Industry:
Synthetic Textiles
Polyester Fiber Processing
Problem:
Fasteners holding drums that stretch the fibers to gain strength
were
breaking. When all fasteners holding a drum broke, the drum would be thrown
to the floor and production would be stopped.
Injuries:
None, but this is because they've had very good luck.
Damages:
The exterior surfaces of the drums are highly polished. When a drum
is
thrown onto the concrete floor they sustain significant surface damage
whereby they then must be repaired or replaced. When notified of the problem,
more than one drum per week was coming off at a replacement
cost of more
than US$10,000 each.
Observations:
The breaking fasteners were 1 1/4" - 12 x 6 1/4"
socket cap
screws. Failures were occurring at the head to body juncture and in the
threads. Wear marks were directly adjacent to the fillet radius.
Investigation:
There are 23 drums on each processing machine. They're positioned
in a
zigzag fashion — one high and the next one low. Each drum is approximately
2' (0.61m) in diameter and 4' (1.22m) long. There were 12 fiber "ropes"
evenly spaced across the drums, and were each about the width of a small
finger. These polyester fibers streamed through the drums in a serpentine
manner. Each proceeding drum turns a little faster than the one before
it, thus stretching the fibers to maximize their strength. The first drum
rotates at about 250 rpm and then progresses until the last one is turning
at 400 rpm. As the fiber is getting stronger, bending forces increase
on the drums. Each drum is fastened to its individual motor roll shaft
by 16 of the socket screws. They're torqued through the open end of the
drum by an elaborate, custom extension using a torque multiplier. While
the method was sound, the assembly of these large screws in an "as
received" condition may not have resulted in consistently developed
clamp force.
There's a hardened backing ring with 16 holes under the screws'
heads.
The top of the ring is countersunk to protect the fillet radii. Examination
of the ring showed some of the tapped holes weren't perpendicular. Lab
tests on the screws showed proper core hardness and a well-tempered martensitic
structure. A scanning microscope (SEM) view of the fracture
surfaces showed
ductile fatigue failures on all.
Conclusion:
The cyclical bending stresses placed on the screws were too much
for standard
fasteners to absorb. The poor machining practices added to the screws'
demise.
Recommendations:
1. Keep the countersinks concentric with the holes.
2. Drill and
tap the
roll shaft holes perpendicular.
3. Change to an UNJ thread
style and roll
the threads after heat treatment.
4. Undersize the body
and build in a
much larger filler radius, and then roll the fillet after
heat treatment.
5. Use a non-metallic anti-seize compound
on the threads to reduce variation
in clamp force.
6. Use a thread-locking patch to minimize
vibrational
loosening.
7. In the future, if needed, change to alloy
steel that has
nickel content for added toughness.
As appeared in Fastener Technology
International magazine. |
