【作者】 黄绪勇；

【导师】 刘沛； 苗世洪；

【作者基本信息】 华中科技大学 ， 电力系统及其自动化， 2008， 博士

【摘要】 计算机技术、传感器技术、微机电技术和现代无线通信技术的飞速发展,推动了无线传感器网络技术的产生和发展。无线传感器网络提供了一种新的感知世界、信息处理的平台。无线传感器网络由大量的集成数据采集、处理,并且能够自组织的进行协调和通信的节点组成,提供了一种全新的数据获取和处理的方法,广泛应用于军事国防、工农业控制、城市管理、生物医疗、抢险救灾、危险区域远程控制、生产制造等诸多领域。本文结合国家自然科学基金项目(基金号:50677024),以当代电力系统网络化、信息化技术高速发展的现状为前提,开展较大规模的无线传感器网络在电力系统中应用的相关理论和基础研究。本文在以下几个方面进行了基础性研究,为电力系统网络提供另一种可靠、优质的通信平台打下前期的理论基础:(1)基于对QoS(可靠性、延时等)的需求,用于电力系统中无线传感器网络的性能必须得到实验验证。到目前为止,网络系统性能的验证都是通过仿真软件来实现,比如说NS、OPNeT等等。通过仿真软件来验证,有许多实际的因为无法得到考虑,例如:障碍物、电磁干扰对性能的影响等等。本文首先应用NDIS(Network Driver Interface Specification)技术和Passthru框架,构建了一个无线自组网络实验平台,给出了平台的特征和应用前景。通过该平台,可以利用PC、移动PC构建一个无线自组网络,在现场环境下实验网络的性能。(2)无线传感器网络拓扑控制技术是无线传感器网络的关键技术之一。无线传感器网络拓扑控制控制的主要目标是降低网络的能量损耗,增加网络的生存周期,减少节点间的通信干扰,为路由提供基础。本文提出了一种基于局部信息的离散分布式拓扑控制算法(A Localized Discrete Distribution Topology-control Algorithm,LDDTA)。LDDTA可以大大地减少无线传感器网络的冗余链接。其基本思想是:首先,将无线传感器网络离散成许多分离网络子单元,称为孤立网络单元(Isolated Network Island, ISI)。这些ISI都具有有限节点度,且具有最佳的连接质量。然后,通过优化算法将各个ISI连接起来,形成一个完成的拓扑结构。仿真结果显示,该算法具有良好的性能。(3)能量约束是无线传感器网络在实际应用中面临的关键挑战之一。实际应用中的无线传感器网络节点的供电有两种方式:电池供电或者从周围环境中取得能量。通过光电池、振动或者磁感应线圈等装置可以从环境中取得能量。无论是电池供电还是从环境中取得供应能量,由于无线传感器节点尺寸的限制,节点的平均供应功率是十分有限的。通过分析用于电力系统中的传感器节点的工作模式,建立了一种无线传感器节点的能量耗费模型,基于一种混合式遗传算法,对节点能量耗费进行了优化,从而为应用于电力系统中的无线传感器网络的工作模式提供了一种有效的参考。同时分析了基于GA的无线传感器网络最优化能量任务分配。(4)在电力系统故障检测和定位方面,给出了一种用于电力系统故障检测的无线传感器网络体系结构、安装和配置方案。分析了基于无线传感器网络的电力系统故障检测机制、定位方法和节点的采样机制,研究分析了用于电力系统故障检测和定位的无线传感器网络可靠性。提出了一种面向事务的可靠性机制。(5)无线传感器网络是一个分布式网络系统,时间同步是分布式系统协同工作的重要基础。本文基于应用于电力系统故障检测的无线传感器网络系统,给出一种新的时间同步方案。当电力系统在发生故障时,在线路上会产生向线路两端传播的行波。利用故障时的行波信号,设计了一种新的无线传感器网络时间同步方案。仿真实验结果表明,该时间同步方案能满足电力系统故障检测对时间同步性的要求。(6)协同问题是无线传感器网络在实际应用中的关键问题之一,根据无线传感器网络自身的特点,结合多智能体理论,研究了无线传感器网络的协同方法。同时,给出了用于电网故障诊断的无线传感器网络组织结构,基于无线传感器网络分布式决策和协同,给出了利用无线传感器网络进行电网故障诊断的方法。更多还原

【Abstract】 The fast evolution of computer technology, sensor technology, MEMS and modern wireless communication technology has impelled development of wireless sensor networks. Wireless sensor networks provide a new platform to sense the world, process the data. It is widely applied in lots of fields, such as military defense, industry control, agriculture control, city management, biomedicine, rushing to deal with an emergency, rescue, remote control to danger zone, manufacture et al. Wireless sensor networks are consist of a large number of nodes which can be capable of communicating, computing and cooperation in ad hoc model. Combined with national natural science foundation of China and based on the status quo of the power system informationization , some basis research of large-scale wireless sensor networks applied in power system are carried through..In this dissertation, the following topics are investigated for providing a new reliable, high quality communication platform:(a) For the requirement of QoS (Reliability, delay etc) in power system, performance of wireless sensor networks should be testified by experiments. Till recently, performance of communication networks are mostly tested by simulation software, such as NS, OPNET etc. But a lot of practical factors can’t be taken into account in simulation software. For examples: obstacles, electromagnetic interferences, and so on. Utilized NDIS(Network Driver Interface Specification) intermediate driver and Passthru architecture, a simulation testing platform of wireless ad hoc networks is constructed. The characteristic and application prospects of the platform are introduced. By the platform, the performance of a wireless ad hoc network can be validated only adopting PC or portable computer.(b) Topology control is one of key problem of a wireless sensor network. The primary aim of the topology control is to reduced the energy consumption, increase the lifetime, decrease the communication interference of the network, and provide groundwork for routing protocols. A localized discrete distribution topology-control algorithm (LDDTA) of wireless sensor networks is presented. The LDDTA strongly decreased the redundancy links of wireless sensor networks. Firstly, the topology links of the wireless sensor networks is discreted and formed a number of isolated network island topologies, which are called isolated network island (ISI). These network island have the bounded node degree and the best link qualities for each node. Then, all isolated network islands are linked by intersubgraph algorithm which minimizes the power energy consumption between the isolated network islands. Simulation results show advantages of the algorithm.(c) Energy constraint of wireless sensor network is also one of the most demanding challenges in actual applications. Generally, there are two ways to supply the power of wireless sensor network: supply by batteries or obtain energy from surroundings. Energy can be obtained from the surroundings by photovoltaic cells, vibration, or an iron core winding which surrounds a line conductor. For the constraint of node’s size, the average supplying power of nodes is very limited. By analyzing the working model of sensor nodes in power system, an energy consumption model is constructed and a hybrid genetic algorithm is used to optimize the energy consumption. The model provides an effective reference of working pattern for wireless sensor networks in power system. At the same time, optimal energy method for task allocation is analyzed by GA.(d) An architecture, installation and deployment scheme of wireless sensor network applied in power system is presented for power fault detection and location. Fault detection, location and sampling mechanism of wireless sensor networks are investigated in power system. The reliability of wireless sensor networks in power system is studies and a reliability scheme oriented to transactions is presented. (e) Time synchronization is an important infrastructure for coordinated work to a wireless sensor network because it is a distribution system. Aimed at the wireless sensor network applied in power system fault detection, a new time synchronization scheme is presented. When a fault occurs in the distribution power system, traveling-waves will come into being which spread along the power lines. The new time synchronization scheme is based on the traveling-waves. Performance analyses and simulation results demonstrate that the proposed scheme meet the requirements in power system.(f) Collaboration is one of the key problem for a wireless sensor network in practice. Based on the characteristic of the wireless sensor network and multi-agents theory, collaboration technique is studied. Utilized organization of the wireless sensor network and based on distribution decision-making, a new power system fault diagnostic scheme is presented.更多还原