03.50 · Thermal warping: unsymmetric versus symmetric cross-ply

A laminate can develop strain and curvature without an applied mechanical load when its layers respond differently to temperature, moisture or an enabled process strain. These exercises use controlled changes in stacking order, exposure or loading to distinguish that response from ordinary applied bending.

Open this exercise in Workbench

Prepare the baseline

Inspect ply order, total thickness, reference states and applied loads before the baseline run. Record the environmental state actually consumed by the structural calculation. For a symmetric or reversed-stack comparison, preserve thickness and exposure as instructed; otherwise the change in curvature cannot be attributed to stacking order alone.

Worked procedure

1. Inspect the 16-ply [0₈/90₈] stack, 2 mm thickness and zero applied mechanical loads. Run the thermal and structural simulation.

2. Review the thermal history and the structural state actually used. Record B-matrix terms and curvatures; if the transferred state is near the reference temperature, select a warmer supported state or end the cycle at a warm dwell and rerun.

3. Replace the layup with [0/0/0/0/90/90/90/90]s, preserving the same 16 plies and total thickness. Rerun and compare B and curvature.

4. Reverse the unsymmetric stack and repeat at the same thermal state. Compare curvature direction.

Review checkpoints

The symmetric stack has negligible extension–bending coupling.

Zero applied force does not require zero thermal deformation.

A nearly uniform state at the reference temperature may give little curvature; this is not a solver failure.

Model limits

Teaching inputs, not qualified allowables. Sequential coupling transfers the reported thermal/moisture state to laminate mechanics; it is not simultaneous 3D multiphysics. Inspect the transferred state and reference values before interpreting stress or curvature. Free laminate CLT curvature, not fixture-constrained distortion, a full tool-release simulation or calibrated cure-shrinkage prediction.

Interpret the comparison

Examine coupling terms and curvature together with ply stresses. Reversing a stack or removing applied loads is a diagnostic comparison, not a reason to assume every stress must disappear. If the exercise includes cure exotherm, distinguish the evolving process state from the final cooled state before calling a stress residual.

How information passes between models

Micro → Laminates: Predicted ply stiffness, strength, density and expansion properties.

Materials → Micro: Constituent stiffness, strength, density and thermal / moisture properties.

Mechanical → Simulation: SIMULATION selects this case and its analysis model; the case owns its applicable cycle and input references.

Laminates → Mechanical: Ply angles and thicknesses, stiffness, mass and ply properties.

Geometry → Mechanical: Part shape and dimensions, thickness or section definition, and model-specific geometric inputs. Each selected case consumes only the dimensions its model supports.

Thermal → Simulation: SIMULATION selects this case and its analysis model; the case owns its applicable cycle and input references.

Laminates → Thermal: Ply angles and thicknesses, stiffness, mass and ply properties.

Models → Micro: Applied model assignment: Halpin–Tsai. Model parameters and formulation are used by Micro.

Models → Mechanical: Applied model assignment: CLT · Linear static with failure indices. Model parameters and formulation are used by Mechanical.; Maximum stress

Models → Thermal: Applied model assignment: 1D transient heat transfer. Model parameters and formulation are used by Thermal.

Further reading and evidence

• Connect a simulation
• Edit laminate materials, angles and thicknesses
• Run and review a model
• Cure exotherm and evolving modulus
• Enter a process cycle and boundary conditions
• Connected inputs and result freshness
• Review effective properties
• Apply symmetry and balance
• Related case study: NASA process cure

Review the recorded validation scope. Retain the original inputs and solver notices with the results. Representative teaching data are not design allowables.