|
| |
|
Return to
Failure Analysis Case Histories
|
Oxygen Corrosion of Carbon Steel
Boiler Tubes
|
|
ENVIRONMENT:
|
Condominium Complex |
|
EQUIPMENT:
|
Hot Water Boilers |
|
MATERIAL: |
Carbon Steel Tubes |
|
SERVICE TIME: |
11 years |
|
FAILURE MODE:
|
Oxygen Corrosion |
|
|
|
Background
|
|
The condominium has two identical horizontally
oriented hot water boilers, connected in parallel, for
providing heat to the residences. As we understand, only one of the
boilers (typically #1, set at 150°F)
is normally used to provide heat; the other boiler (typically #2, set at
120°F) is kept warm to act as a
backup in the event #1 boiler goes out of service. The boilers are
connected in parallel with two chillers, which are themselves connected in
parallel; therefore, the same water circuit is used for both heating and
cooling.
|
|
Nameplate information indicates that each boiler
was manufactured 32 years ago. For each boiler, the maximum working
pressure for use in hot water heating is shown as 100 lbs. Each
boiler is comprised of a cylindrical central firebox surrounded by a
tube-in-shell heat exchanger. Each boiler is fitted with 194 carbon steel tubes.
The tubes are 2.5-inch outside diameter (OD)
0.135-inch wall thickness and
are 13-feet, 10-inches long. The boiler #1 was completely retubed
eleven years ago (#1) while boiler #2 was retubed ten years ago. The boiler tubes are rolled and welded into the
return pass tube sheet (where the failures occur), but are only rolled
into the tube sheet at the other end (where the burner is located). The
boiler tubes are unsupported between the two tube sheets.
|
|
For the past three years, one to three shutdowns
have occurred each season to replace failed tubes. This season, nine
shutdowns have occurred on #1 boiler and four shutdowns on #2 boiler.
Reportedly, the tube failures have been identical: A waterside groove
developed in the failed tubes at the return pass tube sheet, followed by
through-wall cracking of the tubes. All failures have occurred in
the lower quadrant of the heat exchanger, and most failures have typically
occurred between the 10 and 2 o’clock position on the tubes.
|
|
This heating season a heating riser piping
replacement project has been underway which has necessitated draining and
refilling parts of the heating/cooling circuit. In addition, the records
of the firm that handles water treatment indicate that a recirculating
pump was leaking, though the magnitude of the leak was not described.
|
| |
|
Findings
|
|
Examination of Failed Boiler Tubes
|
|
Ten (10), approximately 2-inch long, failed tube
ends were provided to CTL for examination. Also provided was one (1)
approximately 6-inch long tube end from the burner end of boiler (where no
failures had occurred). Each tube end was a partial circumferential
section, part of each tube having been cut away to facilitate removal from
the boiler.
|
|
Each of the failed tube ends displayed the groove
and crack failure features, described above, immediately adjacent to the
˝-inch wide band where the tube had been rolled into the tube sheet, Figure
1. On each tube the groove was approximately 1.5-inches long,
extending only part way around the circumference. In each case, the groove appeared to be
the result of corrosion, mainly due to its irregular surface, rather than
mechanical deformation.
|
|
Six (6) boiler tubes that had failed were available
for examination at the condominium. It was noted that the general
appearance of the tubes fell into two categories: Tubes that had a red,
rusty appearance and tubes that had a black, shiny appearance.
Closer examination of a rusty tube revealed the presence of significant
pitting, accompanied by rust-colored mounds (known as “tubercles”) along
its length, Figure 2. |
|
|
|
 |
 |
|
Figure 1. End of boiler tube showing
typical features of failure: groove and crack adjacent to band where
tube was rolled into tubesheet.
|
Figure 2. Pits and tubercles observed on
failed boiler tube examined at the condominium. Note rusty streaks
oriented at right angles to the tube length, which indicate active
corrosion under stagnant conditions.
|
| |
| Metallography |
|
One of the failed tube ends supplied to CTL was sectioned
longitudinally through the groove and crack for metallographic
examination. The presence of thick black oxide on the groove and
parts of the crack surface, Figure 3, as well as lack of deformation in
the microstructure, confirmed that corrosion was the cause of the
grooving. Deformation of the microstructure at the crack surfaces
indicated that the final failure was by ductile tearing. The microstructure itself consisted
of pearlite in an equiaxed ferrite matrix, typical of low carbon steel, Figure
4. There were no indications of overheating of the tube. |
| |
|
 |
 |
|
Figure 3. Polished metallographic longitudinal cross-section
showing oxide-filled groove and crack. Light-colored material
indicated by white arrows is oxide. Yellow arrow indicates crevice attack on
part of tube rolled into tube sheet. (18X Original Magnification)
|
Figure 4. Microstructure of failed tube showing
pearlite in equiaxed ferrite matrix. (2% nital etch) (125X Original
Magnification)
|
|
|
|
Review of Water Treatment Procedures
|
|
CTL was provided with the records of eight (8)
service calls made to the condominium by the water treatment provider
within the last six months. The following items were noted:
|
|
· A steady drop in nitrite
inhibitor levels from 840 ppm to approximately 300 ppm within 6
monthd.
|
|
· The addition of molybdate
inhibitor to the water treatment regimen, presumably to combat
tuberculation.
|
|
· The notes of a “recirculating pump” leak thought to be responsible for a drop in
nitrite inhibitor levels.
|
|
· The note of
lower-than-expected nitrite levels possibly being the result of “water
loss or oxygen in the system that is ‘eating-up’ the chemical.”
|
|
Chemical Analysis of System Water
|
|
A sample of water was obtained from #1
boiler during CTL’s visit. The sample was analyzed by CTL for dissolved oxygen,
which had a concentration of 5ppm. |
|
|
|
Discussion
|
|
The boiler tube failures
were caused by oxygen corrosion of the tubes produced by dissolved oxygen
in the boiler water. This was based on the rusty appearance of
most of the failed tubes, the presence of pits and tubercles (classic
oxygen corrosion features) along the lengths of some of the failed tubes,
and the thick oxide present on the metallographically prepared failure
site. Oxygen corrosion of the tubes at the failure locations led to the
grooving described earlier. The groove reduced the tube wall thickness
and subsequently acted as a stress-raiser during normal thermal cycling of
the boiler. Stresses from thermal cycling eventually produced the final
failure of the tubes by cracking.
|
|
The reason the failures occurred at the return end
of the boiler was that the tube sheet and tubes were hottest at this end,
which produced localized boiling of the oxygen-laden water. Boiling of
the water produced a scouring effect on protective oxide films, which led
to localized grooving. (It cannot be ruled out that tubes outside this quadrant could
also be damaged, although the corrosion may have been occurring at a
slower rate. It is fairly certain, however, that other boiler tubes,
besides the ones that failed, have suffered pitting and may have the
groove damage.)
|
|
It is possible that occluded cell (crevice)
corrosion played a role in the grooving. In this scenario, a differential
aeration cell is set up between the tube adjacent to the return end tube
sheet and the tube just under the edge of the tube sheet (assuming leakage
of boiler water under the tube sheet.). This cell will lead to corrosion
of the tube just under the tube sheet, which accounts for the observed
grooving.
|
|
Under normal (ideal) operating conditions boiler
water is deaerated (i.e., less than 1 ppm). Under such conditions,
low residual oxygen produces a layer of black iron oxide (magnetite),
which protects steel tubing. Thermal cycling can fracture the
magnetite layer, which exposes underlying bare steel to the boiler water.
In the presence of excessive dissolved oxygen (greater than 2 ppm) in the
boiler water, accelerated corrosion of the steel tubes occurs. Our
analysis shows that the boiler water contained dissolved oxygen at levels
greater than 5 ppm.
|
|
Under the current situation, aeration of the boiler
water occurred through the frequent additions of make-up water to the
system after drain-downs for the riser replacement project and repairs to
the boiler after tube failures, and as a result of the recirculator pump
leakage noted in the water treatment records. The steady drop in
nitrite inhibitor over the last 3 months without a simultaneous rise in
nitrate levels (as indicated by our water tests) provides support for this
assertion. [If no make-up water had been added to the system, nitrate
(oxidized nitrite) levels would be expected to be much higher than
measured.]
|
|
The current water treatment regimen is inadequate
to prevent oxygen corrosion; boiler tube failures will continue to occur
as long as dissolved oxygen is present in the boiler water.
|
|
