基于ACO-PAM综合算法的电力负荷聚类分析

负荷特性分类与综合是实现负荷模型实用化的关键.为建立合适的变电站负荷模型,将聚类方法引入负荷特性分析,提出一种基于ACO-PAM的综合聚类算法.该综合算法是PAM算法对蚁群的历史最优位置进行聚类分析,将此位置代替PAM的参考点,作为新的聚类中心,数据将自适应地加入到适合它的聚类中.ACO算法具有全局搜索能力强、易于与其...     

大气电场仪观测结果的修订

大气电场仪是测量大气电场的重要设备,其测量结果受多种因素影响.通过一系列对比试验表明:电场仪采用探头正置或倒置结构时的观测结果基本一致,倒置结构有利于避免雨水带来的影响.不同的安装环境对测量结果的影响很大,为了消除这种影响,提出了对观测数据进行地面修订的方法,得到不同天气条件下的修订系数.所得结果对于实现多站电场仪联网观测的数据一致性有重要实际意义.     

水网密集区粘土路基当量回弹模量的预估分析

基于非饱和土基本理论,利用基质吸力及土水特征曲线的最新研究成果,对受地下水位控制的粘土路基的平衡湿度状态进行了预估分析;同时采用室内重复动三轴试验方法,建立了非饱和粘性路基土动回弹模量的双线性本构经验预估模型。将上述研究成果相结合,创建了综合考虑路基湿度和应力状态的路基当量回弹模量的预估方法,并结合实体工程建立了以地下水位、路基填筑高度为核心的路基当量回弹模量预估方程。通过与试验路上传统测试方法测试结果的对比分析,表明该评价方法具有准确、可靠特点。研究成果从平衡湿度和应力状态耦合控制的角度,对水网密集区粘土路基当量回弹模量的确定提供了新的视角和途径。     

High-g impact resistance for epoxy molding compound

The behavior of resistance high-g impact of EMC (epoxy molding compound) with two package models, small outline package (POS) and Globtop, was evaluated by experimental method used Hopkinson bar. At 120,000 g (generated in the Hopkinson bar with widths about 70 μ s) no damage in either the POS devices or the Globtop devices was observed. In order to enhance the EMC’s ability of resistance high-g impact, buffering effect of epoxy resin was also studied. The experimental results above all show that EMC has a better performance of impact resistance at about 120,000 g, and epoxy resin can absorb the stress wave to have the protected ability. The study of this paper could serve as a basis for selection packaging materials and enhance its reliability in high-g impact environment.     

基于神经网络的面波迭代反演应用研究

根据滩浅海近地表结构特征,尝试利用地震记录中的面波进行近地表结构研究,以便了解滩浅海地区的近地表地层介质结构变化,为深层油气勘探提供准确的低降速带资料.针对面波频散反演已有方法存在的不足,引入一种基于BP神经网络的迭代反演方法对面波的频散曲线进行拟合迭代,用于反演预测滩浅海低降速带地层参数.由于神经网络具有很强的自学习、自适应、自组织和容错能力,它的反演预测能力非常强大,能够较精确地预测出所要求解的目标数据,同时结合传统迭代反演方法的优点,增强了该方法的反演预测能力.通过对滩浅海近地表结构模型试算,获得好的效果,同时进一步对实际记录进行了计算,也取得了比较满意的结果.     

向家坝二期截流设计及施工

向家坝水电站二期截流流量大、河床覆盖层深厚、龙口流速较大、抛投强度高,截流施工难度大。根据工程特点采取上游单戗双向进占的截流方式,以右岸进占为主,左岸为辅,龙口平抛护底。通过水力学计算成果,预留龙口宽80m,龙口分3个区进占,采取合理的堤头进占方式,保证抛投强度,确保了二期截流一举成功。     

基于直觉模糊推理的网络舆情预警方法*

为解决网络舆情预警等级问题,提出了一种基于直觉模糊推理的网络舆情预警方法。借鉴战场态势分析思想,对网络舆情态势分析的原理进行了阐述,选取了适合计算机实现的七个网络舆情态势分析模式对预警等级进行判断。选择七个舆情话题进行实验,实验结果表明,该方法能够准确地估计出威胁等级,符合专家经验判断,说明该方法是可行的。     

一种基于圆形角点的瞳孔定位算法*

提出了一种利用圆形角点定位瞳孔的算法。该算法主要在定位到人眼区域的基础上,利用改进的SUSAN算法寻找圆形角点;然后根据梯度放缩计算角点向量的梯度值,最后通过筛选函数输出瞳孔中心位置及亮瞳面积,准确地定位到瞳孔,并根据亮瞳的大小定位到瞳孔区域中心。实验结果表明本算法是准确可靠的。     

Effect of coagulation temperature on structures and properties of poly(<Emphasis Type="Italic">p</Emphasis>-phenylene benzobisoxazole) (DHPBO) fibers during dry-jet wet-spinning process

The effect of coagulation temperature on the morphology, microstructures and mechanical properties of dihydroxy poly(p-phenylene benzobisoxazole) (DHPBO) fibers was investigated during dry-jet wet-spinning process, in which the coagulation bath concentration and drawn ratio were kept as 10 wt% of PPA in water and 1.7, respectively. The structures and mechanical properties of the as-spun DHPBO fibers were characterized by FTIR, XRD, SEM, and single fiber tensile testing. The results indicated that in PPA/H₂O coagulation system, when the coagulation temperature was 25°C, highly crystallized DHPBO as-spun fibers possessing fine crystallites, circular and smooth morphology, and excellent mechanical properties could be achieved. Supported by the National Natural Science Foundation of China (Grant No. 50673017), Shanghai Leading Academic Discipline Project (Grant No. B603) and the Program of Introducing Taleuts of Discipline to University of People’s Republic of China (“111” Program) (Grant No. 111-2-04)