Report by KKParthiban On Boiler Explosion of A Shell Type High PR Boiler
Report by KKParthiban On Boiler Explosion of A Shell Type High PR Boiler
Report by KKParthiban On Boiler Explosion of A Shell Type High PR Boiler
The visit was made to the KPR rice mill on 1st July 2019, to inspect the exploded boiler. The purpose
of the visit was to diagnose the cause of the failure. Further on 2nd July 2019, visit was made to identical
installation at MGR rice Industries. This boiler was kept in shut condition for inspection of
undersigned. On 2nd July 2019, boiler installation at Surya Sri rice mill was visited. This boiler was in
operation. Further boilers at Sri Balaji rice mill and Sri Vijaya Laxmi rice industry were visited. All
the boilers are high pressure cogeneration boiler installations. The purpose of visiting other high-
pressure installations is to gather more knowledge on operational practices.
The boiler is a husk fired AFBC boiler. The boiler has bed coils at the combustor, water wall enclosure
accommodating the horizontal superheaters. A shell type evaporator is provided in place of usual boiler
bank tubes. This shell is provided with two downcomers and two risers. An economiser bank is
provided followed by the airheater for heat recovery. The waterwalls have two side walls, front
waterwall and rear waterwall. The bed coils are provided with separate downcomer and riser system.
The two side walls have a central collecting header. The outlet header has a direct connection to the
steam drum. In addition, risers are provided from top header to steam drum. The front wall has
independent downcomers and independent riser. Similarly, the rearwall has independent downcomers
and independent risers. There are thus 5 circulation circuits.
1. The shell is 5000 mm long. This shell is purely an evaporator. This shell does not have a steam
space unlike other shell configurations used in two pass shell boilers or combi pack cases. The
shell will produce more steam at the front and less steam at the rear. The risers are placed at the
rear part of the shell. The first riser from front tube sheet is at 3400 mm away. See photo 25 in
annexure 1. There is an elliptical inspection door with frame welded at the centre. The steam
production is by pool boiling. There should be risers distributed along the shell length for good
circulation. There is no well-defined circulation as in the water tube evaporator. The steam
production will be high in front part of the shell. The steam vapor would generally collect at the
top part of the shell. Then the vapor needs to travel horizontally towards the rear. Ideally the shell
should be provided with multiple outlets. The riser near the front tube sheet should be larger so
that steam produced will be easily rise and connect to the steam drum.
2. One of the risers at shell outlet has less flexibility. Or it is required to share the weight of steam
drum. The downcomers to side waterwall would also share the weight of the steam drum on thermal
expansion. The downcomers to side waterwall need flexibility.
3. The downcomers for the shell are placed almost at the central zone of the tube nest. Generally, the
downcomers are to be distributed to have a good & uniform water chemistry. Otherwise there can
be zones of different water chemistry. The downcomers lack flexibility. Or they are required to
share the weight of steam drum.
4. A boiler with frequent start / stops operation would generate corrosion product till the pH is
stabilized. Also, when the boiler is stopped, there will be some corrosion due to oxygen. Both the
steel corroded due to oxygen and low or high pH conditions accumulate in lower parts of the
boiler. These are the bed coil bottom header, all waterwall bottom headers and the shell bottom
zones. These corrosion sludges are to be removed periodically from the bottom points of the boiler.
For this purpose, a mud hole is provided at the shell bottom. The blow down location is at the rear.
This blow down is not designed for removal of debris. Ideally the blow down should be placed at
both front & at rear ends close to tube sheet.
5. The mud hole at bottom of shell is placed 3200 mm away from front tube sheet as per drawing.
See photo 25 in annexure 1. Ideal location for mud holes would be close to the tube sheets, so that
manual cleaning will be possible.
6. The boiler is supplied with two reflex type level gauges. These level gauges cannot be viewed from
operating floor. No remote indication instrument such as electronic water level Indicator ( EWLI)
or remote liquid level indicator RWLI is available.
7. The steam drum is provided with a level transmitter. There is a single element drum level controller
that would adjust the feed pump rpm. The same controller would generate alarms for low & very
low drum levels. The drum level controller can be put on manual mode too.
8. A feed control valve is provided in this boiler. However, the boiler is provided with positive
displacement pumps. The feed control valve is generally kept full open by logic.
9. The boiler is provided with one safety valve at main steam line and two safety valves at steam
drum. However, the exhaust pipings are too long and of the same dia as in safety valve outlet. The
pressure drop in line would cause chattering of the safety valve. The safety valve make is Darling
Muesco. This safety valves usually have 10% blow down.
10. The steam drum baffle system is with perforated baffle plates. Wire mesh is provided inside the
baffle box. The baffle box is provided with holes of 6 mm dia above normal water level and with
10 mm holes below the normal water level. However, the baffle box has full bottom opening to a
height of 150 mm or so. It is likely that some steam would travel along with water to the
downcomer zone. This can retard the circulation.
11. The downcomers for the shell are tapped at the bottom of the steam drum. Only downcomers for
front & rear wall are taken from sides. However, they are inside riser baffle box. Steam bubble
could retard the flow to downcomers.
12. It is learnt that the boilers operate with 7.5 % oxygen levels. There is no online oxygen analyser
available. The rice husk is pure and is directly conveyed to fuel bunkers.
13. The gusset stays at shell are placed only in top part of the tube sheet. See photo 25 in annexure 1.
In bottom part of the tube sheet it is not provided. As per IBR calculation a gusset was not required.
14. The gusset stay weldment is not a full penetration weld. It does not offer proper support. See photo
7 & 31 in annexure 1.
15. The design gas inlet temperature is 487 deg C for the shell. The inside water temperature will be
237 deg C at an operating pressure of 32 kg/cm2 g. When there is even a small layer of steam
blanketing or scale formation on the front tube sheet, the metal temperature of tube sheet would
exceed permissible design temperatures. The inlet tube sheet needed protection with refractory
lining. There have to be ferrules extending beyond tube sheet so that tube sheet ligament cracking
& furnace to tubesheet weld cracks can be avoided.
16. The tube sheet to shell weldment should not be as per IBR weld figure XII 12 or 12 A. Since it is
a tube sheet exposed to flue gas, it should be as per XII 16. The tip of the tube sheet is exposed to
gas temperature of 485 deg C. This thick weldment is exposed to differential temperatures all the
time. The weldment itself would need refractory lining so that differential temperature can be
minimized. The maintenance of the refractory is of utmost importance. The drawing does not cover
the refractory anchor details which are attached to pressure part. The drawing does not specify the
refractory lining thickness.
17. The HP chemical dosing pipe does not have distributor pipe inside the drum. Better arrangement
to distribute the HP dosing chemical is by injecting at economiser outlet pipe.
1. The blow down sample will not be representative in this boiler. The boiler water is to be collected
from the baffle boxes. But the baffle boxes are scattered. Each baffle box shall be provided with a
blow down sample pipe of 25 nb. All the 25 nb pipes shall be connected to a junction box inside
the drum from which the main blow down pipe shall be taken. Then the blow down pH will be
representative & will be of the circulating water. The blow down pipe shall be provided with a
needle valve on a bypass line. Continuous blow down shall be practiced using the needle valve.
This shall be used for pH and TDS control.
18. The chemistry of water in shell can be different from steam drum as the shell evaporator is working
on pool boiling. There have to samples taken at front and rear of the shell for this purpose. At
present there is drain point available at shell rear.
Following are the important observations made at the exploded boiler premises.
1. The shell got opened from the front tube sheet to shell weldment at inside of the shell. The plane
of opening is at the toe of inner fillet weldment. See photos 5,6 & 7 in annexure 1. The location of
failure matches with that of the experiences by other countries. See annexure 7 & 8 wherein two
publications are attached discussing failures in shell boilers.
2. The top row of the stay tubes got pulled out. See photo 4 in annexure 1. The two extreme stay tubes
at top row are found to be inside the tube sheet. However, the bottom stay tubes are found to be
intact.
3. The gusset stays have got pulled out at the tube sheet zone at front end. It is found that the weldment
is not fused properly. It is a not executed as a full penetration weldment. The gusset to tube sheet
weld size is less. See photo 7 in annexure 1. The full penetration weld figure is inserted in this
photo. This figure is from BS code.
4. A fusible plug is provided in front tube sheet of the shell. It did not open up. It appears that fusible
plug was never changed. However, it was found to be intact. See photo 8 in annexure 1. The fusible
plug should have been fitted as high as possible at the tube sheet. Because the interest should be to
keep the top tube sheet to weldment to stay under water.
5. The water content in the shell above the flue tubes is 816 kgs. At an evaporation rate of 2 TPH, the
water level will touch flue tubes at 23.6 minutes. The is quite a long time. Hence fusible plug did
not melt and offer protection.
6. There are some structural tor rods welded at the top of tube sheet and at the bottom of tube sheet
on gas side. These rods would not have been there from production shop, as normally, no quality
engineer would allow a construction TOR steel rod to be welded for holding refractory. The TOR
rods are not welded for the intended purpose. To hold refractory the rods should be shaped to V.
There is no sign of refractory either at top or at bottom tube sheet. See photo 14 & 15 in annexure
1.
7. The secondary crack propagation had taken place along the heat affected zone of the
circumferential weld. See photo 10, 12 & 13 in annexure 1. The failed edge is somewhat brittle
with any stretching of metal. Hydrogen damage generally does not occur here unless the boiler had
been operating on low pH for long durations. The pH of the shell water was never checked. Old
records since commissioning are to be reviewed.
8. The post weld heat treatment procedure & records were reviewed. The shell as a whole was not
put in the furnace for stress relieving. This is not a good practice. Instead heating burner was placed
inside the shell to carry out stress relieving. Cracks can get initiated at the shell to tubesheet
weldment due to improper heating. There was no UT or DP test done after the post weld heat
treatment (PWHT). The heat treatment charts of shell and drum were compared. The reports on
test plate destructive testing were reviewed for three boilers. Both the steam drum and shell plate
test reports were reviewed. The reports do not point out abnormality particularly in the bend tests
of the weld joint. See annexure 2. Hence chances of poor heat treatment are ruled out. However,
in future the failed test piece pattern and photos are to be backed up to confirm fulfillment of
PWHT. Then we can confirm hardening of HAZ takes place due to hostile water chemistry -low
pH.
9. The saddle pad welding came off. There is no proper fusion of the weld. See photo 11. However,
this has nothing to do with the explosion incident.
10. The water level gauges are reflex type. The reflex gauges need operators to climb up the floors to
check in person. Even then these gauges do not indicate properly. It requires good illumination and
small amount of draining is necessary to note the water level.
11. The only information for drum level is from the drum level transmitter. There is only one
transmitter. This transmitter signal is used for level control for adjusting the speed of the feed
pump. Water level deviations are possible as there was only one operator at the time of failure
incident. In case it happens, the shell will open at the crown as observed here.
12. It was learnt that only one operator was in shift during lunch time at which the explosion took
place. The operator who was in shift got injured as he jumped off the operating floor. He was not
available for discussion.
13. Normal set point for drum level seems to be 60% as known from other plants. This is in order.
14. The tubes are found to be free from scales. See photo 3 in annexure 1. Even the tube sheet was free
from scale. See photo 7.
15. The logbook showed that the pH of boiler water was at 10 -10.5. The pH was being tested both by
titrant liquid and by pH meter.
16. Eloguard 86 is used as a treatment chemical in the boiler. However, the quality is questionable.
There is no protective layer found on the shell plate or tube sheet or on flue tubes. However, there
was no oxygen pitting too.
17. There has to be water level low & very alarm generated from the drum level controller. Alarm
schedule list and alarm settings are not made available. There are no direct acting level switches.
This boiler was shut and offered for inspection. This was to assist diagnosis of the failure at KPR rice
mill.
See annexure 3 for photos & comments. The summary of observations is as below.
1. There is possibility of steam bubbles entry in to the downcomers. The chances are more when the
drum level gets lower. See photo 3 in annexure 3. As the density water entering downcomer
reduces, there will be reduction in circulation to evaporators. To positively eliminate this
possibility, the recommendations given to another plant are provided in photo 4 & 5 in annexure
3.
2. The steam drier needs improvement in construction to prevent mechanical carry over. See photo 6
in annexure 3. Mechanical carry over can cause tube failures in superheater sections. See photo 7
showing our recommendation given to another installation.
3. The steam drum internals show presence of corrosion products. The steam drum is seen with
considerable foaming due to corrosion impurities. See photo 8 & 9 giving a comparison of a steam
drum at this plant & a good steam drum from another plant. Also see photo 15 showing the scale
& corrosion in the shell. See photo 16 & 17 showing the presence of accumulated corrosion
products in the boiler. The pH of the feed water & boiler water may not be as per requirement.
The feed water & boiler water chemistry report is reproduced in photo 19 of annexure 3.
4. The blow down practice should be to achieve the steady pH of the boiler water. If required a needle
valve may be installed at the shell drain for sampling & for continuous blow down. Inappropriate
chemical can also cause this. Eloguard does not seem to do the job. Insist on supplier to help to
achieve proper protection against corrosion. It is advisable to go for basic chemicals such as
morpholine & trisodium phosphate (TSP). Morpholine must be added at DM water at DM plant
outlet itself. Morpholine addition at deaerator should be for achieving return condensate pH of 8.5.
The TSP must be dosed at boiler water through an injection point in economiser outlet piping.
TSP concentration must be at 20 -40 ppm. The aim is to maintain boiler water pH at 9.5 – 10.0.
Deaerator water temperature should be at 105 deg C for oxygen removal. Additionally, oxygen
scavenger such as hydrazine or carbohydrazide is required at 5 ppm level just to eliminate residual
oxygen. Whenever the boiler is to be stopped, water should be drained out when the boiler is hot,
say at 70 deg C. This will leave the boiler dry. For long term preservations dry instrument air can
be filled in boiler, deaerator and entire steam and condensate piping.
5. The HP chemical dosing pipe does not have distributor pipe inside the drum. Better arrangement
to distribute the HP dosing chemical is by injecting at economiser outlet pipe.
6. See photo 14 in annexure 3. This shows the sign of steam blanketing at the shell top near the front
tube sheet. This is not a good sign. There is a need to release the steam lock from this location.
Additional riser is required here, or the one riser must be shifted to the front.
7. The front tube sheet of the shell is exposed to direct heat. See photo 18 in annexure 3. Though the
gas temperature is not alarming (< 485 deg C), thermal strain would be set up due to differential
temperature between inside & outside of the weldment. It is already pointed out in this report
elsewhere, that the tube sheet to shell weldment is not appropriate. The entire tube sheet along with
weldment must be covered with 100 mm thick refractory. The anchors have to be 75 mm long and
should be of 4mm dia & of SS 304 material pitched suitably. The anchors can be V anchors
twisted to hold the refractory in a better way. The length of ferrules must be increased by 100 mm.
in case of any scaling on inside the tube sheet, the tube sheet temperature will approach gas
temperature.
From the observations made at this plant, we can understand the deviations that can cause corrosion
fatigue at the shell to tube sheet weld joint. Rice mills in general have more cycles of startups / shut
down / banking cycles. The boiler is at risk due to steam blanketing at front side of the boiler. In any
case the ultrasonic testing & DP testing on the shell to tubesheet weldment is advised to eliminate the
risk of explosion.
This boiler was in operation at the time of visit. This was to assist diagnosis of the failure at KPR rice
mill.
See annexure 4 for photos & comments. The summary of observations is as below.
1. The bed temperatures were less than 600 deg C in this plant. There can be fluidisation quality issue
due to which thermocouples show less temperature. Presence of stones is ruled out as the husk is
directly coming from the mill. When the free board combustion is high, the shell inlet temperature
would rise. The shell inlet temperature thermocouple was not working properly.
2. It was learnt that this plant had been operating continuously. The drum level control was working
alright. The direct water level gauges were not working here. As such these gauges are not visible
from operating floor. It is advisable to install a second drum level indicator.
3. The boiler should be inspected in shut down for steam locking in front part of the shell as in MGR
rice industry. Depending on the years of operation, UT should be carried out in the shell to
tubesheet weldment. It is applicable for both front & rear of the shell. UT should be carried out in
the shell circular seam joint too.
4. Refractory lining shall be carried out for the front tube sheet of the shell.
5. Discussions were held with operators for proper feed water & boiler water chemistry and on
continuous blow down requirement.
6. Discussions were held with boiler operators on the method of cooling of the boiler during shut
down. Forced cooling must be stopped. Boiler must be left with crack open vent valve. The cold-
water addition is not required once the fire is put off. Slow cooling will prevent the crack initiation
and its growth.
7. Remote water level indicators are recommended for the drum level indication at operating floor /
control room. The manufacturers catalogs are attached in photo 3 – 6 of annexure 4.
E. OBSERVATIONS AT THE BOILER AT SRI BALAJI BOILED & RAW RICE MILL
This boiler was in operation at the time of visit. This was to assist diagnosis of the failure at KPR rice
mill.
See annexure 5 for photos & comments. The summary of observations is as below.
1. The steam drum did not show evidence for proper water quality maintenance. See photo 1 & 2 of
annexure 5.
2. Recommendations listed under section C4 on water chemistry is very much applicable here also.
In general, it is concluded, that the rice mill boiler operators have not been able to achieve the
required feed water & boiler water chemistry. Combined with frequent cycles of startup & shut
down, the shell will be vulnerable for corrosion fatigue mode of failure.
3. In this site the downcomer weld joint leak is reported. It was not clear that the failure was a crack
or pin hole. However, downcomer / riser flexibility is an issue in all the boilers. The steam drum
must be supported at the same level as that of waterwall. Steam drum support by downcomer of
waterwall is one option. Many boilers are operating with downcomer support. Another
recommendation would be to allow the steam drum to remain in base supported condition and add
flexibility to the downcomers. See photos 4 & 5 showing the recommendations given by us for an
IBL make boiler.
This boiler was in shut down at the time of visit. See annexure 6 for photos & comments. The summary
of observations is as below.
1. The boiler is with pendant superheater. The superheater is non drainable. See photo 1 of annexure
6. Hence care needs to be taken during shut down and during hydrotest. During shut down boiler
water should be drained at 1 kg/cm2 pressure so that metal will dry up itself. This will leave no
condensate inside the SH tube bends. During filling SH must be filled first through a filling line at
MS line.
2. Recently a PSH bend had failed and the failed tube sample was available at the plant. See photo 2
& 3 in annexure 6. There can be internal deposit as steam blowing might not have been carried out
properly. Any foreign material deposit can cause localized overheating. Superheater with trapped
low pH water can cause pin hole failure. Erosion failure is ruled out, as the thickness measured in
all other bends did not show thinning.
3. Refractory lining was not found in the front tube sheet of the shell evaporator. This must be done
immediately.
4. The deaerator water pH is too less as seen in boiler logbook. The boiler feed water system should
be corroding. Once the corroded products are dumped to drum, it will cause foaming and carry
over. Then there will be deposit related failure in superheater. The pH of the DM water shall be
boosted in mixed bed outlet itself by morpholine or by eloguard. Continuous blowdown shall be
practised to avoid wide fluctuations in boiler water pH & TDS.
5. The steam drum is seen heavily deposited with hardness salts or phosphates dosed at HP dosing
chemical in the past. The deposits would be available at shell bottom. It is advised to clean up using
iron dispersant. Iron dispersant will cause removal of the iron from the surface. In the process the
scale would be removed. The boiler must be stopped and thoroughly washed to remove the sludges.
The blow down in shell must be relocated near the front tube sheet. The foaming signs confirms
that the SH bend failed due to localized deposit.
6. Recommendations listed under section C4 on water chemistry is very much applicable here also.
In general, it is concluded, that the rice mill boiler operators have not been able to achieve the
required feed water & boiler water chemistry. Combined with frequent cycles of startup & shut
down, corrosion fatigue can be the most possible failure mechanism.
Shell type boilers suffer extensive damage when operated with low water. For this reason, the shell
boilers were limited to 17.5 kg/cm2g operating pressure by CIB, Tamilnadu in the year 1990, when
we had put up for design approval for 21 kg/cm2 g operating pressure. There are cases of extensive
damage even in the case of low-pressure shell type boiler.
The boiler that is exploded is the case of water tube cum shell configuration designed for a maximum
working pressure of 35 kg/cm2g. Shell here is used as an evaporator. Such design was used by TSPL
and Thermax. But they were with distributed downcomer and distributed riser system.
The following are the possible causes for boiler explosion, considering the boiler had been in operation
over 6 years. The failure mode is corrosion fatigue. The metallography change might have happened
due to hostile water chemistry environments in the past. The cracks might have been initiated during
fabrication or later due to water chemistry. It is the thermal cycles which has led to explosion.
Explosion might have been triggered by low water level or pressurization of the shell or due to growth
of the crack path to a critical point.
• The crown part of shell to tubesheet weld suffers thermal strain in normal operation due to
steam blanketing at the front end. Riser is required at front part of the shell.
• Since volatile chemical is used for pH adjustment, whenever the boiler is cooled the pH of the
boiler can be lower than 8. Only in closed system the pH can be maintained. Otherwise the pH
gets lowered, as the volatile chemical leaves with the vented steam. The boiler needs to be
internally inspected for knowing the effect of water chemistry. It appears there is no adequate
knowledge on the part of the boiler operator. The chemical supplier does not inspect the boiler.
His knowledge may also be limited as they only market the chemical. The failure plane was
mostly along the heat affected zone of the weldments. The edge seemed to be affected by
hydrogen. It is possible that pH was less during initial days of operation or even whenever the
plant is started. The reason is that the DM water pH is low. Only after startup of the plant, the
pH is being adjusted. The DM water is not pH adjusted before storing in DM water tank.
• The hand holes plates of headers were never cut for internal inspection. The mud hole is not
opened, and the shell bottom is never cleaned up. There can be debris accumulation under the
foot of the front tube sheet. This can cause corrosion induced cracks at the heat affected zone
of the tubesheet to shell weldment.
• The refractory work is improper / inadequate at the front tube sheet. Its importance is not known
to boiler operators.
• The weld figure used for the shell to tube sheet is not appropriate. Present weld design adds to
thermal stress and facilitates crack initiation at the toe.
• Post weld heat treatment of the shell was done in a novel way. As per IBR regulation 267 e,
the entire shell is to be put inside the furnace so that every part of the shell would be uniformly
heated. The location of thermocouple should be placed in a manner that both inside and outside
metal surfaces achieve the same temperature while accounting soaking time. The explosion
crack had travelled along the HAZ. The appearance of HAZ is seen to be rough. A circular
seam is generally unaffected by hydrogen as it comes under bulk water. At fillet weld where
water is stagnant, the low pH would easily initiate hydrogen crack.
• One has to check the QAP used for the shell. It is not clear whether the DP test was carried out
on the inside of tube sheet to shell fillet weldment. It was necessary to smoothen out the weld
to avoid crack propagation.
• If a boiler is left with low pH water during storage, the heat affected zone gets affected. The
cracks get initiated. Dry storage is often preferred for long storage periods.
• For filling & hydrotest the boiler may be filled with DM water whose pH is 8.5. There is no
pH boosting arrangement at DM water outlet in all the plants visited. This can cause hydrogen
cracks.
• Frequent shut down and startups cycles can initiate crack propagation.
• Forced cooling with cold water is not advised. In fact, once the fire is out, there is no
requirement to maintain the water level at drum.
• Cooling with FD fan is not advised until the boiler water temperature comes down to 100 deg
C.
j) Failure to carry out ultrasonic flaw detection as per BS 2760 / visual inspection as per IBR
• There is no provision to inspect the status of shell tubesheet weldment any time after
installation. It calls for removal of some flue tubes. But this must be done once in 5 years to
know subsurface cracks. UT is advised at the critical welds. DP test is also advised for
identification of toe cracks. It is necessary to remove insulation for headers, drum seam welds
and main steam piping joints and inspect the weld by DP test. This will ensure safety as required
by IBR regulations 390. The manufacturer / any other QC company may be engaged for this
purpose.
• In the absence of regular inspection of the critical weldments, there is no guarantee against
failures.
• The safety valves must be floated / hand popped annually. This was either not done / records
are not available. When the safety valve is not floated, there can be seizure at the seat. Then
for any eventuality the safety valve will not be available. Pressure excursions can trigger the
explosion from the weak areas.
• There should be record for having done this floating or hand popping. Plant owners must sign
and hang the report in boiler control room. The safety valve exhaust piping is not as per
recommendations by well-known safety valve suppliers. It is not as per ASME code too. The
safety valve escape piping is provided with long piping / bends. This creates pressure drop as
the steam is released. The safety valve would chatter. It is advised to enlarge the steam escape
pipe to larger size ( two steps up ).
Immediate steps required for similar shell configurations irrespective of the pressure.
1. The shell weldments shall be inspected by DP test & UT. The number of years of operation have
already exceeded 5 in all cases. In order to create access for inspection, some flue tubes must be
replaced.
2. The shell & steam drum circular seam and longitudinal seams shall be inspected by UT once.
3. Shell to tube sheet weld design shall be corrected for the front tube sheet.
4. Front tube sheet along with its weldments should be lined with refractory to a thickness of 100
mm. Ferrules length shall be increased by 100 mm. Anchors shall be designed as outlined in the
report above.
5. A second level transmitter is advised immediately for all plants. This is required irrespective of the
boiler design.
6. Remote water level indicator can be provided for direct indication of the water level at the operating
floor. Electronic water level indicator is an option for foolproof indication of the drum level. This
will increase the confidence of the operator.
7. The feed water / boiler water chemical treatment program needs to be closely monitored. The pH
at the shell blow down should be at 10. The difference between drum water pH and the shell water
pH has to be understood.
8. Feed water shall be distributed near the downcomer zone. This way subcooled water will be
available at down comer. It will improve circulation. The present feed distributor piping design
shall be revised accordingly.
9. The front / rear wall downcomer shall be extended so that the steam would not enter these
downcomers.
10. Each baffle box shall be provided with a blow down pipe of 25 nb. All the 25 nb pipes shall be
connected to a junction box inside the drum from which the main blow down pipe shall be taken.
Then the blow down pH will be representative & will be of the circulating water. The blow down
pipe shall be provided with a needle valve on a bypass line. Continuous blow down shall be
practiced using the needle valve. This shall be used for pH and TDS control.
11. The pH of the DM water should be boosted at DM plant outlet itself. The pH shall be 8.5.
12. Morpholine / Eloguard shall be dosed to ensure return condensate pH is 8.5.
13. If boiler water pH is not achieved TSP shall be dosed at HP dosing point. The dosing point shall
be at the economiser outlet piping.
14. The steam drum / shell inside surface should develop magnetite layer with present chemical
treatment. If not change to conventional phosphate-based treatment.
15. The deaerator water temperature should be 105 deg C. The deaerator tower shall be checked for
condition of the spray nozzles and cleanliness of the trays.
16. As the shell produces more steam at the front, a riser should be placed at the front of the shell. The
existing riser shall be shifted to the front.
17. There shall be a in service blow down point near the front tube sheet. The pH shall be tested here
also. This is apart from the present blow down point at the rear. Both shell blow downs shall be
operated once in a week.
18. Flexibility shall be introduced in downcomers & riser to avoid failure of weld joints / stubs.
19. Safety valve shall be hand popped once in 3 months. Record shall be kept.
20. Safety valve exhaust piping dia shall be increased so that the pressure drop downstream can eb
reduced. If the piping is taken up, then an open drain is a must near the safety valve.
21. Drum level transmitters must be calibrated annually. The level indicated by both transmitters must
tally or the difference should be constant.
Permanent remedy
22. As a permanent remedy to take care of the operational cycles in rice industry, go for water tube
bank with an economiser bank to take care of the heat duty required. Refer photo 20 to 30 in
annexure 1. This would be a permanent arrangement for reliability of the boiler.
K.K.Parthiban
ANNEXURE 1: EXPLODED BOILER AT KPR RICE MILL, KOMARIPALEM
Photo 1 & 2: The top
photo shows the
exploded boiler.
Bottom drg shows the
boiler configuration.
The shell below the
drum gave way and it
caused the extensive
damage. It is a good
engineering practice to
avoid shell
configuration. The
cause may be low
water level operation
or pressure excursions
or corrosion due to
improper water
chemistry. Water tube
failures don’t cause
big damages like this.
Photo 3: The above is the P&I of the exploded boiler. The steam drum is with two level gauges. One
drum level transmitter is used for adjustment of speed of reciprocating pump. One number drum
pressure transmitter & one number main steam pressure transmitter are provided for panel indication.
Photo 4: The photo
shows the front tube
sheet. The top row
flue tubes are pulled
out. We can say that
the shell got opened up
from the top of the tube
sheet. The gusset
weldment is not fully
fused. It is not a full
penetration weld. The
failed edge does not
show any stretched or
elongated metal. The
crack propagated from
the toe of the
weldment. The heat
affected zone may not
have been softened by
PWHT. Low pH boiler
water chemistry also
causes hydrogen
damage which is
typical of this kind of
fractured edge.
Photo 5: The photo shows the front tube sheet to shell weldment. Part of the inner weldment is seen
here. The failed plate edge is seen to be brittle. It can happen due to low pH operation. The heat
affected zone is supposed to be too less
on the inside. The weld size is 6 mm
here. It appears part of the weld is not
fused properly. This pH of the water
would be different from the bulk water
at such places. UT test and DP test
should reveal such weak locations.
These tests may not have been done.
Photo 7: The photo shows the front tube sheet to shell weldment. Part of the inner weldment is seen
here. The failure plane is along the toe of the inner weldment. The failed edge is seen to be brittle.
The heat affected zone is small here as the weld size here is only 6 mm. The crack cannot be
attributed to improper PWHT method. It was informed that the burner is placed inside the shell. The
gusset weld is not a full penetration weld. Hence it does not provide support. Hydrogen induced
cracks can be the possibility for failure. Cyclic operation of the boiler would have facilitated the
growth of crack across the thickness of the base metal of the shell. The water chemistry is not good
as seen by the presence of mild scale on the tube at the shell entrance. The rest of the tubes is seen
corroded uniformly. This can be due to low or high pH or due to fluctuations in pH of boiler water.
Photo 8: The photo shows the fusible plug in the front tube sheet is intact. There is very less
possibility of low water level.
Photo 9: The photo shows flue tubes weldment. The flue tubes are of 63.5 dia x 4 thk. The weld size
is adequate to act as stay tubes. The ferrules have been provided to protect heating of tubesheet &
tubes at the entry. However, the tube sheet gets heated from the front. On the other side if scale
occurs, the tube sheet metal temperature will approach gas temperature of 480 deg C max.
Photo 10: The photo shows secondary crack propagation which is at the circular seam weld at the
middle of the shell. The crack has travelled along the heat affected zone. Normally this is not the
place of hydrogen embrittlement in service.
Photo 11: The photo shows the poor-quality saddle to shell weld. There fusion is not complete.
Parent material from shell is not pulled out. However, this has nothing to do with the explosion.
Thick edge crack without
elongation along HAZ
Photo 12 & 13: The top photo shows the cracked edge of the circular seam of the shell. This edge is
seen to be rough. The heat affected zone crack without much of elongation can be attributed to
improper heat treatment. However, the conclusion can be drawn only from the test piece failure
pattern. The photos of the bend ( inside & outside bend ) test piece failures should be of interest to
quality personnel. In future the test coupons must be preserved, or photos must be kept for reference.
Photo 14 & 15: The photos show the tor steel rods welded to top & bottom part of the front tube
sheet of the shell. These anchors would not have offered support for refractory. The alternating heat
input due to no of start-stop cycles in the exposed part of the tubesheet initiates cracks on the inside
weldment and allow crack growth. In any case, shell – tube sheet weldments are subject to cracks
initiation due to water chemistry deviations / start – stop cycles.
Photo 16 &17: The above photos show the weld sketch as per code & as per manufacturing drawing.
This weld is usable for shell to tube sheet weldments where there is no heat direct heat input by flue
gas. This weld figure should not have been used here.
Photo 18: The above figure is the recommended figure for flue gas reversal chambers as per IBR.
This weld reduces the possibility of crack initiation at the toe of inside weldment.
Photo 19 -21: There had been several failures in shell boilers at the shell to tube sheet weldment. A
publication brought out on in service inspection is attached in annexure 7.
Photo 22: The above is the introduction of UT requirement by BS 2790 for the 100% inspection of
the weldments. This is not the case in IBR. 100% DP test should have been additionally
recommended. The smoothening of the toe of the weld is required.
Photo 23: The background of the causes of failure & the importance of UT has been stressed out by
the AOTC committee. This committee has laid out periodical inspection requirements too.
Photo 24: The shell type boilers should be inspected periodically once in 5 years. If thermal cycles
are significantly high, the frequency needs to be increased. Considering the stringent requirement of
this weldment, it is prudent to design water tube bank instead of shell configuration. As such the
shell tube sheet is not have space for in service inspection.
Present gas inlet end
Photo 25: The present shell is with gas inlet from left to right. The steam generation is more at inlet
end. Distributed downcomers and distributed risers would be required for the shell type evaporator as
there is no steam space. The blow down should be located close to high heat transfer point or at least
the mud hole should be close to the front tube sheet for easy inspection of status of the tube sheet /
scale on tube sheet. Debris ( loose iron oxide accumulations or scale accumulations) removal will be
easier with the mud hole at the front end. Regular weldment (inside ) inspection is ruled out as the
space is not adequate, for which the shell dia should have been larger. Otherwise the tubes must be
replaced every time a 100%
UT is to be performed.
Photo 4 & 5: The baffle box should have opening for steam outlet as shown in the sketch. No drier
mesh is required here. Only a canopy is required to direct the steam downwards for first stage
moisture separation.
Photo 6: The steam drier system is directly exposed to carry over of water droplets from the bed coil
riser baffle box. The steam drier should be covered all around so that steam enters from top only as
shown in photo 7 below. The baffle box should have window opening with a canopy to direct the
steam downwards. This will reduce the steam bubbles entry to downcomer.
Photo 7: The steam drier system is directly exposed to carry over of water droplets from the bed coil
riser baffle box. The steam drier should be covered all around so that steam enters from top only. The
concept is shown here.
Photo 8 & 9: The photo on the top is the steam drum at this plant. The one at the bottom is from a
different plant. Reddishness indicates deviation in water chemistry. High erratic blowdown will upset
boiler water pH. DM water pH must be raised to 8.5 separately at mixed bed outlet. In case the
magnetite ( grey) surface is not achieved, the chemical can be the mistake. Morpholine can be used
for raising the condensate pH to 9. Trisodium phosphate can be maintained at 20 – 40 ppm in boiler
water to achieve the boiler water pH. Deaerator water temperature should be 105 deg C. But as of
now, there is deviation. This will cause economiser tube failure. Corrosion can be due to improper
preservation during outages. Instrument air can be filled in boiler and sealed. In fact, it can help to all
system connected to boiler including deaerator, condenser, steam piping and turbine.
Photo 10: The photo shows the chemical injection stub in steam drum. This is not used now as
Eloguard 86 is dosed at deaerator. Trisodium phosphate (TSP) should be injected at economiser
outlet line to steam drum. Then the TSP will be well mixed and get distributed in the steam drum.
Photo 11: It is seen that the downcomer pipe for rear wall is taken from bed coil outlet baffle box.
Here steam – water mixture can enter in to the downcomer. It can be seen that the downcomers of the
side wall and shell are just below the baffle box, from which, as of now steam bubbles can enter in to
downcomers of shell & side waterwalls.
Photo 12 & 13: It is seen
that the safety valve
exhaust pipe is extended
with the same outlet size
from safety valve. The
pipe needs to be enlarged.
Drain holes are required
in case the pipe is lifted
up to prevent condensate
accumulation.
Photo 15: This photo shows the opposite side of the shell ( rear part). The shell internal surface is
coated with iron oxides & scale. The fluctuations created by intermittent blow down can be the cause
for reddishness. Continuous blow down practice has to be implemented immediately. The whitish
scaling tendency is due to deviations in water treatment chemical.
Photo 16 & 17: These photos show the bottom side of the shell. The water treatment is bad, or the
shell is not cleaned for long time. Direct visual inspection is not possible due to frame protrusion.
Blow down pipe is required near the tube sheet area. Annually the shell bottom should be washed.
The color of the tubes is not good. There is good amount of corrosion going on.
Photo 18: The photo shows the front tube sheet from the gas side. There are side, top & bottom
sections of weldments directly exposed to high temperature flue gas. These are to be covered to
reduce the impact of thermal stresses that would be set during every heating or cooling cycle. It is
advisable to have 100 mm thick castable refractory over the entire tube sheet. The ferrule length
must be increased accordingly over the present length.
Photo 19: The DM water pH must be boosted at mixed bed outlet itself. The feedwater pH has to be
9. The blow down pH should be 10 – 10.5. TDS is under control. Hence blowdown has to be
minimum. If necessary, a needle valve may be installed for blowdown control. Or else only when
boiler water TDS goes higher than 300 ppm, blowdown must be resorted. During such time the pH
chemical has to be increased. The pH must be measured from shell blowdown. Whenever the boiler
is shut the header and shell drains must be operated.
ANNEXURE 4: BOILER INSPECTION AT SURYA SRI RICE MILL, KOPPAVARAM
Photo 1 & 2:
Photo 4: The above are the remote drum level indicators available for visual observations at
operating floor. This was used in BHEL industrial boilers.
Photo 5 & 6 : The above photos show the conductivity meter type. This will work only if the water is
free from hardness- Ca, Mg salts. This indicator is used in thermal power plants.
ANNEXURE 5: BOILER INSPECTION AT SRI BALAJI BOILED & RAW RICE MILL
JAGGAMPETA
Photo 1 & 2 : The photo on the top is the steam drum here. The colour of the drum is not good here.
There must be iron sludge accumulation at the shell. The mud hole must be opened, and all sludge
must be washed out. Water treatment is to be corrected to achieve the surface as in lower photo.
Reddishness indicates deviation in water chemistry. High erratic blowdown will upset boiler water
pH. DM water pH must be raised to 8.5 separately at mixed bed outlet. In case the magnetite ( grey)
surface is not achieved, the chemical can be the mistake. Morpholine can be used for raising the
condensate pH to 9. Trisodium phosphate can be maintained at 20 – 40 ppm in boiler water to
achieve the boiler water pH. Deaerator water temperature should be 105 deg C. Corrosion can be due
to improper preservation during outages. Instrument air can be filled in boiler and sealed. In fact, it
can help to all system connected to boiler including deaerator, condenser, steam piping and turbine.
Photo 3: The boiler should be inspected in shut down for steam locking in front part of the shell as in
MGR rice industry. Depending on the years of operation, UT should be carried out for the shell to
tubesheet weldments.
• Refractory lining shall be carried out for the front tube sheet.
• Forced cooling must be stopped. Boiler must be left with crack open vent valve. The cold-water
addition is not required once the fire is put off. Slow cooling will prevent the crack growth.
• Waterwall downcomer bends leakage was reported here. The operators reported only pin holes at
the downcomer weld joint. Pin hole corrosion at weldment is possible on account of poor water
chemistry.
• Separate waterwall panels have been added above the bed coils in this plant. There had been stub
cracks or leaks. No photo was available at the plant. The circulation is generally not good in
small heating surface zones. It can cause water chemistry issues as the water management is not
good as seen by the reddishness inside the steam drum. Instead front part of the waterwall could
have been extended downwards as required (partly). However entire waterwall should be
supported at one level. This is possible by supporting the boiler at brackets attached to fins for
the extended part of the waterwall. Ideally all waterwall heights should have been increased.
Then both circulation and supports can be taken care, of course, flexibility is required for
downcomers.
Photo 4 & 5: Shell downcomers & risers need more
flexibility in downcomer and riser piping. This can
be achieved by crisscross layout. The downcomer
tapping at left of drum should feed to right side of
the waterwall and vice versa.
Photo 2 & 3 : SH coil bend had failed at the 10th row from side and 5th from rear. There can be
deposit as steam blowing might not have been carried properly. Any foreign material deposit can
cause localized overheating. Superheater with trapped low pH water can cause pin hole failure.
Erosion failure is ruled out, as the thickness measured in all other bends did not show thinning.
Photo 3: The boiler should be inspected in shut down for steam locking in front part of the shell as in
MGR rice industry. Depending on the years of operation, UT should be carried out for the shell to
tubesheet weldments.
Photo 4: Refractory lining shall be carried out for the front tube sheet.
• Forced cooling must be stopped. Boiler must be left with crack open vent valve. The cold-water
addition is not required once the fire is put off. Slow cooling will prevent the crack growth.
Photo 5 & 6: The deaerator water pH is too less. The boiler feed water system should be corroding.
Once the corroded products are dumped to drum, it will cause foaming and carry over. Then there
will be deposit related failure in superheater. The pH of the DM water shall be boosted in mixed bed
outlet itself by morpholine or by eloguard. Continuous blowdown shall be practised to avoid wide
fluctuations in boiler water pH & TDS.
Photo 7: The steam drum is seen heavily deposited with hardness salts or phosphates dosed at HP
dosing chemical in the past. The deposits would be available at shell bottom. It is advised to clean up
using iron dispersant. Iron dispersant will cause removal of the iron from the surface. In the process
the scale would be removed. The boiler must be stopped and thoroughly washed to remove the
sludges. The blow down in shell must be relocated near the front tube sheet. The foaming signs
confirms that the SH bend failed due to localised deposit.
ANNEXURE 7: AOTC – SAFED PUBLICATION IN SERVICE OF SHELL TYPE BOILERS
Shell BoilerS
Guidelines for the examination
of boiler shell-to-endplate and
furnace-to-endplate welded joints
(SBG 1)
Cut-away section through Macrograph
a typical shell boiler of Shell/Endplate
and tube connection
showing defect
propagation in
susceptible
region
Ground polished
and etched surface Endplate
Defective
region
Boiler
shell
Smoke tube
Shell/End plate
weld (set-in)
Section removed for
metallurgical investigation
Foreword
A revised edition of the Associated offices Technical Committee (AOTC) Guidance Booklet
GN4 was prepared by a working group comprising: representatives from the Safety Assessment
Federation (SAFed), the Health and Safety Executive (HSE) and other interested parties.
This edition was revised in 2012 by SAFed to update references to standards and regulations
where appropriate and to incorporate some changes made by SAFed since its initial publication.
The guidance addresses a particular aspect of boiler examination and is primarily aimed at those
with duties under The Pressure Systems Safety Regulations (SI 2000 No. 128) - including
owners, users and Competent Persons.
The procedures described in this document represent what is considered to be good practice
and incorporate experience gained over many years in assessing defects at shell-to-endplate and
furnace-to-endplate welds in shell boilers.
The Safety Assessment Federation - SAFed - represents the interests of companies engaged
in independent inspection and safety assessment of engineering and manufacturing plant,
systems and machinery.
First published by Safety Assessment Federation Limited, London UK in 1997, this edition 2012
1 INTRODUCTION 5
2 BACKGROUND 6
3 SCOPE 7
5 PREPARATION 12
6 ASSESSMENT OF DEFECTS 15
7 REPAIRS 20
8 ULTRASONIC EXAMINATIONS 21
REFERENCES 24
APPENDIX 25
Since 1993 Inspection Organisations have noted a marked increase in the number of shell
boilers exhibiting cracks at shell-to-endplate and furnace-to-endplate welds; records show
that over 10% of boilers tested in 1995 exhibited cracks at the furnace and access tube-to-
endplate welds* whilst 2.5% of boilers had cracks at the shell-to-endplate welds; these
figures represent a threefold increase since 1992/3. While some of the increased incidence
of cracking may be attributable to Inspection Organisations taking greater control of the
testing of boilers, the overall incidence of cracking appears to be increasing as the general
age of boiler plant increases; changes in operational practices and the demand for higher
output from smaller boilers are factors which could be contributing to this finding.
The Health and Safety Executive's Guidance Note PM 361 provided guidance on weld defect
acceptance levels for set-in endplates to furnace and shell connections of shell type boilers;
this SAFed document supersedes PM 36 and additionally addresses problems of in-service
cracking in the region of such junctions.
2
The 1982 edition of BS 2790 introduced specific requirements for ultrasonic flaw detection
testing of flat plates to furnace and shell welds. The probability of unfused lands and other
significant defects remaining undetected within these junction welds, in boilers constructed
to this Standard, should have been greatly reduced as a result; however for boilers
constructed in accordance with other codes, the non-destructive testing (NDT)
requirements may not be as stringent.
The Pressure Systems Safety Regulations (SI 2000 No 128)3 require that a Written Scheme of
Examination includes any additional testing considered necessary; these guidelines are designed
to assist Competent Persons drawing up or certifying such schemes for shell boilers.
* For the purposes of this document reference to shell-to-endplate and furnace-to-end plate welds includes
access tube attachment welds for which separate acceptance criteria are specified.
Experience has shown that the majority of in-service cracks emanate from the toe of the
internal fillet weld at the junctions of the endplate to shell and endplate to furnace tube.
Such cracks would have developed due to fluctuating stress levels at these junctions and
would have been influenced by the internal environment within the boiler. The stresses are
due to a combination of pressure loads, differential thermal expansion and local
temperature gradients. Oxygen impurities in the boiler water/steam and inadequate pH
(water chemistry) control, among other contributing factors, cause a hostile environment
within the boiler confines.
The prime purpose of the ultrasonic examinations was to detect cracking from the toe of
the internal seal weld of shell and furnace-to-endplate attachment welds; however the NDT
procedure was such that other gross defects in the main body of the weld - such as lack of
fusion, incomplete penetration and internal cavities - were also discovered.
In 1984 the Health and Safety Executive published Guidance Note PM 36 which addressed
the problem of defects within the body of the weld, gave guidance on the assessment of
these defects and, based on a fitness for purpose philosophy, on the determination of future
ultrasonic examination intervals. By 1990 it was recognised that PM 36 was in need of
revision and that a broader approach to the subject would be appropriate; as a
consequence these guidelines additionally address such subjects such as the first in-service
and subsequent examinations, preparation and repairs.
pressurised hot water shell boilers - as well as steam boilers - suffered from cracking
the requirement for the first in-service ultrasonic examination after ten years operation
appeared to be based on a requirement of the Factories Act - for removal of
insulation/brickwork - rather than on technical or statistical considerations.
The likelihood of buried defects, such as unfused land in the body of the weld, being
propagated under boiler operating conditions and rendering the junctions unsafe has been
evaluated; the likely magnitude of stresses in these areas and the capability of the cyclic
conditions to initiate significant fatigue crack growth within the specified inspection period
have been considered. In particular, stresses induced due to pressure, differential expansion
and thermal cycling - especially in the furnace-to-endplate attachments where these stresses
can be severe - were the subject of a series of finite element analyses4 using different
models and conditions. The results indicated that, provided defects were within the limits
specified in these guidelines, significant crack propagation was unlikely under normal
operating conditions. The assessment procedure - Chapter 6 - has been derived from the
foregoing and simplified acceptance/rejection criteria for buried defects have been adopted
as shown in Graph 1 (page 18).
The guidance on preparation, ultrasonic examination and repair has been based on the
collective experience of the Inspecting Organisations participating in the preparation of the
document and is not intended to be definitive.
Pressurised hot water boilers (eg BS 8555) operating pressure > 3 bar g
AND
boiler diameter > 650 mm
AND
working temperature > 100oC
Some designs of vertical boilers possess similar features to horizontal, multi-tubular boilers
and include furnaces, tube nests and large areas of unsupported tube plates. Under
circumstances where all these features are present, this guidance will also be applicable to
vertical boilers.
The recommendations - which are applicable to the welded joints at set-in and set-on
endplates of furnace and shell connections - are for the detection of:
and specify the frequency of tests for shell-to-endplate and furnace-to-endplate connections
- assuming that the boiler is used at all times within its safe operating limits.
Other conditions - such as overheating or failure under loss of water conditions - are
beyond the scope of this document.
The first, in-service, ultrasonic flaw detection testing of the endplate attachment welds for
boilers within the scope of this document should be undertaken no later than FIVE years
after the boiler has been taken into use.
Previously, the requirement was for the first in-service ultrasonic flaw detection examination
to be undertaken after ten years; however, this appears to have been based on a
requirement in the Factories Act for the removal of insulation/brickwork and experience
has shown that many failures from cracking have occurred in boilers considerably less than
ten years old.
Irrespective of the recommendation that the first in-service ultrasonic flaw detection
examination should be undertaken no later than five years after the boiler has been taken
into use, if during the course of normal inspections of the boiler there is reason to suspect
that one or more of the attachment welds may have suffered cracking or be otherwise
defective, the Competent Person may request that the boiler be examined earlier.
excessive corrosion
significant grooving or undercut
development of cracks
loss of water - during operation
overheating
Thermal cycling can have a significant effect on the propagation of cracks especially where
the boiler is fitted with an ‘on/off’ rather than a modulating burner; significant thermal
cycling may influence the decision of the Competent Person when setting the ultrasonic
inspection frequency.*
The periodicity of subsequent NDT should be determined using the assessment procedure
given in Chapter 6. However, this periodicity should in no case exceed FIVE years.
The endplate attachment welds should be subjected to NDT to detect any cracks
propagating from the toe of the waterside fillet weld or from the root of a single-sided
weld. In addition, the front and rear furnace to end plate seam (or the front furnace to end
plate seam on reversed fired boilers) should be tested to detect cracking from the tubeplate
toe of the internal filet weld. At the first in-service ultrasonic examination of these welds,
the testing should include scans for buried defects. Where the buried defect sizes are below
the threshold given in Graph 1 (page 18) this testing need not be repeated; however, if the
buried defect sizes exceed the threshold, the test should be repeated every FIVE years. (See
Section 6.2.2.1)
* Reduction of inspection intervals should be considered where boilers operate with excessive thermal cycling -
ie burner on/off 50 times per day.
Ultrasonic flaw detection over the entire circumference of the welds in the areas indicated
by circles in Figure 1 - should be carried out in accordance with Chapter 8.
Wet Back
Reversal
Chamber
Wet Back
Reversal
Chamber Plate
Access Tube
Reversal Chamber
Furnace
Key to figure 1
Semi Wet
Back Reversal
Chamber
Furnace
C
COAL FIRED BOILER
F
Furnace
Key to figure 1
The six o’clock position where the furnace is located centrally near to the bottom of
the boiler
Areas of shell adjacent to the furnace where the furnace is off-set or for twin furnaces
In some cases - eg small diameter boilers - it may be more practical to prepare the
entire circumference of the weld for examination.
The following ultrasonic testing should be carried out, in addition to that already contained
within SBG 1, to detect cracking from the toes of the internal fillet weld for the full
circumference of the shell. The following scans are required:
Angle beam scans from the endplate/shell to test the weld toe regions
Note
Where smokebox, dryback reversal chamber or other attachment welds obstruct access to
the endplate, opposite the shell attachment weld, it may not be possible to carry out a 0°
scan from the endplate to test the weld root region. In these circumstances it may be
necessary to check for root penetration and fusion by alternative methods. An angle scan
from the boiler shell and/or dressing off the attachment weld reinforcement for a 0° scan
from the shell may offer practical solutions.
If any of the areas examined reveal defects which exceed the acceptance criteria of these
guidelines, the entire circumference of the weld should be examined.
Note
In some instances the extent of ultrasonic examination recommended by these guidelines
may exceed the requirements of the original Code of Construction. It should be borne in
mind that the requirements of the code pertain to inspections during construction; the
recommendations in these guidelines - for the detection of cracks at the endplate
attachment welds - are for boilers in service.
Where NDT reveals defects which are equal to or greater than the acceptance criteria in
Chapter 6, the ultrasonic operator should advise the owner/user immediately so
that repairs or an assessment can be discussed with the Competent Person.
The owner/user should prepare the boiler for examination in advance of the attendance of
the ultrasonic operator/technician.
Chipping -hammers and needle guns should not be used as they may produce a rough
surface which will impede coupling and damage the probe.
Lagging/insulation should be removed to a distance at least 200mm from the position of the
weld(s) to be tested (see Section 4.2) as indicated in Figures 2a and 2b.
Since ultrasonic testing of these welds will be carried out from the shell surfaces, the
exposed areas should be thoroughly cleaned - of scale, flaking paint etc - to sound metal.
200mm
Waterside Waterside
Shell Furnace
450 - 700
The fireside of the furnace tube adjacent to the weld(s) to be tested should be thoroughly
cleaned of scale and combustion products for a width of at least 200mm around the entire
circumference as indicated in Figures 2a and 2b.
The tubeplate should be thoroughly cleaned of scale and combustion products for a
distance of up to 200mm from the toe of the furnace attachment weld for the full
circumference of the furnace as shown in Figures 2a and 2b.
Ultrasonic testing will be carried out from the fireside of the furnace.
5.4 ATTACHMENTS
If, in order to permit examination of a weld, any attachment has to be cut away and
subsequently re-welded to the boiler, re-welding should be carried out to an approved weld
procedure by an approved welder.
The areas on the boiler under consideration for NDT are the various attachment welds
connecting cylindrical sections to flat plates. These comprise:
Front tube plate and rear tube plate to shell 20% UT (except on configurations with
attachment weld flanged ends)
Front tube plate to furnace 100% UT
attachment weld
Rear plate to furnace attachment weld 100% UT
Anchor tubes from rear tube plate to Visual only
reversal chamber
Furnace tube to reversal chamber UT to the extent that the geometric
attachment weld configuration allows
MPI at the first examination
Reversal chamber to shell Visual only
attachment weld
Gusset stays between front tube plate/rear Visual only
tube plate and the shell
This arrangement incorporates two furnace tubes that extend from the front tube plate fully
through to the rear tube plate. The rear of the furnaces contain refractory plugs which
include a removable section for man access. The furnaces can be either corrugated or
completely plain (ie no bowling hoops). The reversal chamber is roughly oval in cross section
and connects to each furnace tube via a short circular tube. This link tube is approximately
75% of the diameter of the furnace tube. The reversal chamber wrapper connects directly to
the rear tube plate, thus forming the dry back. This is insulated with refractory. There is an
‘inset’ into the front tube plate which forms part of a reversal chamber where the flue gas
crosses over from the second to third pass. The areas of front tube plate unsupported by
the tube nests are supported by gusset stays attaching to the shell as opposed to the more
conventional stay bar arrangement. The lower part of the rear tube plate is supported by
gusset stays in an identical manner to the front but on the upper area the gusset stays are
attached to the top of the reversal chamber wrapper plate, thus providing support to both
flat surfaces.
The areas on the boiler under consideration for NDT are the various attachment welds
connecting cylindrical sections to flat plates. These comprise:
Front tube plate and rear tube plate to shell 20% UT (except on configurations with
attachment weld flanged ends)
Front tube plate to furnace 100% UT
attachment weld
Rear plate to furnace attachment weld 100% UT
Furnace tube to gas pass out tube and gas UT on both ends of tube to the extent that
pass out tube to reversal chamber the geometric configuration allows
attachment welds
Reversal chambers (front and rear) Visual only
attachment welds to tube plates
Gusset stays between front tube/rear tube Visual only
plate and the shell/wrapper plate
These guidelines assess two different categories of defect: surface defects and buried
defects.
Surface defects can occur at the toe of the internal fillet weld or at the root of single-sided
welds as indicated in Figure 3. There are essentially two types of surface defect:
undercut
root undercut
root concavity
overlap
lack of root fusion
a t a t
L L = 6mm Min b
Incomplete
Defect at root of penetration of
single sided weld single sided weld
(Fig 3c) (Fig 3d)
T T
Buried defects occur within the body of the welded area and the heat-affected zone as
indicated in Figure 4.
The types of buried defect considered here are common planar defects - such as lack of
side wall fusion, and incomplete weld penetration - and are considered to be in the plane of
the endplate-to-furnace or endplate-to-shell interface.
Other defects - such as hot and cold weld cracking, weld shrinkage cracks, laminar tearing
and volumetric defects (cavities) - may be amenable to similar consideration but in these
cases additional consideration should be given to the reasons for these defects and their
likely influence on the integrity of the joint
2a t
Notation for Figure 4
Unwelded land
T = endplate/tubeplate thickness (mm)
weld root crack
t = furnace or shell plate thickness (mm)
L L = 6mm Min 2a = depth of buried defect (mm)
L = fillet weld leg size (mm)
T
Where surface cracks from the toes of the fillet welds or from the root of single-sided-
welds are detected, these should be repaired. The high local stress levels in these areas can
result in very rapid crack growth when combined with environmental factors.
Shallow grooving or slight undercut at the toe of the fillet weld may increase the risk of
crack formation because of the effect of local stress concentration. Such areas should be
carefully tested to ensure that there is no cracking associated with the surface defect and
they should be rigorously monitored at future inspection intervals.
For single-sided welds on boilers built to BS2790 incomplete penetration or lack of root
fusion up to a maximum of 2.5mm is acceptable - see Figure 3d (For access tubes see
section 6.2.2.3)
Buried defects more than 4mm from the surface should be assessed using Graph 1(page 18).
Buried defects need not be measured at future ultrasonic examinations if the owner/user
can provide documentary evidence that the defect has been previously tested in accordance
with the recommendations in these guidelines and the defect size - 2a - falls below the
'Monitor' line in Graph 1.
If the defect size falls below the appropriate T/t line in Graph 1 but is larger than T/4, the
defect size should be monitored at each future test of the weld - ie at 5 yearly intervals. If
the defect size falls on or above the appropriate line, further assessment and/or repairs are
required. Examples of the use of Graph 1 are given in the Appendix.
If buried defects are closer to the surface than 4mm (fillet weld leg lengths less than 6mm)
there is a possibility that the remaining ligament may fail due to shear; in this case the
allowable defect size - 2a in Figure 4 - should not exceed 4mm. (For defects greater than
4mm refer to Section 6.2.3)
In the case of set-on endplates the acceptance criteria should be subject to special
consideration.
20
(37.5;14.0)
14
(36.9;12.5)
T > 1.2
t
(24;11.8)
12 T > 1.1
t
(36.25;11)
Size 2A(mm)
T > 1.0
(24;10.6)
t
10
Monitor
(Para 6.2.2.1)
(21.25;8.8)
8 (16;8)
(32;8)
o r
nit
6
Mo
(13;5)
Notes
4 1 Linear interpolation between values of T/t
(11;3)
is permitted.
(12;3)
2 P1 and P2 are nominal dimensions and are only
included to ensure that buried defects are at
2
least 4mm from the surface. ie L ≤ 6mm
(see 6.2.2.1) and p ≈ 0.7L.
3 When T/t < 1 the value of T/t = 1 should
be used.
(Co-ordinates of points)
0
5 10 15 20 25 30 35 40
In the particular case of access tubes between the wetback reversal chamber plate and the
rear tubeplate, which are fully stayed, different acceptance criteria may be applied. In this
case, because of the reduced thermal loading, defect size - a - in a single-sided weld of 0.4T
is acceptable - Figure 3d (page 16).
Note
The assessment methods outlined above are designed for application to defects situated
in the plane of the shell-to-endplate or furnace-to-endplate interface. For defects inclined
to these planes the projected widths - established by ultrasonic techniques - should be
used in the assessment.
A more refined assessment - such as that contained in BS EN 79106 - may be carried out
where the defect size exceeds the recommended limits in Section 6.2.2 of these guidelines.
In such cases:
stresses due to pressure, thermal expansion and temperature gradients should be taken
into account in a detailed engineering critical assessment
the remaining combined weld throat depth should not be less than the thickness of the
furnace or shell plate (t)
the nominal shear stress on the net weld area should not exceed 100 N/mm2.
(Stresses due to pressure, thermal expansion and temperature gradients should also be
considered if a monitoring exercise is to be carried out).
All repairs including subsequent NDT should be carried out to the requirements of the
Competent Person certifying the boiler. Workmanship and materials should be to a
standard equivalent to the original code of construction. For further information on repairs
see SAfed Guidance PSG 15, ‘Repairs or modifications to pressure systems’.
In such cases deviations should be agreed by the Competent Person and any additional
testing to ensure integrity should be specified.
7.2 WELDING
BS EN 287 Part 17
BS EN ISO 15614-18
ASME IX9
In most cases the preferred type of repair for a crack at the toe of a fillet weld or root of a
single-sided weld (Figures 3a and 3c) is to use a ‘D’ patch. For cracking into the tube plate
(defect (b) in Figure 3b) it will generally be necessary to remove a ‘D’ patch from the
furnace to gain access to the crack for a grind and weld repair.
When the crack is in a small diameter furnace, or extends for a substantial proportion of
the circumference of a furnace, it may well be more practical to replace the complete ring.
Shallow cracking or grooving (<t/4 deep) at the toe of the fillet welds may be repaired by
grinding and welding.
When deep repairs are required, the metallurgical changes and residual stresses from
welding can lead to early recurrence of cracking in these areas.
Gouging out and weld repair of cracks from the outside is not recommended. This is
because of the difficulty of:
Again, the metallurgical and residual stress effects could lead to early recurrence of cracking
in these areas.
Ultrasonic tests covered in this document will normally be carried out as part of an
examination under The Pressure Systems Safety Regulations (SI 2000 No 128). These
Regulations place duties on Competent Persons and require them to ensure that proper
procedures and precautions are followed - even when some aspects of the work are
carried out by other organisations.
In many instances the tests will be undertaken by the Competent Person who has the
necessary experience and expertise to perform them; in cases where the ultrasonic testing
is carried out by another organisation, the Competent Person will be responsible for
determining the acceptability of:
the NDT organisation and the ultrasonic operator contracted to do the work
the procedures employed
the reports produced
the results.
The recommendations outlined in the following sections are designed to assist the
Competent Person in making the above decisions and are regarded as the minimum criteria
for such examinations.
8.3.1 Qualifications
8.3.2 Training
8.3.3 Experience
A current SAFed/BINDT qualification for NDT of shell boilers ensures that the above
requirements are met and should satisfy the Competent Person’s obligations to ensure that
proper procedures and precautions are followed.
8.4 PROCEDURE
Testing should be carried out in accordance with formal controlled documentation specific
to the boiler(s) to be examined; these documents should be produced by a level III
qualified person who meets the training and experience guidelines in Sections 8.3.2, 8.3.3
and 8.3.4 above.
Where an organisation other than the Competent Person is carrying out the tests, it is the
owner/user responsibility to provide to the Competent Person the documentation in
support of Sections 8.2, 8.3 and 8.4. This documentation should be supplied at least two
weeks prior to the examination and should include:
The Competent Person should review the documentation against the recommended
guidelines and may also require to witness/audit the tests as they are performed; in such
cases, NDT testing should be arranged to coincide with the Competent Person's visit.
8.6 REPORTS
On completion of the ultrasonic tests a full written report should be submitted which
should comprise as a minimum:
In consultation with the Competent Person, other methods eg magnetic particle inspection
13
to BS EN ISO 9934-1 may be applied to supplement ultrasonic testing.
3 Safety of Pressure Systems, Pressure Systems Safety Regulations (SI 2000 No 128)
Health and Safety Commission Approved Code of Practice L122, HSE books
5 BS 855, Specification for welded boilers for central heating and indirect hot water
supply (rated output 44KW to 3MmW ), British Standards Institution, London.
9 W elding and Brazing Q ualifications, Boiler and Pressure Vessel Code, ASME Section IX,
American Society of Mechanical Engineers, New York, USA.
Case study 1
25 6
T = 32mm
13.5
t = 20mm
T/t = 1.6
20
2a = 13.5mm
L1 = 6mm
L2 = 25mm
32
The defect size lies above the T/t >1.2 line in Graph 1 and therefore requires a repair or a
more refined assessment as per Section 6.2.3.
Note
A maximum defect of 13mm in this configuration would be acceptable but would be
required to be subjected to an ultrasonic examination every 5 years.
Case study 2
7 T = 16mm
13 8 t = 10mm
T/t = 1.6
10
2a = 7mm
L1 = 8mm
L2 = 13mm
16
The defect size falls below the T/t >1.2 line in Graph 1 and is therefore acceptable. It also
lies above the 'monitor' line and requires to be subjected to an ultrasonic examination every
5 years.
18 4 T = 20mm
t = 20mm
T/t = 1.0
20
2a = 4mm
L1 = 6mm
L2 = 18mm
20
The defect size falls below the T/t >1.0 line and also below the 'monitor' line in Graph 1. It
is therefore acceptable and the buried defect does not require periodic ultrasonic testing
in future.
Note
The joint should be subjected to periodic ultrasonic examination to test for cracks at the
toe of the weld.
Case study 4
18 5 T = 16mm
t = 12mm
12
T/t = 1.3
2a = 5mm
L1 = 5mm
L2 = 14mm
16
The defect is NOT acceptable because L l < 6mm (ie P1 < 4mm) and 2a > 4mm. (See
Section 6.2.2.1)
Ground polished
and etched surface Endplate
Defective
region
Boiler
shell
Smoke tube
Shell/End plate
weld (set-in)
Section removed for
metallurgical investigation
Safety Assessment Federation
Unit 4, First Floor
70 South Lambeth Road
Vauxhall
London
SW8 1RL
www.safed.co.uk
ANNEXURE 8: PUBLICATION BY WELDING TECHNOLGY CENTRE, AUSTRALIA
TGN-PE-03
Rev: 0
Cracking in Fire-tube Boilers
Date: 1 March 2006
ABN 69 003 696 526
Pages: 6
1. SCOPE
This Note provides information on cracking in welded fire-tube boilers, including location of cracks,
failure modes, causes, detection, repair and prevention. The necessary actions to be undertaken
when cracking is suspected or confirmed are provided. The information provided is based on global
experience.
2. BACKGROUND
Cracking in fire-tube boilers at welded joints is a frequent, costly and potentially dangerous
occurrence. The shortest recorded time for serious cracking to leak is 3 years, representing less than
20,000 cycles, and occurred on a laundry boiler subjected to frequent and rapid firing.
A number of cracked boilers have exploded resulting in major damage and fatalities. Fortunately,
recent improvements in materials, welding and non-destructive testing (NDT) together with a greater
awareness of the potential for cracking have greatly reduced the incidence of failures. As current
boilers age, more cracking is likely to occur.
3. FIRE-TUBE BOILERS
This Note covers cracking in “economic” type fire-tube boilers, where the “dry back” or “water back”
boilers represent the largest use in Australia. The information presented is applicable to virtually all
fire-tube boilers made of carbon-manganese steels with diameters up to 3 metres, operating
pressures generally up to 1 MPa and operating temperature up to 180°C. AS 1228 provides further
details on these boilers. Fig. 1 shows typical details.
Figure 1 shows a schematic representation of the side and end elevation of a fire-tube boiler. The
front and rear closure plates and reversal chambers have been omitted for clarity.
Page 1 of 6
Crack Location Crack Description
Page 2 of 6
The failure mode is corrosion fatigue. Slight corrosion occurs as a result of contact with the water.
Fatigue arises from thermal cycling and pressure cycling. The primary causes of furnace tube
cracking are a combination of:
• High thermal stress generated by large temperature or material thickness differences;
• Bending stresses due to pressure;
• Poor weld shape, particularly at the weld root in the lower part of the furnace;
• High number (over 10,000) of pressure and temperature cycles;
• Fracture of the protective magnetite layer due to cyclic stresses. Magnetite forms on the furnace
tubes and acts as a protective layer but it is brittle and subject to spalling under cyclic stresses.
Its fracture exposes unprotected surfaces to further corrosion;
• Un-removed slag from furnace tube to tube plate welds providing corrosion initiation sites.
For short cracks, the most common type, the resulting failure has generally been leakage. For longer
cracks, the result can be large scale fracture with a dangerous explosion.
Depending on the age and fracture toughness of the tube plate, material crack extension can occur
suddenly by brittle fracture when the boiler cools down to ambient temperature. High local residual
stresses can trigger brittle fracture in heavily cold worked and aged steel. This occurred with a unique
case at location 5 from a 6 mm deep fatigue crack.
5. CRACK DETECTION
Good access is required to visually detect cracks and surfaces should be clean for 50 mm each side
of the weld where cracking initiates.
Page 3 of 6
Visual examination with the aid of lights can detect cracks over 5 mm in length and over 1 mm deep
depending on adequate surface cleanliness. Endoscopes and digital cameras can be used to aid
detection (particularly with low slung boilers), with computers to record information.
Magnetic particle testing (MT) and penetrant testing (PT) are more sensitive than visual inspection if
the suspected crack area is accessible for examination. Ultrasonic testing (UT) is probably the best
method to detect serious cracking.
Ultrasonic testing should be carried out within 10 years from the construction date in normal
circumstances or more frequently under harsh conditions. Similarly if there are significant changes in
operating temperature or pressure, ultrasonic testing should be carried out more frequently e.g. after
initial 10,000 cycles. The ultrasonic testing program should include a reasonable length of weld at
both ends of the furnace and at the top, bottom and sides of the weld circumference at locations 1
and 2. Extra care should be taken with tubes near to stay tubes or near the shell.
Increased inspection frequency should also be implemented if the furnace was manufactured from
steel with Rm>460 MPa, or if the design strength value used is above 110 MPa.
Only personnel with proven expertise and experience should undertake a fracture mechanics
assessment.
Experience with early ductile, low strength steels indicates that furnace tube cracking can be tolerated
up to the lower of 2 mm and 30% of the furnace tube wall thickness. Operation changes should be
implemented to eliminate some of the primary or secondary causes of cracking and de-rating the
boiler output may be required. If crack depths are 50% or more through the furnace wall, the boiler
should be isolated for repair or replacement.
Page 4 of 6
Determination of remaining safe life should take into account:
• Quality and properties of the steel, direction of rolling, presence of inclusions and age. (Prior to
about 1985, most steel in Australia was ingot-poured, with the associated risk of occasional centre-
line inclusions and laminations);
• Severity of future cycling;
• Worst crack length and depth;
• Ratio of furnace tube to tube plate thickness;
• Age of the boiler (gives an indication of the probable origin and properties of the materials used
and the number of cycles experienced);
• Measures to be taken to reduce further corrosion and thermal cycling;
• Preparedness, time and available resources to carry out the repair;
• Feasibility of crack repair;
• Results of any fracture mechanics analysis;
• Residual stress and risk of brittle fracture;
• The measures taken to avoid gas explosions and low water failures. Both can result in severe
plastic straining across the crack leading to furnace tube rupture. These measures are essential
where cracks over 2 mm deep have been detected.
• Outcomes of a risk assessment
• Management responsibility in the event of a failure leading to an explosion.
Given that the crack growth mechanism is corrosion fatigue, the number and extent of thermal cycles
to which the boiler will experience is the primary issue in determining the length of time the boiler can
be operated. In assessing the consequences of failure, it must be determined if the failure mechanism
is likely to be a through wall leak of a boiler tube or a catastrophic rupture of the furnace or tube plate,
potentially leading to an explosion.
Successful repairs rely on competent welders, good weld shape, low hardness, negligible defects and
competent NDT technicians and importantly, proof by tests or previous work that the inside root profile
is as shown in Figure 4.
Figure 4. Local repair technique for furnace tube cracks located at toe of original installation weld
Page 5 of 6
8.2 Replacement of One End of Furnace Tube
This option should be considered where there is extensive cracking at one end and there is good
access from inside the furnace. The repair will involve using procedures, inspection and testing
practices similar to the original construction.
Documenting the number of operating cycles the boiler undergoes together with the severity of those
cycles provides the baseline that will ultimately dictate the frequency of inspection and remaining life
of the boiler.
DISCLAIMER: While every effort has been made and all reasonable care taken to ensure the accuracy of the material contained herein, the authors, editors and publishers of this publication shall not be held to be
liable or responsible in any way whatsoever and expressly disclaim any liability or responsibility for any injury or loss of life, any loss or damage costs or expenses, howsoever incurred by any person whether the
reader of this work or otherwise including but without in any way limiting any loss or damage costs or expenses incurred as a result of or in connection with the reliance whether whole or partial by any person as
aforesaid upon any part of the contents of this publication. Should expert assistance be required, the services of a competent professional person should be sought.
Page 6 of 6
NDNP Document No: 1
TECHNOLOGY 9.4.4 – QR - 002
DIFFUSION Revision No: Rev 0
ABN 69 003 696 526
ACTIVITY # 27 Page 1 of 2
NATIONAL DIFFUSION NETWORKS PROJECT
TECHNOLOGY QUESTIONNAIRE Date: 18 Nov 2005
Pressure Equipment Industry Group
“Cracking in Fire-tube Boilers”
As part of the WTIA National Diffusion Networks Project the Pressure Equipment Industry Sector has identified the need
to provide guidance, identify and remedy cracking in fire-tube boilers. The WTIA has prepared a Technical Guidance Note
“Cracking in Shell Boilers” to explain the features, control and repair of such cracking. As a valued technology expert in
this area we would like you to be part of the Technology Expert Group to review this note. Please complete this
questionnaire so that we can gauge the success of meeting this need.
Objective 1: Identify the need to understand, control and remedy shell cracking
As the Australian fire-tube boilers become increasingly older, there is an increasing need to understand and control the
degradation mechanisms that can lead to failures of such equipment. This guidance note is intended to provide the
Pressure Equipment Industry understanding of fire-tube boiler cracking so an informed decision can be made on
inspection frequencies and repair strategies. How well does the document explain inspection and repair of shell cracking?
poor average good very good
Comments:
Objective 3: Identify current best practice for inspection and repair of shell boilers
The document was written to reflect current best for service practice for inspection, operation and repair of fire-tube
boilers. Do you envisage opportunities for the use of this practice in the industry?
yes no
If not, why?
Objective 5: Broad dissemination of technology to the Pressure Equipment Industry
Please indicate how best to disseminate this Technical Guidance Note to the appropriate Pressure Industry Recipients
Other format?
The WTIA has joined forces with industry and government to create a 3.5 million dollar Technology Support Centres Network. This network will assist industry to identify and exploit world’s best
technology and manufacturing methods to establish a vibrant Australian industry beyond 2006. Together we will be implementing a step by step process which will lead to ongoing viability and greater
profitability for all concerned:
(1) Determine your technological and manufacturing needs;
(2) Identify world’s best practice;
(3) Draw upon the network to implement world’s best practice at your site