Purpose-built WAAM 316L benchmark walls for thermo-mechanical calibration and strain-based hardness prediction

Abstract

Wire Arc Additive Manufacturing (WAAM) of 316L generates complex thermal cycles, distortion and work hardening that are difficult to validate in 3D. Hence, in this paper, we design, build and fully characterise two benchmark walls (two-layer and three-layer) expressly for calibrating and validating open-source thermo-mechanical simulations in MOOSE with emphasis on hardness prediction. The geometry combines unfinished passes, controlled overlap and steady-state regions so that a single build exercises transient and quasi-stationary behaviour. A multi-zone thermocouple layout (top and bottom) enables a robust thermal calibration that reproduces peak temperatures and cooling histories. Full-field 3D scanning yields distortion maps condensed into a normalised warping index; MOOSE reproduces longitudinal/transverse distortion with absolute errors ≤ 0.10 mm. Microhardness maps of two cross-sections are compared to hardness fields reconstructed from the simulated accumulated equivalent plastic strain using a single Tabor-type law. Across lines V1/V2/H1, the full J 2 -invariant surrogate achieves MAE ≈ 6–10 HV (comparable to test scatter), while a reduced “normal-only” strain measure shows MAE ≈ 25–35 HV and misses overlap-band peaks by ~ 20–40 HV, demonstrating the need for the full deviatoric invariant in WAAM 316L. Unlike prior WAAM validation studies that consider either single tracks or thermal/distortion only, this work delivers combined thermal, distortion and hardness datasets on multi-pass walls and an open, element-birth MOOSE workflow that enables full-field 3D hardness mapping from a calibrated strain surrogate. The benchmarks expose how layer count and pass pattern modulate hardness via reheating and constraint, and provide a reusable testbed for heat-source, boundary-condition and hardness-law comparisons.