A key problem in gravity dam design is providing enough stability to prevent slide, and the difficulty increases if there are several weak structural planes in the dam foundation. Overload and material weakening were ...A key problem in gravity dam design is providing enough stability to prevent slide, and the difficulty increases if there are several weak structural planes in the dam foundation. Overload and material weakening were taken into account, and a .finite difference strength reserve method with partial safety factors based on the reliability method was developed and used to study the anti-slide stability of a concrete gravity dam on a complicated foundation with multiple slide planes. Possible slide paths were obtained, and the stability of the foundation with possible failure planes was evaluated through analysis of the stress distribution characteristics. The results reveal the mechanism and process of sliding due to weak structural planes and their deformations, and provide a reference for anti-slide stability analysis of gravity dams in complicated geological conditions.展开更多
针对多无人艇编队避障问题,对静态避障的路径消耗问题进行建模分析,在动态避障时提出一种偏置人工势场法使策略符合艇群国际海上避碰规则(swarm International Regulations for Preventing Collisions at Sea,sCOLREGS)。本方法首先对...针对多无人艇编队避障问题,对静态避障的路径消耗问题进行建模分析,在动态避障时提出一种偏置人工势场法使策略符合艇群国际海上避碰规则(swarm International Regulations for Preventing Collisions at Sea,sCOLREGS)。本方法首先对传统人工势场法进行改进,定义符合艇群会遇态势判断需求的sCOLREGS,通过速度障碍法实时判断碰撞风险,然后利用偏置斥力区域的改进人工势场法实现对规则的遵守。仿真实验表明,本文方法在障碍物与编队大小相当时可显著减少避障路程,在确保避障实时性的同时,较好地遵守了国际海上避碰规则相关条例。研究结论可为海面无人艇集群安全航行提供参考。展开更多
针对现阶段用电设备状态监测技术存在的处理速度较慢、准确率较低等问题,文中基于多突变点检测和模板匹配策略提出了一种用电设备在线状态监测方法。该方法在缓冲区模型和滑动窗口模型的基础上,利用多路搜索树突变点检测(Ternary Search...针对现阶段用电设备状态监测技术存在的处理速度较慢、准确率较低等问题,文中基于多突变点检测和模板匹配策略提出了一种用电设备在线状态监测方法。该方法在缓冲区模型和滑动窗口模型的基础上,利用多路搜索树突变点检测(Ternary Search Tree and Kolmogorov-Smirnov,TSTKS)算法形成窗口维度和缓冲区维度的特征向量,通过两种维度的模板匹配实现用电设备的运行状态匹配和状态切换时刻定位。基于家用电冰箱的仿真实验结果表明,所提方法具有检测速度快、准确率高等优点,可为用电设备状态监测领域提供参考。展开更多
针对在大规模时序医疗数据的分析中现有检测方法检测精度低、检测速度慢等问题,文中提出了一种基于深度学习的时序病变数据段分类方法。该方法在TSTKS(Ternary Search Trees and modified Kolmogorov-Smirnov)算法和滑动窗口理论的基础...针对在大规模时序医疗数据的分析中现有检测方法检测精度低、检测速度慢等问题,文中提出了一种基于深度学习的时序病变数据段分类方法。该方法在TSTKS(Ternary Search Trees and modified Kolmogorov-Smirnov)算法和滑动窗口理论的基础上,利用深度学习技术实现了对病变数据段的快速准确分类。文中以利用该方法对病变数据段进行分类的结果作为依据,实现了滑动窗口大小的动态调整。通过对真实癫痫脑电信号(Electroencephalogram,EEG)进行分析,证明了所提病变数据段分类方法和基于该分类方法的滑动窗口动态调整机制具有检测速度快、精度较高等优点,可以为大规模时序数据的快速分析研究提供一种新选择。展开更多
基金supported by the Innovation Program for College Graduate of Jiangsu Province of 2007 (Grant No. CX07B_133Z)
文摘A key problem in gravity dam design is providing enough stability to prevent slide, and the difficulty increases if there are several weak structural planes in the dam foundation. Overload and material weakening were taken into account, and a .finite difference strength reserve method with partial safety factors based on the reliability method was developed and used to study the anti-slide stability of a concrete gravity dam on a complicated foundation with multiple slide planes. Possible slide paths were obtained, and the stability of the foundation with possible failure planes was evaluated through analysis of the stress distribution characteristics. The results reveal the mechanism and process of sliding due to weak structural planes and their deformations, and provide a reference for anti-slide stability analysis of gravity dams in complicated geological conditions.
文摘针对多无人艇编队避障问题,对静态避障的路径消耗问题进行建模分析,在动态避障时提出一种偏置人工势场法使策略符合艇群国际海上避碰规则(swarm International Regulations for Preventing Collisions at Sea,sCOLREGS)。本方法首先对传统人工势场法进行改进,定义符合艇群会遇态势判断需求的sCOLREGS,通过速度障碍法实时判断碰撞风险,然后利用偏置斥力区域的改进人工势场法实现对规则的遵守。仿真实验表明,本文方法在障碍物与编队大小相当时可显著减少避障路程,在确保避障实时性的同时,较好地遵守了国际海上避碰规则相关条例。研究结论可为海面无人艇集群安全航行提供参考。
文摘针对现阶段用电设备状态监测技术存在的处理速度较慢、准确率较低等问题,文中基于多突变点检测和模板匹配策略提出了一种用电设备在线状态监测方法。该方法在缓冲区模型和滑动窗口模型的基础上,利用多路搜索树突变点检测(Ternary Search Tree and Kolmogorov-Smirnov,TSTKS)算法形成窗口维度和缓冲区维度的特征向量,通过两种维度的模板匹配实现用电设备的运行状态匹配和状态切换时刻定位。基于家用电冰箱的仿真实验结果表明,所提方法具有检测速度快、准确率高等优点,可为用电设备状态监测领域提供参考。
文摘针对在大规模时序医疗数据的分析中现有检测方法检测精度低、检测速度慢等问题,文中提出了一种基于深度学习的时序病变数据段分类方法。该方法在TSTKS(Ternary Search Trees and modified Kolmogorov-Smirnov)算法和滑动窗口理论的基础上,利用深度学习技术实现了对病变数据段的快速准确分类。文中以利用该方法对病变数据段进行分类的结果作为依据,实现了滑动窗口大小的动态调整。通过对真实癫痫脑电信号(Electroencephalogram,EEG)进行分析,证明了所提病变数据段分类方法和基于该分类方法的滑动窗口动态调整机制具有检测速度快、精度较高等优点,可以为大规模时序数据的快速分析研究提供一种新选择。
