【作者】 郑维刚；

【导师】 朴在林；

【作者基本信息】 沈阳农业大学 ， 农业电气化与自动化， 2014， 博士

【摘要】 我国电网建设飞速发展,用电负荷增长迅速,用电保障方面也有了很大的提高,不过相比于发达国家还有一定差距。国内经济发达地区和欠发达地区的保障水平也有很大区别。红外测温技术是及早发现设备外部过热故障和内部绝缘故障的重要手段,也是进行检查电网设备状态的先进检测方法。应用红外测温技术发现电力设备外部连接点的发热隐患,在电力系统中已有了广泛应用。当故障一旦发生,就要以恢复电网正常为第要务。要及时的检查,避免故障的发生,科学的安排电网线路检修计划,真正的了解电网,发现隐患,减少运行风险是应该放在首位的。要解决电网巡检中人工作业和常规检测的不足,提出了无人机系统搭载红外影像设备对电网进行飞行巡检的方案。由于无人机具有机动性强,速度快,到位率高,可控性强等特点,更适合现代智能电网下的巡检要求,该系统引入小型无人直升机为平台,搭载摄像设备和红外影像采集设备进行影像数据搜集工作。综合应用ARM嵌入式技术对无人机搭载的设备进行融合,设计并实现了无人机端数据发送与接收,红外影像的采集与传输,地面站的监控与指令发送等功能,达到巡检时能够实时保持三维地理信息系统和无人机端的信息同步的要求。飞行巡检带来方便的同时也有一定局限性,需要根据其飞行成本,结合区域电网的特性准确建模,建立约束条件来控制成本,发挥一次飞行的最大效率。对此论文提出以电网维修收益最大和检修费用最少为目标,在电力线路的故障率与检修状态上建立关系,构建了多目标双层优化的无人机巡检电力线路的模型,将传统的多目标转化为单目标求解的方法改进为多目标粒子群优化(MOPSO)算法,直接得到基于经济目标和可靠目标的双层优化模型的解集,优化解集收敛过程。利用线路风险值和维修费用的双重约束来确定无人机的巡检范围和巡检的顺序,为检修人员提供科学的辅助决策。通过Matlab对模型的仿真和算法求解,并对比NSGA-Ⅱ算法在同一区域的求解,验证了该模型和算法在求解线路优化巡检方式上能够准确快速得到最优解集。对于以往红外采集设备多采用人工和固定位置的方式监控和拍摄电力设备的传统,以及无人机端使用红外影像设备产生的不稳定的红外图像采集状态,论文提出利用质心的目标图像定位,红外图像分割和目标设备图像中心偏移递归修正法来处理动态多幅图像采集。对于初步处理的图像可以利用直方图的均衡化和局部特征提取的Retinex算法进行进一步降噪处理。提出了通过分析多种常见设备的外形特征和热故障特征形成故障类型数据库,作为判断设备故障的重要辅助依据。运用的最近邻规则聚类算法可以区分大部分常见的电力设备轮廓,并进行取样分析,能够提高软件的故障识别能力和判断的准确性。针对无人机系统的后台控制分析部分,论文搭建了地面站系统软、硬件环境,并利用Microsoft VisualStudio2010为开发平台进行功能模块的设计。包含实时模拟飞行仪表盘,遇险安全性设计等实用功能,超视距网络控制,数据传输,自定义飞行功能和图像的采集等模块化的设计与功能实现。论文提出了地面站的建立应用最新的地理信息系统开发平台所采用的二三维一体化的技术,建立了配电网电力线路拓扑关系,并设计了能够通用的电力线路和设备的空间数据库,建立与无人机的实时通讯连接,实现实时显示无人机状态的巡检三维显示地面站系统。系统显示无人机的方位和安全距离,也能够大大提高其安全性。对于该论文提出的无人机终端ARM系统的硬件集成,各个模块的通讯,地面站三维显示系统的程序都编写了代码并通过测试。图像的降噪、特征提取和粒子群双层优化的算例通过了matlab的测试分析,结论验证了程序的可行性与算例的有效性。更多还原

【Abstract】 The rapid development of China’s power grid construction, electricity load growth is rapid, electricity security has been greatly improved, but compared to the developed countries have certain disparity. Security level in developed or less developed areas of the country also has great distinction. Infrared temperature measurement technology is an important means of early found outside the equipment overheating fault and internal insulation fault,is also a method of advanced detection equipment status.Application of infrared temperature measurement technology to find the heat problems of power equipment external connection point, in the power system has been widely applied.When a fault occurs, it is necessary to restore the normal operation of the grid as the first important task. To timely inspection,avoid the occurrence of the fault,we need for scientific arrangement of transmission line maintenance plan, understand the real power grid, find the hidden trouble, reduce the operation risk should be put in the first place.In order to solve the problem of manual operation and routine detection of network inspection,put forward the UAV system equipped with infrared imaging equipment for scheme of flight inspection on grid.Because the UAV has characteristic of high maneuverability,fast speed,high arrival rate,strong controllability,more suitable for the inspection of modern smart grid environment,the system bring in the small unmanned helicopter as a platform, carrying camera and infrared image acquisition equipment for image data collection. Application of ARM embedded technology fusion of UAV with equipment,it designs and realizes of data sending and receiving on UAV, acquisition and transmission of infrared image, monitoring and command sending functions in the ground station,to achieve the requirement that real-time maintain3D geographic information system and unmanned machine terminal information synchronization in inspection.Flight inspection is convenient but also has certain limitations, need to analyze the cost of flying,combined with the characteristics of accurate modeling of regional power grid,to establish the constraints to control costs, play a maximum efficiency of flight.For these the dissertation put forward to the cost that repair and maintenance of network maximum profit aim is to minimize, build relationships in the failure rate and maintenance state on power lines,then Establishes the model of UAV inspect power line based Multi objective optimization, the improved traditional solution of multi-target into single objective for the multi-objective particle swarm optimization (MOPSO) algorithm,get the solution set of bilevel optimization model based on target of economic and target of reliability directly,optimize convergence process of solution. The dual constraint that line using the risk value and repair costs to determine the UAV patrol and inspection order, provide auxiliary decision science for maintenance personnel.The validation of the model and algorithm can get the optimal solution set in optimization of line patrol mode quickly and accurately. For the previous infrared acquisition equipment, manual and the fixed position of the traditional way to monitor or shoot power equipment,and UAV using infrared imaging equipment caused by the instability of the infrared image capture state, the paper brings forward the target image centroid location, infrared image segmentation and image of target device center offset recursive correction method to handle acquisition of dynamic image.For the image pre processing, can be used for further noise reduction using histogram equalization and retinex algorithm based on local feature extraction.Put forward that through the analysis of shape features many common devices and infrared feature, formating the database of fault type, as an important auxiliary basis judgment equipment failure. Using the clustering algorithm by nearest neighbor rule can distinguish the device profile of most common power, and sampling analysis, so can improve the software to determine accuracy and the capability of fault identification.According to the part of analysis of UAV background control system,the paper builds a soft system of ground station and hardware environment,then design module is designed by using Microsoft VisualStudio2010as the development platform. Contains the design of real-time flight simulation instrument panel,function the safety utility in distress, oth network control, data transmission, the design and function of image collected module, function of flight customization and realizated these.This paper presents the application of two or three dimensional integration technology that the newest technology of Geographic information system development platform to build the ground station, established a distribution network power circuit topology,and designed to spatial database of power lines and equipment in general, to establish a real-time connection with the communication on UAV, so completed the inspection3D display ground station system that could display the UAV state in realtime. System display range and the safe distance of UAV,that also could greatly enhance the safety.In this paper, the hardware integration of UAV terminal ARM system, each module of the communication,the programs of the ground station3D display system are written code and test.Image noise reduction, feature extraction and examples of particle swarm optimization through the analysis of MATLAB test, the conclusion verified the effectiveness and feasibility of the example program.更多还原