Australian Navy orders more metal 3D printed warship prototype parts

Wire-arc additive manufacturing builds a metal part by depositing weld wire layer by layer under a controlled arc. The feedstock is standard welding wire, so material cost stays low compared with powder-bed processes. Deposition rates are high, often several kilograms per hour, which is why the process suits large near-net shapes rather than small intricate ones.

For a warship prototype, that profile fits stiffener brackets, mounting lugs, ducting adapters and frame nodes. These parts are often several hundred millimeters across and carry modest loads. Buying them as castings needs a pattern and a long lead time. Machining them from solid plate wastes most of the stock. WAAM deposits only what the shape needs, then a CNC pass brings critical faces into tolerance.

The process also shortens the loop between design and hardware. A revised bracket geometry can be printed in days rather than waiting on a foundry slot. That matters when a prototype is still being iterated and the drawing changes between builds.

• 1
  Good fitLarge brackets, lugs and nodes with a modest surface finish requirement
• 2
  Poor fitThin walls, fine internal channels and parts under 100 mm
• 3
  MaterialStainless and low-alloy steel wire grades are the common choice
• 4
  Post-processStress relief, then CNC finishing on all mating and sealing faces

An as-deposited WAAM surface is not a finished surface. Expect roughly 0.5–2 mm of stock on a face, with local variation from heat distortion. That is fine for a non-critical web and unacceptable for a bolted joint or an O-ring groove. The practical answer on most defense prototypes is a hybrid route: print near net, then machine the interfaces.

On our 5-axis centers we hold ±0.005 mm on critical features and reach Ra 0.8–1.6 μm on sealing faces. A printed bracket that arrives with 1.5 mm of stock on a mounting pad is a straightforward setup. A printed part with 0.2 mm of stock and a warped flange is not, because there is no clean reference surface left to pick up.

So the print parameters and the machining datum need to be planned together. Leave a machining allowance on every surface that will be clamped, and leave a boss or pad that can be faced flat as the first operation. If the print shop and the machine shop are the same supplier, that conversation happens before deposition starts.

A repeat order is a statement about repeatability. The first build proves the geometry. The second build has to prove that the same geometry comes out again, with the same material properties, from a documented process. That is a paperwork problem as much as a manufacturing one.

For any defense-adjacent prototype, plan for traceability. Wire heat number, deposition parameters, heat treatment cycle, machining setup and final inspection results all need to sit in one record tied to the part serial. Without that, a printed part is just a shape.

Inspection is where printed and machined parts converge. Both need dimensional reports, and both need a documented check on material condition. Our standard flow is raw material check, in-process monitoring and 100% inspection before shipment, with reports on request. On a printed part we also want to see the deposit before machining, so a defect in the build is caught before value is added.

Printing is not automatically the cheaper route. For a bracket under 150 mm, machining from 6061 or 17-4PH bar is usually faster and gives a better surface without a second operation. The economics flip when the part gets large, when the buy-to-fly ratio on a billet is poor, or when the geometry has a lot of removed material.

There is also a porosity question. Wire-arc deposits can carry small voids, and on a pressure boundary or a fatigue-critical fitting that is a real risk. If the part sees cyclic loading at a stress concentration, a wrought billet is the safer starting point. Printing earns its place on stiffness-driven shapes and on one-off geometry that would otherwise need tooling.

We run both routes under one roof, so the recommendation is not tied to a single process. If printing does not make sense for your part, we will say so and quote the machined version.