Edge Failure Criterion
The Edge criterion, the stress-based Grant-Sanders method, for a composite ply: five per-mode checks, the two interactions, the safety factor, and one ply worked by hand.
What the Edge criterion predicts
The Edge criterion is the stress-based Grant-Sanders method, an industrial design method from British Aerospace that took part in the World-Wide Failure Exercise1,2. It checks failure with separate conditions per mode, and only two of them carry any interaction. Shear interacts with transverse tension in matrix failure, and with fiber compression in fiber failure.
Because the checks are discrete, the failure envelope is a polygon with sharp corners rather than a smooth curve, and every corner is a change of governing mode. Each prediction names its mode, which was a design goal of the method.
Edge belongs to the failure-mode family. The cards draw each family's failure envelope for the same ply in the σ1-σ2 plane, with no shear. Edge runs in the dashboard only, so its card names it without a curve.
Each stress, or each strain, is checked against its own limit. Max stress draws a box with sharp corners. Max strain draws a slanted box, because the Poisson effect puts some σ1 into ε2.
All stresses go into one quadratic expression. The envelope is one smooth, ellipse-like curve, and it does not say whether the fiber or the matrix failed.
Fiber and matrix failure get separate checks. The envelope is pieced together from one curve or line per mode.
Also in this family, not drawn: Cuntze, Sun, Edge.
When to use Edge
Use it when plies carry shear with fiber compression, which is where its linear interaction cuts the allowable most. It also suits a design check that has to name one mode per ply. For smooth interaction across all stress states, run Tsai-Wu next to it.
Edge formula
The ply fails when any of the five conditions holds1:
The linear shear-compression interaction reflects that shear softens the matrix that supports the fibers. The compressive failure level drops, while the mechanism stays fiber-dominated.
Edge gives the shear-compression interaction its own allowables, and his own work used 1.5 × S12 for the shear term1. The dashboard uses S12, which makes that interaction about 33% more severe in shear. That is the more conservative reading, chosen to agree with independent implementations.
Failure index and safety factor
Four of the five conditions are linear in the load, so their safety factor is 1/FI. Matrix tension is quadratic, so its safety factor is 1/√FI. The ply's safety factor is the lowest of the five, and that condition names the mode. The overview covers safety factors and margins for all criteria.
Worked example: one ply
Take the +45° ply of the quasi-isotropic laminate example. It is T300/5208 carbon/epoxy, with S1c = 1500 MPa, S2c = 246 MPa and S12 = 68 MPa. Its stresses are σ1 = −93.31 MPa, σ2 = −6.69 MPa and τ12 = −40.01 MPa. Both normal stresses are compressive, so the fiber compression, shear and matrix compression conditions apply.
| Fiber compression, −σ1/S1c + |τ12|/S12 | 0.6506 |
| Shear, |τ12|/S12 | 0.5884 |
| Matrix compression, −σ2/S2c | 0.0272 |
The fiber ratio alone is only 0.06221, yet fiber compression governs, because the shear term adds to it. FI = 0.6506 and SF = 1/FI = 1.537. That is lower than the maximum stress safety factor of 1.7, where shear governs alone.
Limitations
- With only two interactions, stress states away from those two corners are treated as independent checks.
- The World-Wide Failure Exercise follow-up found that the shear-compression interaction overpredicts for glass/epoxy laminates2.
- It predicts first-ply failure only. What the laminate carries after that is a progressive failure question.
Beyond the Live Calculator
The free calculator does not run Edge. The app runs it with eight other criteria on every ply of a laminate, with first ply failure and progressive failure.
Create free accountFrequently asked questions
References
- Edge, E.C., "Stress-Based Grant-Sanders Method for Predicting Failure of Composite Laminates", Composites Science and Technology, Vol. 58, 1998, pp. 1033-1041. doi:10.1016/S0266-3538(96)00139-X
- Edge, E.C., "A Comparison of Theory and Experiment for the Stress-Based Grant-Sanders Method", Composites Science and Technology, Vol. 62, 2002, pp. 1571-1589.