LP-FNO Inference Demonstraction

Controls

Power (W) -

Scan Speed (m/s) -

Normalized Enthalpy H (-) -

Adjusts scan speed at fixed power

Selection Maps Drag directly on either map

Interactive P-V Enthalpy Map Drag to set power + scan speed directly

Interactive H-P Map Drag to set power + enthalpy (scan speed derived)

Plot Mode

3D uses backend surface rendering

Render Mode

Hi-Res sends final renders during live interaction

Field Y Index T Contour (K) Show T contour

Selected P-

Selected Vscan-

H*-

H_model = P/sqrt(|V|)-

H_norm (server)-

About This Demo

This demo lets you explore two LP-FNO surrogate models for melt-pool prediction in laser processing. Use the sliders or interactive maps on the left to set scan speed, laser power, and normalized enthalpy, then view the predicted 3D temperature field and melt-pool geometry in real time.

Classical LP-FNO is a purely classical Fourier Neural Operator trained end-to-end on high-fidelity simulation data. Hybrid LP-FNO augments the classical architecture with quantum layers. Switch between the two models with the toggle above the result, or use Error Compare to see their prediction errors side by side against the ground-truth simulation. Since ground truth is available only for a discrete set of (P, V) combinations, the comparison uses the nearest available sample.

Result

none Awaiting first render

View

No result yet

Citation

Benoit, A., Ivas, T., Papierz, M., Sagingalieva, A., Melnikov, A., and Iseli, E., A Fast and Generalizable Fourier Neural Operator-Based Surrogate for Melt-Pool Prediction in Laser Processing, arXiv preprint arXiv:2602.06241, 2026. Available at: https://arxiv.org/abs/2602.06241.