DCNN深度特征与交替方向乘子的相关滤波跟踪

针对采用相关滤波的判别式目标跟踪遇到的瓶颈问题：由于目标快速移动引起边界效应，使得相关滤波器在学习与更新过程中可能会引入错误，最终错误的累积将导致跟踪失败。在采集深度学习特征与样本相似性度量的基础上，提出一种引入交替方向乘子方法的改进相关滤波目标跟踪算法，选择DCNN深度特征有效地表征待跟踪目标的初始状态，通过在线分类过程中样本相似性比对与半监督学习，辅助解决相关滤波器在学习过程中存在的自学习问题。所提目标跟踪算法特别适合训练样本为持续获得的、同时存储空间较小的机器学习过程，提高目标在快速运动与部分遮挡等复杂情况下的跟踪成功率，针对VOT2016标准测试视频的实验表明：当目标面临快速运动时，对比CN、SAMF、STC算法，所提DA-CFT跟踪算法将跟踪成功率分别由60.4%～73.4%、67.2%～82.9%、80.9%～88.1%提升至85.6%～91.0%。     

Numerical modeling of effect of slot on bubble motion in aluminum electrolytic process

A transient three-dimensional (3D) model was established to understand the bubble motion in an industrial electrolytic process. An anode with a new design was tested. It incorporates two slots that allow an efficient removal of gas bubbles. The electromagnetic fields were described by solving Maxwell's equations. The bubble movement was studied with two-way coupling Euler–Lagrange approach. The interplay of current density and bubble nucleation rate was included. The collision and coalescence of bubbles were considered. Random walk module was invoked for involving the chaotic effect of the turbulence. The numerical results were validated by experimental measurements. The results indicate that the current distribution and the bubble nucleation periodically change. Due to the slot, the bubble elimination heavily increases. The contribution of the slot to the bubble removal exceeds 50% in the case of three currents, and the promotion of the slot decays with increasing the current.     

基于VOF模型的层叠器流动过程分析及流道结构优化

基于FLUENT软件的流体体积(VOF)模型研究了不同壁面滑移程度以及不同流道结构参数下4层等厚熔体通过层叠器倍增为8层熔体时,流道中熔体的分层情况变化。结果表明,壁面滑移程度的降低会促使熔体在上下壁面聚集,进而导致上下壁面处熔体层厚增加;汇流段与出口段间圆角半径、汇流段扩压角和平衡段长度这3个结构参数则只会影响熔体在左右壁面的聚集,进而影响各层熔体的尺寸精度,且其中扩压角的影响程度最大,平衡段长度次之,圆角半径的影响较小。     

Analysis and optimisation of particle mixing performance in fluid phase resonance mixers based on Euler/Lagrange calculations

A novel mixing principle utilising oscillating liquid columns was analysed numerically with regard to particle dispersion characteristics. For producing fluid oscillations a pipe (diameter 100 mm) was immersed centrally into a vessel (diameter 450 mm) filled with liquid (filling height 700 mm) and periodically pressurised (frequency 1.2 Hz). The outlet geometry of the central pipe, just ending near the vessel bottom, has a strong effect on mixing and was optimised in this study. The principle of a FPR-mixer does not require rotating stirrers and in the turbulent regime it has power numbers comparable to propellers. The numerical calculations were conducted by a Euler/Lagrange approach neglecting two-way coupling as well as inter-particle collisions for clarity in order to only focus on the effect of interfacial forces on particle dispersion. The continuous phase was calculated in an unsteady way based on the Reynolds-averaged equations combined with the k-ω-SST (shear stress transport) turbulence model. Lagrangian tracking was conducted considering all relevant forces; drag, gravity/buoyancy, fluid inertia, added mass, Basset force and transverse lift forces due to shear and particle rotation. The importance of these forces was analysed with respect to the turbulent particle Stokes number (considered range 0.004 < St < 10.0) and particle/liquid density ratio (i.e. 1.05, 1.5 and 2.5). Finally, the significance of Basset force and shear-rotation lift force (i.e. Magnus effect) on the dispersion process was quantified by mixing parameters.     

Measurement of heat transfer coefficients for steam condensation on a vertical 21.5-mm-O.D. tube in the presence of air

ABSTRACT

Governing the rate of heat transport by condenser tubes in the passive containment cooling system (PCCS), the steam condensation over a vertical cylinder in the presence of air was investigated experimentally. The main objective of this study was to explore if the condensation heat transfer coefficient relies on the tube dimension, which has been a variable missed in most condensation models or has been embraced without experimental demonstration under phase change environments. The mean heat transfer coefficient was measured in the condensation test facility named JERICHO (JNU Experimental Rig for Investigation of Condensation Heat transfer On tube). The outer diameter of the condenser tube used in this study was set to 21.5 mm. The measured heat transfer coefficients were compared to those obtained from the 40-mm-O.D. tube, and a multiplier to correct the variation of the heat transfer coefficient with the tube diameter was proposed for its application to Lee correlation. The proposed correlation was further validated against another set of experimental data obtained from a separate test facility housing the 31.8-mm-O.D. tube.     

Fractional fixed-structure H∞ controller design using Augmented Lagrangian Particle Swarm Optimization with Fractional Order Velocity

The optimization of a fixed-structure controller is in general computationally intractable, posing a challenge to control system engineers. This paper introduces a novel technique for designing such control techniques by formulating it as a nonlinear non-convex constrained problem. A novel PSO algorithm, namely the Augmented Lagrangian Particle Swarm Optimization with Fractional Order Velocity (ALPSOFV), is proposed for improving the convergence rate. The structure of the controller is selectable and, therefore, the Fractional Order Proportional Integral Derivative (FOPID) algorithm is chosen in the scope of the study. The results for three test examples show the good performance of the ALPSOFV.     

斜坡稳定性安全系数的统一解

斜坡是山区、冻土地区,以及黄土地区广泛发育的一类地质形态.斜坡具有滑动深度小、滑动面易于确定、易受降雨和滑动面上水分影响的浅层滑坡特点.一般地,将地下水流线概化成与斜坡互相平行的直线和流线相互平行的水平线.从两类特殊渗流条件下的孔隙水压力入手,应用有效应力原理得出了无限斜坡稳定性安全系数的统一解,并通过工程实例对所得出的统一解进行了验证.文中提出的考虑渗流影响的无限斜坡稳定性分析方法对类似滑坡分析有一定的参考价值,尤其对于复杂渗流条件下的、流场研究程度较高的重要道路工程地段及桥涵附近边坡稳定性分析、坝基抗滑稳定性分析与验算等问题的研究具有实践指导意义.     

Geometric Parameterization Technique for Continuation Power Flow Based on Quadratic Curve

The electricity sector, especially in emerging countries, has experienced several transformations, mainly resulting from the increase of electricity demand. This encourages more investment in the generation sector and causes increasing concerns with the development and improvement of tools for static voltage stability analysis of electrical power systems. This paper presents a new geometric parameterization technique for continuation power flow (CPF) that works based on the addition of a parabola that passes through three points in the plane formed by the variables of total real power losses and loading factor. This technique eliminates the Jacobian matrix singularity at the maximum loading point, which allows obtaining the solution trajectory (P–V curve) without any need to change the parameter, which is a very common procedure in the currently available CPFs. Intending to define a simple and efficient step size control procedure, the total real power losses values are normalized by its base case value. The results obtained by applying the proposed technique to the IEEE-300 bus system and two real large systems of 638 and 787 buses show its effectiveness.