arXiv subjects

S. Peng

Publications and source records attributed to S. Peng.

5 recordsLinked to original sources

arXiv2026

Submesoscale and boundary layer turbulence under mesoscale forcing in the upper ocean

The interaction among quasi-geostrophic mesoscale eddies, submesoscale fronts, and boundary layer turbulence (BLT) is a central problem in upper ocean dynamics. We investigate these multiscale dynamics using a novel large-eddy simulation on a \qty{100}{\kilo\meter}-scale domain with meter-scale resolution. The simulation resolves BLT energized by uniform surface wind and convective forcing. A front interacts with BLT within a prescribed, spatially inhomogeneous mesoscale eddy field, representing a canonical eddy quadrupole. Using a triple flow decomposition, we analyze the dynamic coupling and kinetic energy budgets among the large-scale field, submesoscale field, and the resolved BLT. Our analysis reveals significant heterogeneity in the structure and intensity of submesoscales and BLT under varying mesoscale forcing. Turbulent kinetic energy and production rates can vary by an order of magnitude along the front, creating distinct turbulent hotspots whose locations are tied to the underlying large-scale flow. The region under stronger mesoscale convergence holds stronger horizontal and vertical geostrophic shear productions for BLT, and stronger self-production and BLT-destruction for submesoscales. In contrast, the region under dominant mesoscale divergence holds dramatic distortion of the front isotherm, along with dominant submesoscale vertical buoyancy production and self-destruction. Within this idealized framework, these results provide a controlled process-level characterization of how prescribed mesoscale heterogeneity modulates BLT and submesoscales in the ocean mixed layer, which can inform future parameterization developments.

physics.flu-dyn↗

arXiv2025

Neighborhood-Adaptive Generalized Linear Graph Embedding with Latent Pattern Mining

Graph embedding has been widely applied in areas such as network analysis, social network mining, recommendation systems, and bioinformatics. However, current graph construction methods often require the prior definition of neighborhood size, limiting the effective revelation of potential structural correlations in the data. Additionally, graph embedding methods using linear projection heavily rely on a singular pattern mining approach, resulting in relative weaknesses in adapting to different scenarios. To address these challenges, we propose a novel model, Neighborhood-Adaptive Generalized Linear Graph Embedding (NGLGE), grounded in latent pattern mining. This model introduces an adaptive graph learning method tailored to the neighborhood, effectively revealing intrinsic data correlations. Simultaneously, leveraging a reconstructed low-rank representation and imposing $\ell_{2,0}$ norm constraint on the projection matrix allows for flexible exploration of additional pattern information. Besides, an efficient iterative solving algorithm is derived for the proposed model. Comparative evaluations on datasets from diverse scenarios demonstrate the superior performance of our model compared to state-of-the-art methods.

cs.LG↗

arXiv2012

On the threshold-width of graphs

The GG-width of a class of graphs GG is defined as follows. A graph G has GG-width k if there are k independent sets N1,...,Nk in G such that G can be embedded into a graph H in GG such that for every edge e in H which is not an edge in G, there exists an i such that both endpoints of e are in Ni. For the class TH of threshold graphs we show that TH-width is NP-complete and we present fixed-parameter algorithms. We also show that for each k, graphs of TH-width at most k are characterized by a finite collection of forbidden induced subgraphs.

math.CO↗

arXiv2003

Discrete molecular dynamics simulations of peptide aggregation

We study the aggregation of peptides using the discrete molecular dynamics simulations. At temperatures above the alpha-helix melting temperature of a single peptide, the model peptides aggregate into a multi-layer parallel beta-sheet structure. This structure has an inter-strand distance of 0.48 nm and an inter-sheet distance of 1.0 nm, which agree with experimental observations. In this model, the hydrogen bond interactions give rise to the inter-strand spacing in beta-sheets, while the Go interactions among side chains make beta-strands parallel to each other and allow beta-sheets to pack into layers. The aggregates also contain free edges which may allow for further aggregation of model peptides to form elongated fibrils.

cond-mat.soft↗

arXiv2001

SNS Standard Power Supply Interface

The SNS has developed a standard power supply interface for the approximately 350 magnet power supplies in the SNS accumulator ring, Linac and transport lines. Power supply manufacturers are providing supplies compatible with the standard interface. The SNS standard consists of a VME based power supply controller module (PSC) and a power supply interface unit (PSI) that mounts on the power supply. Communication between the two is via a pair of multimode fibers. This PSI/PSC system supports one 16-bit analog reference, four 16-bit analog readbacks, fifteen digital commands and sixteen digital status bits in a single fiber-isolated module. The system can send commands to the supplies and read data from them synchronized to an external signal at up to a 10KHz rate. The PSC time stamps and stores this data in a circular buffer so historical data leading up to a fault event can be analyzed. The PSC contains a serial port so that local testing of hardware can be accomplished with a laptop. This paper concentrates on the software being provided to control the power supply. It includes the EPICS driver; software to test hardware and power supplies via the serial port and VME interface.

cs.OH↗