A study on determining hardening curve for sheet metal

The hardening curve for sheet metal can be determined from the load–displacement curve of tensile specimen with rectangular cross-section. The previous researches, however, have paid little attention to its use in large deformation. Moreover, it varies with materials, deformation conditions and so on, and there are not enough hardening curves available in manuals. In order to study metal behavior, it is very important to establish a method to create a large strain hardening curve based on the normal mechanical test. In this paper, two new kinds of specimens are proposed, one is a multi-stepped specimen that can be traced to the three-stepped specimen, and the other is a tapered specimen which decreases the complexity of the multi-stepped specimen in manufacture. In this study, circle grids are imposed on the specimen surface to calculate true strains at different positions of the specimen. It is found that the load added to the different segments of specimen just before fracture can be determined by the maximum load and breaking load. True strains and corresponding true stresses can be determined after the specimen is pulled to fracture, so the hardening curve can be easily achieved. After a great deal of experiment, the results show that the tapered specimen has almost the same key parameters as the multi-stepped specimen, and the former is more easily used. Meanwhile, the work-hardening exponent (n) and the coefficient of normal anisotropy (r) can be obtained conveniently, and the forming limit line can also be approximately induced.     

基于接纳控制的智能电网需求响应

采用效用函数刻画用户的用电满意度, 将需求响应问题建模为一类凸优化问题. 针对电力供应商的供电量不能满足用户最小用电需求的问题, 结合分布式用电量调度和实定价, 设计两类接纳控制算法. 仿真结果表明, 通过接纳控制, 满足了购电用户的最小用电需求, 保证了用户的用电质量, 能够实现电网的供需平衡.     

改进的计算网格域内实体信任模型设计

传统基于信任域的网格信任模型以自治域内的实体数目作为信任值计算复杂度的唯一参数,安全性较低。在现有计算网格信任模型基础上,提出一种改进的自治域内实体信任模型,将时间衰减和惩罚因子引入到模型中,建立GSP对用户的直接信任度和推荐信任度计算模型,通过仿真实验验证其具有更高的抗Whitewashing攻击能力。     

窄带电力噪声干扰下的G3标准分组检测算法的研究

G3标准OFDM(正交频分复用)多载波PLC(电力线通信)技术应用于智能电网时,必须解决恶劣电力环境下的数据同步问题。当电力信道存在电力背景噪声及随机脉冲噪声时,经典的延时相关分组检测算法对数据起点检测存在较大误差。针对电力噪声的特点,为提高检测的准确度,将微分算法应用于延迟相关算法,提出新的分组检测算法。在电力噪声模型下,通过对经典算法及新的算法进行性能评测比较,验证了新算法对恶劣环境的适应性。     

面向属性的参数化建模在Revit API中的应用

为改进目前构件级建模过程存在的效率低问题,提出面向属性的参数化建模过程.此方法使建模师不需要重复传统建模操作,通过设置模型属性参数就可以实现相关数据的统计和计算从而快速建模.以球面网格结构为例,将参数化建模思想融入到Revit二次开发技术中,借用VS开发工具实现三维模型的快速生成.在操作时,用户只需通过用户界面输入矢高、跨度和环数等结构参数即可以完成模型的快速建立.通过此程序,还可以实现肋环形和施威德勒型2种类型的选择,实时监测设置的参数是否满足规范要求.     

未来大数据环境下的配用电通信网虚拟网络架构及应用

根据智能配电网的通信需求,本文提出了基于无线 Mesh 网和电力线通信网的智能配电网通信框架。在此框架的基础上,设计了一个遗传算法和启发式算法,为不同类型的业务分别建立虚拟网络并将其映射到异构的物理网络中,通过增加传输分集来保证实时业务的可靠性。通过仿真,验证两种映射算法在保证实时业务可靠性的基础上,使尽力而为型业务的吞吐率达到最大化。     

浅谈110 kV变压器的维护与检修

作为电网的核心电气设备,变压器会对电网的运行产生重大影响。为了确保变压器的可靠、安全运行,电力企业需要对变压器进行日常维护与检修。阐述了变压器常见异常状况和日常维护的具体内容,以供参考。     

超级计算中心客户细分及差异化服务策略研究

本文以中国科学院计算机网络信息中心超级计算中心（以下简称超级计算中心）客户服务工作为研究对象,运用 K-means 算法对客户进行细分,进而对每类客户群提出相应的差异化服务策略。实施差异化服务策略可以更好地分配资源、提供更有效的客户服务。