天津大学学报(自然科学与工程技术版) /zrb/oa &nbsp;封一 /zrb/oa/darticle.aspx?type=view&id=202009000 2020年09月01 00:00 2020年09 0 792433 目录 /zrb/oa/darticle.aspx?type=view&id=202009001 2020年09月01 00:00 2020年09 0 132014 &nbsp;酸性土壤环境中多龄期埋地钢管地震易损性分析 /zrb/oa/darticle.aspx?type=view&id=202009002 <a name="_Hlk19517788"> </a> <a name="_Hlk18850986"> <span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">在酸性腐蚀土壤环境中,埋地钢管的抗震性能会随着服役龄期的延长而显著降低</span> <span> <span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">.为了研究酸性土壤环境中多龄期埋地钢管的抗震性能,对钢材进行人工加速腐蚀试验与拉伸破坏试验,建立了钢材屈服强度、极限强度、弹性模量及伸长率与失重率之间的关系;在全面腐蚀模型的基础上,揭示了酸性土壤环境中埋地钢管力学性能与几何尺寸随服役龄期的变化规律,并建立时变本构模型.基于</span> <span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">ANSYS</span> <span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">有限元分析软件,建立了酸性土壤条件中考虑锈蚀钢材几何尺寸及力学性能退化的不同服役龄期有限元模型;采用增量动力分析法及云图法,建立了描述地震动强度与结构地震响应之间关系的概率地震需求模型,结合管道破坏状态的划分和概率意义上极限损伤限值的确定,对埋地钢管进行地震易损性分析,从而建立了考虑锈蚀耐久性能退化的不同服役龄期埋地钢管时变地震解析易损性模型,表征了管道单元损伤指标与地震动强度指标之间的关系</span> <span> <span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,并进一步给出了不同管径多龄期埋地钢管的地震易损性曲线.</span> <span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">结果表明:在酸性腐蚀土壤环境中,随着服役龄期的延长,埋地钢管的几何尺寸及力学性能均不断削减,出现不同震害状态的概率呈现不同幅度的升高;在其他条件相同的情况下,随着管径增大,出现不同震害状态的概率有一定程度的降低.</span> </span> </span> </a> 2020年09月01 00:00 2020年09 881 889 1517115 &nbsp;贺金川<sup>1</sup>,韩 峰<sup>2, 3</sup>,郑山锁<sup>2, 3</sup>,谢孝奎<sup>2</sup>,蔡永龙<sup>3</sup> &nbsp;不同排水条件下桶形基础上拔承载特性研究 /zrb/oa/darticle.aspx?type=view&id=202009003 &nbsp;<span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">桶形基础因其安装方便、可重复使用等优点而广泛应用于海洋工程中,在服役期间或者退役回收时会受到上拔承载力作用,准确计算桶形基础的上拔承载力是保证基础稳定工作和顺利回收的重要前提,桶形基础的上拔承载力与桶的几何尺寸、上拔速率、土体参数以及在自重等预压荷载下的固结情况等因素有关.本文采用有限元的方法对桶形基础上拔承载特性进行研究,通过对不同长径比、上拔速率、土体参数及预压荷载等影响因素进行正交组合计算,研究了其不同上拔速率下破坏模式及负压发展规律,提出归一化上拔速率</span><i><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">vs</span></i><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">/<i>C</i><sub>v</sub></span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">其中</span><i><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">v</span></i><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">为实际上拔速率,</span><i><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">s</span></i><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">为排水路径,</span><i><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">C</span></i><sub><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">v</span></sub><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">为固结系数</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">对桶形基础上拔过程的排水条件进行划分.根据破坏模式及抗拔力各组分比重提出了桶形基础在不排水和完全排水条件下上拔承载力计算公式,在此基础上建立了归一化的上拔承载力与桶形基础几何形状</span><i><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">L</span></i><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">/<i>D</i></span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">、归一化上拔速率</span><i><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">v</span></i><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">s/<i>C</i><sub>v</sub></span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">以及土体强度之间的对应关系,并提出了部分排水条件下上拔承载力计算方法.结合算例给出了该成果在工程上的应用步骤和方法,成果可直接用于工程设计.</span> 2020年09月01 00:00 2020年09 890 899 1447636 &nbsp;孙立强<sup>1</sup>,刘政卿<sup>1, 2</sup>,齐玉萌<sup>1, 2</sup>,封晓伟<sup>3</sup> &nbsp;智能振动碾压机的自抗扰循迹控制方法 /zrb/oa/darticle.aspx?type=view&id=202009004 &nbsp;<span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">循迹控制是实现智能振动碾压机自动化工作的重要环节.为了改善智能振动碾压机在非结构化路面环境下的循迹效果,解决现有控制算法存在的调参工作量大以及鲁棒性低等问题,本文提出了一种基于扰动观测的反步控制方法.首先,建立了振动碾压机的整车运动学模型和一种具有预测功能的误差动态模型.然后,基于一台改装的智能振动碾压机平台研究了外界扰动对车辆的铰接角度、航向角度及位置的影响规律,发现外界扰动对碾压机的铰接角、航向角度的影响需要实时观测估计,而对碾压机轮位置的影响可通过姿态信息修正补偿.最后,在以上研究的基础上设计了扩张状态观测器用于实时估计系统内外总扰动,从而将误差动态模型简化为串联积分型系统,并基于李雅普诺夫函数设计控制器的反馈控制律.</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Simulink</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">和硬件在环的仿真结果表明:在车辆的铰接角、航向角同时受内部、外界扰动作用的情况下,循迹连续弯道曲线中距离、角度和曲率误差的最大值分别为</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">0.08<span style="LETTER-SPACING: 2pt">8</span>m</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">、</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">0.11<span style="LETTER-SPACING: 2pt">8</span>rad</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">和</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">0.04<span style="LETTER-SPACING: 2pt">2</span>m</span><sup><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正书宋简体">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">1</span></sup><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">.说明基于扰动观测的反步控制方法展现了良好的控制精度和鲁棒性.</span> 2020年09月01 00:00 2020年09 900 909 1686630 &nbsp;谢 辉,赵龙同,阮迪望 &nbsp;Tandem-GMAW 电弧增材制造基层成型宽度研究及热过程分析 /zrb/oa/darticle.aspx?type=view&id=202009005 &nbsp;<span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">电弧增材制造网格壁板以逐层堆积的方式直接在基板上沉积网格结构,能够极大地提升生产效率并降低成本.在网格壁板的成型过程中,基层的成型和整体的热过程决定了上层是否能成功完成成型.本文针对不同板厚作为基板的基层成型不规律问题,利用不同环境因素下的试验得到其成型熔宽及余高数据,根据得到的数据建立有限元填充模型.采用有限元模拟方式研究基板厚度、预热温度、距边缘距离等关键参数对电弧增材制造网格壁板的影响规律.试验结果表明:随着基板厚度的增加,基层成型宽度明显减小;随着预热温度的升高,在</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">8<span style="LETTER-SPACING: 2pt">0</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">℃</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">温度以内宽度增加不明显,当预热温度超过</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">12<span style="LETTER-SPACING: 2pt">0</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">℃</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,基层成型宽度出现明显的增加;在边缘距离大于</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">2<span style="LETTER-SPACING: 2pt">0</span>mm</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">时,边缘距离对成型宽度基本没有影响;边缘距离小于</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">2<span style="LETTER-SPACING: 2pt">0</span>mm</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">时,随着边缘距离的减小,成型宽度逐渐加大,在边缘距离为</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">1<span style="LETTER-SPACING: 2pt">0</span>mm</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">时整个边缘基板的温度达到</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">500°</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,基板塌陷倾向明显.最后通过上述试验结果,应用热力学理论构建基层成型数学模型,结合统计学最小二乘法,辅助确定热源参数后采用有限元模拟的方法可以进行不同工艺参数及板厚的热过程研究,这种创新性结合热力学解析解、统计学方法和简单实测解,可以有效地减少电弧增材基层热过程分析时间,提升成型研究效率.</span></span></span> 2020年09月01 00:00 2020年09 910 916 1538744 &nbsp;石俊彪<sup>1, 2</sup>,赵 昀<sup>1, 2</sup>,陈树君<sup>1, 2</sup>,迟 杏<sup>3</sup> &nbsp;微通道流动沸腾过程中异态相干沸腾的强化传热研究 /zrb/oa/darticle.aspx?type=view&id=202009006 &nbsp;<span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">随着电子器件的集成化和小型化,其散热量超过</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">1<span style="LETTER-SPACING: 2pt">0</span>MW/m<sup>2</sup></span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">将成为现实,这超出了目前大功率系统中使用的单相冷却方案的上限,所以必须再次开发新的冷却方案.克服单相传热局限性的一种方法是转变为两相沸腾传热,而临界热流密度又是所有沸腾传热的上限值.因此,为了提高微通道内流动沸腾传热的临界热流密度,本文设计开发了非均匀导热性传热板.通过将两种不同导热性能的材料</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">铜和聚四氟乙烯</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">交替布置在靠近传热表面的传热板内,实现了传热表面的非均匀温度分布和异态相干沸腾模式</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">核态沸腾与膜态沸腾共存且相互干涉的状态</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">.同时搭建了微通道流动沸腾实验系统,其微通道截面尺寸为</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">1.8<span style="LETTER-SPACING: 2pt">4</span>mm</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: Symbol">&acute;</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">70.0<span style="LETTER-SPACING: 2pt">0</span>mm</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,通道长度为</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">280.<span style="LETTER-SPACING: 2pt">0</span>mm</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,传热板表面尺寸为</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">10.<span style="LETTER-SPACING: 2pt">0</span>mm</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: Symbol">&acute;</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">10.<span style="LETTER-SPACING: 2pt">0</span>mm</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,流体工质为去离子水.在不同入口流速</span><i><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">v</span></i><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">=</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">0.<span style="LETTER-SPACING: 2pt">1</span>m</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">/s</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">、</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">0.<span style="LETTER-SPACING: 2pt">2</span>m</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">/s</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">、</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">0.<span style="LETTER-SPACING: 2pt">4</span>m</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">/s</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">和不同过冷度</span> <span style="FONT-SIZE: 9pt; FONT-FAMILY: Symbol">D</span><i><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">T</span></i><sub><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">sub</span></sub><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">=</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">10.<span style="LETTER-SPACING: 2pt">0</span>K</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">、</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">20.<span style="LETTER-SPACING: 2pt">0</span>K</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">、</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">30.<span style="LETTER-SPACING: 2pt">0</span>K</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">条件下,研究了非均匀导热性传热板在微通道流动沸腾中的传热强化效果.结果表明,相对于单纯的核态沸腾状态,异态相干沸腾状态能够有效地提升流动沸腾传热的临界热流密度.此外,改变入口流速和过冷度对临界热流密度有明显影响且趋势相同,减小入口流速和过冷度都会增大临界热流密度的提升比例.在本文的实验条件范围内,在水的流速</span><i><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">v</span></i><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">=</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">0.<span style="LETTER-SPACING: 2pt">1</span>m</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">/s</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">、过冷度</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: Symbol">D</span><i><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">T</span></i><sub><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">sub</span></sub><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">=</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">10.<span style="LETTER-SPACING: 2pt">0</span>K</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">的条件下,实现了最高约</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">43.4</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">的临界热流密度提升比例.</span> 2020年09月01 00:00 2020年09 917 923 1135308 &nbsp;宇高義郎 <sup>1, 2, 3</sup>,郭嘉翔<sup> 1, 2</sup>,陈志豪<sup>1, 2</sup> &nbsp;高硅铝合金高温微观力学行为研究 /zrb/oa/darticle.aspx?type=view&id=202009007 &nbsp;<span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">高硅铝合金以其优异的性能在电子封装领域具有广阔的应用前景,准确地测量其高温微观力学行为具有重要意义.本文基于数字图像相关</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">DIC</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">方法,对</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Al</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">27</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Si</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">、</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Al</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">42</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Si</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">、</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Al</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">60</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Si 3</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">种不同硅含量高硅铝合金的拉伸试样开展了扫描电子显微镜</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">SEM</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">下的高温原位拉伸实验研究.分析了在</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">20</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">~</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">30<span style="LETTER-SPACING: 2pt">0</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">℃</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">下测得的</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">3</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">种合金的应力</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">应变曲线、微尺度全场应变分布规律以及拉伸试样的断口形貌.结果表明,硅的含量和温度对高硅铝合金拉伸力学行为具有显著的影响.随着温度的升高,</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">3</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">种合金的应变量均逐渐增大,其中</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Al</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">27</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Si</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">的应变量变化最大.</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">3</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">种合金的抗拉强度随温度的升高均近似呈线性趋势降低,常温下</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Al</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">27</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Si</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">最高,</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">20<span style="LETTER-SPACING: 2pt">0</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">℃</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">以上时</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Al</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">42</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Si</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">最高,</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Al</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">60</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Si</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">合金的抗拉强度最低.在</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Al</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">27</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Si</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">与</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Al</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">42</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Si</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">合金的应变场中的铝基体相内部出现了明显的应变集中现象,而</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Al</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">60</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Si</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">的应变分布较均匀.温度对</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">3</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">种合金的微尺度拉伸变形场分布规律影响不大.合金的拉伸断口形貌表明,随着硅含量的增加,高硅铝合金主要的拉伸断裂机制由铝相的韧性断裂逐渐转变为硅相的脆性断裂,而温度对其影响较小.</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">3</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">种高硅铝合金在不同温度下拉伸时均无明显的屈服现象,也未出现颈缩现象.</span></span></span> 2020年09月01 00:00 2020年09 924 931 4315355 &nbsp;陈金龙<sup>1, 2</sup>,王 滢<sup>1</sup>,孙翠茹<sup>1, 2</sup> &nbsp;基于改进SOM 的壁画图像裂缝自动识别与修复 /zrb/oa/darticle.aspx?type=view&id=202009008 &nbsp;<a name="_Hlk15129045"><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">针对壁画中存在的裂缝这一常见病害,提出一种基于人工神经网络</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">ANN</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">的自组织映射</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">SOM</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">图像修复算法,将人工智能技术应用于古建筑壁画修复领域.基于壁画裂缝本身的线性结构特征,对图像进行多尺度形态学边缘梯度检测提取边缘信息,使得裂缝边界区域灰度变化剧烈,从而达到边界突出的效果;对变换后的图像进行自适应阈值分割处理,以保证图像中每个像素点都属于目标区域;选取面积作为目标区域的连通规则进行度量以去除虚假目标,达到精确提取的目的,实现对破损像素的自动识别和标注;对壁画中已标注的破损区域采用改进的</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">SOM</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">算法进行修复,通过</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">SOM</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">聚类对图像进行分层,在单个图层中迭代计算出破损像素的值,实现对图像的并行化分层修复,在保障修复精度的同时提升修复的速率;合并图层,完成标注区域修复部分;最后通过对</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">3</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">种类型裂缝的壁画修复,本文所提出的改进</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">SOM</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">算法在修复图像峰值信噪比</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">PSNR</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">、特征相似度</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">FSIM</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">等</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">4</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">类指标显著提升,并且修复时间平均缩短</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">40.34</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,表明方法对于古建筑壁画裂缝修复的有效性和优越性.</span></a> 2020年09月01 00:00 2020年09 932 938 2941272 &nbsp;杨 挺<sup>1</sup>,王双双<sup>2</sup>,盆海波<sup>1</sup>,王兆霞<sup>3</sup> &nbsp;基于LightGBM 和DNN 的智能配电网在线拓扑辨识 /zrb/oa/darticle.aspx?type=view&id=202009009 &nbsp;<a name="OLE_LINK7"></a><a name="OLE_LINK4"></a><a name="OLE_LINK3"><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体"><span>为提高电网的安全运行水平和经济性,灵活的可重构的网络拓扑结构是未来智能配电网的基本特征.而配电管理系统</span><span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">DMS</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">中的大部分功能,如状态估计,潮流计算和电压控制等,都基于网络当前的拓扑结构.因此,拓扑辨识是</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">DMS</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">的基础功能之一,研究更为高效和准确的智能配电网拓扑辨识方法具有重要意义.结合配电网的结构和运行特点,建立了基于机器学习的智能配电网拓扑辨识框架,并提出了基于</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">LightGBM</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">和深度神经网络</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">DNN</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">的配电网在线拓扑辨识方法.该方法借助</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">LightGBM</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">实现特征选择,筛选出对配电网拓扑辨识最有效的少量量测,以深度神经网络实现配电网运行断面量测数据与其拓扑结构间的映射.</span></span><a name="OLE_LINK10"></a><a name="OLE_LINK9"><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体"><span><span>考虑实际应用中可能存在量测数据丢失的情况,提出了基于最小方差的缺失值填补方法.同样利用样本间的最小方差进行未知拓扑的甄别,并借助增量学习机制,通过增量训练</span></span><span><span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">DNN</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">模型实现拓扑知识库的更新.</span></span></span><span><span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">与现有方法相比,本文提出的拓扑辨识方法仅需要配电网中少量节点的运行断面量测数据,同时适用于辐射状和弱环网结构,计算效率可支持在线拓扑辨识.通过</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">IEEE 33</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">节点配电网和</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">PG&amp;E 69</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">节点配电网验证了所提方法的有效性与优越性,<a name="OLE_LINK12"></a><a name="OLE_LINK11">并分析了对于不同噪声水平情况、量测特征值缺失和存在未知拓扑的适应性</a>.</span></span></span></span></a></span></a> 2020年09月01 00:00 2020年09 939 950 1802133 &nbsp;裴宇婷,秦 超,余贻鑫 &nbsp;基于分布式能源的煤矿带式输送机与蓄电池协同经济调度策略 /zrb/oa/darticle.aspx?type=view&id=202009010 &nbsp;<span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">带式输送机是煤矿中运输煤炭的重要设备.为了解决煤矿带式输送机电费成本过高的问题,基于带式输送机变速控制技术,将分布式可再生发电设备引入到带式输送系统运行过程当中;通过对带式输送机带速、煤流量的协调控制,实现设备的满载运行和分布式能源出力的本地消纳;进而引入分时电价,通过对带式输送机运行参数和蓄电池充放电功率的协同优化引导转移设备高电价时期负荷,实现带式输送机能耗和运行成本的降低.首先对含风机、光伏和蓄电池的煤矿带式输送机系统进行建模,建立带式输送机动态能耗模型,构建以运行成本最低为目标、以分布式能源发电消纳为约束的带式输送机与蓄电池协同经济调度策略.算例验证表明:应用所提控制策略可显著减少带式输送机能耗和设备磨损,并节省设备电费成本达</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">38.39</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">.在此基础上分析了蓄电池容量和带式输送机带速变化率对调度结果的影响,结果表明合理的储能容量和带速变化率取值对带式输送系统经济运行至关重要.</span> 2020年09月01 00:00 2020年09 951 958 798532 &nbsp;穆云飞<sup>1</sup>,姚泰昂<sup>1</sup>,贾宏杰<sup>1</sup>,赵 波<sup>2</sup>,张雪松<sup>2</sup>,倪筹帷<sup>2</sup> &nbsp;国家安全视角港口基础设施保障能力评价 /zrb/oa/darticle.aspx?type=view&id=202009011 &nbsp;<a name="OLE_LINK16"></a><a name="OLE_LINK15"><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">港口</span><span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">本文中的港口均指沿海民用港口</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">作为连接陆与海战略通道的重要节点,在国家安全面临威胁时极可能被征用,其集疏能力与保障的时效性关乎国家安全军事行动的进程乃至成败</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">.本研究定位于沿海港口应对领土与主权纠纷等特殊时期、特殊任务需求的运输保障能力评价.针对该领域缺少定量研究及需求分析的现状,笔者通过“需求与供给”的定量对比分析,从而明确国家安全视角的港口基础设施保障能力,综合特殊时期的任务需求、特殊的运输品类,以及对港口基础设施的特殊要求,设定了“供给</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">需求”双侧指标评价体系,确定了港口特殊评价内容,选定了评价指标,建立了“供给</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">需求”双侧指标评价模型,并以沿海</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">4</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">个区域港口作为案例开展实证研究,<a name="OLE_LINK4"></a><a name="OLE_LINK3">计算得到了各区域港口的大规模兵力投送能力指数、军民综合运输保障能力指数及疏港交通能力指数.通过比较分析</a><a name="OLE_LINK6"></a><a name="OLE_LINK5">,一方面验证了方法的有效可用性,另一方面确定了各区域港口基于国家安全的运输保障能力的短板.针对短板制定了相应的改进方案</a>,局部改进后,北部区大规模兵力投送能力指数从</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">0.444</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">提升到了</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">0.747</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">;南部区该指数从</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">0.368</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">提升到了</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">0.776</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,二者的军民综合运输保障能力指数调整后可接近最大值</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">1.000</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,该方案的提出为下一步港口基于国家安全的规划建设提供决策依据.</span></span></a> 2020年09月01 00:00 2020年09 959 966 949455 &nbsp;张 静<sup>1, 2</sup>,李小冬<sup>1</sup>,张智慧<sup>1</sup>,王 瑞<sup>3</sup> &nbsp;基于改进MOGOA 的无人机群航迹规划研究 /zrb/oa/darticle.aspx?type=view&id=202009012 &nbsp;<span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">针对电子侦察系统中反辐射无人机群进行辐射源无源</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">定位时机群</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">的</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">编队形式会对定位精度产生影响的问题,将克拉美</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">罗界</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Cramer-Rao lower bound</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">CRLB</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">作为定位精度方面的优化目标,与其他优化目标、约束一起引入机群的航迹规划中,使无人机群运动过程中保持良好编队,确保无源定位精度.文中针对多优化目标复杂环境中航迹规划算法寻优能力不高的问题,提出了一种基于改进多目标蝗虫算法</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">IMOGOA</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">的无人机群</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">3</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">维航迹规划方法,通过对</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">MOGOA</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">的选择方式、收敛参数进行改进从而提高算法的收敛性能以及全局搜索性能.首先,建立无人机群航迹规划的运动学模型,并引入距离约束,除定位精度以外还引入了路程、威胁代价等作为航迹规划的优化目标函数,然后,对改进多目标蝗虫算法进行详细说明,最后设计基于</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">IMOGOA</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">的无人机群航迹规划方案的算法流程,并在设定场景中对该算法的性能进行了仿真分析.结果表明,所提出的</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">IMOGOA</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">能够成功地规划出无人机群从初始位置到辐射源位置处的</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">3</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">维航迹,同时使无人机群在运动过程中保持良好的定位精度,经</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">IMOGOA</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">规划的机群编队定位精度最高可达</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">1.2</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,性能明显优于正方形编队和随机编队,并通过将</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">IMOGOA</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">与原始蝗虫算法</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">GOA</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">、原始多目标蝗虫算法进行对比,结果表明</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">IMOGOA</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">的收敛速度比</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">MOGOA</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">快</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">11.1</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,搜索性能相较</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">GOA</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">提升</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">13.8</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">.</span> 2020年09月01 00:00 2020年09 967 975 1593575 &nbsp;陈 涛<sup>1, 2</sup>,李由之<sup>1, 2</sup>,黄湘松<sup>1, 2</sup> &nbsp;嵌入DenseNet 结构和空洞卷积模块的改进YOLO v3 火灾检测算法 /zrb/oa/darticle.aspx?type=view&id=202009013 &nbsp;<span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">为解决现有火灾检测算法无法同时满足高检测率、低误报率以及高实时性的检测需求的问题,提出了一种基于卷积神经网络的改进</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">YOLO v3</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">目标检测算法,通过深度卷积神经网络自动提取火焰特征对全图进行多尺度特征图预测</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">. </span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">首先,针对网络公开火灾数据集数量较少、场景种类受限、火焰尺度单一等问题,自建了一个包含</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">1<span style="LETTER-SPACING: 2pt">3</span>573</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">张火灾图片的火灾数据集用于对模型进行训练和测试,其中训练集图片</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">1<span style="LETTER-SPACING: 2pt">0</span>014</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">张,测试集图片</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif; LETTER-SPACING: 2pt">3</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">559</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">张</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">. </span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">接着,为了提升网络对于多尺度目标</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">(</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">尤其是小尺度目标</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 宋体">)</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">火焰的特征提取效果,通过在原</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">YOLO v3</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">的特征提取网络</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">Darknet</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">53</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">中嵌入空洞卷积模块以充分利用上下文信息,扩增感受野的同时保证不丢失特征图的分辨率</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">. </span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">此外,在特征提取网络中加入</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">DenseNet</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">密集型连接网络结构单元,以增强特征复用,同时缓解深度卷积神经网络在特征传播过程中的梯度消失问题</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">. </span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">该改进的特征提取网络相比原网络层数进一步加深,网络参数量显著减少</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">. </span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">结合火灾检测任务需求实际,简化了损失函数,加快了网络的收敛速度</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">. </span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">实验结果表明:该算法检测速度快,检测精度高,不仅能够实时检测大尺度火焰,对于火灾发生初期的小尺度火焰也同样检测灵敏,其检测速度可达</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">26.0</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">帧</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">/s</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,精确率可达</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">97</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,且在多种复杂光照环境下均能良好地抑制误报</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">.</span> 2020年09月01 00:00 2020年09 976 983 1685742 &nbsp;张 为,魏晶晶 &nbsp;基于改进VGG-16 模型的英文笔迹鉴别方法 /zrb/oa/darticle.aspx?type=view&id=202009014 &nbsp;<span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">笔迹鉴别是通过对待测文本和样本笔迹的相似度进行比较,来判定笔迹是否相同的一种检验技术,其在司法鉴定、法庭科学以及金融领域合同确认等多个领域都有广泛的应用.传统英文笔迹鉴别方法是通过比对被鉴别文本与模板的相似程度来实现,效率低,准确度差.近年来,随着深度神经网络技术的飞速发展,利用其自主学习的优势提取相关特征,可以大大提高笔迹鉴别的准确率.传统</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">VGG</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">16</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">模型在图像分类上一直表现良好,但由于网络结构一直采用顺次连接的方式,导致训练时间过长,参数调整难度大,且不能很好地提取图像的细微特征,因此对笔迹鉴定的效果不够理想.本文通过对传统</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">VGG</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">16</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">卷积神经网络模型进行改进,提出了一种</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">CC</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Times New Roman&quot;,serif">-</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">VGG</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">网络模型,利用复合卷积层替换部分卷积层,实现了手写体英文笔迹的自动鉴别.在公开的</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">CVL</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">和</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">ICDAR2013</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">数据集上,该模型取得了较好的鉴别效果,平均正确率分别达到</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">92.7</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">和</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">86.9</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">%</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,与现有算法相比准确率均有所提高.此外,建立了一个包含</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">130</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">类、共</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">2<span style="LETTER-SPACING: 2pt">6</span>000</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">张图片的手写英文笔迹图像数据集</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: &quot;Time New Roman&quot;,serif">EI130</span><span style="FONT-SIZE: 9pt; FONT-FAMILY: 方正楷体简体">,在该数据集上该模型也取得了较高的准确率.与其他算法的对比实验证明了本文算法在训练时间上具有优越性;此外,在多个数据集上的实验结果也证明了本文算法的有效性和先进性.</span> 2020年09月01 00:00 2020年09 984 990 1764002 &nbsp;何 凯,马红悦,冯 旭,刘 坤