基于YOLOv3与注意力机制的桥梁表面裂痕检测算法

为了实现桥梁表面裂痕的快速准确检测和及时修复,在目标检测网络YOLOv3的基础上,结合深度可分离卷积与注意力机制,提出实时检测桥梁表面裂痕的轻量级目标检测网络.使用深度可分离卷积操作替换YOLOv3的标准卷积操作,达到降低网络参数量的目的.同时为了解决深度可分离卷积操作带来的网络精度下降的问题,引入MobileNet v2的反转残差块.卷积块注意力模块同时关注图像的通道注意力和空间注意力,较好地进行特征的自适应学习.实验表明,文中算法可实现对桥梁表面裂痕的实时检测.相比YOLOv3,具有更高的检测精度和检测速度.     

Situational variables in online versus offline channel choice

Since the advent and rapid diffusion of the Internet, the subject of consumer channel choice has attracted a large amount of research, mainly focused on the influence of channel, consumer and product category characteristics as its drivers. The interaction between channel choice and the purchase situation has been largely ignored, however. This paper is an attempt to fill this gap by identifying the key purchase situation variables and conducting an experiment to assess their impact on the choice between the traditional retail outlet and the online store. The results show that the key determinants of channel choice relate to time and distance. Distance-to-store and time pressures are among the factors affecting the probability of online purchase. Using a conceptual framework to explore differences in the impact of situational variables across product categories (high/low involvement, search/experience good), we show that distance-to-store has more influence on the likelihood of online purchase in situations involving search goods, while social variables are found to play a role only in the context of high-involvement goods.     

唐墓室壁画裂缝的自动虚拟修复方法

裂缝是壁画中存在的一种常见的病害,为了去除壁画中的裂缝提出一种对壁画中裂缝进行自动虚拟修复的方法。对图像进行高帽变换提取裂缝信息,对变换后的图像进行阈值分割得到裂缝的二值图像。根据裂缝的颜色和结构信息利用基于连通域度量的方法去除虚假目标,实现自动识别和标注。利用基于样本的数字图像修复技术对壁画中裂缝进行自动修复,在修复中对样本区域进行重构大大缩短了修复的时间,提高了修复效率。实验仿真表明该方法能够有效去除壁画中影响人眼视觉的较大裂缝,能够实现对壁画中裂缝的自动虚拟修复。     

Numerical investigation of micro- and nanochannel deformation due to discontinuous electroosmotic flow

Large pressures can induce detrimental deformation in micro- and nanofluidic channels. Although this has been extensively studied for systems driven by pressure and/or capillary forces, deflection in electrokinetic systems due to internal pressure gradients caused by non-uniform electric fields has not been widely explored. For example, applying an axial electric field in a channel with a step change in conductivity and/or surface charge can lead to internally generated pressures large enough to cause cavitation, debonding, and/or channel collapse. Finite electric double layers within nanofluidic channels can further complicate the physics involved in the deformation process. In order to design devices and experimental procedures that avoid issues resulting from such deformation, it is imperative to be able to predict deformation for given system parameters. In this work, we numerically investigate pressures resulting from a step change in conductivity and/or surface charge in micro- and nanofluidic channels with both thin and thick double layers. We show an explicit relation of pressure dependence on concentration ratio and electric double layer thickness. Furthermore, we develop a numerical model to predict deformation in such systems and use the model to unearth trends in deformation for various electric double layer thicknesses and both glass and PDMS on glass channels. Our work is particularly impactful for the development and design of micro- and nanofluidic-based devices with gradients in surface charge and/or conductivity, fundamental study of electrokinetic-based cavitation, and other systems that exploit non-uniform electric fields.     

Simulation of air flow in the typical boiler windbox segments

Simulation of turbulent air flow distribution in CFBC furnace, wherein primary air is entrained through inlet duct system called windbox, is attempted through state of art CAD/CFD softwares. Establishment of flow in windbox channel, distributed plate nozzle and combustor is complicated, due to sharp turns and presence of several solid boundaries makes the fluid flow highly turbulent. Hence, the simulation process is aimed in different parts to understand the flow behavior in each of the component associated with windbox. Towards this, the present paper develops the basic understanding for airflow distribution in windbox channel, wherein air exit is considered only through 6×3 array of distributed plate nozzle bottom faces. This analysis also highlights that recirculation flow, at several locations of windbox channel/distributed plate nozzle, which aids to generate high pressures zones and severe turbulent fluctuations. These effects in turn leads to unequal air-flow at exit, which are unable to carry the incoming crashed coal particles and lime stones to furnace for efficient combustion.     

Thin crack observation in a reinforced concrete bridge pier test using image processing and analysis

In reinforced concrete (RC) structural experiments, the development of concrete surface cracks is an important factor of concern to experts. One conventional crack observation method is to suspend a test at a few selected testing steps and send inspectors to mark pen strokes on visible cracks, but this method is dangerous and labor intensive. Many image analysis methods have been proposed to detect and measure the dark shadow lines of cracks, reducing the need for manual pen marking. However, these methods are not applicable for thin cracks, which do not present clear dark lines in images.This paper presents an image analysis method to capture thin cracks and minimize the requirement for pen marking in reinforced concrete structural tests. The paper presents the mathematical models, procedures, and limitations of our image analysis method, as well as the analysis flowchart, the adopted image processing and analysis methods, and the software implementation. Finally, the results of applying the proposed method in full-scale reinforced concrete bridge experiments are presented to demonstrate its performance. Results demonstrate that this method can capture concrete surface cracks even before dark crack lines visible to the naked eye appear.     

Microcavity Plasma Devices and Arrays Fabricated by Plastic-Based Replica Molding

The fabrication of lightweight and optically transparent arrays of microplasma devices having microcavities fabricated by a polymer-based replica molding process is reported. This process enables arrays of microcavity plasma devices and connecting channels with feature sizes as small as 20 mum to be produced inexpensively and precisely over surface areas of at least tens of square centimeters. Devices having transparent electrodes and substrates (including glass and flexible plastic) have been operated successfully at rare gas pressures up to 700 torr. Individual microplasma pixels with cross-sectional dimensions of 200times200 mum² and 50times50 mum², as well as plasma channels having widths of 20 and 150 and aspect ratios (channel length:width) as large as 10⁴:1, have been fabricated and tested. Representative voltage-current data for these structures and lifetime measurements for a 20times20 pixel array are presented.     

Computational investigation and parametrization of the pumping effect in temperature-driven flows through long tapered channels

The temperature-driven rarefied gas flow and the associated pumping effects through long channels with linearly diverging or converging cross sections are computationally investigated. The implemented kinetic modeling is well known and relies on the infinite capillary methodology coupled with the mass conservation principle along the channel. The net mass flow rate and the induced pressure difference between the channel inlet and outlet are parametrized in terms of the geometrical and operational data including the channel inclination and the inlet pressure. Specific attention is given to the diode effect. The investigated flow setups include (a) the maximum pressure difference scenario with zero net mass flow rate (maximum pumping effect), (b) the maximum net mass flow rate scenario with equal inlet and outlet pressures and (c) all intermediate flow cases where both the net mass flow rate and the pressure difference are different than zero. In the first limit case, the pressure difference is always increased with the channel inclination and, depending on the inlet pressure, it may be larger for either the diverging or converging channel. In the second limit case, the mass flow rate is always decreased when the channel inclination is increased and it is always higher for the diverging channel. In both limit cases, optimum operation scenarios, in terms of the diode effect and the overall performance, are extracted. For intermediate cases, the characteristic curves of the net mass flow rate versus the pressure difference have been developed, indicating that the mass flow rate is inversely proportional to the pressure difference. The results strongly depend on the channel inclination. The present work may support decision making on the suitability of tapered channel flow to meet certain pumping specifications and the design of cascade-type thermally driven micropumps.     

Flows of rarefied gaseous mixtures in networks of long channels

A method is presented to calculate the steady flows of rarefied gaseous mixtures in networks of long channels. The approach is based on the kinetic level. First, the McCormack linearized kinetic model is solved to obtain the local flow properties in the channels in a wide range of gaseous rarefaction and mole fraction. Second, the global flow properties including the flow rates and the distribution of the pressure and the mole fraction are deduced. An integral equation is introduced in order to determine the flow rates as functions of the differences of the partial pressures between the two ends of each channel. The conservation of mass at the nodes of the network results into a system of linear algebraic equations. The overall mathematical problem is solved iteratively. Pressure driven flows of He/Xe and He/Ar through an example network of circular tubes are calculated at intermediate values of the gaseous rarefaction. The results of the flow rates and the pressures and the mole fractions at the nodes in the whole system and the representative distributions of the pressure and the mole fraction along the channels are presented and commented on.     

Micro T-mixer as a rapid mixing micromixer

In this paper, micro T-mixers are fabricated and tested to investigate their feasibility as a rapid mixing micromixer. The micro T-mixers are fabricated out of a silicon substrate and bonded to a Pyrex glass plate to enable their mixing performances be observed and characterized. The mixing is characterized using a blue dye and a colourless liquid, the results are further verified by the hydrolysis reaction of dichloroacetyl phenol red. Different pressures are applied onto the inlets of the micro T-mixers and their corresponding mixing performances are observed with an optical microscope. Liquid streams break up into striations at progressively higher Reynolds number of flow and there exists a Reynolds number, between 400 and 500, when these striations disappear into uniform concentration across the mixing channel. The observations are further supported by computer simulations, which enable the fast mixing to be explained by the asymmetrical flow conditions at the inlets, in addition to the generation of vortices and secondary flow at the junction. It is shown that for a micro T-mixer with a mixing channel having a hydraulic diameter of 67 μm, an applied pressure of 5.5 bar is sufficient to cause complete mixing within less than a millisecond after the two liquids make contact.     

Microfluidic pump based on the phenomenon of electroosmosis of the second kind

A micropump based on strong polarization of ion-exchange beads and corresponding actuation by electroosmosis of the second kind was designed and fabricated. Experimental results from operation with AC and DC voltage showed a close to second order relationship between flow and voltage, in good agreement with theory. The difference between experimental and theoretical flow rates and pressures is attributed to the hydrodynamic resistance of the channel network. A modified pump design which should yield higher flow rates and pressures was suggested.     

Fundamental frequencies of annular plates with internal cracks

The Rayleigh method with a simple sub-sectioning technique has been used to obtain fundamental frequencies of annular plates with internal concentric cracks. Single-term deflection functions which satisfy the respective boundary conditions for the bending of an annular plate are used as admissible functions. The predicted fundamental frequencies are in good agreement with the reported analytical results for annular plates without any internal cracks. Numerical results for fundamental frequencies are presented for annular plates with internal cracks simply supported and clamped on both the outside and inside edges of the plates.     

Flow over a membrane-covered,fluid-filled cavity

The flow-induced response of a membrane covering a fluid-filled cavity located in a section of a rigid-walled channel was explored using finite element analysis. The membrane was initially aligned with the channel wall and separated the channel fluid from the cavity fluid. As fluid flowed over the membrane-covered cavity, a streamwise-dependent transmural pressure gradient caused membrane deformation. This model has application to synthetic models of the vocal fold cover layer used in voice production research. In this paper, the model is introduced and responses of the channel flow, the membrane, and the cavity flow are summarized for a range of flow and membrane parameters. It is shown that for high values of cavity fluid viscosity, the intracavity pressure and the beam deflection both reached steady values. For combinations of low cavity viscosity and sufficiently large upstream pressures, large-amplitude membrane vibrations resulted. Asymmetric conditions were introduced by creating cavities on opposing sides of the channel and assigning different stiffness values to the two membranes. The asymmetry resulted in reduction in or cessation of vibration amplitude, depending on the degree of asymmetry, and in significant skewing of the downstream flow field.     

结合NSCT和图像形态学的路面裂缝检测

路面图像的复杂性及裂缝信息的弱信号性导致对路面裂缝进行检测非常困难,为此提出一种基于非下采样contourlet变换(NSCT)和图像形态学的路面裂缝检测算法.首先对图像进行NSCT得到不同尺度、不同方向上的变换系数,在NSCT域中根据变换系数自适应地确定阈值,并应用广义非线性增益函数来增强较弱细节的局部对比度;然后对增强处理后的变换系数进行反变换;最后用图像形态学方法和中值滤波实现裂缝检测及孤立噪声点去除.通过对实际的高速路面图像测试表明,与直方图增强、小波变换及contourlet变换相比,该算法能更有效地增强弱对比度的细小裂缝,克服了常规算法易受离散噪声点以及光照条件等干扰的问题,具有较强的鲁棒性且高效实用.     

Numerical study of the anisotropy of wave responses from a fractured reservoir using the grid-characteristic method

This work is aimed at studying the problem of subsurface disturbance propagation in a massive rock containing various inhomogeneities: empty or filled cracks. Numerical solutions to the problems of wave propagation in such substantially inhomogeneous media have been obtained. The dependence that the parameters of the response from a fractured reservoir have on the problem parameters is studied. The latter parameters are the density of the crack location, the fractured reservoir extent, amount of cracks, the initial disturbance location, crack inclination, and disturbance frequency. The concept of response anisotropy is introduced and the dependence that the anisotropy has on the abovementioned parameters is studied.     

The phase relations in the FeO-MgO-Al2O3-SiO2 system: assessment of thermodynamic properties and phase equilibria at pressures up to 30 GPa

Thermodynamic parameters for minerals in the FeO-MgO-Al₂O₃-SiO₂ (FMAS) system have been assessed from phase equilibria at pressures up to 30 GPa. All solid solution phases except orthopyroxene have been described by the sublattice model. Calculated phase diagrams of end member systems, exchange reactions in the FMAS system, phase equilibria for fixed compositions at pressures up to 30 GPa are in a reasonable agreement with the experimental data. The most critical data in the MgO-Al₂O₃-SiO₂ (MAS) system were compositions of garnet in equilibrium with Mg-rich ilmenite at 1273 K. To reproduce the experimentally observed Fe-Mg partition between perovskite and magnesiowustite, it is necessary to take into account the presence of Fe³⁺ ions in perovskite. The phase diagram of the FMAS system for pyrolite composition has been calculated in the temperature range of 1500 to 2400 K at pressures between 12 to 32 GPa.     

基于空间聚集特征的沥青路面裂缝检测方法

沥青路面裂缝自动检测是制约公路养护科学决策的最主要瓶颈.针对现有裂缝检测算法在大规模应用特别是广地域、多路况等复杂环境下算法稳定性、可靠性及实时性等方面存在严重不足问题.本文在观察大量实际工程路面图像基础上, 对路面裂缝特征进行全新定义, 提出了一种基于空间聚集特征的沥青路面裂缝检测方法, 参考裂缝的空间分布、灰度、几何等特征, 以子块图像为处理单元, 采用逐步求精的策略对子块图像进行分割, 快速定位空间聚集区域, 再对聚集区域进行评估得到信度高的裂缝候选区域; 最后以裂缝候选区域为种子区域, 在准确估算裂缝发展趋势的基础上, 结合裂缝片段聚集及相似性等特性, 去除噪声同时合并连接断裂的裂缝, 实现了裂缝区域较为完整的检测.通过测试多路况、多采集环境下近万样本, 并采用不同的方法对测试结果进行评估, 结果显示, 算法对不同类型路面图像中具有不同特征的裂缝区域均具有良好的检测性能, 裂缝定位准确性达到95%以上, 裂缝区域检测的完整性达到90%以上.     

瓦斯抽采钻孔漏气通道检测装置研制及应用

瓦斯抽采钻孔孔周裂隙和封孔段空隙通道是造成钻孔漏气失效的主要原因。为有效检测钻孔漏气通道,基于管流流体力学理论和漏气检测判别方法,研制了瓦斯抽采钻孔漏气通道检测装置。通过检测不同钻孔深度气样参数并分析其分布规律和突变情况,确定抽采钻孔失效原因和漏气通道位置;检测装置采用高稳压阻式压力传感器、激光甲烷传感器和荧光氧气传感器实现抽采负压、瓦斯浓度和氧气浓度检测,并采用1.5 m/节快接式25 mm薄壁不锈钢管作为取气管件,钻孔检测深度达30 m。现场应用结果表明,抽采管段检测参数变化稳定,说明抽采管未发生破损或接口漏气等,抽采管密封效果较好;在封孔段,距孔口9~18 m范围内存在多处不同程度的突变点,最大漏气通道在距孔口9~12 m范围内,说明原封孔深度不足,原封孔工艺无法有效密封漏气通道。将封孔深度增加至12 m,并采用“两堵一注”带压注浆封孔工艺,进行对比试验,结果表明,改进后试验钻孔整体抽采效果大幅改善,孔口瓦斯体积分数提升至55%以上,在距孔口12 m以深范围内瓦斯体积分数变化稳定,氧气体积分数近乎为0,漏气通道减少。试验结果验证了瓦斯抽采钻孔漏气通道检测装置能够有效检测漏气通道,为有针对性地调整封孔方式和相关参数及后续改进工作提供依据。     

Failure mechanisms of pressurized microchannels: model andexperiments

MicrochanneIs were created by fusion bonding of a Pyrex cover to a thermally oxidized silicon wafer, which contained anisotropically etched grooves. Such channels are frequently used in microfluidic handling systems, for example, in chemical analysis. Since in some of these labs-on-a-chip, in particular those used in liquid chromatography, the channels are subjected to high pressures of up to a few hundred bar, it is important to have information about the mechanical stability of the channel chip, in particular of the wafer bond involved in it. The latter is the subject of this paper. The maximum pressure that can be applied to several different channel chips was investigated experimentally. In order to find the relation among this maximum pressure, channel geometry, materials elasticity, and bond energy, an energy model was developed that is generally applicable to all types of wafer bonds. It was shown that the model is substantiated by the experimental pressure data, from which it could be calculated that the effective bond energy increased from 0.018 to 0.19 J/m² for an annealing temperature ranging from 310 to 470°C     

Decision on economical rail grinding interval for controlling rolling contact fatigue

Rail players around the world have been increasing axle loads to improve the productivity of freight and heavy haul operations. This has increased the risk of surface cracks at curves because of rolling contact fatigue. Rail grinding has been considered an effective process for controlling these cracks and reducing risks of rail breaks. The complexity of deciding the optimal rail grinding intervals for improving the reliability and safety of rails is because of insufficient understanding of the various factors involved in the crack initiation and propagation process. This paper focuses on identifying the factors influencing rail degradation, developing models for rail failures and analyzing the costs of various grinding intervals for economic decision making. Various costs involved in rail maintenance, such as rail grinding, downtime, inspection, rail failures and derailment, and replacement of worn‐out rails, are incorporated into the total cost model developed in this paper. Field data from the rail industry have been used for illustration.