the catalyst bed and inert
typically this ratio is between 2.5 - 3:1. This ratio material indicated that if the flame has deviated from
means that only 80% of the available ‘fuel’ in the proc- the vertical, it would not have been stopped by the re-
ess stream is combusted and the maximum temperature fractory lining.
achieved in the secondary reformer is moderated by the
mixing of the un-combusted primary effluent at around There have been a number of cases [7, 8, 9 and10]
700-820ºC (1,292-1,508°F). of refractory failure followed by vessel rupture due to
flame impingement, resulting from poor design or mis-
The air to feed ratio is always a critical trip in the alignment of the burner. Less severe examples of these
front end, however, instances have occurred whereby problems can lead to localised hot spots on the shell.
this ratio was not maintained. In one incident 1000mtpd Refractory can then vaporize in areas of extreme tem-
(1,100 stpd) plant was operating at full rates but needed peratures and subsequently condenses on the cooler
to reduce rate by 30%, this occurred but the feed rate secondary catalyst, coating it with alumina and in some
was reduced before the air rate. As a result the ratio of cases will forming pink rubies [11] as transition metals
feed (‘fuel’) to air was nearer to the stoichiometric com- are incorporated into an alumina spinel.
bustion ratio so virtually all of the feed (fuel) gas was
combusted. After this incident a shutdown inspection

AMMONIA TECHNICAL MANUAL 140 2005

 Over the years improvements in the understanding dry the whole bed and in so doing, the catalyst at the
of secondary reforming[12] combined with increased top and middle of the bed became dehydrated.
catalyst activity have allowed a reduction in installed
catalyst volumes and therefore a reduction in the bed After nitrogen drying the operator switched to
height. This increases the combustion volume and 100% steam against advice of only using 5%. A re-
therefore reduces the potential for the hot combustion hydration exotherm was observed and then due to an
gases impacting on the catalyst and thereby reduces the unrelated plant trip (site power failure), the catalyst was
problems of alumina vaporisation and catalyst fusion held under these conditions (100% steam) for a period
of time.
The issues associated with burners have been the
subject of many papers over the years of these sympo- The exotherm reached 530ºC (986°F) and took sev-
sia with many of the earlier generation burners being eral hours to cool down with N2. Thankfully the final
prone to failure leading to flame impingement onto the activity when the HTS catalyst was brought on-line was
catalyst bed [13-16]. Burner designs have improved acceptable and the pressure-drop was low indicating
over the years, especially following the introduction of that little damage had been done to the catalyst.
computational fluid dynamics (CFD) and other design
tools so that the true performance of all the components Upstream problems causing HTS damage
of the secondary reformer can be modelled in detail
[17]. Upstream waste heat boiler leaks are certainly one
of the most common damage mechanisms for HTS cata-
lysts and the best catalysts are designed to withstand
Lessons: additional stresses imposed on the catalyst as a result.
• Check that air isolation systems are in good work- Waste heat boiler issues are familiar to most in the in-
ing order and have high integrity dustry the selection of prior references [19-25] indicates
• Monitor the approach to equilibrium and pressure how widespread the problems of maintaining equip-
drop and look for evidence of burner problems if ment on this service are.
these parameters consistently rise above normal
expected values Recent research completed by Johnson Matthey has
• Monitor the secondary shell temperature via heat shown it is both the drying process and pore structure
indicative paint or the rate of water boil off. of the catalyst that dictates how well a catalyst survives
• During turnarounds inspect the refractory and a wetting incident, which is an all too frequent occur-
burner for damage rence as a result of boiler leaks.
• If concerned, use CFD to model the burner per-
formance
In addition to stress placed on catalyst due to the
wetting and subsequent dry out, additional hydraulic
High Temperature Shift pressure build-up and damage to the catalyst frequently
may occur due to accumulated solids either from trace
Oxidation solids evaporated from the BFW or refractory materials
dislodged from within the WHB and associated lined
This has been covered in the secondary reformers transfer main.
section.
An example of this is shown in the following case
Start-up Exotherms (Dehydration) study from a large ammonia plant, Figure 8 shows the
normalized pressure drop
After years of debate this phenomena was re-
searched and explained by Johnson Matthey [18]. In a
recent case study a small hydrogen plant left a fresh
charge of HTS catalyst in a reactor open to the atmos-
phere during the winter months so the catalyst absorbed
water. Start-up required nitrogen heating for 2 days to

2005 141 AMMONIA TECHNICAL MANUAL

 masking the active sites. The use of an inert high sur-
4
face area support such as SHIFTSHIELD™ from
HTS Reactor
Johnson Matthey can help reduce the impact of con-
Skimmed tamination
3
Lessons:
PD Bar

2 • Beware of hydration effects on dry cata-

lysts.
1
• If you have a severe boiler leak, carefully
and thoroughly dry the catalyst bed before
bringing the bed back online.
0
• Regular monitoring of shift converter per-
0 500 1000 1500 2000
formance can help identify problems such
Figure 8 : Dayspressure
HTS on-line drop trend as boiler leaks early. Inert high surface
area support can reduce impact of con-
tamination and are a cost effective way of
Due to a load limit dictated by the catalyst prolonging the HTS catalyst life.
support outlet grid, once the pressure drop exceeded
3.5 bar (51 psi) the plant rate had to be reduced. This
restriction justified a short outage to skim the reactor
removing around 400mm (16”) of material from the
top of the catalyst bed..

Samples removed during the bed skim were sub-

jected to a semi-quantative XRF analysis for major
components which revealed a deposit gradient as fol-
low.

Component Top of Bottom

% skim of skim
Fe2O3 31.1 87.2
CuO 0.9 3.4
Al2O3 39.2 0.9
SiO2 10.5 0.3
NiO 4.8 0.1
Na2O 3.3 0.1

Table 1 : HTS skim analysis

The analysis shows the degree to which the top of
the bed had become fouled with foreign material;
mainly refractory materials, catalyst displaced from the
secondary reformer and the WHB immediately up-
stream. The presence of 10% silica (SiO2) indicated
that extensive damage to refractory had occurred expos-
ing a lot of the ‘cold face’ insulating materials which
are normally protected by a ‘hot face’ of high alumina
material

At the conditions within the HTS bed, silica (SiO2)

is mobile and over time will be carried into the bed
causing deactivation of the HTS catalyst due to silica

AMMONIA TECHNICAL MANUAL 142 2005

 Low Temperature Shift (LTS) Case 1 :

The risks [26] associated with catalyst start-up are A large ammonia plant had to carry out a top skim of its
well known; LTS reduction is very exothermic. LTS bed due to high pressure drop; the main cause was
poor atomization of quench water.
CuO + H2 → Cu + H2O ∆H =-81kJmol-1
Case 2:
The procedure is only completed when a new
charge of catalyst is commissioned and this is increas- The photograph below shows the failure in 2003 of
ingly becoming an infrequent operation due to longer pipework at the position of a quench nozzle upstream of
catalysts lives. an LTS bed on a hydrogen plant (converted ammonia
plant). Erosion caused by the water flash had reduced
The reduction of a new charge is usually well moni- the 8”NB pipe wall thickness from 10mm (0.4”) to less
tored and controlled normally with the assistance of the than 2mm (0.08”).
catalyst vendors.

Re-oxidation

If a plant were to lose the feed to the reformer but

the air compressor remains on line, the eventual result
would be an exotherm guaranteed to rupture the LTS
vessel. However, the HTS is likely to experience an
exotherm first as outlined above. Re-oxidation of the
LTS catalyst needs to be carefully controlled at the time
of discharge.

Wetting/Condensation Figure 9: Failed LTS line at quench

It is known that water can either be either hero or Other cases of similar failures [27, 28] at quench
villain in the shift reactors [18]. If the correct process nozzles have been reported earlier.
conditions are applied. it participates in the WGS reac-
tion heroically producing H2! However, if the catalyst Lessons:
is wetted either by steam, either by being allowed to • Closely monitor hydrogen concentration and tem-
condense on the catalyst or it is wetted due to a waste perature rise during reduction.
heat boiler leak, then depending on the intrinsic • When using natural gas as a carrier for reduction
strength of the catalyst, the pellets may collapse. The keep the bed temperatures below 230°C (446°F).
moisture will also leach chlorides down the bed, deacti- • Make sure trip systems that prevent process air get-
vating the rest of the catalyst. ting to the LTS converter are working.
• Ensure quench systems are on shutdown inspection
In many older plants an upstream quench is used to lists and are operating properly with good atomisa-
add additional steam to encourage the WGS reaction. tion.
Without regular inspection and maintenance the spray • Make sure the isolation valves do not allow wet syn
attemporators can fail to atomise correctly leading to gas to pass during front end start-ups.
catalyst wetting or more seriously critical damage to the
pressure shell itself.

2005 143 AMMONIA TECHNICAL MANUAL

Methanation Support Media

In operation, the main concern regarding the The duty of the support media for the various cata-
methanator is the level of inlet carbon oxides as tem- lyst vessels on an ammonia plant is frequently over-
perature runway is possible with high levels of carbon looked in preparation for catalyst change outs.
oxides. It is important to consider how quickly the trip
system will respond to changes in bed temperature, as Below the bed: Supports do a key job holding the
shown in [29] where due to delays in heat transfer from loads imposed from above while preventing catalyst
the catalyst bed to the thermocouple bed temperatures from passing towards the outlet collector while generat-
reached approximately 750ºC (1,382°F) despite the ing a low pressure drop.
inlet temperature trip setting being 400ºC 752°F).
Above the bed: Support media on top of the bed
The main hazards when methanation reactors are protects catalysts from the high inlet gas velocities thus
shutdown are nickel carbonyl formation [1], and self- preventing disturbance and milling. In some cases, the
heating when exposed to air. support media captures entrained solids and therefore
prevents larger pressure drop increases across the bed.
This example of self-heating comes from a small
methanator on an olefin plant. Case study 1

The plant was shutdown and purged with nitrogen; Figure 10 below shows a bed of ZnO that had been
the inlet and exit valves and thermocouples were re- operated without any hold down material such that the
moved for repair. Open ends of the pipes were covered particles had been milled, which had been picked-up
with plastic sheet with the catalyst in reduced state. circulated and deposited forming a central mound.

During the shutdown on reconnection of the ther-

mocouples a reading of 454ºC (1,013°F) was seen. To
cool the vessel, a nitrogen purge was established which
gave a reading of 649ºC (1,200°F) on the bottom ther-
mocouple.

Clearly, the plastic sheet is not an adequate isola-

tion allowing air to enter the vessel containing the re-
duced catalyst; a chimney effect due to the exotherm
then drew in additional air which in turn increased the
exotherm.
Figure 10: Mound formed from ZnO milling
Lessons:
• Reduced methanation catalyst becomes very As mentioned in the HTS section high voidage me-
hot when exposed to air. dia can also be used to improve the desired properties
• Secure isolation and inert purge is essential for of the support e.g., pressure drop reduction or addi-
maintenance on vessels containing reduced tional screen capacity.
catalyst.
• With little or no gas flow thermocouples do not Case study 2
show the peak temperature.
A large plant had recently changed a full charge of
The final chapters of this paper concern the loading 60m3 LTS catalyst, which was reduced without any
of catalysts and therefore apply to a number of the reac- problems. However when operating at full rate the start
tor duties across the flowsheet. of run pressure drop was measured at 2.1 bar (30 psi)
and rising, where the previous charge had only been 1
bar (14.5 psi) at end of run.

AMMONIA TECHNICAL MANUAL 144 2005

 A catalyst problem was suspected and addition downstream and collected on the tubes of the LTS inlet
pressure tapping (taps) were added to the catalyst drop exchanger – which required regular shutdowns to clean.
out pipe of the vessel to help characterize the pressure
drop in different parts of the vessel. As a result of this A recent enquiry associated with HDS and HTS
measurement the pressure drop was quantified as 0.3 catalysts simply specified ‘ceramic balls’.
bar (4 psi) from catalyst bed and 1.8 bar (26 psi) below
the catalyst bed. • Lessons:
• Store support media to the same standard as cata-
The vessel was emptied and a detailed investigation lysts the costs will be the same if they fail!
completed. Investigation showed that a number of dif- • Only use high purity alumina support above 300ºC
ferent types of inert balls had been used and that one of (572°F) in steam environments
the types installed had failed catastrophically, Figure
11. • Reuse of support balls is not uncommon but these
incidents highlight the dangers inherent in using
material were the history is not clear.

Catalyst Loading

As catalyst improvements have led to longer run

lengths for catalyst charges changing catalysts is be-
coming an infrequent task.

Changes in personnel may mean that many of the

responsible staff may not have conducted a catalyst
change before.

In these circumstances detailed understanding of

Figure 11 : Failed support balls the design intent and arrangement of the catalyst bed is
important as is detailed records of what to do at the
The cause of the failure was believed to be the catalyst change and records including photographs of
rapid drying of the balls which had become wet during what was actually done.
long term storage. However, it proved very difficult to
simulate in tests the extreme conditions of heat transfer Case study 1
and depressurisation leading to vaporization of trapped
moisture within a 25mm (1”) ball as was thought to A plant implemented a plant up-rate and as part of
have occurred within the plant reactor. this increased the HTS volume (due to advances in cata-
lyst the supplier had suggested installing a lower vol-
The investigation also showed that some of the ume).
balls had previously been used in other reactors i.e.,
HTS and HDS. When mixing inert balls between reac- In order to maximise the catalyst volume the hold-
tors great care is needed to ensure that silica-alumina down system was removed. Subsequent milling in-
supports are not used in the pre-reformer, secondary re- creased the pressure drop.
formers or HTS where silica migration can be an issue.
Case study 2
Case Study 3
In another case in a tubular cooled reactor, the plant
In another case a plant decided to use silica-alumna experienced an increasing pressure drop after a catalyst
balls in a HTS bed at what was thought to be a tempera- change out.
ture that was low enough for silica migration not to be
an issue. This was not the case and silica migrated

2005 145 AMMONIA TECHNICAL MANUAL

 The plant had to be shutdown inspection showed Finally, before conducting any activity that deviates
catalyst had been loaded on top of the tubesheet as well from what has been done in the past, “Stop, Think and
as in the tubes. Removing the catalyst on top of the re- Carefully Consider” the effect of the deviation on the
actor and 150mm (6”) down the tubes restored the pres- plant.
sure drop to normal.

Lessons: References

Gas distribution and bed protection requires careful [1] Catalyst Catastrophes
design along with the rest of a plant during an uprate. Dr. Mike P. Roberts
• Check the design, loading job method and Vol 44 AIChE Symposium: Orlando, 2003
previous loading reports.
[2] Primary Reformer Failure
Conclusions D H Timbres and Mark McConnell
Vol 42 AIChE Symposium: Montreal, 2002
It is unfortunately the case that with increasing fre-
quency history is repeating itself and the hard won [3] Common Problems on Primary Reformers
learning of the last 50 years has been lost and now is Bill Cotton and Peter Broadhurst
being relearned. Vol 45 AIChE Symposium: Denver, 2004

Activities such as shutdowns and start-ups are be- [4] Failure and Damage of HT Shift Converter
coming infrequent practices due to increased plant and D Subagio, L A Pratisto and M Slamet
catalyst reliability. Additionally manning levels are re- Vol 43 AIChE Symposium: San Diego, 2002
duced with many of the experienced plant staff, moving
into new roles, leaving the operating company or retir- [5] Downgraded High Temperature Shift Converter
ing. This all reduces the available corporate memory and Explosion in Air Line to Secondary Reformer
and problems that were long since resolved are reoccur- M. N Othman, K. Niwa and P K. Bakkerud
ring. Vol 43 AIChE Symposium: San Diego, 2002

Many, but not all of the problems detailed, relate to [6] Rupture of a Process Line Exit the HTS converter
plant operation under transient conditions such as start- J Veazey and J Winget
up and shutdown. In such circumstances, plant instru- Vol 29 AIChE meeting Seattle, 1988
mentation may not provide representative readings
leading to a false impression of what is actually hap- [7] Failure of Secondary Reformer Vessel
pening. In many cases, due to the increased plant and M Sterling and A Moon
catalyst reliabilities that are achieved today, a particular Vol 17 AIChE symposium: Salt Lake City. 1969
operating shift may not have started up or shut down
the plant for many years. [8] Shell Rupture of a Secondary Reformer
G M Lawrence
Adherence to standard operating procedures (SOP) Vol 24 AIChE symposium: Denver, 1981
is absolutely vital; failure to do so will place the plant
in conditions that have not been considered or catered [9] Commissioning O2 Blown Secondary Reformers
for in the plant design and operating procedures. This G Shaw, H deWet and F Hohmann
can result in increased risk of a significant incident. Vol 35 AIChE symposium: Vancouver, 1994
The SOP’s represent part of the corporate knowledge of
a company and the learning that this represents should [10] Failure & Repair of Secondary – An Experience
not be ignored without due care and attention to the V.K. Bali, O.P. Kathuria and R.K. Srivastava
consequences. Vol 43 AIChE Symposium: San Diego, 2002

As mentioned two years ago, checklists should be

provided to cover un-familiar tasks.

AMMONIA TECHNICAL MANUAL 146 2005

[11] Hot Spot & Ruby Formation in Secondary Re- [22] Repair and Re-tubing of Reformer Gas Boiler
former A Tipnis, D Deshmukh and N Bandopadhyay
D Pasaribu, I. Primdahl and C Speth Vol 34 AIChE symposium: Orlando, 1993
Vol 43 AIChE Symposium: San Diego, 2002
[23] Failure & Repair of Two Primary Make Gas
[12] Secondary Reforming: Theory and Operation Boilers
P W Farnell P Bailey and J MacDonald
Vol 34 AIChE symposium: Orlando, 1993 Vol35 AIChE symposium: Vancouver, 1994

[13] Failure of Secondary Catalyst & Air Burner [24] Failures in Waste Heat Boilers
Catalyst H Thielsch, F Cone and J Freeman
S Mukherjee, R Jaitely, S Chatterjee and K Aggar- Vol35 AIChE symposium: Vancouver, 1994
wal
Vol 28 National meeting, Minneapolis, 1987 [25] Failures of Secondary Waste Heat Boilers
J Singh, S.L. Varma and B.M. Patel
[14] Secondary Reformer Air Mixer Failure Vol 44 AIChE Symposium: Orlando, 2003
P Rocha de Avila and A S Neto
Vol 34 AIChE symposium: Orlando, 1993 [26] A Problematic LT Shift Catalyst Reduction
I R Barton
[15] Catalyst Fusion Vol 36 AIChE Symposium: Arizona, 1995
K Wright and G Haney
Vol 14 AIChE symposium: New Jersey,1971 [27] Failure of Process Gas Line at LT Shift
C Trotter
[16] Inspection & Repair of a Secondary Reformer Vol 31 AIChE symposium: San Diego, 1990
K A Vick
Vol 27 AIChE symposium: Boston,1986 [28] Natural Gas Line Failure
D H Timbres and Mark McConnell
[17] Investigation & Resolution of Secondary Re- Vol 42 AIChE Symposium: Montreal, 2002
former Burner Failure
P W Farnell [29] Temperature Runaway in a Methanator
Vol 41 AIChE Symposium: Tuson, 2000 A Janssen, N Siraa and J Blanken
Vol 23 AIChE Symposium: Portland, 1980
[18] Water & Your Shift Converter – Hero or Vil-
lain?
Frances Lynch and Sue Appleton
Vol 44 AIChE Symposium: Orlando, 2003

[19] WHB’s Problems and Solutions

P Hinchley
Vol 19 AIChE Atlantic City, 1976

[20] Problems with Process Gas Cooler

R Vikus and M Severin
Vol 29 AIChE symposium Seattle, 1988

[21] Failure of Waste Heat Boiler

E Lebold, J Reinghaus and R schober
Vol 33 AIChE symposium: San Antonio, 1992

2005 147 AMMONIA TECHNICAL MANUAL