Case Study

Optimization on a Vessel

Holding the bonding temperature within 1.5 °C across the cylinder and the domes of a pressure vessel, where curvature alone would swing it by 80 °C. Today the same optimisation runs in about 10 seconds in the browser.

Thank you for reading this post, don't forget to subscribe!
80 °Cbonding-temperature swing at the cylinder–dome transition without control
±1.5 °Ctemperature held around the target after optimisation
30–175%range of total laser power needed to compensate the curvature
~10 sfor the optimisation in AFP Web today, down from 3 minutes

The problem: curvature changes the heat input

When a tape is wound over a pressure vessel, the tooling curvature changes continuously from the cylinder into the elliptical dome. The laser sees a different surface angle at every moment: the peak laser intensity rises by about 80%, and the bonding temperature shifts by about 80 °C at the cylinder–dome transition. With constant laser power, part of the vessel overheats and degrades while another part bonds poorly.

AFP process optimisation framework for a pressure vessel: optical-thermal model, curvature effects and laser power control
AFP process optimisation framework (graphical abstract of the published study).

The solution: optical-thermal optimisation

Model the winding path

A winding path over the cylindrical and ellipsoidal (dome) sections, with the tooling curvature varying along it.

Trace the laser and solve the heat

The numerical optical-thermal model predicts where the laser energy lands and the transient bonding temperature at every position.

Optimise the laser power

Total laser power is adjusted between 30% and 175% of the reference case to cancel the local curvature effects.

Result

The process temperature stays within 1.5 °C of the desired value along the whole path, improving consolidation and the mechanical performance of the vessel.

Optical localisation of laser intensity on curved positions of the vessel
Optical localisation on curved positions.
Effect of the optimised temperature on mechanical performance
Effect on mechanical performance.

Unsteady optical-thermal events on a curved surface: the pressure vessel case.

Today: the same optimisation in seconds, in the browser

What once took minutes on a workstation now runs in the public AFP Web Digital Twin demo, up to about 8× faster after the latest GPU optimisation. That matters in AFP because heat from each new layer flows into the layers below: the faster a full lay-up can be simulated, the sooner defects from accumulated heat can be found and designed out.

Optical analysis (100 steps)20 s
3 s
Thermal simulation35 s
4 s
Optimisation3 min
10 s

Benchmark hardware: a laptop GPU with 80 W TGP and 16 GB GDDR6, as announced on OTOMcomposite on LinkedIn.

AFP Web digital twin: optimisation in a few seconds. Watch on YouTube

Winding a vessel, pipe or tank?

Try the optimisation yourself in the public demo, or let us run it on your geometry.