Fatigue Reassessment of a Vibrating Screen Bracket
An existing 6061-T6 aluminium bracket supports a vibrating screen motor. The load cycles between 0.4 and 5.8 kN at 340 mm from the wall for a design life of 2.1 million cycles. The bracket is a 12 mm plate, 80 mm deep at the wall, with a 6 mm fillet at a 90° step. A static FEA using nominal properties gives a peak von Mises stress of 168 MPa at the fillet, below the 276 MPa yield strength, and the reviewer approved it with a safety factor of 1.64. The fabricator hand-finishes the fillet; dye penetrant testing shows no indications. Material data: ultimate strength 310 MPa, fully reversed fatigue strength at 5×10^8 cycles 96 MPa. The notch sensitivity for the fillet radius is 0.82 and the elastic stress concentration factor is 2.35. Nominal bending stress at the net section, away from the fillet, ranges from 7 MPa to 95 MPa over the cycle. The client now requests a 30% increase in motor load with no change to the bracket geometry or material.
1. Was the static FEA approval sufficient? Calculate Kf and the fatigue-equivalent stress, and state whether infinite life is demonstrated at the current load.
Read the worked answer
Static approval was not sufficient. The 168 MPa FEA peak is below yield, but it is also lower than Kt × nominal stress = 2.35 × 95 = 223 MPa. That discrepancy suggests the fillet mesh is too coarse, the nominal section is misdefined, or the peak is not captured correctly. The static safety factor may therefore be overstated before fatigue is even considered.
For fatigue, Kf = 1 + q(Kt − 1) = 1 + 0.82 × 1.35 = 2.11.
Nominal alternating stress = (95 − 7)/2 = 44 MPa. Nominal mean stress = (95 + 7)/2 = 51 MPa.
Goodman equivalent alternating stress: σa_eq = 44 / (1 − 51/310) = 52.3 MPa.
Apply the fatigue notch factor: 2.11 × 52.3 = 110 MPa.
That is above the 96 MPa endurance limit. Infinite life is not demonstrated. The bracket is in finite life, and the FEA does not give a reliable fatigue clearance. The dye penetrant result only shows no existing crack; it says nothing about initiation life.
2. Re-evaluate for the requested 30% load increase using the same nominal stress range scaled by 1.3. What is your recommendation?
Read the worked answer
Scale the nominal stresses by 1.3: maximum 95 × 1.3 = 123.5 MPa, minimum 7 × 1.3 = 9.1 MPa.
Alternating stress = (123.5 − 9.1)/2 = 57.2 MPa. Mean stress = (123.5 + 9.1)/2 = 66.3 MPa.
Goodman equivalent: σa_eq = 57.2 / (1 − 66.3/310) = 72.8 MPa.
Apply Kf = 2.11: 2.11 × 72.8 = 154 MPa.
This is far above the 96 MPa endurance limit, and the increase is substantial, not marginal. The local static stress also rises: 123.5 × 2.11 = 261 MPa, only about 6% below yield, leaving essentially no static margin at the notch.
My recommendation is to reject the 30% load increase on the current geometry and material. A new fatigue assessment with a refined fillet model is needed, but the result is unlikely to reverse the conclusion without a design change. The client should be told that the bracket cannot be approved for the higher load as-is, and that a redesign or load reduction is required.
3. What design and verification changes would you require before accepting any load increase?
Read the worked answer
The dominant variable is the notch. Priority changes:
- Increase the fillet radius or replace the sharp step with a tapered transition to reduce Kt.
- Increase plate thickness or add a gusset to lower nominal bending stress.
- Consider a different alloy or steel if weight allows; 6061-T6 has a modest fatigue strength relative to its yield strength.
- Specify shot peening or a controlled surface finish only after the geometry is improved; surface treatment does not compensate for a poor notch.
Verification should include:
- A refined submodel of the fillet with mesh convergence and a documented nominal stress extraction.
- A fatigue calculation using the actual load spectrum, mean stress correction, and appropriate factors on stress and life.
- Strain-gauge measurement at the fillet or a representative test bracket if the calculated margin is close.
- A dye-penetrant or eddy-current inspection after a fatigue test, not only before service.
I would not accept a desktop re-run alone. The earlier FEA missed the notch severity, so the revised analysis needs independent review and physical verification before any load increase is approved.