Advisor: Prof. Karl Jacob

will propose a doctoral thesis entitled,

Manufacturing-Aware Deviation Encoding for Robust Inverse Design of Additively Manufactured Lattice Metamaterials

On

Friday, June 26th, 2026

1:00 pm EST

Zoom Link: https://gatech.zoom.us/j/98511662676

Abstract

Additive manufacturing (AM) of lattice metamaterials enables fabrication of diverse structures with highly tunable mechanical responses, but process-induced geometric deviations and variability create uncertainty when comparing as-fabricated performance to predictions. Conventional inverse design and topology optimization frameworks often assume ideal reproductions of as-designed geometries, while methods that address uncertainty often treat manufacturing variability as stochastic noise to be managed probabilistically rather than as structured geometric information to be directly encoded. As a result, current workflows struggle to deliver robust, uncertainty-aware designs where predicted properties remain accurate under real process variability. This dissertation introduces the Manufacturing-Aware Deviation Encoding (MADE) vector: a compact, empirically-derived descriptor that encodes the process-specific geometric deviation fingerprint, which can be embedded directly into surrogate modeling and the inverse design loop. MADE is derived from a minimal set of CT-extracted geometric deviation features that explain dominant variance in mechanical performance over a range of lattice geometries. Demonstrated on fused filament fabrication (FFF)-printed thermoplastic polyurethane (TPU) re-entrant auxetic lattice samples, this framework enables surrogate models that propagate manufacturing uncertainty through the inverse design loop, helping bridge the gap between computational design intent and reality.

Committee