|
| |
|
Return to Failure Analysis Case Histories
|
Corrosion Fatigue of Copper Tubes in a Subcooler
|
|
ENVIRONMENT:
|
Refrigerant and water |
|
EQUIPMENT:
|
Condenser |
|
MATERIAL: |
Copper |
|
SERVICE TIME: |
5 to 6 months |
|
FAILURE MODE:
|
Corrosion Fatigue |
|
|
|
Summary
|
|
Longitudinal
corrosion fatigue cracks developed in preexisting defects (notches) on the outer
surface of the tube near the lands. The corrosion fatigue cracks served as
precursors for the circumferential fatigue cracks that lead to the ultimate
failure. Corrosive chemical species in the refrigerant on the outside of the
tube contributed to the failure.
|
| |
|
Background |
|
The submitted tube was removed
from the condenser/subcooler in the chiller unit that had been in service for
five to six months. The refrigerant was circulated on the shell side
of the tube and cooling tower water was on the tube side.
|
| |
|
Description of Material |
|
CTL received seven sections of
pipe, Figure 1. In service, four lands about 2-inch long supported the
tube and occurred about every 45-inch. A flat section on the bottom
characterized the lands and top of the tube and two indentation marks 2-inch
apart on each side of the land. The overall length of the tube was about
16-ft. |
|
|
|
 |
|
Figure 1.
The relative location of the lands and tube sheet with respect to the
copper tubing is revealed. |
|
| A crack
occurred toward the end of the second land; the final separation of the
crack tube occurred during the removal process and was characterized by wall
thinning and a ductile tear. It was assumed that this tearing occurred
as the tube was being extracted. The side of the tube had a circumferential
thick crack. |
|
| Findings |
|
The outside of the tube was of a uniform new copper
color. The subject fracture occurred on the second land, Figure 1.
An examination of the fracture under a stereoscopic microscope revealed that
through-wall longitudinal cracks were present, Figure 2. In this area it
was not possible to determine which cracks occurred first or their origin.
|
|
Fine longitudinal cracks were found on the third
land on the bottom side of the tube. Circumferential cracks could be found
originating from some of the longitudinal cracks.
|
|
At higher magnifications, shallow
pits a few mils in depth could be found on the outside surface of the tube in
the same area as the longitudinal cracks, shown in Figure 3.
|
|
A cross section
of the copper tubing was taken near Land #3 in the same area that the
longitudinal cracks were found. Small defects in the outer surface of the
copper tube were detected that have the general form of v-shaped notches.
Cracks initiated at the bottom of many of these notches.
|
| One of the
cross sections with longitudinal cracks discussed above was etched.
The longitudinal cracks contained a large amount of oxides, Figure 4. |
|
|
 |
 |
|
Figure 2. Fracture
surface with the stereoscopic microscope. Note the longitudinal
cracks. (11X original magnification) |
Figure 3. Shallow pits near the longitudinal cracks on Land #3
(125X original magnification) |
|
|
|
 |
|
Figure 4. Etched crack section.
Crack full of oxide corrosion products. (1250 X Original
Magnification) |
| |
|
DISCUSSION
|
|
Several modes of failure were in
progress simultaneously to reduce the life of this tube. The mode that
resulted in the failure of the tube was a fatigue crack that ran, for the most
part, circumferentially around the tube. It was assumed that the ductile
part of this fracture occurred as the tube was being removed.
|
|
The nature of the longitudinal
cracks supports corrosion accelerated by fatigue, or corrosion fatigue.
The small mechanical defects, assumed to be present from the drawing of the
tube, served as the nucleation site of the corrosion/fatigue. The oxide
corrosion product in these cracks and the branching, transgranular structure of
the cracks provided further support that corrosion was active. The pits,
although not directly implicated in the failure, provided additional evidence
that corrosive conditions existed on the outside of the tube. Ultimately,
the longitudinal cracks became the stress concentrators for the circumferential
fatigue crack. The fatigue crack started in one of the branches of the
longitudinal cracks in the very high stress area next to the land. Once
the fatigue crack started, failure was rapid in the tube already weakened by the
longitudinal cracks.
|
|
Chemical analysis both inside and
outside of the tube provided additional confirmation that the cracks originated
on the outside of the tube. The particles found in the tube contained some
chlorides and chemicals normally associated with treated cooling water (Si, P,
Ca, Al). In contrast, the corrosion products in the crack only contained
aluminum, oxygen, copper, and sulfur.
|
|
