Ryan Davies | Searcharxiv

arXiv2026

Bi-stable Nonlinear Energy Sinks (BNESs) for Response Mitigation and Drag Reduction of Subsea Cables Undergoing Vortex-induced Vibrations

A methodology for passive mitigation of vortex-induced vibrations (VIVs) in subsea dynamic power cables is developed using optimized, strongly nonlinear bi-stable mass-spring-damper attachments, termed bi-stable nonlinear energy sinks (BNESs), within the open-source MoorDyn library. A fully three-dimensional time-domain framework captures cable dynamics, Morison-type hydrodynamic forcing, nonlinear vibration-mitigation mechanisms, and spatially and temporally varying currents. The BNESs are consistently integrated into the cable model, allowing treatment of highly non-stationary VIVs. A data-driven optimization study samples the BNES design space across multiple current profiles and shows that properly tuned configurations substantially reduce peak-to-peak cable vibration amplitudes. The BNESs also produce significant and robust reductions in cumulative VIV-induced energy intake from the surrounding flow and in drag energy. To the authors' knowledge, passive nonlinear attachments are shown for the first time to reduce both energy intake and drag amplification in a subsea cable, with potential benefits for short- and long-term fatigue. Time-frequency wavelet analysis reveals targeted nonlinear energy transfers and scattering from dominant high-amplitude, low-frequency cable modes to lower-amplitude, higher-frequency modes. This modal redistribution promotes rapid dissipation through hydrodynamic damping and internal structural losses, explaining the simultaneous reductions in vibration amplitude and cumulative energy intake. The results demonstrate that BNESs can provide effective and robust VIV mitigation for subsea power cables under realistic unsteady operating conditions and motivate future studies involving combined current-wave loading, platform-induced motion, and fatigue-life assessment.

physics.flu-dyn↗

arXiv2019

GEMSEC: Graph Embedding with Self Clustering

Modern graph embedding procedures can efficiently process graphs with millions of nodes. In this paper, we propose GEMSEC -- a graph embedding algorithm which learns a clustering of the nodes simultaneously with computing their embedding. GEMSEC is a general extension of earlier work in the domain of sequence-based graph embedding. GEMSEC places nodes in an abstract feature space where the vertex features minimize the negative log-likelihood of preserving sampled vertex neighborhoods, and it incorporates known social network properties through a machine learning regularization. We present two new social network datasets and show that by simultaneously considering the embedding and clustering problems with respect to social properties, GEMSEC extracts high-quality clusters competitive with or superior to other community detection algorithms. In experiments, the method is found to be computationally efficient and robust to the choice of hyperparameters.

cs.SI↗