Damage evolution analysis in mortar, during compressive loading using acoustic emission and X-ray tomography: Effects of the sand/cement ratio

This paper explores the use of acoustic emission (AE) and X-ray tomography to identify the mechanisms of damage and the fracture process during compressive loading on concrete specimens. Three-dimensional (3D) X-ray tomography image analysis was used to observe defects of virgin mortar specimen under different compressive loads. Cumulative AE events were used to evaluate damage process in real time according to the sand/cement ratio. This work shows that AE and X-ray tomography are complementary nondestructive methods to measure, characterise and locate damage sites in mortar. The effect of the sand proportion on damage and fracture behaviour is studied, in relation with the microstructure of the material.     

Using Electrical Capacitance Tomography to Investigate Gas Solid Contacting

An ECT system has been shown to be useful in observing differences in particle behaviour in a bubbling fluidised bed. An image analysis technique is further described that utilises the construction of solids concentration profiles at three key values, x = 0.2, x = 0.5 and x = 0.8, which have been identified as important solids concentrations in a new Bubble Structure Model. Significant differences in the bubble structure are shown to result when the concentration of fine particles is slightly increased in a bubbling fluidised bed. Changes of this type would seriously alter the gas solid contact efficiency in the fluidised bed, which would significantly influence selectivity of in‐bed catalytic reactions and gas residence time.     

The voidage function and effective drag force for fluidized beds

Here, an experimental investigation on the effective drag force in a conventional fluidized bed is presented. Two beds of different particle size distribution belonging to group B and group B/D powders were fluidized in air in a diameter column. The drag force on a particle has been calculated based on the measurement of particle velocity and concentration during pulse gas tests, using twin-plane electrical capacitance tomography. The validity of the voidage function “correction function”, (1−ε_s)ⁿ, for the reliable estimation of the effective drag force has been investigated. The parameter n shows substantial dependence on the relative particle Reynolds number , and the spatial variation of the effective static and hydrodynamic forces. It is also illustrated that, a simple correlation for the effective drag coefficient as function of the particle Reynolds number (Re_p), expressed implicitly in terms of the interstitial gas velocity, can serve in estimating the effective drag force in a real fluidization process. Analysis shows that, the calculated drag force is comparable to the particle weight, which enables a better understanding of the particle dynamics, and the degree of spatial segregation in a multi-sized particle bed mixture. The analogy presented in this paper could be extended to obtain a generalized correlation for the effective drag coefficient in a fluidized bed in terms of Re_p and the particle physical properties.     

盲环境下数字图像的自动修复算法研究*

当前所有的数字图像修复算法都是一种半自动的方法,因为在对图像修复前,都是由用户给出想要修复的区域,而算法无法自动获取照片的破损所在。提出了盲环境下图像自动修复的概念,以基于数字图像的样本修复为基础,改进了基于纹理合成的图像修复算法。首次运用数字图像可信性评估体系于图像修复,并通过可信性综合度量模型判断确定待修复区域。实验证明,该算法最终达到了盲环境下的图像自动修复效果。     

Micro-computed tomography scanning approaches to quantify,parameterize and visualize bioturbation activity in clogged streambeds: A proof of concept

Fine particle clogging and faunal bioturbation are two key processes co-occurring in the hyporheic zone that potentially affect hyporheic exchange through modifications in the sediment structure of streambeds. Clogging results from excessive fine sediment infiltration and deposition in rivers, and it is known to decrease matrix porosity and potentially reduce permeability. Faunal bioturbation activity may compensate for the negative effect of clogging by reworking the sediment, increasing porosity, and preventing further infiltration of fines. Although both processes of clogging and bioturbation have received significant attention in the literature separately, their combined effects on streambed sediment structure are not well understood, mostly due to the lack of a standard methodology for their assessment. Here, we illustrate a novel methodology using X-ray computed tomography (CT), as proof of concept, to investigate how, together, clogging and bioturbation affect streambed porosity in a controlled flow-through flume. By visualising gallery formations of an upward conveyor macroinvertebrate; Lumbriculus variegatus as a model species, we quantified bioturbation activity in a clogged streambed, focusing on orientation, depth, and volume at downwelling and upwelling areas of the flume. Gallery creation increased the porosity of the streambed sediment, suggesting a potential improvement in permeability and a possible offset of clogging effects. We illustrate the promising use of X-ray CT as a tool to assess bioturbation in clogged streambeds, and the potential role of bioturbation activity supporting hyporheic exchange processes in streambeds, warranting further studies to understand the extent of bioturbation impacts in natural systems.     

地震波层析成像探测技术与水平定向钻拉管技术在地下障碍物探测中的应用

顶管施工属于非开挖施工方法,是一种不开挖或者少开挖的管道埋设施工技术,该项技术彻底解决了管道埋设施工中对城市建筑物的破坏和道路交通的堵塞等难题,在稳定土层和环境保护方面凸显其优势。但是,在顶管施工过程中一旦遇到地下障碍物,将导致顶管顶力突然增大,顶管顶进方向偏离设计轴线,严重影响顶管施工的顺利进行。福州市某污水顶管施工需要穿越河道,通过前期资料收集发现,河道两侧可能存在护岸抛石等地下障碍物。为了查明护岸抛石的分布情况,综合应用了地震波层析成像探测技术与水平定向钻拉管技术进行探测,既为后期顶管施工的顺利进行提供了参考依据,又降低了探测成本,提高了探测效率。     

基于特征矢量化的肺结节特征选择算法*

针对肺结节良恶性分类模型中特征选择过程无法避免特征多样性不受破坏的问题,提出一种将肺结节特征矢量化处理的特征选择方法。首先,假设每个肺结节特征都是由数据、类型构成的一个矢量。然后,按照特征类型添加特征到相应的特征子集,并分别利用Relief算法评价特征、特征子集的分类重要性。最后,通过动态阈值的方式筛选得到优化后的特征子集。在150个肺结节样本的分类实验中,采用本文算法所取得的敏感性为94.7%、特异性为93.7%、虚警率为5.2%、受试者工作特性曲线下面积为97.3%。分析表明,本文算法几乎不破坏肺结节特征的多样性,能够显著提高肺结节良恶性分类的准确性。     

基于非参数假设检验的拓扑推断算法

针对基于门限比较的二叉树修剪拓扑推断算法稳健性差的问题, 提出了一种基于非参数假设检验的网络拓扑推断算法. 该算法首先应用经典的二叉树拓扑推断算法获得树状网络的二叉树结构, 然后应用维尔科克森秩和检验算法逐个判断二叉树中的每条内部链路是否需要修剪, 最后修剪二叉树, 删除所有需要修剪的内部链路, 得到真实的树状拓扑. 由于该算法使用统计检测的方法, 无需设置门限, 相对门限比较法具有更好的稳健性. 仿真实验表明, 该算法相比基于门限的二叉树修剪算法具有更高的推断精度.     

SADE-ELM电容层析成像流型辨识算法

针对电容层析成像反问题流型识别较难的问题,提出了一种新的ECT流型辨识算法——差分演化优化极端学习机算法,进而提出了基于自适应差分演化优化极端学习机(Sa DEEML)的ECT辨识算法.在论述极端学习机算法的基础上,结合差分演化算法对极端学习机算法进行优化,自适应差分演化算法中的关键参数,通过训练得到各类流型的分类器的参数,构造分类器进行精准与快速分类.实验结果表明:该算法能有效克服极端学习机算法的缺点并提高了局部与全局收敛能力,通过与BP、SVM算法比较,该算法具有竞争力,并为电容层析成像流型辨识的研究提供了新算法.     

一种多小波电容层析成像图像融合方法

针对电容层析成像系统的"软场"效应和病态问题对重建图像精度的影响,在分析电容层析成像基本原理和成像算法的基础上,基于滤波LBP和神经网络RBF特性,提出了一种基于多小波变换的ECT图像融合方法.该方法首先对待融合图像进行多小波分解,再分别采用最大值法和区域梯度法融合规则对分解后图像低频部分和高频部分进行图像的融合,最后经过多小波重构得到最终融合图像.仿真实验结果表明,融合后成像精确度得到明显提高,减少了误差,图像更接近原型,为ECT图像重建的研究提供了一个新的方法.