Automotive fuel cell sloshing under temporally and spatially varying high acceleration using GPU-based Smoothed Particle Hydrodynamics (SPH)

Understanding how fuel sloshes in a fuel cell, as a vehicle races around a circuit, is an important but mostly unexplored factor when designing fuel containment systems. Cell designs are based on knowledge of how liquids slosh in other containers, with the design and placement of structures, such as weirs, based on engineering judgement.This work aims to provide better understanding for this difficult problem with a view to improve future designs. A Graphics Processing Unit (GPU) based Smoothed Particle Hydrodynamics (SPH) model is presented to simulate the fuel sloshing problem, with results from a simplified and real fuel cell geometry shown and compared against real data recorded in a vehicle. The vehicle motion and accelerations are included in the SPH simulations using a body force within the momentum equation. Results show good agreement between the simulation and the real fuel movement, with bulk motion captured well for accelerations up to 5 times gravity.Focus is placed on the practicality of the method for use as part of an industrial design process, therefore the amount of time needed to compute results is considered throughout. Computational performance is found to be within acceptable limits, while numerical accuracy is actively considered through the use of Kahan compensated summation. It is concluded that the model is successful in capturing the necessary fluid dynamics for it to be useful in fuel cell design. It is expected that the method will provide insight into current cell designs and highlight where improvements can be made.     

一种基于群组判断离差的偏好集结方法

针对群决策偏好集结中违反Pareto 最优性的情况, 设计一种基于群组判断几何离差的同质性集结方法. 该方法在集结前进行几何离差测试, 以确定个体决策信息的离差水平. 离差较小时, 可基于几何平均集结; 对于离差较大且修正复杂度较高的决策信息, 采用主成分分析(PCA) 从高维决策信息中提取大多数相关信息, 在不依赖主观分析的情况下进行加权集结. 仿真实验表明, 所提出的方法能够在不违背Pareto 最优性的基础上集结离差较大的群决策信息.

     

基于地址-事件表示的高速二值连通域标记方法

为提高二值连通域标记的速度,将地址-事件表示AER（Address Event Representation）思想引入到二值图像处理,提出了一种基于事件对等价标号的二值连通域标记方法。该算法无需多次遍历图像中的背景点和冗余目标点,首先将待标记的连通域以AER“事件对”的方式编码保存,通过“事件对”的遍历生成临时标号和等价标记表;然后根据等价表修改临时标号;完成标号映射后最终实现连通域标记。整个算法只处理极低冗余的事件信息,避免了对全图像素的重复扫描与处理。实验结果表明,图像以AER“事件对”方式存储,数据量仅为全帧图像的10%~35%,有较高的压缩比;且该算法速度快,可达到了传统基于等价标号算法的1.5~8倍。     

基于SAM-CS和SOFM的胃上皮肿瘤细胞图像识别

针对胃上皮肿瘤细胞图像(以下简称肿瘤细胞图像)黏结严重和信息冗余的特点,提出了一种将自适应观测矩阵的压缩感知(SAM-CS)和自组织特征映射(SOFM)神经网络相结合的算法。该算法将肿瘤细胞图像拉成列向量,然后利用通过自适应过程产生的观测矩阵,基于压缩感知理论对图像信息进行观测,产生线性观测向量,最后利用SOFM神经网络的学习算法对观测向量进行训练和分类,实现对肿瘤细胞图像的识别。实验表明,相比常用算法,该算法至少提高了4.2%的识别准确率和5.7%的运算速度。     

在线考试时间控制技术的设计与实现

在线考试必须实现时间显示、定时存盘、自动交卷等时间控制功能,时间控制技术是在线考试的关键技术。传统的Web技术难以实现时间控制功能,使用AJAX技术可以方便地实现时间控制功能。文章介绍了在线考试时间控制技术的设计与实现,具体说明了AJAX组件的实现方案和关键代码。     

Networked Estimation of Multi‐Agent Systems Subject to Faults and Unreliable Information

In this paper, a novel framework for networked estimation of multi‐agent systems subject to presence of actuator faults is proposed. This framework is developed based on the notion of sub‐observers where within a group of sub‐observers each sub‐observer estimates certain states that are conditioned on a given input, output, and other state information. We model the overall estimation process by a weighted estimation (WE) digraph. By selecting an appropriate path in the WE digraph, an assigned supervisor can select and configure a set of sub‐observers to successfully estimate all the system states. In the presence of large intermittent disturbances, noise, and faults certain sub‐observers may become invalid, and consequently the supervisor reconfigures the set of sub‐observers by selecting a new path in the estimation digraph such that the impacts of these uncertainties are confined to only the local estimators. This will prevent the propagation of uncertainties on the estimation performance of the entire multi‐agent system. Simulation results provided for a five satellite formation flight system in deep space confirm the validity and applicability of our proposed analytical work.     

Optical complexities of living cytoplasm--implications for live cell imaging and photo-micromanipulation techniques

The ability to manipulate the intracellular environment within living cells and to monitor the cytosolic chemical changes which occur during cell stimulation has lead to major advances in our understanding of how cells read and respond to their environment. Perhaps the most powerful suite of techniques for achieving these dual objectives is based on the use of light (photons). Because cells are 'transparent', light has been used to both interrogate and manipulate the chemistry inside living cells, exploiting technical advances in both the physical and biochemical sciences. However, cells are neither transparent nor homogeneous with respect to their optical properties. The interface between light and the living cell cytoplasm thus represent an important, yet largely ignored, interface. There has been no review of the optical properties of cytoplasm and little discussion about how the optical properties of living cytoplasm influence the outcome of such measurements and manipulations. In this short review, we discuss the importance of understanding the optical properties of cytoplasm for such techniques and how imperfections in experimental interpretation can arise.     

Parametric modeling of microwave passive components using combined neural networks and transfer functions in the time and frequency

A novel parametric modeling technique is proposed to develop combined neural network and transfer function models for both time and frequency (TF) domain applications of passive components, where the neural network is trained to map geometrical variables to the coefficients of transfer functions. Built on our previous work, a new order‐changing module is developed to enforce stability of transfer functions and simultaneously guarantee continuity of coefficients. A constrained optimization strategy is introduced to enforce passivity of transfer functions through a neural network training process. A general equivalent circuit for two‐port passive components is generated directly from coefficients of arbitrary‐order transfer functions. Once trained, the parametric model can provide accurate and fast prediction of the electromagnetic behavior of passive components with geometrical parameters as variables. Compared to our previous work, the proposed method enables models to work well in the time domain providing good accuracy in challenging modeling applications. Two parametric modeling examples of spiral inductors and interdigital capacitors, and their application in both time and frequency domain simulations of a power amplifier are examined to demonstrate the validity of the proposed technique. © 2012 Wiley Periodicals, Inc. Int J RF and Microwave CAE , 2013.     

移动办公平台在天门烟草公司专卖管理的应用

以往智能终端机只是利用于卷烟销售方面,用于专卖工作的还不曾提及,是行业的盲点。基于3G网络的专卖移动办公平台以IMS为网络基础,以J2EE与J2ME为技术架构,使用成熟的中间件和可复用的业务组件实现了系统的跨平台性、伸缩性、可移植性和可维护性,为"卷烟上水平"提供有力的市场保障。     

结合水平集方法和PL主曲线分析的重叠细胞检测

提出了一种重叠细胞检测的新方法,该方法针对细胞图像处理中因细胞重叠而引起的过分割与欠分割现象这一难点,引入基于Mumford-Shah模型的水平集方法得到细胞的初始轮廓,采用EP（erosion-propagation）和PL主曲线相结合的图像分析方法得到目标的精确轮廓线。使用基于目标形状的图像配准方法和基于目标轮廓的分段特征分类器完成重叠细胞的检测。应用该方法对600张随机抽取的细胞图像中9374个细胞进行检测,实验正确率为91.83%,结果表明该方法能够有效地判断细胞是否重叠和重叠类型,为后续分割工作提供依据。