Fatigue Failure Through BendingDavid Burnette

ME 498

Overview

• Objectives of experiment

• Importance and theory

• Experimental details

• Result

• Conclusions and recommendations

Objectives

• To become familiar with fatigue testing procedures

• Develop fatigue data for AA 6061-T6 specimens

• Extrapolate the endurance limit from the S-N curve (at 5x10^8 cycles)

• Compare estimated endurance limit and cycles to failure to known

• Evaluate the surface characteristics of fatigue failure

What is Fatigue?

Crack Propagation

Examples of Fatigue Factors

• Size, loading types• Stress concentration factors• temperature, corrosion

125.7 mm

Applied load + self loading weight

Bearing

Cantilever arm

ø5.00

Set screw b = 1257. P r

I (N/mm2)

I = d4

64

N

aapplied b

1

a ultimate

endurance

0 81 2.

endurance

ultimate9.0log

X

1b

]Nlog[2

1X reference

Desired Stress Stress

(MPa)

Required Load

(g)

.9 247.5 2337.7

.8 220 2190

.7 192.5 1916

.6 165 1641

Testing Procedures - Application of Stresses

y

y

y

y

Test Specimens – Cycles to Failure Comparison

Aluminum Alloys:Nearly pure (>95%), precipitation hardening, tempering, lack of carbon

Experimental Setup

1

2

3

4

DC

B A

Cantilever Arm

6061-T6 specimen

Motor

LOAD Bending Stress

Results

• Endurance limit for 6061-T6 alloy at 5x108

• Predicted Cycles to Failure v. Observed

• Fracture Surface

Results - Chauvenet

N d/σ5 1.656 1.737 1.818 1.869 1.9110 1.9612 2.0414 2.1016 2.1518 2.2020 2.24

1 data point removed with Chauvenet’s:

D=σ *(d/σ)

Stress Allowed Deviation (cycles)

.9 46659

.8 142573

.7 331482

.6 1367672

y

y

y

y

y

Load (kg) Stress (MPa) Mean Cycles Predicted % Difference 2.34 247.5 45423 67900 33.1 2.19 220 124550 203000 38.6 1.92 192.5 389750 702000 44.5 1.64 165 1269375 2940000 56.8

Results - Predicted Cycles

Causes of Error

eccentricity (set screw), yield strength, diameter, number of points (12)

Fatigue Failure for AA 6061-T6

y = -0.1147x + 8.927

R2 = 0.9998

8.10

8.15

8.20

8.25

8.30

8.35

8.40

8.45

2.0000 3.0000 4.0000 5.0000 6.0000 7.0000

Log (Cycles to Failure)

Log

(A

ppli

ed s

tres

s)

σe = 85 MPa

Results - Diameter Uncertainty

I = d4

64

Radius

(mm)

Endurance limit(at 5x10^8 cycles)

Difference from published values

2.50 85 MPa 10.5%

2.45 89 MPa 6.3%

Results - Surface conditions

• Fatigue failure versus dynamic failure

• Crack Propagation

Conclusions• Fatigue failure is very different than static Fatigue failure is very different than static or dynamic failuresor dynamic failures

• A small change in diameter can A small change in diameter can significantly increase the stresssignificantly increase the stress

• Wide range of deviations (Factor of Safety)Wide range of deviations (Factor of Safety)

• Difference of only 10.5% with eccentricity Difference of only 10.5% with eccentricity (human error), diameter uncertainty, alloy (human error), diameter uncertainty, alloy uncertainty, etcuncertainty, etc

Recommendations

• Replace set screws with chuck or threaded specimens

• Increase size of aluminum specimens (fewer points)