配电网故障诊断系统

发布时间：2018-06-18 11:19

本文选题：GRNN神经网络 + 配电网故障诊断　；参考：《福州大学》2014年硕士论文

【摘要】：随着经济社会的发展和人民生活水平的提高,电力用户对于电力系统的安全性和可靠性提出了更高的要求。然而爆炸性增长的用电负荷、不断扩大的电网规模都使得电网拓扑结构越来越复杂进而导致电网故障诊断的难度也在随之上升。配电网络是电力系统中发电-输电-配电环节的最后一环,直接关系到用户的切身利益。在自然、人为因素(如工地施工不规范和自然灾害)的等问题干扰下,配电网的安全问题日益突出,导致配电网事故频发。当电网发生故障后,迅速的确认故障源和故障原因是电力公司排除故障进而进行恢复性供电的前提,因此研究行之有效的配电网故障诊断系统,辅助运行人员快速准确地找出故障源并修复故障,对于提高配电网可靠性和安全性具有重要的意义。现如今大规模遥信和遥感设备的投入,使得电网的自动化水平有了很大的提高,这也为研究配电网故障诊断系统提供了基础。当前电网故障诊断方法研究的难点是如何克服故障诊断过程中的不确定因素的干扰(即保护装置和断路器的动作可靠性、调度中心接收到的警报信息的完整性和正确性),如何实现大型电网络的在线故障诊断以及如何应对电网拓扑变化带来的诊断难题等问题。本文对目前国内外主流的电网故障诊断方法进行了总结和分析,针对现今电网故障诊断方法存在的主要问题,提出结合采用面向元件的FOA-GRNN(果蝇优化算法-广义回归神经网络)神经网络的故障诊断模型和基于因果时序网络的警报信息甄别模型的电网故障诊断方法,并依此为基础开发配电网故障诊断专家系统。当配电网发生故障时,首先系统自动识别停电区域并建立可疑故障元件集合；其次SCADA(数据采集与监控系统)和SOE(事件顺序记录仪)所收到的警报信息经过警报信息甄别系统剔除错误警报,最后将保护装置和断路器动作信息上传给FOA-GRNN神经网络进行故障诊断,得到配电网中故障元件集的故障可信度,最终由专家系统实现故障综合诊断并给出解决方案。经算例验证,该方法能够有效提高运行人员故障处理效率,快速并准确的实现配电网的故障诊断。相比于其他故障诊断方法,该方法能够较好的处理故障诊断过程中存在的各种不确定因素以及配电网拓扑变化,对配电网复杂故障有较好的诊断能力。
[Abstract]:With the development of economy and society and the improvement of people's living standard, electric power users have put forward higher requirements for the security and reliability of power system. However, the explosive growth of power load and the increasing scale of power grid make the topology of power network more and more complex, which leads to the difficulty of fault diagnosis. Distribution network is the last link of generation, transmission and distribution in power system, which is directly related to the vital interests of users. Due to the interference of natural and man-made factors (such as construction irregularities and natural disasters), the safety problems of distribution network become more and more prominent, resulting in frequent distribution network accidents. When the fault occurs in the power network, it is the premise for the power company to remove the fault and restore the power supply, so the effective fault diagnosis system of the distribution network is studied. It is of great significance for auxiliary operators to find fault sources and repair faults quickly and accurately for improving the reliability and safety of distribution network. Nowadays, with the input of large-scale remote communication and remote sensing equipment, the automation level of power grid has been greatly improved, which also provides a basis for the study of distribution network fault diagnosis system. How to overcome the disturbance of uncertain factors in the process of fault diagnosis (that is, the operational reliability of protective devices and circuit breakers) is the difficulty of the current research on fault diagnosis methods of power network. The integrality and correctness of the alarm information received by the dispatching center, how to realize the on-line fault diagnosis of the large electric network and how to deal with the problem of diagnosis caused by the topology change of the power network are discussed. In this paper, the main methods of power network fault diagnosis at home and abroad are summarized and analyzed. A fault diagnosis model based on component oriented FOA-GRNN (Drosophila Optimization algorithm-Generalized regression Neural Network) neural network and an alarm information screening model based on causal sequential network are proposed. Based on this, an expert system for fault diagnosis of distribution network is developed. When the distribution network fails, the system automatically identifies the blackout area and establishes the set of suspicious fault elements. Secondly, the alarm information received by SCADA (data acquisition and monitoring system) and SOE (event sequence recorder) is eliminated by the alarm information screening system. Finally, the protection device and circuit breaker action information are uploaded to the FOA-GRNN neural network for fault diagnosis, and the fault reliability of the fault component set in the distribution network is obtained. Finally, the fault comprehensive diagnosis is realized by the expert system and the solution is given. The example shows that this method can improve the efficiency of fault treatment and realize the fault diagnosis of distribution network quickly and accurately. Compared with other fault diagnosis methods, this method can deal with all kinds of uncertain factors in the process of fault diagnosis and topology change of distribution network, and has better ability to diagnose complex faults of distribution network.
【学位授予单位】：福州大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM727

【相似文献】