FAILURE AND REPAIR OF THE SECONDARY REFORMER – AN

EXPERIENCE (a)
V. K. Bali*, Indian Farmers Fertiliser Cooperative Limited (IFFCO), India

1. Summary

The paper addresses the failure of large magnitude experienced in the Secondary
Reformer of an Ammonia Plant after giving satisfactory service of about 13 years.
Massive in-situ repair and the various examinations to establish the probable cause of
the failure have also been discussed.

2. Introduction

Located at Bareilly Uttar Pradesh in India, Indian Farmers Fertiliser Cooperative Ltd, operates
two ammonia plants, each with name plate capacity of 1350 MTPD of ammonia. Both of
these plants have been designed based on Haldor Topsoe technology with steam reforming of
natural gas and/or naphtha.

Ammonia-I was commissioned in the year 1988 and has been operating normal since then
without any major breakdowns.

Failure of large magnitude has been experienced in Oct. 2001 in secondary reformer of
Ammonia Plant-I, after giving 13 years of service.

The paper highlights the incident of the failure, subsequent detailed inspection to establish the
cause and extent of damage and in-situ repair of large magnitude in secondary reformer.

3. Secondary Reformer

The secondary reformer forms an integral part of the synthesis gas generation section of a
conventional ammonia plant, and performs two useful functions within the ammonia
production process. Firstly it is the point of introduction of the elemental nitrogen from the
air, from which the product ammonia is synthesized and secondly it achieves a very high
methane conversion, giving a low methane content gas for further processing before the final
ammonia syntheses reaction.

Functionally, the secondary reformer consists of two distinct regions, the combustion zone
and the fixed catalyst bed. The combustion is non-premixed, and since the feed gas is above
its auto-ignition temperature ignition occurs automatically as the two streams mix.

*Executive Director (nitrogenous fertilizer)

Email: [email protected]
(a) received on 29/04/02
 2

4.2 The Process

Reformed gas from the primary reformer enters secondary reformer at the top portion. The
gas enters at 800ºC and 32 kg/cm2g and flows down through the annular space. The gases are
directed downwards through a deflector and guide ring to enter the combustion zone of the
reformer. Preheated air at 550ºC is introduced to the air mixer burner. The process gas passing
downward, comes in contact with the air jets. The air jet entrains the process gas on top and
bottom side. It also entrains a large quantity of recirculating fluid, flowing upwards from the
centre of the bed, which acts as a flame stabliliser ensuring ignition of process gas and air
mixture. After the ignition, the hot gases at a temperature of about 1200ºC pass through the
catalyst bed.

The combustion reaction supplies the heat required for the endothermic reforming reaction
that takes place in the catalyst bed. The temperature at the outlet of reformer is about 990 ºC
and pressure is 31 kg/cm2 g. Methane at the exit is 0.3 mole %.

4.2 The Details

The secondary reformer for the plant was supplied in 1986 by Hyundai Heavy Industries,
South Korea. The reformer operates under the most severe conditions of the process having a
pressure of 32 kg/cm2 (design shell pressure 38 kg/cm2) and extreme temperature reaching to
1200 deg. C. The normal shell temperature is about 160 deg. C. with design temp. of 350 deg.
C. The shell of the reformer is made of low alloy steel CS SA204 Gr B (C-1/2 Mo) having a
thickness of 80 mm and conical portion has a thickness of 92 mm. The thickness of neck
portion and bottom dished end is 50 mm and 92 mm respectively. Due to high operating
temperature. The shell is refractory lined from inside to minimise the heat loss and to keep the
shell temperature less than 350 deg. C. Figure 1 shows the internal details of secondary
reformer. Further, to warn against any possible refractory failure and high shell temperature,
the shell is coated with Thermo Indicative Paint which changes colors at high shell
temperature. Figure 2 shows the temperatures profile across the cross section of refractory and
shell of secondary reformer.

3..1 Packing Details

Packing details from bottom to top of secondary reformer are as given below:

Catalyst support:
The catalyst support in the secondary reformer is a high alumina firebrick diffuser cone.
Around the diffuser cone alumina lump of size ranging from 100-200 mm at the bottom and
50-100 mm at the top are placed. Above this to a height of 300-mm alumina lumps of size 25-
50 mm are placed.
Catalyst:
Above alumina lumps, 39 m3 of nickel based catalyst is charged. The height of catalyst bed is
2800 mm.

Alumina lumps:
 3

Above the catalyst bed there is 250 mm high packing of 25-50 mm alumina lumps.

Distribution tiles:
On the top of alumina lumps, there is a 50-mm thick hexagonal refractory alumina tiles to
distribute the gases.

In the middle of this arrangement of tiles, 37 hexagonal tiles are without holes. Around this,
are arranged 1150 pieces of hexagonal tiles with nine hole each. On the periphery 92 tiles are
arranged depending on the space left after arranging the hexagonal tiles.

Circle bricks:
Around this set of hexagonal tiles, 60 standard circle bricks are placed. The circle bricks are
fitted tightly against the reformer lining.

3..2 Refractory Lining:

The first layer in contact with secondary reformer vessel is 200 mm thick insulating castable
refractory (50-60% alumina). The next layer, which is in contact with insulating layer on one
side and the high temperature process gas on the other side, is high alumina castable
refractory (94% alumina). All these refractory are held together by stainless steal anchors
which are welded to the vessel (Refer Figure 3).

4. History of Secondary Reformer

Secondary reformer of Ammonia-I has been operating satisfactorily since commissioning in

the year 1988 without any major breakdown. Brief history of the vessel describing the repairs
carried out on different occasions is as given below:

4.1 The Refractory

Refractory of conical and neck portions was inspected during every turnaround. Complete
refractory of the secondary reformer was inspected during 1990,1992 and 1995 after emptying
out all the catalyst. Most of the times the condition of refractory was found to be alright and
only minor repair work was carried out except in 1992 as described below:

In 1992 the topmost ring of conical portion was found badly damaged and hot face refractory
of about half of the conical portion had fallen down. There were five vertical cracks about 5-6
mm wide in topmost layer of cylindrical portion and there were gaps behind every segment of
hot face refractory in conical portion. Total hot face refractory in the conical portion was
replaced. Refractory was repaired in topmost segment (top 900 mm) of cylindrical portion and
in top neck up to 200 mm from bottom of the neck.
 4

4.2 The Catalyst

The catalyst RKS-2 7H (Vol-39m3) supplied by HTAS was charged in secondary reformer
when the plant was commissioned in 1988. During 1990, 1992 and 1995 all the catalyst was
removed and screened for dust removal to achieve reduction in pressure drop. Complete
charge of original catalyst was replaced with new catalyst C-14-2 supplied by UCIL, in
August 98 after achieving longevity of around 10 years.

4.3 The Air Mixer

The only major maintenance carried out on the air mixer was during 1992 when the mixer
was found in heavily damaged/eroded condition. The holes of the air mixer had been enlarged
and were having multiple cracks. The mixer assembly was cut and spare assembly was
welded.

5. Failure of Secondary Reformer

5.1 Description of incidence

Ammonia-I was running normal at production rate of 1500 MTPD when on the evening of
18th Oct. 2001, gas leakage from one of the longitudinal weld joints in the conical portion of
secondary reformer was observed. On close observation, it was found that the gas was leaking
from a 10mm long crack and a pinhole at a distance of 10mm from the crack on the edge of
one of the longitudinal weld seam of the cone. Figure 4 depicts the various longitudinal and
circumferential weld joints of secondary reformer.

Considering the critical process conditions of very high temperature and the process gas
containing hydrogen, the Ammonia-I was immediately shut down for thorough inspection and
repair.

5.2 Description of Failure

Initial inspection immediately after shutdown:

5.2.1 Inspection carried out from outside

• Magnetic particle test: On inspection by magnetic particle testing, a crack of

approximately 65-mm length (10 mm surface and 55 mm sub surface) was detected on
the longitudinal seam L-1.
• Ultrasonic testing: On inspection by ultrasonic testing (UT) from outside, the length of
the crack with varying depth was found to be around 1650 mm on the longitudinal seam
L-1.
After having detected a long crack on L-1 seam, the other two longitudinal petal joints (L-2
and L-3) of the conical shell were also checked by UT from outside and similar cracks were
found. The length of cracks was detected to be 1350 mm on L-2 seam and 2360 mm on L-3
seam with varying depth.
 5

5.2.2 Inspection carried out from inside

After cooling of the secondary reformer, the refractory from the conical portion was removed
for detailed inspection from inside. The crack on L-1 seam was visible with the naked eye to
the same length as was indicated by UT test from outside. The other two longitudinal petal
joints (L-2 & L-3) of the conical shell were also checked from inside after removing
refractory and similar cracks were found as was indicated by UT test from outside.

5.2.3 Detailed inspection after removal of refractory

Results of inspection (DP/MPI/UT) on various weld seams after removing refractory from
inside the vessel are summarised as below:

Weld seam No. L-1

• 1650 mm long crack at weld edge with depth varying from 10 mm to 92 mm from inside.
Weld seam No.L-2
• 1350 mm long crack at weld edge with depth varying from 12 mm to 62 mm from inside.
Weld seam No. L-3
• 2360 mm long crack at weld edge with depth varying from 12 mm to 62 mm from inside.
Weld seam No.L-4
• A 65 mm long crack observed in UT at a distance of 290 mm from the lower edge of the weld seam.
The same was confirmed by DP/MPI from inside. The depth of the crack was 6 mm to 7 mm.
Weld seam C-3
• 39 Nos transverse micro cracks were observed during DP test from inside on the weld and Heat
Affected Zone. Depth of cracks was 5 to 6 mm.
Weld seam C-4
• 20 Nos circumferential cracks of varying length on the upper edge of the weld seam were revealed
during DP test from inside which are distributed all along the circumference. Length of the cracks
varied between 60 mm to 1020 mm totaling to 5045 mm. Many of the cracks were joined together
by subsurface cracks as revealed during further testing by MPI. Similarly 16 Nos circumferential
cracks were also observed at lower edge of the weld seam during DP test. The length of the cracks
varied from 30mm to 1110 mm. The depth of these cracks ranged between 15 to 30 mm on average
as estimated by UT from inside using angle probe. In addition to above, several fine transverse
cracks were also observed on the weld. During UT of the circumferential cracks, the depth was found
to be varying between 8 mm to 12 mm from inside.

Parent Metal
• Several branched cracks observed on the internal surface of the shell which propagated to different
depths varying from 6 mm to 15 mm. These branched cracks were mainly between L seams of the
cone and near the C-3 and C-4 seam on knuckle area.

From the above observation, it was evident that the failed weld joints (L-1, L-2, L-3 & C-4)
had cracks from inside at the weld edge i.e. junction of the weld seam and the parent material
in the HAZ area. However the upper circular weld seam C-3 had only transverse cracks and
the lower circular seam C-4 had developed some transverse cracks in addition to major
longitudinal cracks on the upper and lower weld edges. Photographs (PH-1, PH-2 & PH-3)
show the cracks on longitudinal seams and photographs (PH-4 & PH-5) show the cracks on
C-4 seam and the parent metal.
 6

5.2.4 Inspection after dehydrogenation of long seams

Dehydrogenation and stress relieving of long seams (L-1, L-2 and L-3) was carried out for
improving the weldability of joints. Subsequent to dehydrogenation treatment (DHT) of long
weld seams at 620 Deg C, the original cracks further increased to 3135 mm, 2630 mm and
3135 mm respectively out of total length of 3250 mm of each seam. Figures 5 and 6 depict
sectional view of the cone indicating cracks on longitudinal seams before and after
dehydrogenation. The number of transverse cracks on C-3 and C-4 weld seams increased. The
length of circumferential cracks on the edges of the C-4 weld also increased as some of the
adjacent cracks joined together and few new cracks developed. Also several clusters of
branched cracks were observed on the cone parent metal as well. The concentration of cluster
of cracks was more in the lower ‘knuckle’ portion of the cone about 680 mm height above the
C-4 weld. Also cluster of transverse cracks were observed in a 200 mm wide band below the
C–4 weld. Average depth of cracks was 6-25 mm. Figure 7 shows the cracks on longitudinal
weld seams L-1, L-2 and L-3 and Figures 8, 9 and 10 show the cracks on circumferential
seams C-3, C-4 and on the parent metal in the conical portion of secondary reformer.

6. Repairs of secondary Reformer

6.1 Secondary Reformer Shell

After initial inspection, the actual repairing job could be started after about 8 days from plant
shutdown with the start of heat treatment for dehydrogenation. It was planned to do the
dehydrogenation at higher temp of about 650 0C in place of temperature of around 350 0C
normally used for dehydrogenation in order to improve the weldability of material.

Total affected area was considered for dehydrogenation. Complete heating, soaking and
cooling cycle of about 99 hrs followed. During the dehydrogenation, the load of transfer line
and neck portion of secondary reformer was supported on a mobile crane with chain blocks to
facilitate free movement (thermal expansion) of the conical portion during heating. To avoid
any undue stresses on weld joints between the transfer line and the vessel, the transfer line
was jacked from bottom in proportion to the thermal expansion of the vessel during heating.

After completion of dehydrogenation, NDT inspection (full DPT and random MPT) of
complete cone was taken up. It was observed that length of cracks on L-seams (L-1, L-2 and
L-3) had increased covering almost full length. In addition to this, minor branched cracks
were also found on the parent metal of cone.

The extent of damage and the repair job requirement was firmed up by detailed inspection
carried out after the dehydrogenation.

Though the defects were mainly from inside but considering the major repairs requirement on
the longitudinal seams it was decided to cut open all the weld seams to full thickness and
carry out the welding from outside on account of difficulty in welding from inside in confined
space and high temperature conditions. Figure 11 depicts the details of weld groove design
adopted for repair.
 7

Repair job was started with gas cutting of weld seams (L-1, L-2 and L-3) by automatic air
gouging machine and weld edge preparation by grinding. After cutting, the cut edges and the
adjacent parent metal surfaces showed that cracks are extended into the parent metal also.
These cracks were ground off till there was no further indication of any remaining crack.

Fitting of back strips, compression plate, moon plate and stay pipes was carried out from
inside on all three L seams (L-1, L-2 and L-3) to facilitate the welding up to depth of 50 mm
from outside and to avoid distortion during welding (Refer Figure 12).

The conical portion was preheated up to 200ºC before the start of welding as per the welding
procedure. Buttering of 3 mm thick layer was applied on the grooves from outside to establish
the weldability and achieve a homogeneous defect free surface for further weld deposition.
MPT of buttering layer was done to confirm that it was defect free.

All the three grooves for L-1, L-2 and L-3 were welded by Manual Metal Arc (MMA)
procedure up to 50 mm depth from outside using stringer bead technique. After welding,
intermediate stress relieving of L-seams was carried out at 620-650 ºC temperature with a
soaking period of 1 hour. Stay pipes, moon plates and compression plates were removed after
completion of 50 mm welding from outside.

Back strip was also removed by gouging, grinding and welding was carried out from inside to
fill up the grooves. Subsequently DP test from inside and outside was done. Along with jobs
on L-1, L-2 and L-3, grinding and welding of other parent metal cracks were done from inside
of the vessel. Gouging, grinding and welding of longitudinal and transverse cracks of
circumferential welds C-3, C-4 and heat affected zone were carried out simultaneously.

On completion of all welding jobs as described above, post weld heating of the cone was done
at 300-350ºC. After cooling down of the cone, radiography of weld joints was done.

After verifying the soundness of weld joints, balance outside welding was done followed by
post weld heat treatment at 300-350ºC. Finish grinding was done on all the weld joints for
NDT.

After finish grinding DP & UT of all repaired joints were carried out from outside and inside
of the secondary reformer. Final stress relieving was done at a temperature of 600-650ºC.

Finally after stress relieving, the detailed inspection (PT, UT and hardness tests) of the
complete repaired area were carried out and found satisfactory.

6.2 Refractory Lining

Considering the criticality of refractory lining as a safeguard for shell against high
temperature, the procedure for refractory lining was finalized in consultation with the experts
in the field. External agencies were involved for carrying out the application along with
technical assistance and supervising the refractory lining job.

Refractory lining job was carried out in top neck, conical portion and 450 mm of cylindrical
shell and 800 mm inside the transfer line. Total job was carried out in eight stages for cold
face refractory (200 mm thick layer) and another eight stages for hot face refractory (100 mm
 8

thick layer) (Refer Figure 13). Refractory lining and dry-out of top dished end of secondary
reformer was done separately.

Additional old hot face refractory lining to the extent of 125-150 mm (vertical height) was
removed to make tongue and groove joint between old and new refractory. Complete cleaning
of old refractory and the parting plane of old and new refractory was carried to achieve error
free refractory surface. Casting of cold face refractory was carried out first and hot face
refractory was applied over the cold face refractory layer after natural cooling of first layer.
Studs were fixed to the bush by threading and tack welding. The tips of studs were wrapped
with PVC tape and bituminous point was applied on the studs for expansion clearance. Tape
was also wrapped on the retainer ring for clearance.

After casting cold face refractory up to 300 mm below the neck it was decided to complete
hot face layer of lower portion first since further application of cold face refractory in neck
portion would restrict the movement of men and material. Accordingly stud tips were opened,
V portion of anchors was welded, PVC tape wrapped on V-tips, bituminous applied on the V
portion and casting of the hot face layer started. After completion of job up to neck, the
casting of refractory in neck and other areas was taken up in similar manner.

Total job was carried out in stage wise manner starting from the bottom using original
available steel shuttering of 6 mm thick plates and 40 x 40 mm angle after modifying to meet
the specific requirement. However during execution of jobs, it was felt that use of heavy metal
shuttering within a restricted space was not the ideal choice for maintaining precise lining
thickness as well as in controlling of leakage of water from the castable mix.

Total consumption of cold face refractory (200 mm thick layer) was around 16 MT and that of
hot face refractory (100 mm thick layer) was around 17 MT.
The hot face refractory surface was finished by filling the gaps at joints between different
stages and chipping of refractory wherever spilled out. Hot face refractory in neck portion
required chipping thickness to achieve ID of 1080 mm. The cracks of old lining were patched
using Tabcast-94 castable.

Hot curing of refractory was carried out after complete application of both cold face and hot
face lining. A burner was provided at the bottom of secondary reformer and natural gas was
used as fuel in the burner. A chimney was provided at top of secondary reformer to vent out
the flue gases. The steps in the hot curing included rise of temperature up to 560 oC in steps
with soaking in between followed by controlled cooling up to 150 oC and then natural cooling
as per the laid down procedure. Thermocouples were provided for continuous monitoring of
temperature during curing. The only difficulty experienced during refractory lining job was
during removal of shuttering which was stuck to the refractory surface of hot face lining.

The total repair of secondary reformer comprising of i) Inspection, ii) Heat Treatment iii)
Shell Repair and iv) Refractory lining, as carried out at various stages are summarized below:
 9

INSPECTION
Initial inspection immediately after shutdown (without breaking refractory) and subsequent refractory
removal from inside.
a) Inspection carried out from outside
• Dye penetration test
• Magnetic particle test.
• Ultrasonic testing
• Inspection carried out from inside
• Visual inspection
• Magnetic particle test (MPT)
• Dye penetrant test (DPT)
• Ultrasonic flaw detection (UFD)

Inspection after dehydrogenation treatment to establish the final welding requirement.

• Dye penetrant test (DPT)
• Ultrasonic flaw detection (UFD)

Inspection after buttering of 3mm thick layer of L-1, L-2 & L-3
MPT of buttering layer at 200ºC to establish the weldability and to achieve a homogeneous defect free
surface for further weld deposition.
Inspection after filling 50mm depth of weld groove.
• Intermediate Stress Relieving.
• DPT after grinding

Radiography (RT) of weld joints (L-1, L-2 & L-3) after completion of 50mm welding from outside and
complete welding from inside.

Inspection after completion of welding and post heat treatment.

• DPT & UT from outside
• DPT & UT from inside

DPT & UT of all repaired ‘L’ & ‘C’ seams after final Stress Relieving and Hardness testing
 10

HEAT TREATMENT
(Coils and thermocouple arrangement for heat treatment are depicted in Fig-14 & 15)
Dehydrogenation after initial inspection at a temperature of 620 – 650º C with 1.5 hrs
soaking.
Preheating up to 200ºC before starting any welding job.
Post weld heating up to 300-350ºC after each stage of welding for 3 hours.
Intermediate stress relieving of L-1, L-2 & L-3 after 50 mm welding from outside and weld
edge preparation from inside at a temperature of 620- 650 ºC for period of 1 hour.
Cooling down to facilitate inspection after each stage of welding.
Final stress relieving at 600-650 ºC after completion of all welding jobs with 2.5 hours
soaking time and subsequent controlled cooling.

REPAIRS
Gas cutting of weld seam (L-1, L-2 & L-3) by automatic air gouging machine and weld edge
preparation by grinding.
Fitting of back strips, compression plate, moon plate and stay pipes to avoid distortion of
cone during welding.
Buttering of 3 mm thick layer at 200ºC from outside
Welding by Manual Metal Arc (MMA) procedure up to 50mm depth from outside on L-1, L-
2 & L-3 after preheating up to 200 ºC followed by intermediate stress relieving.
Removal of stay pipes, moon plates, compression plate after welding of 50 mm depth from
outside and post weld heating up to 300-350 ºC.
Back strip gouging & grinding.
Welding from Inside by Manual Metal Arc (MMA) procedure after preheating up to 200 ºC.
Gouging, grinding & welding longitudinal and transverse cracks of C-3 & C-4 weld and
HAZ.
Grinding and welding of other parent metal cracks on cone away from the weld.
Balance welding from outside of L seams after preheating up to 200 ºC.
Finish grinding of weld surface from inside and outside for NDT (DPT, UT & Hardness)
after completion of all welding jobs & final stress relieving.
 11

REFRACTORY LINING

Refractory lining job was carried out in top neck, conical portion and 450 mm of cylindrical
shell and 800 mm inside transfer duct.
Removal of additional hot face refractory lining of 150 mm to make tongue and groove joint.
Fixing of studs in the bushes by threading and tack welding.
Application of cold face refractory in cylindrical and conical portions up to 300 mm below
neck.
Welding of V-portion of anchors.
Application of hot face refractory in cylindrical and conical portions up to 300 mm below
neck after natural cooling of cold face refractory
Application of cold and hot face refractories in balance areas of neck, cone and transfer line.
Finishing of hot face refractory surface.
Hot curing of refractory.

6.3 Resource Mobilisation

Initial inspection of the secondary reformer revealed that there was extensive damage in the
conical portion of the equipment and the same shall involve detailed inspection to ascertain
the extent of repairs required, machining, heat treatment and refractory lining jobs.

It may be noted that there was no spare material available on-site, when the incident occurred,
either for repair of the shell or for the refractory lining, all of which was to be arranged from
outside to enable the job to be carried out.

Keeping in view the volume of job a separate site office was created with availability of
engineers and other technical staff on round the clock basis. The job was monitored at highest
level at plant as well as corporate office.

Outside parties were mobilized for supply and installation of refractory, supply of welding
electrodes, DP test material, skilled manpower, hiring of radiography equipment, material and
manpower for heat treatment, etc. Experts were contacted for carrying out the diagnosis and
supervision of repairs.

Total repairing job was carried out involving manpower from different disciplines. In the
Mechanical discipline 1500 man days of supervisory staff, 670 man days of welders, 200 man
days of fitters and fabricators and 725 man days of riggers were utilized. At a time six welders
(two on each L-seam) were involved in welding.

In the Inspection discipline a total 400 man days were utilized which include the man hours of
inspection staff from other fertilizer units also.
 12

In the Civil discipline 200 man days of both supervisory and subordinate staff were utilized
for removal and installation of refractory.

The Electrical discipline personnel were utilized for providing lighting and other power
connections for heat treatment, welding machines, grinders etc.

12 welding machines, 300 sandering discs, 300 grinding wheels, 2108 kg of electrodes, 800
M3 of BSL scaffolding material, and 2 cranes were utilized in the repair job.

In addition to above, DP material (60 cans of penetrate, 150 cans of developer & 300 cans of
cleaner), 3 ultrasonic flaw detectors, 2 magnaflux MPI units, 30 curie strength source
(Iridium-192) in a remote operated radiography camera were utilised for inspection activities.

The job involved extensive arrangement of heating coils, heating machines and controllers for
heat treatment job. Thermocouples were provided at all critical points and the readings were
brought to a central place for precise control of temperature. For this purpose, 81 heating coils
(Total length – 80 meters) and 42 themocouples were employed. Peak power consumption
(220 Volts supply) for heat treatment during final stress relieving was to the tune of 1 MW.

7. Schedule of Repairs

Not many references of in-situ repair of such a failure in secondary reformer were available.
In the absence of that tentative schedule of various activities was planned.

The overall repair time from the date of incident to the secondary reformer going back into
service was estimated to be around 50-55 days whereas the actual time achieved was 73 days
as per the breakup given below:

Cooling down, removal of catalyst and refractory lining and initial inspection 8 days
Dehydrogenation and detailed inspection to identify the extent of damage. 12 days
Preparation and repairs including heat treatment and inspection 33 days
Refractory lining, catalyst loading and startup 20 days
Total duration 73 days

To carry out repairs in the shortest possible time, adequate resources were mobilized and the
repair work was carried out on round the clock basis. The services of experts in the field,
available in India, were requisitioned and all out efforts were made to repair the equipment
satisfactorily by following the laid down procedure.

The enormity of the job can be judged by the fact that around 2108 Kg of electrodes were
consumed in the welding of cracked weld seams and defects in parent metal. In addition to
above around 33000 Kg of refractory material was consumed in relining of repaired portion of
secondary reformer.
 13

8. Probable Causes of Failure

Before analysing the probable causes of failure, it will be interesting to note the various facts
regarding the secondary reformer:

• Secondary reformer service in an ammonia plant can be characterised as an easy catalytic

duty under difficult mechanical conditions.
• The failure of refractory, damage to air burner and RG boiler downstream of the reformer
are common but incidences of shell cracking and weld failure are very much isolated.
• Secondary reformer operating conditions are basically defined by the process, leaving the
operator little flexibility to change operating conditions.
• Normally monitored parameters in secondary reformer typically provide no indication of
developing internal problems.
• No specific requirements of inspection for secondary reformer have been stipulated by the
Consultant/supplier of the equipment, other than inspection of Air Mixer Nozzle,
Refractory and Catalyst.
• Secondary reformer operating in high temperature and high pressure hydrogen
environment is susceptible to damage due to hydrogen picked up by the metal.

8.1 Effect of hydrogen environment on metals

Any damage to equipment operating in high temperature high pressure hydrogen environment
is due to hydrogen picked by the metal. In addition, any hydrogen dissolved in the metal
during the welding process also adds to it. The two major forms of hydrogen damage which
are of concern are hydrogen attack and hydrogen embrittlement. While the former occurs
during prolonged operation at high temperature, the latter adversely affects the material at
lower temperatures, i.e., during start up and shut down.

8.1.1 High temperature hydrogen damage

Hydrogen attack of steel results in grain boundary fissuring and/or decarburisation. Both these
damages result in the weakening of the metal. The susceptibility to this form of attack
increases with increase in hydrogen partial pressure and temperature. To increase resistance to
attack, the metallurgy is upgraded by the addition of chromium and / or molybdenum. For the
present reactor C-0.5Mo was used as the design temperature of the shell was 350 ºC.
Intergrannular fissure / crack and methane bubbles are detected when metal is damaged by
hydrogen attack.

8.1.2 Hydrogen embrittlement

Hydrogen embrittlement is a phenomenon whereby ferritic steels fail in the presence of

hydrogen and stresses in the metal. The failure is by cracking followed by further growth of
cracks. The hydrogen picked up during welding and also during service can cause this type of
damage. The susceptibility to cracking increases with increase in strength of material,
hydrogen content and stresses. The welds and HAZ have maximum hardness (higher strength)
and therefore the weldments are most susceptible to this type of damage. The mode of failure
can be both transgranular or intergranular.
 14

C-0.5 Mo during welding can develop transformation products (like bainite) which are more
susceptible to hydrogen embrittlement than the base metal. However, with proper weld
procedure and PWHT, the structure can be tempered to reduce its strength (i.e. hardness)
within the non-susceptible range. As far as excessive hydrogen in the metal is concerned,
there are two possible sources, namely (i) hydrogen dissolved during welding because of
inadequate pre-heating and use of consumables containing high amount of moisture and (ii)
hydrogen picked up from the process gas reaching the metal surface through damaged
refractory lining.

At the operating hydrogen partial pressure in the present case, a stable crack may grow over a
period of time under the influence of both the residual and operating stresses and dissolved
hydrogen. The embrittlement effect is likely to be more at lower temperatures. As the
reformer shell temperature is considerably low (around 160º C max), the cracks will grow
along the susceptible path. Once the less susceptible areas of low hardness/ low strength are
reached, the cracks are likely to be arrested.

8.2 Inspection carried out to find the cause of failure

Insitu metallography and replica testing were carried out on the cracked area of longitudinal
seams of conical portion to find out the cause of the failure. The replica taken at the four
locations were examined at higher magnification under microscope as described below:

8.2.1 Outer Surface

On the cracked area of longitudinal weld seam and HAZ microstructure was examined.
Uniform grains were observed in the weld metal. HAZ showed fine grains of ferrite and
pearlite. The crack was observed at 1.5 mm from the weld fusion line on the HAZ. The crack
was parallel to the weld seam and was of transgrannular in nature. The microstructure of base
material near the longitudinal seam showed banding of ferrite and pearlite.

8.2.2 Inner Surface

On the cracked area of longitudinal seam the microstructure showed main crack with
secondary crack. The crack was longitudinal and located in the HAZ.

The microstructure of base material near the longitudinal seam showed banding of ferrite and
pearlite and was similar to the base material microstructure as observed on the outer surface.

The weld metal microstructure from inside showed incomplete dendrites of ferrite and carbide
which are not properly tempered. The HAZ showed bainitic structure. Also, the polished and
etched cross section of the weldment showed that weld bead profile and HAZ were not visible
on outer diameter side whereas the weld bead profile and HAZ were visible on inside
diameter side which indicated that the outside portion was normalized whereas the inside
portion had not undergone complete normalizing heat treatment. Thus it is probable that
PWHT, re-normalizing and tempering treatment of the hot formed conical shell portion of the
secondary reformer might have been inadequate due to non-uniformity of temperature along
the length and across the depth of the weld and the base metal thickness which made the cone
more susceptible to hydrogen assisted cracking during service.
 15

Besides in-situ metallography, the samples of defective material were taken for metallurgical
examination. Photographs PH-6 and PH-7 show several cracks on the sample piece cut from
the weld and the HAZ of longitudinal seam L-1.

The following tests were carried out on the sample pieces to ascertain the cause of the
problem and the condition of shell material

1) Visual examination
2) Metallography examination
3) Hardness test (Weld, HAJ & parent metal)
4) Microhardness test
5) Chemical analysis
6) Tensile and impact test
7) Fractrography

Visual examination of the fracture surface of the samples showed brown to dark brown color
and brittle type of failure. The direction of fracture observed on sample removed from inside
surface of longitudinal seam showed that the cracks had started from inner surface of cone
and propagated towards the outer surface. Also, there were radial marks starting inside surface
which are characteristic of brittle fracture. After the brittle fracture, some quasi-cleavage type
of fracture was observed. Metallography examination revealed that cracks had started from
heat affected zone at the toe of the weld and propagated into base material. The cracks are
fine and transgranular in nature indicative of hydrogen assisted cracking.

The hardness in HAZ of sample removed from the inside surface of the longitudinal seams
showed very high hardness values. The hardness in the outer surface and inner surface of
weld metal and HAZ shows a large variation. Hardness of the weld metal on outer surface of
longitudinal seam is lower than the hardness of weld metal on inner surface of longitudinal
seam. Hardness of the HAZ is very high. With increase in hardness, the notch sensitivity of
material increases making it more susceptible to brittle failure, especially in hydrogen
environment where temperature and pressure are high. Figures 16 and 17 depict the hardness
and micro-hardness test results on weld, HAZ and base metal.

The polished and etched cross section of the weldment showed that weld bead profile and
HAZ were not visible on OD side whereas the weld bead profile and HAZ were visible on ID
side which indicated that the outside portion was normalized whereas the inside portion had
not undergone complete normalizing heat treatment.

Spectroscopic chemical analysis carried out on inside and outside surfaces of both base and
the weld material met the specification requirement of SA 204 Gr. B. The carbon content of
base material is found to be on maximum limit i.e. 0.25 %. The carbon equivalent calculated
showed value of 0.57 which make the steel highly hardenable. High hardness was observed in
the banded area of base material which makes steel susceptible for cracking.

Tensile test for both weld and base material were carried out and were meeting the
requirement as per SA204 Gr B. Transverse impact test specimens were machined from weld
metal. The impact energy values were found to be satisfactory. However, it was found that
impact energy of weld metal near ID shows much lower values (22-27 joule) as compared to
weld on OD (around 55 joule) and base material (around 45 joule), indicating embrittlement
effect due to hydrogen.
 16

Fractography of the sample using Scanning Electron Microscope was made to study the
nature of cracking. The fractographs of the sample indicated that the main cracking which
resulted in failure was of brittle mode and the cracks were intergrannular in nature indicating
towards hydrogen embrittlement as cause of the failure.

Based on various in-situ metallography tests and metallurgical examinations, cracking due to
hydrogen embrittlement or hydrogen assisted cracking appears to be the most likely cause of
failure which is evident from the facts as summarized below:

1) Microscopic examination of specimens from the cracked knuckle area and the welds
indicate that the cracks had started from the HAZ and weld regions from inside surface of
the vessel and propagated further into base material and to outer surface by formation of
voids/bubbles, a mechanism typical of hydrogen assisted cracking.

2) Non-uniformity of temperature control during PWHT, normalizing and tempering of the

hot formed cone during fabrication making it susceptible for in-service cracking.

3) Cracking had occurred on the inside surface of the weld having low hardness values i.e.
213 to 238 VHN.

4) The presence of hairline cracks on many locations on weld, HAZ and base material even
at low hardness levels indicates the severity of hydrogen attack.

5) The tensile test of the base material indicated reduction in the percentage elongation from
37% to 27% on the mid thickness sample of outside surface.

6) Impact testing showed reduction in impact values on the inside surface weld samples as
compared to outside surface samples indicating embrittlement effect due to hydrogen pick
up.
7) The fractrography of the sample indicated the brittle mode cracking.
 17

9. Conclusions

While the failure of refractory, damage to air burner and problems in RG boiler down-stream
of reformer are fairly common, the problem encountered in the secondary reformer of IFFCO
Aonla (India) is certainly an isolated case if not unique.

Considering all the available information, it is difficult to come to a very definite conclusion
regarding the cause of the failure. However, cracking due to hydrogen embrittlement or
hydrogen assisted cracking appears to be the most likely cause of failure.

Our experience brings out the following facts about operation and maintenance of secondary
reformer which are worth considering by all operators of ammonia plant:

• Visual inspections of the hot face lining are misleading and give practically no indication
of the condition of the refractory lining behind.
• It is important not to wait for a catalyst change out or actual refractory damage before
inspecting the secondary reformer.
• Spray of water on the outer surface of secondary reformer during normal operation is
likely to misguide regarding the condition of inside refractory as the hot spot may not be
visual during initial stage of refractory damage.
• Time and money spend in inspection and maintaining the vessel in healthy condition will
always be paid back. Further it can save you from expensive if not catastrophic failure.
• Periodic inspection of the secondary reformer shell and its weld joints from outside can
give fairly good idea about the health of the vessel.
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36