Liquid membranes for gas/vapor separations

A review on developments of liquid membranes (LMs) in the field of gas and vapor separation of the last 16 years is presented. Liquid membrane configurations employing supports, i.e. immobilized, supported and contained liquid membranes are focussed and detailed information on the respective materials, i.e. supports (supplier, type, thickness, pore width, porosity, tortuosity), liquids and carriers, are presented together with their specific separation tasks. Performance of different LMs in terms of permeability and selectivity as well as stability (duration of testing, applied differential pressures) are compared and discussed. Finally, different preparation methods of LMs are illustrated.     

基于改进的CNN-LSTM的DGA域名检测算法

近年来,网络安全问题层出不穷,其中僵尸网络是造成网络瘫痪的重要原因之一。僵 尸网络利用域名生成算法（DGA）生成大量恶意域名进行网络攻击,对网络安全造成威胁。现有的 DGA域名主要分为字典型和字符型,传统的深度学习方法无法同时检测出两种类型的DGA域名, 尤其是无法检测出基于字典的DGA域名。针对这个问题,本文提出了改进的CNN-LSTM的DGA 域名检测算法,该算法融合了卷积神经网络（CNN）、注意力机制和双向长短时记忆网络（BiLSTM）, 可以同时检测出两种类型的DGA域名。最后进行了不同算法的对比实验,实验结果表明,与其他 深度学习模型相比,该算法提高了DGA域名的二分类和多分类的准确率和F1值。在多分类实验 中,通过改进损失函数,提高了小样本数据的域名检测率。     

最小代数的同调和上同调

计算出了域上的最小代数的同调和上同调,并作为应用发现了比∧代数还要简单的DGA与Steen- rod代数具有相同的上同调．     

Compact corigid objects in triangulated categories and co-<Emphasis Type="Italic">t</Emphasis>-structures

In the work of Hoshino, Kato and Miyachi, [11], the authors look at t-structures induced by a compact object, , of a triangulated category, , which is rigid in the sense of Iyama and Yoshino, [12]. Hoshino, Kato and Miyachi show that such an object yields a non-degenerate t-structure on whose heart is equivalent to Mod(End()^op). Rigid objects in a triangulated category can the thought of as behaving like chain differential graded algebras (DGAs). Analogously, looking at objects which behave like cochain DGAs naturally gives the dual notion of a corigid object. Here, we see that a compact corigid object, , of a triangulated category, , induces a structure similar to a t-structure which we shall call a co-t-structure. We also show that the coheart of this non-degenerate co-t-structure is equivalent to Mod(End()^op), and hence an abelian subcategory of .      

变压器在线监测技术的应用

作为一项成熟的检测技术变压器油中溶解气体气相色谱分析（DGA）在有效检测变压器各类故障方面发挥了毋容置疑的作用。与此同时变压器在线溶解气体检测技术的应用已成为确保变压器安全运行、状态检修工作有效开展的一个重要前提,对此本文简要介绍了某厂目前变压器在线溶解气体监测技术的应用状况及成功监测到#1主变故障分析。     

基于Passive DNS的速变域名检测

利用Passive DNS采集校园网真实运行环境的域名访问记录, 从域名的多样性、时间性、增长性和相关性等方面构建18个特征集, 提出基于随机森林算法来识别速变域名的模型。交叉验证实验表明, 所构建的模型对域名分类的准确率超过90%。在所采集的数据集上, 所构建的模型比FluxBuster 能更有效地识别速变域名。     

一种基于深度学习的快速DGA域名分类算法

提出了一种基于深度学习的CNN-LSTM-Concat快速DGA域名分类算法,使用多层一维卷积网络对域名字符进行序列化处理,LSTM网络层用于强化获取字符间长距离依赖关系。通过将LSTM的多序列输入转化为单向量输入,在保证检测性能的前提下,能够大幅提高训练和检测速度。实验证明,我们的方法对DGA域名分类的准率在公开数据集上达到98.32%。同时,在准确率相比主流的LSTM方法更高的情况下,检测时间比LSTM方法快6.41倍。     

Americium and plutonium separation by extraction chromatography for determination by accelerator mass spectrometry

A simple method was developed to separate Pu and Am using single column extraction chromatography employing N,N,N′,N′-tetra-n-octyldiglycolamide (DGA) resin. Isotope dilution measurements of Am and Pu were performed using accelerator mass spectrometry (AMS) and alpha spectrometry. For maximum adsorption Pu was stabilized in the tetra valent oxidation state in 8 M HNO₃ with 0.05 M NaNO₂ before loading the sample onto the resin. Am(III) was adsorbed also onto the resin from concentrated HNO₃, and desorbed with 0.1 M HCl while keeping the Pu adsorbed. The on-column reduction of Pu(IV) to Pu(III) with 0.02 M TiCl₃ facilitated the complete desorption of Pu. Interferences (e.g. Ca²⁺, Fe³⁺) were washed off from the resin bed with excess HNO₃. Using NdF_3, micro-precipitates of the separated isotopes were prepared for analysis by both AMS and alpha spectrometry. The recovery was 97.7 ± 5.3% and 95.5 ± 4.6% for ²⁴¹Am and ²⁴²Pu respectively in reagents without a matrix. The recoveries of the same isotopes were 99.1 ± 6.0 and 96.8 ± 5.3% respectively in garden soil. The robustness of the method was validated using certified reference materials (IAEA 384 and IAEA 385). The measurements agree with the certified values over a range of about 1–100 Bq kg⁻¹. The single column separation of Pu and Am saves reagents, separation time, and cost.     

电力变压器热故障的在线诊断技术

提出了基于变压器表面的红外热图像对变压器热故障进行在线诊断的方法。利用红外热像仪实时拍摄变压器表面的热图像,建立其温度分布场。根据变压器的表面温度、内部结构参数、物理特性参数和环境温度,建立其传热模型,对变压器内部铁芯故障点温度进行反演计算,通过与离线DGA诊断结果进行比较,证明了该方法的可行性,实现了变压器热故障的非接触、在线诊断。