SOURCE-LINKED INTELLIGENCE

High-Fidelity Fabrication of HAR Microlens Arrays via Adaptive Toolpaths, Intelligent Feedback, and Elliptical Vibration Cutting

CORDIS · observation · Publication date unknown

ues on standard 3-axis ultra-precision machines. This project develops a fully integrated closed-loop framework combining curvature-adaptive volute spiral toolpaths, FEM simulations, physics-informed machine learning (PIML), and elliptical vibration cutting (EVC). Adaptive toolpaths ensure constant cutter engagement and uniform scallop height, while inverse modeling and geometric compensation minimize deflection and misalignment errors. Thermo-mechanical FEM simulations, validated experimentally on a KERN Evo system with piezo-dynamometer monitoring, capture cutting forces and stress fields to guide optimization. FEM outputs and force signals train PIML models embedding physical laws to predict tool wear, chatter, and surface deviations under sparse data. These models drive semi-automated adaptive corrections, targeting 85% chatter suppression, ≥85% tool wear detection, and >70% burr reduction. Final validation with high-resolution profilometry and CMM confirms reproducibility across MLA geometries. Research will be conducted under Prof. Erhan

Read original source ↗ Open in workspace

recordType

award

status

SIGNED

region

EU

value

165205.2

unit

EUR