Crack growth problem

1
Crack growth problem in a riveted lap-joint panel
Paulo F P de Matos, Pedro M G P Moreira, Paulo M
S T de Castro IDMEC, University of Porto,
Portugal
2
Objectives

• Crack growth prediction in complex structures
• 2) Compare numerical/analytical results with the
• experimental measurements.

3
Specimen geometry
Rivet type MS20426D5-6 in Al
2017-T31
Specimen BJ3, UPisa
Load conditions
Al 2024-T351
smax110 MPa R0.1
4
Lap-joint mesh
Lap-joint mesh / layer1 and layer 2
Riveted lap joint mesh, where triangles represent
rivets
5
Lap-joint panel stress analysis

• Boundary conditions
• - Top edged of layer 1 fixed along
• X and Y directions.
• The bottom edge of layer 2 was
• fixed along X and a constant
• tensile stress of 110 MPa was
• applied.

y
x
6
Crack tips location and crack shape
Inspected sites for crack initiation
Crack shape
7
Crack scenario at 137950 cycles
Stress sy at 137950 cycles near crack tips 1 to 4
Crack scenario at 137950 cycles
8
Algorithm for crack propagation
C7.46E-8 m3.333
9
Comparison of results
Continuous lines represent the FRANC2D/L
results Bullets concern to experimental results
10
Comparison of results
Continuous lines represent the FRANC2D/L
results Bullets concern to experimental results
11
Comparison of results
Continuous lines represent the FRANC2D/L
results Bullets concern to experimental results
12
Final remarks

• The main conclusions of the present work are
• 1) Crack growth modeling was done in a lap-joint
  panel from 132150 to 138450 cycles
• 2) A good agreement was found between the
  experimentation and the prediction of FRANC2D/L
  based on the Paris Law.
• Although the good results obtained there are
  important
• limitations that must be mentioned
• 1) As previously mentioned cracks grow with an
  elliptical shape and not as through cracks

13
Final remarks
2) The rivets are treated as spring elements. It
is only a simplification that has some
limitations. The typical stress distribution
of a hole is not considered 3) FRANC2D/L
allows only a 2D model of a joint, so the
secondary bending expected in a typical lap joint
cannot be modelled.