智能建筑供配电故障诊断与定位研究

发布时间：2018-06-13 23:31

本文选题：智能建筑 + 供配电网　；参考：《华东交通大学》2017年硕士论文

【摘要】：随着智能建筑在我国的不断发展,人们对智能建筑的供配电网络的可靠、安全、稳定的运行提出了更高的要求。若智能建筑的供配电网络发生了故障,不仅会影响楼内用户的正常日常生活,严重是会造成巨大的财产损失及人身安全问题。因此,对智能建筑供配电网络的故障诊断、定位与恢复的研究就十分的有重要意义。本文结合智能建筑供配电网络的特点,对其进行拓扑结构分析及故障特征的分析,确定了其故障类型。提出了一种基于模式识别技术的智能建筑供配电网络故障诊断的算法。再对其进行故障定位,提出了一种基于蜻蜓算法的小波神经网络的智能建筑供配电网络故障定位算法。将其与其他算法比较分析,本文提出的两种算法能够较好的实现建筑供配电网络的故障诊断与定位。最后,对智能建筑供配电网络提出了相应的供配电恢复策略。其中智能建筑供配电网络的特征分析及提取,故障诊断与定位算法的研究是问论文的重点研究内容。本文的主要研究内容如下:(1)根据智能建筑供配电网络的特点,对其网络拓扑结构进行了描述及分析。(2)对智能建筑供配电网络常见的故障进行故障特征分析,主要采用三序分量法对其进行分析,最终确定四种供配电网络故障类型为单相接地短路、两相相间短路、两相接地短路、三相短路及两种电网设备故障为断路器故障、母线故障。(3)提出了一种基于模式识别技术的智能建筑供配电网络故障诊断的算法。主要是将幅值量,相角量及能量对其各种进行相应的小波变换后,按各种权重组合的电气故障特征量以及由简单贝叶斯推论的得到各元件开关量相融合的方法,作为贝叶斯网络的特征信息进行学习训练,可以很好的完成供配电网络的故障诊断。(4)提出了一种基于蜻蜓算法的小波神经网络的智能建筑供配电网络故障定位算法。将小波神经网络的各连接权值和各阈值作为蜻蜓算法里蜻蜓个体位置向量,小波神经网络的权值和阈值按照蜻蜓个体位置算法寻求最优值。相比于其他算法,该算法能够快速准确的进行故障的定位。(5)根据智能建筑供配电网络故障的特点,对提出了相应的策略。根据现有的供配电网络恢复算法,对其进行供配电网络恢复算法进行了分析,并将其进行分类,总结了各类算法的有缺点。
[Abstract]:With the continuous development of intelligent building in China, people put forward higher requirements for the reliable, safe and stable operation of intelligent building power supply and distribution network. If the power supply and distribution network of the intelligent building breaks down, it will not only affect the normal daily life of the users in the building, but also cause huge property losses and personal safety problems. Therefore, the research on fault diagnosis, location and recovery of intelligent building power supply and distribution network is of great significance. According to the characteristics of intelligent building power supply and distribution network, this paper analyzes its topological structure and fault characteristics, and determines its fault type. A fault diagnosis algorithm for intelligent building power supply and distribution network based on pattern recognition technology is proposed. Then, a fault location algorithm of intelligent building power supply and distribution network based on dragonfly algorithm and wavelet neural network is proposed. Compared with other algorithms, the two algorithms proposed in this paper can achieve better fault diagnosis and location in building power supply and distribution networks. Finally, the recovery strategy for intelligent building power supply and distribution network is put forward. The feature analysis and extraction of intelligent building power supply and distribution network, the research of fault diagnosis and location algorithm are the key contents of this paper. The main contents of this paper are as follows: (1) according to the characteristics of intelligent building power supply and distribution network, the network topology is described and analyzed. Three sequence component method is used to analyze the fault. Finally, it is determined that four types of fault types of power supply and distribution network are single phase grounding short circuit, two phase interphase short circuit, two phase grounding short circuit, three phase short circuit and two kinds of power network equipment faults as circuit breakers. This paper presents an algorithm for fault diagnosis of intelligent building power supply and distribution network based on pattern recognition technology. After the corresponding wavelet transform of amplitude, phase angle and energy, the electrical fault characteristic quantity of each weight combination and the method of getting the switch quantity of each component by simple Bayesian inference are obtained. As the characteristic information of Bayesian network, learning and training can complete the fault diagnosis of power supply and distribution network very well. This paper presents a fault location algorithm of intelligent building power supply and distribution network based on dragonfly algorithm and wavelet neural network. Each connection weight and threshold of wavelet neural network are taken as dragonfly individual position vector in dragonfly algorithm. The weight and threshold of wavelet neural network seek the optimum value according to dragonfly individual position algorithm. Compared with other algorithms, this algorithm can locate faults quickly and accurately. According to the fault characteristics of intelligent building power supply and distribution network, the corresponding strategy is put forward. According to the existing restoration algorithm of power supply and distribution network, the algorithm of power supply and distribution network restoration is analyzed and classified, and the shortcomings of all kinds of algorithms are summarized.
【学位授予单位】：华东交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TU852

【参考文献】