High Pressure Carbamate Condenser

Leak Detection and Control

SAFCO has experienced a leak in the High Pressure Carbamate Condenser of the Urea plant, twice.
As a result, a technical in-house team has studied and developed a methodology for detecting and
monitoring the leak until it can be fixed. The developed method has helped in enabling operation of
the HPCC equipment in the first instance, for more than 14 months and in the second, for more for
10 months.

Khaled Abdul Aziz Al-Khuraimi

SAFCO

Introduction Problem Statement

S AFCO is one of the leaders in the

fertilizer sector, worldwide. It has four
urea plants - three of them are
Stamicarbon licensed with a total design
capacity of 8,700 MTPD. The SAFCO-2 urea
In June 2012, a leak was detected in the HPCC.
It was kept in service under very close
monitoring until its scheduled replacement
during the turnaround in April 2013.

plant is a Stamicarbon licensed plant built by

Chyoda. The name plate capacity of this urea Process Description
plant is 1,800 MTPD, and the normal plant load The SAFCO steam/ condensate system is a
is 108% to 111%. semi-closed loop, as shown in Figure 1. The
source is from extraction of steam turbine
The high pressure carbamate condenser (HPCC) (TS102). Steam is generated in the HPCC
of SAFCO-2 has been in service since (E202) steam drum vessel (V904A/B) and
commissioning in 1992. The total number of condensed in the stripper (E201) steam drum
on-stream time since commissioning is close to (V905), rectifying heater (E302), and evaporator
21 years. (E401) steam drum (V903). The SAFCO-2 urea
plant is the only one of the SAFCO plants that
uses condensate from the desorber as feed to the
steam drum, V-904A/B.

2013 65 AMMONIA TECHNICAL MANUAL

 Figure 1. SAFCO-2 steam/condensate circuit

Conductivity in the system is measured online condensate had a direct relation with the
in two locations – the turbine condensate and desorber outlet conductivity.
the desorber outlet (C802). Routine monitoring  From December 2011 until May 2012, the
of the system is done using the turbine turbine condensate conductivity and
condensate meter, the desorber outlet meter, and desorber conductivity increased above the
the steam drum, V-904A/B blowdown lab limit of 15 and 20 μS/cm·s, respectively.
analysis.  On June 18, 2012, the hydrolyzer level
transmitter was replaced. The desorber
conductivity returned back to its normal
Background History and value of less than 15 μS/cm·s, however the
Observation turbine condensate conductivity did not
decrease.
The background and history can be summarized
in the following points:
 In December 2011, the hydrolyzer level Leak Detection Methodology
transmitter malfunctioned, resulting in
turbine and desorber conductivity increasing All observations in the previous paragraph
above the limit of 10 and 15 μS/cm·s, indicated a leak in the system. The following
respectively (turbine condensate and steps were taken to identify the leak:
desorber conductivity are normally less than 1. List all the possible causes.
10 and 15 μS/cm·s, respectively). 2. Trace the leak.
 Until May 2012, it was observed that 3. Prepare an action plan.
increasing conductivity in the turbine

AMMONIA TECHNICAL MANUAL 66 2013

 Possible Causes CO2. Results were normal as pH was 8.2,
conductivity was 4.6 μS/cm·s, and no NH3
All possible causes were brainstormed and or CO2 was detected.
listed. The list was classified into minor and 2. Tube side leak from rectifying column –
major points. As the leak amount was small and E-302 steam inlet and condensate outlet
only slowly increasing, it was assumed that it samples were analyzed for pH, conductivity,
was from a minor point such as non-return NH3, CO2 and urea. Results at the inlet and
valves (NRV), steam jacket, the nitrogen:carbon outlet were the same. Some results showed
ratio (N/C) meter, or steam traps. It was also higher NH3 at the inlet compared to the
considered that the leak could be from a major outlet, which was due to temperature
source such as equipment or the feed itself. difference. To double check, the outlet
condensate from the rectifying column was
The minor leakage points are as follows: isolated from V-903 and diverted to the
1. Urea melt header into the steam jacket, waste water tank (T-901), but conductivity,
2. N/C meter into the steam jacket, NH3 and CO2 measured in V-903 remained
3. Carbamate pump into the steam jacket, and in the same range.
4. Urea solution transfer line1 steam jacket. 3. Tube side leak from evaporator – The
evaporator tube side operates under vacuum,
The major leakage points are as follows: so in the case of a tube leakage, steam will
1. Make up water quality, go into the process side diluting the urea
2. Tube side leak from the rectifying column, melt. Thus, there is no chance for urea to
3. Tube side leak from the evaporator, come out to the shell side under normal
4. Desorber outlet condensate quality, operation. To double check, a sample was
5. Tube side leak from the HPCC, and taken and the results showed pH,
6. Tube side leak from the stripper. conductivity, NH3, and CO2 were the same
as the rectifying column and no urea was
Tracing the Leak found in V-903. Also, urea melt going from
the evaporator to the granulator was normal
Leak source detection was done, starting from
and there was no change in urea solution
the minor points, by checking the suspected area
concentration.
physically, isolating the suspected source to
4. Desorber outlet condensate quality –
check the effect on the system and taking lab
Desorber outlet condensate was affected to
analyses. The investigation was conducted as
some extent, but when the desorber
follows:
condensate supply to V-904A/B was
isolated, the conductivity reduced from 21 to
Minor Points:
8.6 μS/cm·s and then remained constant.
1. The health of steam jackets and N/C meter
5. Tube side leak from HPCC – Condensate
were checked and found to be in normal
inlet and steam outlet were analyzed for pH,
condition.
conductivity, NH3, and CO2. Concentrations
of CO2 and NH3 in V904A/B were higher
Major Points:
than the other places. Repeated samples for
1. Make up water quality – Make up steam
HPCC inlet and outlet revealed a minor leak
condensate from V-905 via stripper was
in the equipment.
analyzed for pH, conductivity, NH3, and
6. Tube side leak from stripper – Make up
steam condensate coming from V-905 via
1
In SAFCO there is a grid connection between all urea the stripper was analyzed for pH,
plants that is used for transferring urea solution in the case conductivity, NH3 and CO2. The results
of a shutdown.

2013 67 AMMONIA TECHNICAL MANUAL

 were normal and no NH3 or CO2 was 2. A leaking tube-to-tube sheet (TTS) weld,
detected. caused by stress corrosion cracking due to
chloride attack.
Inspection Report 3. A leaking welded tube plug from a
previously plugged tube.
The last inspection for the HPCC was done in
November 2010. The inspection found stress In order to trace the source of the leak an on-line
corrosion cracks in the material of the tube due inspection of the HPCC top and bottom tube
to chloride. The chloride was from upsets in the sheet’s outer periphery was conducted by
condensate coming from utilities, which Stamicarbon. The inspection result was found
occurred in the past, causing chloride to stay in to be acceptable.
the crevice. A tube crack is as shown in Figures
2 and 3. Impacts and Mitigation Action Plan
The following are the impacts of operating
under the conditions described above:
 Corrosion in steam and condensate circuit.
 Corrosion in the turbine.
 HPCC failure before turnaround.

Based on the existing situation and the previous

experienced with a HPCC leak in SAFCO-3, it
was agreed to operate the plant under close
monitoring and control of the conductivity and
ammonia content. The measured values of
conductivity and ammonia indicated a very
Figure 2. HPCC tube crack small leak, so it was believed that there was a
minor risk for corrosion and further damage.

The following action plan was developed and

implemented to manage the potential impacts:
1. Monitor the top and bottom HPCC lab
analysis for NH3 and CO2 pick-up, pH and
conductivity.
2. Monitor the turbine condensate conductivity
analyzer and lab analysis for NH3 and CO2
pick-up, pH and conductivity.
3. Divert the admission steam to vent when the
cation conductivity crossed the limit of
5 μS/cm·s.
4. An on-line inspection of the HPCC top and
Figure 3. Microscopic picture of the crack bottom tube sheet’s outer periphery was
done. The inspection result was found to be
Based on the inspection report, the leak source acceptable.
was assumed to be from the following: 5. Clear limits were also agreed upon, so that a
1. A leaking tube, caused by stress corrosion shutdown has to be taken in case those limits
cracking due to chloride attack.

AMMONIA TECHNICAL MANUAL 68 2013

 were exceeded. These limits were as equipment was ordered after the crack was
follows: found in the last inspection).
 A sudden change in conductivity
(increase or decrease).
 An increase in the NH3 from the HPCC
Monitoring and Control
above 50 ppm (based on experience with It was agreed to operate the plant under close
a high pressure scrubber leak where the monitoring and control of the conductivity and
leak could not be detected below this ammonia content. The measured values of
value). conductivity and ammonia indicated a very
 Keep the pH of the steam condensate small leak, so it was believed that there was a
system above 8.5 (preferably above 9) to minor risk for corrosion and further damage.
minimize the risk for corrosion in the
steam condensate system. Equipment leakage was increasing at a very
 In order to save the turbine, the cation small rate, reaching a value of 60 μS/cm·s. The
conductivity limit was 5 μS/cm·s (also leak continued in the same increasing rate until
based on experience). October 2012 when a crash shutdown occurred.
6. In case the leak crossed the limit, the After the shutdown, the leak quantity declined,
equipment has to be repaired or replaced if but still had the same increasing rate, as shown
the new equipment was delivered (new in Figure 4.

HPCC Top
pH Conductivity NH3 CO2 Poly. (Conductivity)

300 250
Low Load
250
200

NH3 & CO2 pick up
pH & Conductivity

200
S/U after 
unplaned S/D 150
150 Normal Operation
100
100

50
50

0 0

Figure 4. HPCC leak trend after Oct 2013 crash shutdown

Conductivity had increased sharply during the conductivity was possibly due to presence of
low load period due to fluctuation in the corrosion products in the crevice, which
differential pressure across the HPCC. After partially blocked the leak point.
increasing the plant load, the reduction in

2013 69 AMMONIA TECHNICAL MANUAL

 SAFCO-3 Case Benchmarking Regardless of the source of the leak, action
plans to fix the problem were developed as
The HPCC leak in SAFCO-2 was the second follows:
case of such a scenario within SAFCO plants. 1. On short notice until a plant shutdown,
Therefore, the SAFCO-2 HPCC leak case was monitor the conductivity, NH3, CO2 and Fe
benchmarked with the previously experienced in the HPCC.
leak in the SAFCO-3 plant. 2. It is not mandatory, but advisable to treat the
unit with 2% trisodium phosphate (TSP), to
The SAFCO-3 HPCC exchanger leak started on slow down stress corrosion.
September 9, 2010, and contaminated the entire 3. During a shutdown, pull a suspected tube for
steam and condensate circuit. About 30 MTPH further investigation to confirm the presence
(33 STPH) of steam was vented to atmosphere of chloride stress corrosion cracking.
as admission steam, while about 40-45 MTPH 4. During a shutdown, find the leak and repair
(50 STPH) of desorber condensate was dumped. as there is a risk of corrosion of the C-steel
Even so, the leak amount had crossed tube sheet.
2000 µS/cm·s and NH3 pick up across the
HPCC was more than 700 ppm. The plant load Results after the shutdown and repair were as
was maintained at full load (114%). follows:
 A total number of 16 tubes were plugged
The plant was closely monitored for pH, with a reduction in the heat transfer of only
conductivity, NH3 pick up from the HPCC, and 0.5%.
iron (Fe) content. As the leak was high but only  The vented admission steam of 30 MTPH
increasing by a small rate, the decision was was recovered.
made to keep the plant running with close  30 MTPH (33 STPH) of desorber water
monitoring for the mentioned aspects. The leak contaminated with ammonia was recovered.
behavior was fluctuating up and down following  Total recovery of steam as well as process
the system pressure fluctuation. By the end, condensate was 60 MTPH (STPH).
equipment ran around 14 months until a planned
shutdown was taken to inspect and fix the The main cause of the leak was the TTS weld
problem. joint pin hole (Figure 5). Carbamate passed
through this pin hole with high pressure,
Based on the last inspection report, it was initially eroded the stainless steel layer (8 mm,
expected the leak source was a leaking tube. 5
/16 in thick) and then corrosion started at the
Even so the following scenarios were assumed: carbon steel portion (475 mm, 18.7 in thick).
1. A leaking tube caused by stress corrosion The cavity was up to 350 mm (13.8 in) deep and
cracking due to chloride attack. Chloride 125 mm (4.9 in) wide (Figure 6). The
analyses of steam condensate (V-904) for remaining portion of the CS intact was only 125
the last three years remained less than mm (4.9 in).
0.05 ppm, except on September 11–12,
2010, when it reached as high as 3.3 ppm
due to an upset in the utility
demineralization unit.
2. A leaking tube to tube sheet weld due to
thinning of tubes was unexpected to happen
as the tube to tube sheet weld was found to
be satisfactory.
3. A leaking tube plug.

AMMONIA TECHNICAL MANUAL 70 2013

 TTS pinhole
leak from
tube 36 N 1W

Figure 5. Initial pinhole leak where cavity Figure 7. The welded repair area
started
Findings and Results
The equipment in SAFCO-2 has been operating
under close monitoring with a known leak,
keeping leakage as a benchmark. The continued
operation and planning has been based on a
similar experience with the SAFCO-3 HPCC.

Unfortunately, the plant experienced a shutdown

twice during the leak period and the leak
behavior changed. Even so, the plant was kept
running until a forced shutdown was taken when
the NH3 pick up increased sharply crossing the
Figure 6. Corroded area after grinding value of 500 ppm. This amount of ammonia
had an effect on the evaporation section vacuum
The present repair (Figure 7) is only a and the plant was shutdown.
temporary measure. SAFCO inspection and
Stamicarbon indicate the HPCC is to be
replaced at the soonest possible time. Based on Conclusion
the inspection report issued, procurement
SAFCO has good experience in plant
actions have started for a new exchanger.
troubleshooting, even though each case is
unique. The monitoring and control plan
succeeded in the SAFCO-3 HPCC but it failed
in the SAFCO-2 case because two crash
shutdowns changed the leak behavior.
Fortunately, the SAFCO-2 shutdown occurred
10 days before the turnaround and the new
equipment was already on-site and turnaround
resources were simply rescheduled.

2013 71 AMMONIA TECHNICAL MANUAL

AMMONIA TECHNICAL MANUAL 72 2013