In this survey paper, we comprehensively examine the ongoing research concerning interaction between pantographs and catenaries, a vital aspect in ensuring uninterrupted electricity supply to trains. Future perspectives for future studies ensure satisfactory performance at 400 km/h above are preliminarily explored. Initially, paper provides an overview of current design assessment system. A systematic on numerical modelling pantograph-catenary is conducted. The applicability quantities...

Previous reinforcement learning (RL) methods suffer significant performance degradation or collapse when deployed to the real world due huge sim-real gap. This article proposes a hybrid offline-and-online meta-RL (HOMRL) algorithm that leverages prior task experience learn and adapt new pantograph active control tasks in real-world applications. The policy process consists of three phases: offline meta-policy pretraining, online adaptation, fine-tuning. First, we construct an approach learns...

The pantograph-catenary system (PCS) is the essential power supply in high-speed railway, but its coupling performance influenced significantly by rapidly increasing train speed. actively controlled pantograph one of promising technologies to suppress fluctuation contact force (PCCF). In this paper, we propose a novel control strategy based on deep reinforcement learning (DRL) overcome complex time-varying characteristic PCS, which distorts identification classical methods. First, non-linear...

Active pantograph control is the most promising technique for reducing contact force (CF) fluctuation and improving train's current collection quality. Existing solutions, however, suffer from two significant limitations: 1) they are incapable of dealing with various types, catenary line operating conditions, changing speeds, contingencies well 2) it challenging to implement in practical systems due lack rapid adaptability a new pantograph-catenary system (PCS) conditions environmental...

To meet the demands of transport development in mountainous regions, extensive efforts have been devoted to ongoing construction long suspension railway bridges. In context electrified railways, crucial responsibility transferring electric current train falls upon pantograph-catenary system. The installation catenary on bridge deck intertwines dynamics bridges with trains, thereby directly influencing dynamic performance system and potentially degrading collection quality. An imperative...

The substantial computational and memory demands of Large Language Models (LLMs) hinder their deployment. Block Floating Point (BFP) has proven effective in accelerating linear operations, a cornerstone LLM workloads. However, as sequence lengths grow, nonlinear such Attention, increasingly become performance bottlenecks due to quadratic complexity. These operations are predominantly executed using inefficient floating-point formats, which renders the system challenging optimize software...

Deep Neural Networks are increasingly leveraging sparsity to reduce the scaling up of model parameter size. However, reducing wall-clock time through and pruning remains challenging due irregular memory access patterns, leading frequent cache misses. In this paper, we present NPU Vector Runahead (NVR), a prefetching mechanism tailored for NPUs address miss problems in sparse DNN workloads. Rather than optimising patterns with high overhead poor portability, NVR adapts runahead execution...

The design and experimental study of a Ka-band gyrotron traveling-wave tube (gyro-TWT) amplifier operating in the circular <formula formulatype="inline" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex Notation="TeX">$\hbox{TE}_{01}$</tex></formula> mode at fundamental cyclotron harmonic are presented. A periodic dielectric loaded structure is adopted interaction circuit gyro-TWT to suppress spurious modes...

Controlling variations in part surface shapes is critical to high-precision manufacturing. To estimate the variations, a manufacturing plant usually employs number of multi-resolution metrology systems measure flatness and roughness with limited information about shape. Conventional research establishes models by considering spatial correlation; however, prediction accuracy restricted measurement range, speed, resolution systems. In addition, existing monitoring approaches do not locate...

The bracing wire is one of the most important components catenary support devices (CSCs) in high-speed railway, which used to connect messenger base and hook for stabilizing contact line receiving current. However, it prone a loose defect due frequent impact pantograph. Therefore, this article, an automatic detection method wires using computer vision technologies proposed. First, images are enhanced through deep convolutional neural network (CNN) called image-adaptive 3D LUTs (3D lookup...

Abstract Background To explore the value of a multiparametric magnetic resonance imaging (MRI)-based deep learning model for preoperative prediction Ki67 expression in prostate cancer (PCa). Materials The data 229 patients with PCa from two centers were retrospectively analyzed and divided into training, internal validation, external validation sets. Deep features extracted selected each patient’s MRI (diffusion-weighted imaging, T2-weighted contrast-enhanced T1-weighted sequences) to...

To clarify the hypothesis tests associated with full covariate modelling (FCM) approach in population pharmacokinetic analysis, investigate potential impact of multiplicity and evaluate simultaneous confidence intervals (SCI) as an to control multiplicity.Clinical trial simulations were performed using a simple one-compartment model. Different numbers covariates, sample sizes, effect sizes correlations among covariates explored. The false positive rate (FPR) power evaluated.The FPR for FCM...

In high-speed railways, the pantograph-catenary system (PCS) is a critical subsystem of train power supply system. particular, when double-PCS (DPCS) in operation, passing leading pantograph (LP) causes contact force trailing (TP) to fluctuate violently, affecting collection quality electric multiple units (EMUs). The actively controlled most promising technique for reducing (PCCF) fluctuation and improving current quality. Based on Nash equilibrium framework, this study proposes multiagent...

The fluctuation of pantograph–catenary contact force seriously affects the current collection quality, maintenance cost, and operation safety high-speed trains. In recent years, agents developed with deep reinforcement learning (DRL) technology have achieved significant success. However, most these jobs are restricted to narrow task distributions stationary environments, which require a large amount training data cannot adapt quickly new task. We propose contrastive learning-based...

This paper presents a method to reduce surface variation in face milling processes based on high-definition metrology (HDM) measurements. Our previous research has found and established the relations between patterns, cutting forces, process variables. Based findings, this compares potential machining methods finds that approaches of varying feed rate lateral cutter path planning are most feasible for control. By combining two approaches, an algorithm is developed force along direction...

Abstract. The phenomenon of low dissolved oxygen (known as hypoxia) in a coastal ocean system is closely related to combination anthropogenic and natural factors. Marine hypoxia occurs the Yangtze Estuary, China, with high frequency long persistence. It primarily organic nutrient enrichment influenced by river discharges physical factors, such water mixing. In this paper, three-dimensional hydrodynamic model was coupled biological simulate analyze ecological East China Sea. By comparing...

The active pantograph is a promising technology to suppress contact force fluctuation in catenary systems (PCS). Recently, the rapid development of reinforcement learning techniques has dramatically facilitated complex system controllers' design. However, low data efficiency problem fatal because collection costly. In this paper, We propose Ensemble Q-functions Model-based Reinforcement Learning algorithm (EQ-MBRL) achieve data-efficient learning. First, we introduce an ensemble...

Abstract The prerequisite for the effective operation of vehicle collision warning system is that necessary not implemented. Therefore, behavior prediction driver should perform when preceding braking key to improve effectiveness system. This study was conducted acquire characteristics in car-following confronted by vehicle, including reaction time and behavior, establish a model. A driving experiment on expressway using devices, such as millimeter-wave radars controller area network (CAN)...