What Machine Learning Can Tell Engineers Before the Next Aluminum WAAM Build-3DPTimes

Home > News

What Machine Learning Can Tell Engineers Before the Next Aluminum WAAM Build

November 6, 2026 News Source: Scientific Reports, November 3, 2026 8 Share to X.com

A Scientific Reports study estimates four mechanical properties from circular ER5356 wire-arc builds, giving engineers a way to choose which deposition settings to test next.

Choosing settings for wire arc additive manufacturing means deciding which combination is worth a physical trial. A study published in Scientific Reports on November 3 offers a way to make that choice for circular builds made from ER5356 aluminum wire. The researchers trained a physics-guided machine learning framework on data from cold metal transfer deposition, a wire-arc process that controls how molten metal transfers from the wire to the build.

The framework estimates four outcomes from the same set of process inputs: tensile strength, elongation, flexural strength and hardness. Those outputs answer different engineering questions. A setting that looks promising on strength alone may be less attractive once the expected elongation is considered. Estimating the responses together gives engineers a way to compare candidate settings before committing time and material to new builds.

The study also treats process selection as a multi-objective problem. Instead of collapsing the four properties into a single score, engineers can examine candidate settings that represent different trade-offs and decide which are worth testing for their application. That is the practical value of the approach: it helps define the next experiment.

The evidence is specific to the alloy, circular toolpath and deposition conditions studied. For a shop considering a different geometry or machine, the next step would be to print and test specimens at the proposed settings and compare the measured results with the predictions. Repeat builds would also help establish how much variation the process introduces.

Primary research: https://www.nature.com/articles/s41598-026-73406-w