3.6.16 · D5Spacecraft Structures & Systems Engineering
Question bank — Classical laminate theory — ABD matrix
Before the drill, we build every symbol the questions use, and anchor each to a picture, so nothing here appears un-earned.
The picture behind the symbols
A laminate is a stack of thin plies. We slice it with a vertical coordinate , measured from the mid-plane (the surface at ), positive pointing "up". Look at Figure 1: each ply occupies a slab between two heights.
Recall The reduced stiffness
in one breath is the transformed reduced stiffness of ply — the law inside that ply. You get it by taking the ply's own-axis stiffness (built from ) and rotating it by the fiber angle (see Reduced Stiffness Matrix Q and Transformation of Stiffness). Its form: . Two plies of the same material but different angle have different .
Figure 2 shows how the three sub-matrices arise from integrating against different powers of — this is the geometric root of the scaling laws the questions probe.
Figure 3 shows why mirror symmetry forces : the signed -lever contributions of mirror-paired plies cancel.
True or false — justify
is guaranteed by any laminate that has the same plies top and bottom.
False — you need mirror symmetry about the mid-plane (same material, angle, AND same distance , as in Fig. 3). A stack like [0/90/0/90] has matching plies but they are not mirror-placed, so .
A symmetric laminate can still bend when you pull on it.
False — for a symmetric laminate, in Eq. (4), so a pure in-plane force produces only mid-plane strain and zero curvature . Bend-stretch coupling is exactly the thing carries, and it has vanished.
depends on the order in which you stack the plies.
False — Eq. (1) weights each ply only by its thickness , not its position. Reshuffling the same plies leaves identical; only and feel the order.
depends on stacking order.
True — Eq. (3) weights each ply by , which grows fast with distance from the mid-plane, so plies at the surface dominate. Moving a stiff ply outward raises .
A balanced laminate always has .
False — "balanced" (equal numbers of and plies) kills and , i.e. shear-extension coupling in . It says nothing about mid-plane symmetry, so can still be non-zero.
The matrix is always symmetric.
True — because appears in both the force row and the moment row of Eq. (4), the full is symmetric (). This is a consequence of a single strain energy governing both, so mixed partials are equal.
If two plies have the same fiber material they have the same .
False — is the transformed reduced stiffness in laminate axes, so a ply's angle changes it. A 0° and a 90° ply of identical material have different (see Transformation of Stiffness).
Adding a thin ply far from the mid-plane changes more than adding it near the mid-plane.
True — the weighting in Eq. (3) means a ply at large contributes hugely to bending stiffness, while one at barely moves (its is tiny).
Spot the error
"The laminate is symmetric, so ."
The claim confuses which matrix vanishes. Symmetry gives , not ; by Eq. (1) is a sum of positive-definite terms and is never zero for real plies.
" has units of stiffness, same as ."
Wrong units. is Pa·m, is Pa·m (it multiplies curvature in 1/m to give force/width ). You cannot add or compare and entries directly.
"To get strains, just multiply the loads by the matrix."
Eq. (4) reads , so to go from loads to strains you must invert: multiply by .
"For a symmetric laminate under only, I still need the full inverse."
When Eq. (4) is block-diagonal, so and decouple. With you only invert the .
" vanishes for every ply, so is always zero."
That integral only vanishes for a ply straddling the mid-plane symmetrically; for a ply lying entirely on one side, . in Eq. (2) is a sum over plies, and it vanishes only when those signed contributions cancel.
"Curvature and strain have the same units, so and are interchangeable."
No — is dimensionless while has units 1/m. That difference is exactly why (Pa·m) and (Pa·m) differ by two powers of length.
"The [0/90] two-ply laminate is symmetric because it has one 0 and one 90."
It is antisymmetric, not symmetric — the mirror of the 0° bottom ply would be a 0° top ply, but the top is 90°. Hence and pulling causes curvature.
Why questions
Why does the same integral show up in both the force and moment equations?
In it comes from integrating (curvature's effect on force); in from integrating (stretch's lever-arm moment). The shared integral of Eq. (2) is precisely why one matrix couples both directions.
Why is the Kirchhoff-Love assumption () linear in ?
"Plane sections stay plane and perpendicular" means a straight fiber through the thickness stays straight after loading, so its strain varies linearly with height — no bulging, no shear warping.
Why does behave like the beam quantity ?
Both measure resistance to bending, and both get their strength from material stiffness times the square-ish moment of area — here the integration in Eq. (3) is the tensor cousin of that makes up .
Why can we integrate ply-by-ply instead of doing one continuous integral?
Each ply has a constant through its own thickness, so the through-thickness integral splits into a sum of clean per-ply pieces (Fig. 2). Discontinuous stiffness forces the summation form.
Why does an asymmetric laminate warp when it merely cools down, even with no applied load?
Uneven ply placement gives , so the thermal contraction strains (a stretching effect) feed into curvature through — the same coupling covered in Thermal and Hygroscopic Effects.
Why do engineers prefer symmetric layups for spacecraft panels?
decouples stretch and bend, so thermal cycling in orbit produces no unwanted warping and the panel stays flat — critical for mirror mounts and antennas.
Edge cases
A single isotropic ply centered on the mid-plane: what is ?
Zero — one ply symmetric about has with , so the term in Eq. (2) vanishes. A lone centered ply is trivially symmetric.
What happens to , , if you double every ply thickness uniformly (scale total thickness by 2)?
scales by 2 (linear in , Eq. 1), by 4 (, Eq. 2), and by 8 (, Eq. 3). Thicker laminates gain bending stiffness fastest.
A laminate with all plies at the same angle (e.g. [0/0/0]): is it a "laminate" in the coupling sense?
Effectively no coupling from angle differences — every is identical, so it behaves like a single thick orthotropic plate and (it is symmetric about its own center).
What is for a laminate that is symmetric in geometry but the two halves are different materials placed as mirror images?
Still — the mirror-image condition requires matching AND matching ; if both halves mirror exactly (same material at mirrored positions), symmetry holds and vanishes.
Zero applied moment but a non-symmetric laminate under pure : is the curvature zero?
No — with , does not force ; solving the coupled Eq. (4) generally gives non-zero . The load bends it even though you applied no moment.
When is the matrix ill-conditioned (hard to invert accurately)?
When entries span wildly different magnitudes — Pa·m and Pa·m differ by the thickness squared, so mixing them in one can produce large condition numbers, especially for thin laminates.
Does a laminate that fails First Ply Failure suddenly change its matrix?
The linear ABD is computed from intact ply stiffnesses; predicting First Ply Failure uses ABD to get ply stresses, but ABD itself only updates if you degrade the failed ply's in a progressive-damage model.
Can classical laminate theory (and thus ) handle a thick sandwich with a soft core?
Poorly — CLT assumes thin plies with no transverse shear; a soft-core sandwich shears through the core, so you need Sandwich Panel Theory or a shear-deformable / FE approach.