期刊文献+
共找到7篇文章
< 1 >
每页显示 20 50 100
对分布式潮流协调计算收敛机理的一种新分析 被引量:2
1
作者 高明 姚建国 +2 位作者 张文斌 杨胜春 袁玲 《江苏电机工程》 2008年第4期58-62,共5页
异步迭代模式下的分布式潮流分解协调计算外层迭代格式的构造取决于对分解协调计算收敛机理的分析和认识。将基于同心松弛概念的池塘理论加以扩展,用来分析边界等值注入误差在潮流计算中对各节点产生的影响。在此思路下提出了4个铺垫引... 异步迭代模式下的分布式潮流分解协调计算外层迭代格式的构造取决于对分解协调计算收敛机理的分析和认识。将基于同心松弛概念的池塘理论加以扩展,用来分析边界等值注入误差在潮流计算中对各节点产生的影响。在此思路下提出了4个铺垫引论,据其对分解协调算法的收敛机理从同心松弛概念这一新角度做出了理论分析,并在此基础上提出了确定协调计算合并参数的新方法——直接选取法,再将其和原有方法进行了比较讨论,给出了已有的对收敛机理理论解释之间的联系。IEEE9、39、118节点系统上进行的数值仿真实验验证了本文的分析以及直接选取法的优点。 展开更多
关键词 分布式潮流计算 协调计算 收敛机理 合并参数
下载PDF
多流多科目实时数据处理算法的设计与实现
2
作者 王灏 山寿 郝朝 《计算机测量与控制》 2020年第3期246-249,共4页
随着飞行试验测试技术的快速发展,单架次需要处理的试飞科目越来越多,而多科目处理参数分布在不同的网络流中,这对机载实时数据处理的效率带来严峻的挑战;利用"合并-处理-分流"的方法,设计一种多流多科目网络数据实时处理算法... 随着飞行试验测试技术的快速发展,单架次需要处理的试飞科目越来越多,而多科目处理参数分布在不同的网络流中,这对机载实时数据处理的效率带来严峻的挑战;利用"合并-处理-分流"的方法,设计一种多流多科目网络数据实时处理算法;该算法将多科目处理参数按照参数名、采样率等信息进行整合形成合并参数列表;在实时处理时根据合并参数信息实时处理多网络流中的数据,并将处理结果存储在数据缓存中;最后将多科目数据进行分流即可得到处理结果;测试结果表明,该算法可实际应用于多流多科目数据处理中,大幅提高实时处理的效率,同时可以得到准确的数据处理结果。 展开更多
关键词 多数据流 多科目 参数合并 实时处理 数据分流
下载PDF
基于YOLOv3的变速箱油封座缺陷检测
3
作者 张晓昇 方凯 徐鑫 《湖北汽车工业学院学报》 2022年第1期70-74,共5页
针对小型汽车变速箱后油封座缺陷检测精度低、检测速度慢、目标不敏感等问题,提出改进YOLOv3检测算法。在YOLOv3算法基础上,将第11层浅层特征与网络深层特征融合,生成尺度为104×104的新特征图层,提取更多小缺陷目标特征;再将BN层... 针对小型汽车变速箱后油封座缺陷检测精度低、检测速度慢、目标不敏感等问题,提出改进YOLOv3检测算法。在YOLOv3算法基础上,将第11层浅层特征与网络深层特征融合,生成尺度为104×104的新特征图层,提取更多小缺陷目标特征;再将BN层参数合并到卷积层,共用连通数据减少占用显存等空间。结果表明:文中算法能有效识别有缺陷油封座,准确率提高了1.25倍;加快了模型前向推移速度,检测速度提高了6 f·s^(-1)。 展开更多
关键词 缺陷检测 YOLOv3 参数合并
下载PDF
Microcalorimetric study on host-guest complexation of naphtho-15-crown-5 with four ions of alkaline earth metal 被引量:1
4
作者 宋明芝 朱兰英 +2 位作者 高希柯 窦建民 孙德志 《Journal of Zhejiang University-Science B(Biomedicine & Biotechnology)》 SCIE EI CAS CSCD 2005年第1期69-73,共5页
Thermodynamic parameters of complexation of naphto- 15-crown-5 with four alkaline earth ions in aqueous media was determined using titration microcalorimetry at 298.15 K. The stability of the complexes, thermal effect... Thermodynamic parameters of complexation of naphto- 15-crown-5 with four alkaline earth ions in aqueous media was determined using titration microcalorimetry at 298.15 K. The stability of the complexes, thermal effect and entropy effect of the complexation is discussed on the basis of the guest ions structure and the solvent effect. The stability constants tendency to vary with ion radius was interpreted. Complex of naphtha-15-crown-5 with calcium ion is very stable due to the synergism of static electric interaction and size selectivity between the host and the guest. 展开更多
关键词 MICROCALORIMETRY Host-guest complexation Naphtho-15-crown-5 Alkaline earth metal Molecular recognition
下载PDF
The Application of the Combined Method for Selection of Optimal Excitation Parameters
5
作者 Roman Petrichenko Inga Zicmane +1 位作者 Michal Kolcun Matus Novak 《Journal of Energy and Power Engineering》 2014年第4期724-728,共5页
The quality as well as reliability of electrical energy transmitted to consumers is one of the main parameters of successful operation of the power system. The searching of optimal coefficient's combination of PSS (... The quality as well as reliability of electrical energy transmitted to consumers is one of the main parameters of successful operation of the power system. The searching of optimal coefficient's combination of PSS (power system stabilizer) is the main goal of this article. The possibility of application of the new combined approach for the optimal excitation's settings search is presented. MC (Monte Carlo) method, in order to search and select the optimal combination of excitation system, was applied. The proposed method has been researched with a mathematical model of the power system. This model has been built using Matlab/Simulink software. Paper shows advantages and disadvantages of the proposed methods. 展开更多
关键词 Power systems stabilizer excitation parameters Monte Carlo method.
下载PDF
基于深度学习的铸件缺陷检测 被引量:8
6
作者 鲍春生 谢刚 +1 位作者 王银 谢新林 《特种铸造及有色合金》 CAS 北大核心 2021年第5期580-584,共5页
提出一种YOLOv3-Mv2的检测模型,对铸件划痕缺陷进行检测。首先,选择更小的骨架网络MobileNetv2替换原来的特征提取网络结构Darknet53,以减少网络的计算参数量,提升检测速度;其次,在深层特征和浅层特征相融合的基础上增加新的检测尺度,... 提出一种YOLOv3-Mv2的检测模型,对铸件划痕缺陷进行检测。首先,选择更小的骨架网络MobileNetv2替换原来的特征提取网络结构Darknet53,以减少网络的计算参数量,提升检测速度;其次,在深层特征和浅层特征相融合的基础上增加新的检测尺度,以增强对小缺陷目标的检测能力;然后将BN(Batch Normalization)层的参数重新计算,与卷积层共用连通区域数据,加快模型前向推断的速度;引入CIoU(Complete IoU)函数提高定位准确度。结果表明,YOLOv3-Mv2与原YOLOv3算法相比,平均精度(mAP)值提高了5.42%,实时性提高了23 f/s。 展开更多
关键词 缺陷检测 MobileNetv2 合并参数 检测尺度 损失函数
原文传递
Multi-objective parameter optimization for a single-shaft series-parallel plug-in hybrid electric bus using genetic algorithm 被引量:4
7
作者 CHEN Zheng ZHOU LiYan +2 位作者 SUN Yong MA ZiLin HAN ZongQi 《Science China(Technological Sciences)》 SCIE EI CAS CSCD 2016年第8期1176-1185,共10页
Recently, the single-shaft series-parallel powertrain of Plug-in Hybrid Electric Bus (PHEB) has become one of the most popu- lar powertrains due to its alterable operating modes, excellent fuel economy and strong ad... Recently, the single-shaft series-parallel powertrain of Plug-in Hybrid Electric Bus (PHEB) has become one of the most popu- lar powertrains due to its alterable operating modes, excellent fuel economy and strong adaptability for driving cycles. Never- theless, for configuring the PHEB with single-shaft series-parallel powertrain in the development stage, it still faces greater challenge than other configurations when choosing and matching the main component parameters. Motivated by this issue, a comprehensive multi-objectives optimization strategy based on Genetic Algorithm (GA) is developed for the PHEB with the typical powertrain. First, considering repeatability and regularity of bus route, the methods of off-line data processing and mathematical statistics are adopted, to obtain a representative driving cycle, which could well reflect the general characteristic of the real-world bus route. Then, the economical optimization objective is defined, which is consist of manufacturing costs of the key components and energy consumption, and combined with the dynamical optimization objective, a multi-objective op- timization function is put forward. Meanwhile, GA algorithm is used to optimize the parameters, for the optimal components combination of the novel series-parallel powertrain. Finally, a comparison with the prototype is carried out to verify the per- formance of the optimized powertrain along driving cycles. Simulation results indicate that the parameters of powertrain com- ponents obtained by the proposed comprehensive multi-objectives optimization strategy might get better fuel economy, meanwhile ensure the dynamic performance of PHEB. In contrast to the original, the costs declined by 18%. Hence, the strat- egy would provide a theoretical guidance on parameter selection for PHEB manufacturers. 展开更多
关键词 multi-objective parameter optimization single-shaft series-parallel powertrain plug-in hybrid electric bus (PHEB) genetic algorithm (GA) driving cycle city bus route
原文传递
上一页 1 下一页 到第
使用帮助 返回顶部