基于BP神经网络的变压器多参数在线监测系统的研究

发布时间：2018-03-11 01:06

本文选题：变压器　切入点：ZigBee网络　出处：《安徽理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：电力系统包含众多电气设备,各电气设备在电力系统中都扮演着重要的角色,其中变压器扮演的角色尤为关键。变压器进行在线监测有利于对变压器工作状态的实时掌握和了解,从而当变压器发生故障时能够及时发现并处理,避免对整个电网造成重大影响,与此同时,也有利于故障发生后及时的对变压器进行检修,以免造成更大危害,从而降低发生故障后造成的损失,使电力系统稳定运行。本文首先简单介绍了变压器在线监测对整个电力行业的影响,对变压器在线监测研究背景以及意义进行了探讨,简述其近年来的发展现状;其次,介绍了几种常用的无线传感器网络技术,分析比较了各自的优缺点,从而选出ZigBee技术用于采集和传输变压器在线监测参数信息,进而对ZigBee技术作了详细介绍;然后对变压器在线监测系统设计进行了讨论,对其总体结构进行了研究,对无线传感器网络数据信息采集和传输进行了硬件电路设计,并分别对变压器在线监测所检测参数即局部放电和油中溶解气体所使用的传感器进行了选型,通过比较从而确定了将脉冲电流传感器和超声传感器用于变压器局部放电信号检测,以及将半导体式气敏传感器用于油中溶解气体检测;最后,对无线传感器网络数据信息采集和传输进行了软件设计,简单介绍了 BP神经网络的工作原理,分别对其输入、输出向量以及隐藏层节点数进行了选定,进而确定其学习样本的训练次数,构建了基于BP神经网络的变压器多参数在线监测模型,对变压器四种故障类型进行判定。
[Abstract]:Power system contains a lot of electrical equipment, each electrical equipment plays an important role in the power system, The role of transformer is particularly important. On-line monitoring of transformer is helpful to grasp and understand the working state of transformer in real time, so that the transformer can be found and processed in time when the transformer is in trouble. Avoiding the major impact on the whole power network, at the same time, it is also conducive to the timely maintenance of the transformer after the failure, so as not to cause more harm, thereby reducing the loss caused by the failure. This paper introduces the influence of transformer on-line monitoring on the whole power industry, discusses the research background and significance of transformer on-line monitoring, and describes its development status in recent years. This paper introduces several commonly used wireless sensor network technologies, analyzes and compares their advantages and disadvantages, and then selects ZigBee technology to collect and transmit transformer on-line monitoring parameter information, and then introduces the ZigBee technology in detail. Then, the design of transformer on-line monitoring system is discussed, the overall structure is studied, and the hardware circuit is designed for data collection and transmission of wireless sensor network. The sensors used in the on-line monitoring of transformers, i.e. partial discharge and dissolved gas in oil, are selected. By comparison, the pulse current sensor and ultrasonic sensor are used to detect partial discharge signal of transformer, and semiconductor gas sensor is used to detect dissolved gas in oil. The software design of data acquisition and transmission for wireless sensor network is presented. The working principle of BP neural network is introduced, and the input, output vectors and hidden layer node points are selected respectively. Furthermore, the training times of the learning samples are determined, and the transformer multi-parameter on-line monitoring model based on BP neural network is constructed, and the four types of transformer faults are judged.
【学位授予单位】：安徽理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM41

【相似文献】