Illustrative: Weld Lack-of-Fusion Defect Leading to Fracture
How a lack-of-fusion defect in a butt weld acted as a crack-like flaw, reducing fatigue and fracture capacity below the design assumption.
Case type
Illustrative engineering case. This case demonstrates how a common weld defect — lack of fusion — acts as a pre-existing crack, reducing both fatigue and fracture capacity.
Illustrative case. The mechanism — a lack-of-fusion defect acting as a crack-like flaw — is well documented in welding and fracture mechanics literature.
1. The system or structure
A full-penetration butt weld in a 20 mm thick steel plate. The weld was fabricated by submerged arc welding (SAW). The structure carries cyclic axial loading.
2. What failed?
A crack was detected at the weld centreline during ultrasonic inspection after 80,000 load cycles. The crack was approximately 25 mm long and had propagated from a lack-of-fusion zone at the weld root.
3. Operating and load environment
Cyclic axial loading with a nominal stress range of 80 MPa and R = 0.1. The design fatigue life for a full-penetration butt weld in this detail class was 500,000 cycles.
4. Failure location
The crack initiated at the weld root, at the location of a lack-of-fusion defect approximately 8 mm long and 2 mm deep. The defect was planar, oriented perpendicular to the applied stress.
5. Physical failure mechanism
A lack-of-fusion defect is effectively a pre-existing crack. It eliminates the crack initiation phase of fatigue life and reduces the total life to the propagation phase only.
- Normal weld fatigue life: initiation + propagation. A sound butt weld in this detail class would be expected to survive 500,000+ cycles
- With lack-of-fusion: the defect acts as an initial crack of approximately 2 mm depth. The entire life is crack propagation
- Paris law prediction: da/dN = C(ΔK)^m. With ΔK = Δσ × Y × √(πa) = 80 × 1.12 × √(π × 0.002) ≈ 7.1 MPa√m, and C ≈ 1.5×10⁻¹², m ≈ 3 for structural steel
- Predicted propagation life from 2 mm to 25 mm: approximately 70,000 cycles — consistent with the observed 80,000 cycles
6. Why did it happen?
- Physical cause: fatigue crack propagation from a pre-existing lack-of-fusion defect at the weld root
- Contributing factor: the welding procedure was not correctly qualified or implemented, producing lack of fusion at the root
- Contributing factor: the radiographic inspection did not detect the defect — planar defects oriented perpendicular to the X-ray beam can be difficult to detect
- Contributing factor: the fatigue assessment assumed a sound weld — no defect tolerance was included
9. Engineering lessons
- Lack-of-fusion defects are crack-like flaws that eliminate the fatigue initiation period — life is propagation-only
- Inspection method must be selected for the expected defect type — ultrasonic testing is better for planar defects than radiography
- Welding procedures must be qualified and implemented correctly — a sound weld is a prerequisite for achieving the design fatigue life
- For critical welds, a fracture mechanics assessment with an assumed initial flaw size provides defect tolerance