Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
To mitigate pathological, disconnected structural layouts that plague classical dynamic-compliance optimization near resonance, this work applies a convolutional Kolmogorov-Arnold network (KATO) for neural-reparameterized topology optimization targeting active input power (AIP) minimization. Evaluated on engine-supporting deck panels and thruster foundation frames under machinery vibration, the framework incorporates Helmholtz PDE filtering and Heaviside projection to enforce strict manufacturing and feature-size constraints. KATO matches GCMMA in dynamic vibration attenuation within 0.5 dB while executing 6.4–10.4x faster on stress-aware formulations and delivering 22–59x lower static compliance following binary re-analysis.
Classical topology optimization breaks down near structural resonance, but neural reparameterization with Kolmogorov-Arnold networks prevents layout disconnection while delivering 10x faster solves and up to 59x stiffer binarized structures.
Ship structural vibrations contribute to noise, fatigue, and equipment damage, while dynamic-compliance topology optimization can produce pathological designs near resonance. This study extends neural-reparameterized topology optimization using a convolutional Kolmogorov-Arnold network (KATO) to forced-vibration design with active input power (AIP) as the objective. Applications include a 100 Hz engine-supporting deck panel and an 18 Hz thruster foundation frame. Helmholtz PDE filtering and Heaviside projection control feature sizes and manufacturing tolerance. Across both deck families, all eight optimized layouts reduce AIP relative to size-optimized references and, after finite-depth extrusion, also achieve lower static compliance. For unrestricted, manufacturing-aware, and stress-aware frame variants, KATO matches GCMMA in AIP within 0.5 dB while yielding 22-36x lower static compliance after matched-volume binary re-analysis. In a near-resonant 300 Hz case, both methods reduce initial AIP by more than 32 dB; KATO maintains a connected design, achieves 59x lower binary static compliance, and reduces maximum AIP over 1-500 Hz by 2.7 dB. KATO runs 6.4-10.4x faster than GCMMA for the implemented stress-aware formulations. The results demonstrate neural AIP-driven topology optimization as an efficient approach for designing connected, feature-size-controlled ship structures with improved forced-vibration performance.