基于人工神经网络的智能化架空线路应变快速解调方法北大核心

为了提高智能化光纤复合架空线路态势感知的实时性,将人工神经网络方法应用于光纤沿线应变解调,确定了神经网络的结构。编程实现了基于洛伦兹模型的最小二乘谱拟合方法和神经网络方法,采用不同信噪比和布里渊频移的布里渊谱训练神经网络,将它们应用于某光纤复合架空线路沿线光纤应变的测量,从不同角度比较了两种方法的计算结果。计算结果表明,神经网络方法能有效获得光纤沿线的布里渊频移进而获得应变,具有与谱拟合方法相似的准确性,但应变解调时间仅约为谱拟合方法的1/20000。研究结果为提高智能光纤复合架空线路态势感知的实时性提供了参考。     

MP PFGun脉冲爆燃压裂技术在阿曼SSM-XX井的应用

简要叙述了MP PFGun技术工作原理、系统结构特点以及PropFrac脉冲压裂模拟器软件基本功能,并对SSM-XX井的模拟计算P-T曲线和井下压力仪实测曲线进行了对比。结果表明:SM-XX井采用MP PFGun电缆输送工艺安全,脉冲压裂持续时间长,且利用PropFrac模拟软件指导施工设计,取得了较好的应用效果。     

基于CiteSpace的资源型产业研究知识图谱分析

资源型产业发展为推进国家经济增长和工业化进程提供了重要保障。为深入了解资源型产业的研究情况，以CNKI数据库为数据源，搜集2000—2020年关于资源型产业的核心及以上期刊论文，利用CiteSpace软件从发文作者与研究机构分布、关键词共现网络和时区图谱等方面，绘制知识图谱，进行可视化分析。研究发现：资源型产业领域的研究成果愈加丰富，但研究群体间联系合作较少，且现有的合作研究主要集中在所处地域资源富集和具有学科优势的研究机构及学者；资源型产业领域的研究热点可概括为产业发展、资源型城市、产业集群、产业结构、产业链和产业集聚等方面；针对资源型产业领域未来可从资源型产业相关理论研究、创新发展模式和可持续发展等方面深入展开。     

Green Measurements for Software Product Based on Sustainability Dimensions

Software is a central component in the modern world and vastly affects the environment’s sustainability. The demand for energy and resource requirements is rising when producing hardware and software units. Literature study reveals that many studies focused on green hardware; however, limited efforts were made in the greenness of software products. Green software products are necessary to solve the issues and problems related to the long-term use of software, especially from a sustainability perspective. Without a proper mechanism for measuring the greenness of a particular software product executed in a specific environment, the mentioned benefits will not be attained. Currently, there are not enough works to address this problem, and the green status of software products is uncertain and unsure. This paper aims to identify the green measurements based on sustainable dimensions in a software product. The second objective is to reveal the relationships between the elements and measurements through empirical study. The study is conducted in two phases. The first phase is the theoretical phase, where the main components, measurements and practices that influence the sustainability of a software product are identified. The second phase is the empirical study that involved 103 respondents in Malaysia investigating current practices of green software in the industrial environment and further identifying the main sustainability dimensions and measurements and their impact on achieving green software products. This study has revealed seven green measurements of software product: Productivity, Usability, Cost Reduction, Employee Support, Energy Efficiency, Resource Efficiency and Tool Support. The relationships are statistically significant, with a significance level of less than 0.01 (p = 0.000). Thus, the hypothesised relationships were all accepted. The contributions of this study revolve around the research perspectives of the measurements to attain a green software product.     

Effect of CdO addition on photon,electron, and neutron attenuation properties of boro-tellurite glasses

This research article aims to study the effect of CdO addition on the radiation shielding characteristics of boro-tellurite glasses in the composition of 50B₂O₃ - (50-x) TeO2- xCdO, where x = 0, 10, 20, 30, 40 and 50 mol%. These glasses were exposed to gamma radiation and the transmitted gamma photons were evaluated for energies varying from 15 keV to 15 MeV using Geant4 simulation toolkit. The number of transmitted photons was then used to characterize the gamma shielding for the studied glasses in terms of linear/mass attenuation coefficients, MFP, Z_eff, and HVL. The simulation outcomes were theoretically confirmed by using Phy-X software. The beta (electron) shielding characterization of the involved glasses was also investigated by determining the projectile range and stopping power using ESTAR software. Additionally, the fast neutron shielding characterization of the glasses was achieved by evaluating removal cross-section (Σ_R). The results reveal that the CdO has a small influence on the shielding performance of the boro-tellurite glasses against gamma, beta, and neutron radiations. The shielding performance of the boro-tellurite glasses was compared with that of common shielding materials in terms of MFP. It can be concluded that the boro-tellurite glasses regardless of the concentration of CdO content have promising shielding performance to be used for radiation applications.     

微纳层叠聚丙烯腈凝胶流动特性的数值模拟研究

建立了层叠流道的三维模型和有限元网格模型，根据流变测试数据，采用Polymat对物料的黏度模型参数进行拟合，并利用Polyflow软件对聚丙烯腈（PAN）凝胶在层叠流道内的三维等温流动过程进行了数值模拟分析。研究发现，当入口流量增大时，层叠流道出口速度的不均匀性增加；沿流动方向流道内压力逐渐降低，并在出口处降低至同一最低值；流道进出口压力差与入口流量大小具有正相关性；在流道的中心截面上剪切速率分布均匀，波动较小。     

基于自适应阈值活动语音检测和最小均方误差对数谱幅度估计的低信噪比降噪算法

针对低信噪比(SNR)环境下传统方法对声信号降噪的局限性，提出了一种联合自适应阈值活动语音检测（VAD）算法和最小均方误差对数谱幅度估计（MMSE-LSA）的实时降噪算法。首先，在VAD算法中通过基于能量概率最大值的概率统计来对背景噪声进行估计，对得到的背景噪声进行实时更新并保存；然后，将实时更新的背景噪声作为MMSE-LSA的参考噪声，并对噪声幅度谱进行自适应更新，最后进行降噪处理。通过在真实场景中对四类声信号进行实验，结果表明，该算法在保证对低SNR声信号的实时处理的情况下，相较于传统MMSE-LSA算法，降噪信号的SNR能够提高10~15 dB，且不存在信号过减的情况，可应用于实际工程。     

数据驱动的机床等待过程节能方法研究

为了降低机床等待过程中的能耗，提出了一种实时数据驱动的机床等待时间预测与节能控制方法。首先，建立了射频识别驱动的生产进度评估方法，并以生产进度数据作为输入，构建了基于堆栈降噪自编码的机床等待时间预测模型；其次，依据预测的机床等待时间，提出了机床状态切换方法，以降低机床能耗；最后，通过一个电梯零部件制造车间的案例分析，表明该方法的预测误差仅为4.1%，同时将机床等待过程能耗降低了57%，实现了制造车间的节能减排。     

热冲压工艺参数对22MnB5马氏体钢汽车B柱性能影响的有限元模拟

通过有限元仿真软件Autoform分析了热冲压过程中工艺参数的变化对22MnB5马氏体钢B柱起皱、回弹、减薄、马氏体量以及强度的影响。结果表明:22MnB5马氏体钢B柱热冲压最优化工艺参数为加热温度930 ℃,冷却速率80 ℃/s。该工艺参数下,热冲压过程各处均完成马氏体转变,硬度分布均匀,材料减薄率较低,热冲压成形效果好,尺寸精度高,冲压件强度均大于1400 MPa。     

R2D2: A scalable deep learning toolkit for medical imaging segmentation

Deep learning has gained a significant popularity in recent years thanks to its tremendous success across a wide range of relevant fields of applications, including medical image analysis domain in particular. Although convolutional neural networks (CNNs) based medical applications have been providing powerful solutions and revolutionizing medicine, efficiently training of CNNs models is a tedious and challenging task. It is a computationally intensive process taking long time and rare system resources, which represents a significant hindrance to scientific research progress. In order to address this challenge, we propose in this article, R2D2, a scalable intuitive deep learning toolkit for medical imaging semantic segmentation. To the best of our knowledge, the present work is the first that aims to tackle this issue by offering a novel distributed versions of two well-known and widely used CNN segmentation architectures [ie, fully convolutional network (FCN) and U-Net]. We introduce the design and the core building blocks of R2D2. We further present and analyze its experimental evaluation results on two different concrete medical imaging segmentation use cases. R2D2 achieves up to 17.5× and 10.4× speedup than single-node based training of U-Net and FCN, respectively, with a negligible, though still unexpected segmentation accuracy loss. R2D2 offers not only an empirical evidence and investigates in-depth the latest published works but also it facilitates and significantly reduces the effort required by researchers to quickly prototype and easily discover cutting-edge CNN configurations and architectures.