When you step onto a plane bound for a holiday, the aircraft’s tapered curves are probably the last thing on your mind. Yet those contours embody a delicate balancing act in aerospace engineering: the trade-off between weight and strength. Every extra kilogram of material adds to an aircraft’s fuel consumption, but trim away too much and a fraction of a millimetre could mean the difference between structural safety and failure.
To walk this tightrope, most aircraft today are built partly from composite laminates. Similar to plywood, these advanced materials comprise overlapping layers or ‘plies’ of strong elastic fibres. However, while they can have impressive strength-to-weight ratios, laminate parts can also pose problems when engineers taper them to reduce weight.
“Conventional laminates use Quad layups, where plies are stacked with their fibres laid at 0°, 90°, or ±45° angles in varying sequences,” said Dan Wang, a Senior Scientist at the A*STAR Institute of High Performance Computing (A*STAR IHPC). “Removing or ‘dropping’ a single ply from a Quad laminate to gradually taper it can alter the material’s stiffness unevenly, often leading to design difficulties.”
In collaboration with Stephen Tsai of Stanford University in the US, Wang and A*STAR IHPC colleagues have been investigating Double-Double (DD) laminate designs as a more versatile option. Built from repeating pairs of fibre layers laid at balanced angles, DD laminates change their stiffness with a metal-like consistency, simplifying the design process and reducing the risk of local buckling and delamination when considering the spacing constraints.
“This smooth transition also enables the use of powerful, gradient-based optimisation methods on DD laminates to create complex aerospace structures,” Wang added.
The team recently published a new computer modelling approach for optimising the buckling resistance of gradually tapered DD laminate designs. Their method combined high-fidelity local models to capture how ply drop-offs behave in gradually tapered parts, then translated that detail into homogenised material properties for a larger global model.
“Our model captures the real effects of tapering in the global search, iterates as needed, then re-validates locally,” said Wang. “By doing so, we can lower computational costs while enabling a cleaner, scalable formulation of thickness variation.”
The team tested their method on a series of benchmark problems, including flat panels and a C-spar structure representative of an aircraft wing. They found that not all tapering strategies were beneficial: a simple linear tapering strategy reduced a DD laminate’s buckling resistance—measured as first buckling load, or λ₁—by over 50 percent. In contrast, an optimised gradual tapering improved λ₁ up to 280 percent without adding weight.
“Our findings illustrate that properly chosen design strategies and defined design spaces are crucial for optimisation to yield effective results,” said Wang.
Wang added that the team plans to experimentally validate their method on scaled composite panels, integrate more manufacturing constraints into its optimisation, and work with industry partners to embed it into existing design workflows.
The A*STAR-affiliated researchers contributing to this research are from the A*STAR Institute of High Performance Computing (A*STAR IHPC).
