Show newer

Comment on "Revisiting the divergent multipole expansion of atom-surface interactions: Hydrogen and positronium, $\alpha$-quartz, and physisorption" (arXiv:2308.04656v3) arxiv.org/abs/2501.14803

High-Performance, Efficient, Low-Fresnel Number Focusing Metamirror in the D-Band arxiv.org/abs/2501.14875

Integrated 3D printing of transparency-on-demand glass microstructure arxiv.org/abs/2501.14888

Physics-Informed Neural Networks for microflows: Rarefied Gas Dynamics in Cylinder Arrays arxiv.org/abs/2501.13108

Physics-Informed Neural Networks for microflows: Rarefied Gas Dynamics in Cylinder Arrays

Accurate prediction of rarefied gas dynamics is crucial for optimizing flows through microelectromechanical systems, air filtration devices, and shale gas extraction. Traditional methods, such as discrete velocity and direct simulation Monte Carlo (DSMC), demand intensive memory and computation, especially for microflows in non-convex domains. Recently, physics-informed neural networks emerged as a meshless and adaptable alternative for solving non-linear partial differential equations. We trained a PINN using a limited number of DSMC-generated rarefied gas microflows in the transition regime with Knudsen number from 0.1 to 3, incorporating continuity and Cauchy momentum exchange equations in the loss function. The PINN achieved under 2 percent error on these residuals and effectively filtered DSMC intrinsic statistical noise. Predictions remained strong for a tested flow field with Kn equal to 0.7, and showed limited extrapolation performance on a flow field with Kn equal to 5 with a local overshoot of about 20 percent, while maintaining physical consistency. Notably, each DSMC field required about 20 hours on 4 graphics processing units (GPU), while the PINN training took less than 2 hours on one GPU, with evaluations under 2 seconds.

arXiv.org

A car-following model with behavioural adaptation to road geometry arxiv.org/abs/2501.13191

A car-following model with behavioural adaptation to road geometry

Understanding the effect of road geometry on human driving behaviour is essential for both road safety studies and traffic microsimulation. Research on this topic is still limited, mainly focusing on free-flow traffic and not adequately considering the influence of curvature on car-following dynamics. This work attempts to investigate this issue and model the adaptation of car-following behaviour to horizontal curvature. For this purpose, the maximum desired speed - which mainly determines the free-flow dynamics - is expressed as a parsimonious function of the curvature. A spatial anticipation mechanism is also included in order to realistically describe the driving behaviour when approaching or exiting from curves. The accuracy of the augmented model is evaluated using the Modified Intelligent Driver Model (M-IDM) and trajectory data from free-flow and car-following traffic (Naples data and Zen Traffic Data). The results show that a significant improvement is achieved in free-flow dynamics. In car-following situations, improvements are mainly observed at high speed and are dependent on the observed driver. Overall, the analysis highlights the lack of sufficiently spatially extended trajectory data to calibrate and evaluate such driving behaviours.

arXiv.org
Show older
Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.