Cause of Damage and Repair of

Reformed Gas Firetube Boiler

The secondary reformer waste heat boiler was damaged due to the high pH level of boiler water.
It was repaired successfully within 4 months because synthesis loop boilers were maintained in good
condition.

Zeng Zhong quan

Jianfeng Chemicals, Puling, Sichuan, Peoples Republic of China
Mark A. Holderman
Brown and Root Petroleum and Chemicals, Alhambra, CA91803

Introduction The feedstock for the ammonia plant is natural gas,

the source of which is located 80 km away from the
ammonia/urea complex. The natural gas is delivered to

T
he Jianfeng Chemical Industry Complex was
authorized to build an ammonia/urea complex the ammonia/urea complex by a pipeline at a nominal
by the China National Planning Commission in pressure of 14 bar g.
1987. The China National Technical Import Co. was As a consequence of improper operation of the dem-
given the task of selecting the engineering and pro- ineralized water unit and continued operation of the
curement contractor. The Sixth Design Institute of secondary reformer waste heat boiler with the pH of
China was chosen as the engineering design contractor. the boiler water too high, the secondary reformer
The Jianfeng Chemical Industry Complex is located waste heat boiler was damaged.
near the city of Puling in Sichuan province in China. This article is concerned with the cause of* damage
The location is close to the Wu River which drains and repair of the secondary reformer waste heat boiler
into the Yangtze River. and improvements in operation to avoid future damage
The chemical complex includes a 1,000 metric ton of all equipment and improve safety in the
per day ammonia plant, a 1,750 metric ton per day ammonia/urea complex.
urea plant, and associated utilities and offsites.
The ammonia plant licensor is C. F. Braun and Co. Steam System Description
(now owned by Brown and Root, Inc.) located in
Alhambra, CA. High pressure steam generation
Construction of the Jianfeng Chemical Complex was
performed by the Seventh Construction Company of Demineralized water is imported from a demineral-
China. ized water unit. The demineralized water is preheated

AMMONIA TECHNICAL MANUAL 1997

and deaerated both thermally and chemically. discovered a problem with the boiler water quality.
The deaerated boiler feed water is pumped by high Jianfeng operating personnel were informed that the
pressure boiler feed water pumps to each of two steam waste heat boilers should not be operated with the
drums. boiler water in its present condition:
High pressure steam is generated in three waste heat
boilers, the secondary reformer waste heat boiler and Boiler Water Specification
two ammonia synthesis loop waste heat boilers. Conductivity < 100 \i S/m (at 25°C)
Demineralized water is imported from the off plot to pH 9.5±0.5
maintain the boiler water level in the steam drums.
The high pressure steam is then superheated in two Boiler Water Actual
stages, first by waste process heat, and second by Conductivity < 100 \i S/cm (at 25°C)
Primary Reformer flue gas (see Figure 1). pH 10.5

Secondary reformer waste heat boiler Jianfeng continued to operate with the boiler water
quality unchanged.
The secondary reformer waste heat boiler is a hori-
zontal firetube boiler designed for natural circulation Shutdown for Maintenance
of boiler water. The steam drum is supported on the
waste heat boiler on the risers and downcomers. On January 21, 1994, both the ammonia and urea
The secondary reformer waste heat boiler is single units were shut down. During the shutdown, mainte-
pass with 404 tubes. Each tube is 44.5 mm outer diam- nance and repair was performed on the following to
eter with a wall thickness of 5.8 mm. The length of make the ammonia unit performance test ready:
each tube is 4,850 mm. The tubesheets are thin wall (1) The ammonia synthesis converter effluent con-
type: denser was replaced with the new heat exchanger.
Tubesheet material: SA387.11 C12 (1 1/4 Cr) (2) The boiler feed water preheater tube bundle was
Tube material: SA213 Til (1 1/4 Cr) removed to perform modifications to improve heat-
Shell material: S A302B transfer rate.
A ferrule, 202 mm long, is inserted into the inlet end (3) New inlet guide vanes for the process air com-
of each individual tube of the secondary reformer pressor were installed.
waste heat boiler to a depth of 100 mm inside the (4) Trays in the water wash section of the Benfield
inner wall of the inlet tubesheet (see Figure 2). solution regenerator were checked and repaired.
(5) The process gas vent valves for the CO2 absorber
Startup and Commissioning overhead and the syngas compressor suction were
replaced.
Construction was completed in June 1993.
Commissioning of the ammonia plant began in late Startup After Maintenance Shutdown
May, and the initial introduction of feed gas occurred
on June 23,1993. The maintenance shutdown was completed on
Startup continued smoothly until August 20, 1993. February 10, 1994. Natural gas was introduced into
By this time, the startup of the ammonia plant had pro- the primary reformer again the next day. The follow-
gressed to the ammonia synthesis loop. ing is a summary of plant operation events prior to dis-
For reasons not related to the secondary reformer covery of the damage to the secondary reformer waste
waste heat boiler, Jianfeng operated the ammonia heat boiler.
plant without the assistance of any outside process February 11: The feed gas compressor lube oil pres-
specialists from September to December of 1993. sure control valve diaphragm failed resulting in the
In early December, the outside process specialists loss of process feed gas.

AMMONIA TECHNICAL MANUAL 1997

 February 13: The recently installed replacement waste heat boiler was found to be well preserved.
process vent valve for the CO2 absorber overhead was The refractory was removed from the tubesheet on
operating abnormally. The ammonia unit shut down to the process gas inlet side of the secondary reformer
calibrate the valve. waste heat boiler. No leaks were found in the inlet
February 15: Secondary reformer process air protec- tubesheet with a hydrostatic test pressure of 5.6 bar g
tive steam line was leaking. Process air was stopped to on the steam/shell side; however, 92 of the 404 tubes
implement repair of the steam leak. were found to be leaking. The location of the leaking
February 17: High pressure steam safety relief valve tubes are shown on Figure 3.
lifts due to mechanical blockage of the high pressure The process gas outlet temperature control damper
to medium pressure steam letdown valve. Process air was removed, and no leaks were found in the outlet
was stopped. tubesheet with a hydrostatic test pressure of 5.6 bar g
February 18: Ammonia unit was shut down. The on the steam and shell side.
steam system was completely shut down to check and
repair the high pressure to medium pressure steam let- Checking secondary reformer waste heat boiler
down valve. on the steam and shell side
February 18: With the ammonia front end under
nitrogen circulation, a large quantity of water was The hand hole at the process gas inlet side of the
found in the liquid drain boot on the boiler feed water steam and shell side was opened. A hard shell of
preheater. Water was also discovered when checking deposits was found on the tube bundle approximately
the drain between the steam superheater and the high 200 mm away from the inlet tubesheet. The hard shell
temperature shift converter. Analysis of the water indi- was removed. Clean corroded pits were found on the
cated the presence of PO43-. A check began of the sec- outer surface of the tubes.
ondary reformer waste heat boiler.
Single tube was removed at location near the
Checking Secondary Reformer Waste centerline of tube bundle
Heat Boiler
A layer of a white adhesive substance was found on
The checking process was divided into mechanical the outer surface of the tube from 0 mm to 1,600 mm
and production groups as follows: away from the tubesheet on the inlet end of the tube.
Mechanical Group: (1) Confirm the location and The thickness of the layer of the white adhesive sub-
extent of the secondary reformer waste heat boiler stance thins with the increase in distance from the
leaks. (2) Develop procedure to repair the secondary tubesheet on the inlet end of the tube. The white adhe-
reformer waste heat boiler. sive substance was characterized by looseness; it was
Production Group: (1) Perform analysis as to the soft and easily peeled off.
reason why the leaks in the secondary reformer waste Under the white adhesive layer were brown corro-
heat boiler have occurred. (2) Develop procedures to sion deposits which cover the surface of the tube from
ensure successful startup of the secondary reformer 0 mm to 100 mm away from the tubesheet on the inlet
waste heat boiler after the completion of the repair. (3) end of the tube, and from 0 mm to 3,250 mm away
Correct the causes of the secondary reformer waste from the tubesheet on the outlet end of the tube.
heat boiler leaks to ensure the leaks do not reoccur. The tube was severely corroded from 100 mm to
540 mm away from the tubesheet on the inlet end of
Checking secondary reformer waste heat boiler the tube on the upper half of the tube (the ferrule
on process gas side extends 100 mm beyond the tubesheet inside the tube).
The tube was most severely corroded at a distance 100
The refractory on the process gas inlet side between mm to 160 mm from the inlet end of the tube.
the secondary reformer and the secondary reformer The severely corroded tube segment was covered

AMMONIA TECHNICAL MANUAL 1997

 ester

BOLBÎFŒDVKMER PUMPS
dCtMJM PRESSURE STEAM HEADER

Figure 1. Simplified steam system flow diagram. Figure 2. Secondary reformer waste heat boiler
tube inlet ferrule.

t
TOP

* SYMBOLIZES
LEAKING TUBE

M
"*ö^"
o" o
'°J *§«•

Figure 3. Location of secondary reformer waste heat boiler leaking

tubes.

AMMONIA TECHNICAL MANUAL 1997

 Table 1. Analysis of Secondary Reformer Waste Heat Boiler Corrosion Deposits

Type Composition (wt.%)

Mg Al Si Ca Cr Mn Fe P
Brown Corrosion Deposits 25.0 2.0 4.6 2.8 0.8 0.6 64.3 Nil
Black Corrosion Deposits 9.6 1.9 2.5 3.8 1.7 0.8 79.7 Nil
White Deposits 58.0 2.3 13.6 8.9 0.2 0.4 12.7 3.9

Notes: (1) The brown corrosion deposits were mainly composed of Fe2O3.
(2) The black corrosion deposits were mainly composed of Fe3O4.

with a layer of thick, hard black corrosion deposits. Checking Related Systems
Each of the black corrosion deposits had a smooth sur-
face and was porous internally. The maximum thick- Ammonia synthesis loop waste heat boilers
ness of a corrosion deposit was 6 mm.
There were also pin-shaped corroded pits on the No water was found on the ammonia synthesis gas
upper half of the tube from 160 mm to 800 mm away side of the ammonia synthesis loop waste heat boilers.
from the tubesheet on the outlet end of the tube (see The hand holes on the ammonia synthesis loop
Figure 4). waste heat boilers steam and shell side were opened.
An analysis of the secondary reformer waste heat The tube surfaces were smooth and clean. The inter-
boiler corrosion deposits is shown in Table 1. nals were covered with a uniform black protective film
with no evidence of corrosion.
Analysis of Boiler Water Prior to Leakage
Steam system letdown valves
Figures 5 and 6 trend the pH and the conductivity of
the boiler water from February 10 to February 17. Hard deposits were found on the inner surface of the
These were the eight days prior to the day before the pipes downstream of each of the three steam letdown
detection of the secondary reformer waste heat boiler valves. The high pressure steam to medium pressure
leakage. steam letdown valves (see Figure 1) that utilize boiler
Operation records of the deminearalized water unit feed water to desuperheat steam letdown from the high
indicate that for the period of time prior to the detec- pressure steam header to the medium pressure steam
tion of the secondary reformer waste heat boiler leak- header (two valves) and steam letdown from the medi-
age, the demineralized water unit was operated with um pressure to the low pressure- steam header were
shortened regeneration cycles due to a high demand checked.
for demineralized water during the startup period of
the ammonia plant. The mixed bed ion-exchange Feed gas compressor turbine
resins were also found to be functioning poorly. The
mixed bed ion-exchange resins were also found to be There was a layer of thin white deposits on the inner
functioning poorly. The mixed bed ion-exchange surface of the medium pressure steam pipe to the feed
resins were subsequently replaced. The high pH of the gas compressor turbine (the medium pressure steam
boiler water was due to excessive addition of NH4OH piping is arranged in a manner so that the feed gas
into the boiler feed water. compressor turbine takes desuperheated steam that is
let down from the high pressure header rather than

AMMONIA TECHNICAL MANUAL 1997

let down from the high pressure header rather than the protective film coating from the tubes leaving the
medium pressure steam from the off plot package boil- tubes prone to corrosion attack.
ers during startup situations). It was theorized that under the high pH boiler water
condition in the steam/shell side of the secondary
Refrigeration compressor turbine reformer waste heat boiler, corrosion developed on the
outer surface of the boiler's tubes. The OH- attacked
The medium pressure steam pipe to the refrigeration the iron in the tubes forming Fe2O3 and Fe3O4. As the
compressor turbine was found to be free of deposits black corrosion deposits had no strength, eventually
(the medium pressure steam piping is arranged in a the corrosion progressed far enough into the tube wall
manner so that the refrigeration compressor turbine that the tube wall was penetrated.
takes medium pressure steam from the off plot pack- The rate of corrosion was temperature related. As
age boilers during startup situations). the boiler tubes became more coated with deposits
during the one week period prior to the failure of the
High temperature shift converter secondary reformer waste heat boiler tubes, the tube
surface temperature increased and accelerated the cor-
A large volume of boiler water was drained from the rosion in the areas of the tubes where the temperatures
outlet pipe from the high temperature shift converter. were the highest.
Repeated rapid temperature excursions during start-
Steam superheater up and emergency shutdown had little effect on the
corrosion.
White deposits were found on the process side of the
steam superheater coils which are located downstream Repair of Secondary Reformer Waste Heat
of the secondary reformer waste heat boiler. Boiler
Secondary reformer waste heat boiler steam The determination of the scope of work to be per-
drum formed was as follows:
(1) By design, a maximum of 10% of the waste heat
Inspection of the secondary reformer waste heat boiler's tubes can be plugged. Since 92 of the 404
steam drum found that all internal parts were in good tubes were found to be leaking, a minimum of 52
condition. tubes must be replaced.
(2) By observation through the hand hole, it has
Explanation of Corrosion Problem been noted that all visible tubes have the same degree
of fouling. It was concluded from this observation that
The corrosion and subsequent failure of the sec- all the tubes were corroded to some degree. All 404
ondary reformer waste heat boiler was caused by poor tubes must be replaced, as those tubes not observed to
boiler feed water quality and associated poor boiler be presently leaking would eventually fail.
water quality. The high pH of the boiler water and the The manufacturer of the secondary reformer waste
high concentration of solids in the boiler water was boiler was contacted. The manufacturer required that
observed to occur for one week prior to the failure of the waste heat boiler be transported to their workshop
the secondary reformer waste heat boiler tubes. High in France for the complete tube replacement proce-
pH was observed to occur for a two month period dure. The repair period would be three months, with
prior to the failure of the secondary reformer waste an additional three months required for transportation
heat boiler tubes. to and from the Jianfeng worksite to France. It was
During startup, the secondary reformer was operated estimated that one month would be required to remove
with a high rate of boiler blowdown due to process and reinstall the secondary reformer waste boiler.
reasons. The high rate of boiler blowdown removed Chinese Nanjing Chemical Machinery Works had an

AMMONIA TECHNICAL MANUAL 1997

established procedure to replace the secondary tubesheet.
reformer waste heat boiler tubes in a field location.
Once the waste heat boiler tubes were accessible, the Post weld heat treatment
expected period required was three months.
The new boiler tubes were manufactured by Chinese (1) After completion of the waste heat boiler tube
Great Wall Steel Works by special order with a deliv- replacement, post weld heat treatment was performed
ery time of 30 days. only on the two individual tubesheets using an electri-
In order to resume production as rapidly as possible, cal heating process.
Jianfeng management decided to do the tube replace- (2) A hardness test for the tubesheets was performed
ment on-site. after the post weld heat treatment.
(3) During the post weld heat treatment for the out-
Preparation for tube replacement let tubesheet, several electrical failures occurred that
resulted in failure to comply exactly with the tempera-
(1) The temperature control damper at the reformed ture decreasing phase of the post weld heat treatment.
gas outlet end of the waste heat boiler was removed.
(2) The spool pieces at each end of the waste heat Restoration after tube replacement
boiler were cut and removed to allow space for tube
removal equipment and welding personnel (Figure 7). (1) The refractory on the inlet tubesheet was
(3) The steam superheater was removed to provide installed.
space for removal of the old tubes and installation of (2) The steam superheater was reinstalled and all
the new tubes (Figure 7). associated piping spools were reconnected.
(4) The individual tube inlet ferrules were removed. (3) The spool pieces at each end of the waste heat
(5) The refractory on the process gas inlet side of the boiler were reinstalled.
tubesheet was removed. (4) The temperature control at the process gas outlet
(6) The secondary reformer waste heat boiler and side of the waste heat boiler was reinstalled.
steam drum remained in place during tube replace- (5) The new refractory in the reinstalled spool pieces
ment. at each end of the waste heat boiler was reinstalled.

Tube replacement procedure Dryout

(1) Two 2.6 KVA mill machines, one located at each After curing of the new refractory, the refractory
end of the tube to be removed, were to be used. The dry-out step was carried out simultaneously with the
mill machines were adjusted both horizontally and startup of the reforming and shift systems, the boil-out
vertically (see Figure 8). of the steam dram and boiler, and the drying out of the
(2) A special milling cutter was used to guarantee high temperature shift catalyst.
that the beveling exactly met the requirements of the
welding procedure. Progress of repair
(3) The tube was replaced in small groups of five
tubes in a specified order to prevent the tubesheets The ammonia unit was shut down due to boiler
from deforming. water leakage from the secondary reformer waste heat
(4) To complete the welding position as per proce- boiler into the reformed gas stream on February 18,
dure, the welding body was preheated to 200°C before 1994. The last manhole cover was reinstalled on June
the welding was started. 13,1994 (115 days later). Table 2 is a summary of the
(5) All tubes had intermediate stress relief (ISR) work performed along with the duration of each phase
treatment after welding. of the work.
(6) Pressure test by inspection was required for the

AMMONIA TECHNICAL MANUAL 1997

 dl
Figure 4. Secondary reformer waste heat boiler tube corrosion locations.

Figure 5. pH trend.

391

19«

Figure 6. Conductivity trend.

Figure 7. Secondary reformer waste heat boiler equipment.

AMMONIA TECHNICAL MANUAL 8 1997

Table 2. Summary of Work Performed in Each Phase

Checking, developing repair procedure and repair preparation 40 days

Removal of steam superheater, the two spool pieces cut and removed lOdays^
Removal of outlet control damper and inlet tubesheet refractory 2 days
Tube replacement 40 days
Post weld heat treatment 7 days
Pressure test 3 days
Reinstallation of steam superheater and associated pipe spools 6 days
Installation of new refractory, curing of the refractory and reinstallation of the
two end spool pieces 5 days
Miscellaneous work and calibration 2 days

Total 115 day s

Preparation of Related Systems for Feed gas compressor turbine

Systems for Startup
To clean the steam line to the feed gas compressor
High temperature shift catalyst turbine wet steam was blown to it before startup.
During the January 1995 scheduled maintenance, the
Strict control of temperature during startup was nec- impellers were checked and found to be in good
essary to ensure that the catalyst was dried properly. condition.
After one month of operation, the high-temperature
shift catalyst bed pressure drop reached 3.3 bar. The Steam systems
catalyst was replaced in January 1995 during the
ammonia unit's scheduled maintenance, with the All three steam systems were blown clean using wet
Chinese Laohe Chemical Works type B205 high-tem- steam before the restart of the ammonia unit.
perature shift catalyst.
Steam superheater

The white deposits on the reformed gas side of the

steam superheater were removed by washing with a
mixture of citric acid and a corrosion inhibitor.

Operation After Repair

The following were measures adopted to extend the

service life of the equipment:
(1) The initial temperature increase after restart of
the ammonia unit shall be strictly carried out per
Figure 8. Tube cutting mill machine. schedule. Special attention was to be given to the low

AMMONIA TECHNICAL MANUAL 1997

temperature stage. (1) All boiler tubes were in good condition.
(2) Avoid emergency shutdowns of the ammonia (2) The inlet tubesheet was in good condition with
unit front end with particular attention to the shutdown no leakage being detected.
of the process air system. Avoid fluctuations of (3) The outlet tubesheet had 34 cracks. It was
process and steam temperatures and pressures. recalled that during the post weld heat treatment stage,
(3) The inlet temperature to the secondary reformer several electrical failures had occurred during the out-
waste heat boiler shall be kept above 350°C. let tubesheet heat treatment stage.
(4) Monitor and control the outlet temperature of the
secondary reformer waste heat boiler. Any adjustments Re-repair of secondary reformer waste heat
must be made slowly. boiler
(5) Monitor the condensate sampling points daily for
the presence and concentration of PO43-. (1) The temperature control damper at the process
(6) Monitor and control to design specification, the gas outlet side of the waste heat boiler was again
quality of the demineralized water, boiler feed water removed.
and boiler water. (2) All 34 cracks were ground out and welded.
(7) Increase frequency and duration of the secondary (3) The entire outlet tubesheet was post weld heat
reformer waste heat boiler intermittent blow down. treated using an electrical heating method. The heat
treatment curve was precisely followed and no electri-
Operation summary after repair cal power failures occurred.

June 13, 1994: The last manhole on the waste heat Operation summary after re-repair
boiler was closed. The primary reformer was fired and
with nitrogen in circulation, the refractory was dried The ammonia unit was put back into operation after
out at the same time as the steam drum boil out and the January 1995 scheduled maintenance. Up to April
high temperature shift catalyst drying. 10, 1996, there have been no abnormal fluctuations of
June 23, 1994: Feed gas was again introduced to the the secondary reformer waste heat boiler reformed gas
primary reformer with the resumption of ammonia outlet temperature. PO43- has not been found in the
production in three days. Steam production and the condensate at the sampling points.
outlet temperature from the secondary reformer waste
heat boiler were normal. Conclusions
Sept. 8, 1994: The temperature of the reformed gas
from the waste heat boiler began to slowly drop. The secondary reformer waste heat boiler was dam-
Sept. 10,1994: PO43- was found in the condensate aged initially by maintaining the pH level of boiler
taken from the sampling points. water too high and thus making the demineralized
Continuing operation found the waste heat boiler water unit fail.
outlet temperature and the concentration of PO43" in The secondary reformer waste heat boiler was suc-
the condensate at the sampling points varying, indicat- cessfully repaired in-place in 115 days. Ammonia pro-
ing the leakage from the waste heat boiler into the duction was resumed three months earlier than if the
reformed gas stream was not constant. waste heat boiler had been returned to France for
repair. The repaired secondary reformer waste boiler
achieved the design heat-transfer rate.
Re-check of secondary reformer waste heat
boiler The ammonia synthesis loop waste heat boilers were
not damaged because:
On January 1, 1995, the ammonia unit was shut (1) The ammonia synthesis loop waste heat boilers
down for scheduled maintenance. The waste heat boil- were not operated using high conductivity boiler
er was opened again to check it's condition: water.

AMMONIA TECHNICAL MANUAL 10 1997

 (2) The protective film coating was not removed adopted to extend the service life of the equipment
from the ammonia synthesis loop boilers as a result of were carefully followed.
operation with a high rate of boiler blowdown. During the maintenance shutdown of the
(3) The process operating temperatures for the ammonia/urea complex in May 1996, the secondary
ammonia synthesis loop waste heat boilers are lower reformer waste heat boiler tubes were checked by
than the operating temperature for the secondary observation through the two hand hole openings. The
reformer waste heat boiler. tubes were found to be covered with a uniform black
After startup of the ammonia plant, all measures protective film with no evidence of corrosion.

DISCUSSION
Wim Roos, DSM'. Can you tell us something more minus 0.5, and the conductivity at 100 micromho per
about the water specifications? meter at 25°C. Concerning the pH, we had analyses as
Holderman: The water specification, which I listed in early as December at 10.5 and as high as 11. Shortly
the article, was maintained at a pH of 9.5, plus or before the failure, the conductivity had reached 700.

AMMONIA TECHNICAL MANUAL 11 1997