Open the calculatorOpen the calculator

Tsai-Wu vs Tsai-Hill: A Worked Example

One quasi-isotropic laminate, one load, two quadratic criteria. They name different first plies. The load envelope further down also runs on three other layups.

Problem

A [0/45/-45/90]s laminate of T300/5208 carbon/epoxy carries Nx = 100 N/mm in tension and Ny = −200 N/mm in compression, with no shear and no bending. Which ply fails first, and at what load? Answer it once with Tsai-Wu and once with Tsai-Hill.

Ply 1 on top, 0.125 mm each, 1 mm in total. The tags mark the plies each criterion says fail first: TW for Tsai-Wu, TH for Tsai-Hill.

Why compare these two

Tsai-Wu and Tsai-Hill are both quadratic interaction criteria. Each folds σ1, σ2 and τ12 into one failure index, both are built from the same ply strengths, and neither tells you whether the fiber or the matrix gave way. On paper they look interchangeable. This example shows a laminate where swapping one for the other moves the first failure to a different ply.

σ1σ2
No interaction
Max stressMax strain

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.

σ1σ2
Quadratic interaction
Tsai-HillTsai-Wu

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.

σ1σ2
Failure modes
HashinPuck

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.

The other criteria check something else. Maximum stress compares each stress with its own strength, with no interaction. Hashin and Puck split failure into fiber and matrix modes. Put one of those next to Tsai-Wu and part of the gap comes from what each criterion checks. Put Tsai-Hill next to Tsai-Wu and the gap comes only from how the quadratic is built: Tsai-Wu has linear terms that use the tensile and compressive strengths together, and Tsai-Hill has none. For maximum stress next to both, see the off-axis strength example, and for all criteria side by side, the failure criteria overview.

Inputs

Layup[0/45/-45/90]s, ply 1 on top
Ply thickness0.125 mm
E1181000 MPa
E210300 MPa
G127170 MPa
ν120.28
S1t / S1c1500 / 1500 MPa
S2t / S2c40 / 246 MPa
S1268 MPa
Nx100 N/mm
Ny−200 N/mm
Nxy0 N/mm

The ply stresses come from Classical Lamination Theory. The free failure criteria calculator takes ply stresses, so each row of the results table has a link that opens it with that ply's stresses and these strengths.

Change the load direction

Each curve is the first-ply failure load of the whole laminate for every ratio of Nx to Ny. Turn the load and see where the two criteria pick a different ply (the shaded sectors).

Pick another layup to redraw both curves for it. The results table below stays on the quasi-isotropic example. The criteria pick different plies only in layups that mix a ±θ ply group with a 0° or 90° group. On [0/90]s they pick the same ply in every direction, but the failure loads still differ: compare |N| in the readout.

On every layup the +θ and −θ plies carry the same stresses under any mix of Nx and Ny, so the readout names ply groups. |N| is the length of the load vector at first-ply failure.

-1200-800-4000400-1200-800-4000400Nx (N/mm)Ny (N/mm)
Tsai-WuTsai-HillExample loadCriteria pick different plies
Layup
[0/45/-45/90]s
[0/90]s
[0/±45]s
[0/±60]s

Direction 296.57° from +Nx: Nx : Ny = 0.447 : −0.894

CriterionFirst ply|N| at failure
Tsai-Wu90° plies343.2 N/mm
Tsai-Hill±45° plies377.7 N/mm

The criteria pick different plies in this direction.

Results

Tsai-Wu says
90° plies fail first
SF = 1.535
Tsai-Hill says
±45° plies fail first
SF = 1.689

The two criteria disagree on where cracking starts, and by 9% on the first-ply failure load. Against the WWFE-I test data (below), Tsai-Wu follows the measured shear strength under transverse compression and Tsai-Hill does not.

Safety factor of each ply under each criterion. The shortest bar is the first ply to fail.

0°3.7453.32145°1.9011.689−45°1.9011.68990°1.5351.877SF = 101234Safety factor
Tsai-WuTsai-Hill

Ply stresses in the fiber axes, in MPa, then the failure index (FI) and safety factor (SF) of each criterion. The stack is symmetric, so plies 5 to 8 repeat the first four.

PliesAngleσ1σ2τ12Tsai-Wu FITsai-Wu SFTsai-Hill FITsai-Hill SF
1, 80°405.9−27.470−0.35033.7450.090653.321Open in calculator
2, 745°−93.31−6.69−40.010.21041.9010.35051.689Open in calculator
3, 6−45°−93.31−6.6940.010.21041.9010.35051.689Open in calculator
4, 590°−592.514.0900.52741.5350.28391.877Open in calculator

Tsai-Wu: the 90° plies fail first, at 1.535 times the load. Tsai-Hill: the ±45° plies fail first, at 1.689 times the load. The stresses and safety factors come from an independent implementation of Classical Lamination Theory. The failure indices come from the free calculator, and its safety factors match the table.

Against test data

Which one is right? The World-Wide Failure Exercise (WWFE-I) published biaxial tests on single plies2. For E-glass/LY556 epoxy under transverse stress σ2 plus shear τ12, moderate transverse compression raises the shear strength from about 61 to 97 MPa. Tsai-Wu bends the same way. Tsai-Hill only falls.

-175-125-75-252575020406080100120σ2 (MPa)τ12 (MPa)
Tsai-WuTsai-HillTest, Soden, Hinton & Kaddour (2002), Table 12

Strengths for both curves are the values given to the exercise's contributors: S2t = 35, S2c = 114 and S12 = 72 MPa (Soden, Hinton & Kaddour 1998, Table 1)1. This is a different material from the example laminate: it is the ply the exercise tested.

What it means

The ±45° plies carry τ12 = 40.01 MPa against a shear strength of 68 MPa. Both criteria give that shear the same weight, (τ12/S12)2 = 0.3462. The difference is everything else.

Across the fibers this ply is six times stronger in compression than in tension. Tsai-Wu sees that through its linear term F2σ2. In the ±45° plies σ2 is compressive, F2σ2 = −0.1401, and the index drops to 0.2104. In the 90° plies σ2 is tensile, F2σ2 = +0.2951, and together with the fiber compression it lifts the index to 0.5274. Tsai-Hill has no linear term. It measures the small compressive σ2 in the ±45° plies against S2c and the tensile σ2 in the 90° plies against S2t, squares both, and the shear in the ±45° plies wins.

So the two quadratic criteria disagree on where cracking starts, and by about 9% on the first-ply failure load. If the tensile and compressive strengths of your ply differ this much, use Tsai-Wu. If you also need to know whether the fibers or the matrix give way, check the plies with Hashin as well.

Related docs:

References

  1. Soden, P.D., Hinton, M.J. & Kaddour, A.S., "Lamina Properties, Lay-Up Configurations and Loading Conditions for a Range of Fibre-Reinforced Composite Laminates", Composites Science and Technology, 58(7), 1011-1022, 1998.
  2. Soden, P.D., Hinton, M.J. & Kaddour, A.S., "Biaxial Test Results for Strength and Deformation of a Range of E-Glass and Carbon Fibre Reinforced Composite Laminates: Failure Exercise Benchmark Data", Composites Science and Technology, 62(12-13), 1489-1514, 2002.
  3. Jones, R.M., Mechanics of Composite Materials, 2nd Ed., Taylor & Francis, 1999.
Open in calculator

Try it on your own ply. The failure criteria calculator opens with the 90° ply's stresses and these strengths.

Open the 90° ply
Written by Rick Schrijver
Share