Sie sind auf Seite 1von 31

SAND2006-4591C

Fatigue Failure of Fasteners: Optical


Metallography and SEM Fractography

Don Susan, Alice Kilgo, Dick Grant, Bonnie McKenzie,


and Gary Zender

Sandia National Laboratories


Albuquerque, NM

39th International Metallographic Society Symposium --


Failure Analysis and Prevention Session

Microscopy and Microanalysis 2006


Chicago, IL
August 2nd, 2006

Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company,


for the United States Department of Energy’s National Nuclear Security Administration
under contract DE-AC04-94AL85000.
Outline
• 4 Case Studies: 3 out of 4 are stainless steel
fasteners, one carbon steel example
1. Free-machining stainless steel bolts, poor
surface finish, wrong heat treat condition,
machined threads
2. Low quality carbon steel, low fatigue strength
3. Annealed stainless steel material used for
fasteners, fastener size too small
4. Bolt strength too low for fatigue application
• Summary
Introduction
Motivation:
• Many high-reliability components designed by
Sandia must withstand fatigue conditions, usually
vibrational fatigue, in service. Common failure
locations during fatigue testing are threaded
fasteners – screws, bolts, tie-rods, etc.

• Many fastener fatigue failures are due to a


combination of factors. Among these are: alloy
selection, mechanical properties/heat treatment,
surface finish, poor microstructure (e.g.
inclusions), and severity of vibration fatigue
environment.

Objective: To determine the cause(s) of failure in


four case studies of fastener fractures.
• Schematic
diagram used
for estimating
fatigue stress
state and
severity.

(ASM Handbook, Volume 11 (2002), “Failure Analysis and Prevention”, 2002,


ASM International, Materials park, OH, pg. 633.)
Case Study #1: Stainless Steel Shipping
Container Bolts
• Macro-appearance of stainless
Rough fracture
steel “eyelet” bolts
• Poor thread surface
finish (chatter)
“Flat fracture”

4 mm
• Fracture appearance
indicative of fatigue
~6mm failure
(0.25 in.)

• Shipping container bolts, vibration tested – 3 hours in each of 2 orientations


• A bolt fractured and several were bent during re-tightening in preparation for a third test sequence
• SEM fractography
• Metallography: 1. new (off shelf) bolts, 2. failed bolts, 3. tested/not failed, bolts
Thread surface
Dark particles – Mn(Cr)Sulfide
inclusions

Close-up
Fracture surface of thread damage
and fracture

• Fatigue crack initiation on


tooling marks/poor surface
finish on threads
SEM Fractography

Rough fracture

“Flat fracture”

• Ductile overload
4 mm region showed
“woody” appearance,
elongated
dimple rupture

Sulfide particles

• Fatigue fracture – no clear fatigue striations


Metallography showed high sulfide inclusion content
of free-machining (AISI 303) stainless steel (as-received bolt)

200x
200x
10% oxalic acid etch,
Longitudinal and transverse electrolytic
sections, As-polished

Thread surface

100x 400x
Microstructure of Broken Bolts

• Machined threads
50X

200X

• Transgranular, secondary fatigue cracks in thread roots of broken bolt


(oxalic acid etch), 200x
• Some cracks opened wide due to orientation during final tearing
Other Tested (but not broken) Bolts

• Several secondary, sub critical fatigue cracks found in thread roots


• 7 secondary cracks found in 5 out of 11 thread roots
Poor Thread Characteristics

• Significant thread machining damage also found


• Surface tearing, subsurface deformation
Case Study #1: Summary
• Microstructure
– Grain size ~25 microns, no microstructural differences between failed bolt and
other bolts analyzed
– Numerous secondary fatigue cracks indicate severe conditions – either
fatigue stress much too high, or material fatigue strength much too low
• Hardness/Strength
– 87.7 ± 0.3 Rockwell B, (~86 ksi tensile strength), is typical for AISI 303
stainless steel in the annealed condition. Material was specified to be “not
annealed”.

• Conclusions
– Improper material condition is the major cause of failure – fatigue strength too
low. As comparison, 304SS in annealed condition has fatigue strength of ~41
ksi (106 cycles), in full-hard condition the fatigue strength is ~100 ksi
(Aerospace Structural Materials Handbook, Code 1303, 1973)
– Poor thread surface finish (easy initiation sites) is a secondary factor, but not
the primary cause of failure

Recommendations:
1. Use AISI 303 or 304 in the high strength, cold-worked condition,
2. Use higher strength stainless steel alloy (e.g. Nitronic 60, 15-5PH, or 13-8PH)
3. Change machining practice to eliminate severe tool chatter damage
4. Inspect bolts prior to installation to ensure good surface finish
5. Perform hardness and/or tensile tests prior to installation
Case Study #2: Threaded Tie Rods

200x, 2% nital etch

• 4 connecting rods (control fin movements


on a projectile) failed in vibration tests
• Failure in 1st free thread -- 2 fatigue cracks
and final overload failure in center: “reverse
bending” fatigue
• Microstructure is ferrite with some pearlite
and high amount of large elongated
inclusions – fairly low quality steel with high
impurity content
• Threads are ‘rolled in’ in this case
• Zinc coated
500x
SEM Fractography – Final Overload Failure

• Some regions of
brittle/intergranular
overload
fatigue appearance,
fatigue indicating “dirty”
low toughness
steel

• Some
secondary
cracking/tearing
and poorly
formed dimples,
indicating low
quality steel

• Final overload failure shows


mainly ductile microvoid coalescence
SEM Fractography – Fatigue Regions

overload
fatigue fatigue

• Evidence of fatigue striations


conclusively shows fatigue
as failure mechanism
Case Study #2: Summary
• Metallography showed the threaded rods
were made of hot-rolled, relatively low quality
steel
• Ferrite/pearlite (not quench and tempered)
• Hardness: 90-92 HRB, implied an unnotched
tensile strength of 88-94 ksi
• SAE Grade 1, the lowest strength level
fastener (2800 lb. breaking load was
reported, corresponds to ~81 ksi)
• NTS/UTS ratio < 1, indicates low
toughness
• Reversed bending fatigue, some striations
found on fracture surface

Recommendations: 1) Switch to higher


strength, Grade 8 fastener material (~150 ksi
UTS). 2) However, design changes also needed
such as increase diameter or reduce the
• Effect of UTS on fatigue limit threaded length, because fatigue endurance
of steel: endurance limit increases with limit increases comparatively little for notched
increase in strength for smooth samples, samples. 3) Eliminate resonances
but only slightly for notched samples
(R.W. Hertzberg, Deformation and fracture
Mechanics of Engineering Materials, 1976, pg. 429.)
Case Study #3: S.S. Screw from Radar System
2 fatigue cracks growing
toward each other • 3 small “No. 6”
annealed 304 stainless
steel screws failed in a
radar system
development test

200 m
Rubbing
damage on
• Fatigue fracture fracture surface
in reverse
bending, two
fatigue fractures
initiated on either
side (in thread
root), final shear
fracture in center

20 m
• Ductile, final overload in center
Damaged/smeared
surface

Fatigue striations

20 m

• Relatively soft austenitic


stainless steel shows good fatigue
striation evidence
Case Study #3: Summary

• Fracture surface predominantly fatigue with clearly defined striations;


Final fracture in center showed ductile dimple rupture

• Microhardness ~ 198 Vickers (500g), consistent with annealed 304

• Striation spacing ~ 1 micron;


Some areas showed much larger spacing, presumably corresponding to
continued high stress fatigue after one or more screws failed.

• Number of cycles to failure ~3000 to 10,000, difficult to determine better


than order of magnitude estimation

Recommendations: Larger (No. 10) and/or higher strength screws should be


used. Customer proposed a change to heat treated carbon steel with
cadmium plate. This would be acceptable, but embrittlement concerns with
Cd plate make high strength stainless steel (precipitation hardened) screws a
better choice.
Case Study #4: Stainless Steel Bolts, Attached a Diagnostics
Kit to a Geothermal Drill

• Original Grade 8 steel bolts failed with brittle fracture at


head/shank interface (1st thread)
• Replaced with Grade 5, 5/16” 304 stainless steel bolts
• Difficulty to torque to specified range, possible damage
during initial tightening
• 14 screws total, several failed in fatigue
• Analyzed broken screws and new (off the shelf) screws

1 14 2 9

6 mm
oxalic acid etch,
Nomarski DIC, 200x

• 304 SS microstructure appeared normal,


good quality bolts with rolled threads
• Hardness: ~24 HRC, 25 HRC/120 ksi UTS min specified for Grade 5 bolts
• Corresponds to “1/4 hard” 304 stainless
SEM Fractography: fracture surfaces showed well-defined
“beachmarks” of non-uniform spacing, as well as and fatigue striations

30 m

Bolt 1 , near origin


• Attempts were made to measure and count the fatigue striations
• Estimate of # propagation cycles (can’t determine # cycles for crack initiation)

Fracture surface

Thread surface
~110 microns from origin
~625 microns from origin
~1.125 mm from origin
~1.625 mm from origin, near fatigue crack tip
1.6

approx. striation spacing (microns)


1.4

1.2

0.8

0.6

0.4
bolt #1
0.2 bolt #9

0
0 500 1000 1500
distance from origin (microns)

Examples of striation spacing data along the fatigue crack length

• Increase in striation spacing is typical for fatigue crack propagation


• Estimate of # propagation cycles ~1000 to 3000, depending upon
the fatigue crack length on each bolt (low cycle/high stress fatigue)
• Differences among bolts could be due to “load shedding”, i.e. as some
bolts fail, the stress intensity increases on the remaining bolts
Case Study #4: Summary

Recommendations:
• Larger bolt diameter is recommended but not possible
due to space restrictions.
• Therefore, the use of high strength stainless is
recommended (17-4PH, 15-5PH, 13-8PH, etc.) – improved
fatigue resistance with good corrosion resistance/better
surface finish than Grade 8 high strength steel.
• High strength bolts with UTS 150-180 ksi should keep
applied stresses below ~0.5 UTS.
– 304 stainless (1/4 hard): YS ~ 75 ksi, UTS ~ 120 ksi
– hypothetical ¼ inch bolts, load capacity is ~3600 lbs. for
yielding, ~5900 lbs. for failure
– Switch to high-strength stainless, H1025 condition: YS ~
175 ksi, UTS ~ 185 ksi
– For ¼” bolt, load capacity is ~8500 lbs. for yielding, ~9000
lbs for fracture. So increase capacity by about a factor of 2.
Summary

• Metallography useful for: 1) determining manufacturing method of


fasteners (e.g. machined vs. rolled threads), 2) overall quality and
cleanliness of the material, 3) finding sub critical/incipient fatigue
cracks.

• SEM fractography needed to definitively determine the failure


mechanism. In some cases (depends mainly on the material),
beachmarks and/or fatigue striations can be used to estimate the
fatigue lifetime and the failure scenario. Fatigue striations more
visible in lower strength, austenitic alloy fasteners.

• High strength PH stainless steel fasteners should be specified in


applications where fatigue resistance and corrosion resistance
are both required. Other recommendations should consider
fastener size and “design for fatigue”.
Acknowledgements

Thanks to Garry Bryant and Robert Wright for


additional metallography and hardness testing.
Thanks also to Dr. Jim VanDenAvyle for careful
review and Dr. Ken Eckelmeyer helpful
discussion.

Das könnte Ihnen auch gefallen