面向文本非结构化数据的输变电系统故障诊断方法

发布时间：2018-04-10 10:11

本文选题：输变电系统　切入点：故障诊断　引自：《电力科学与技术学报》2017年03期

【摘要】：从非结构化数据中提取信息,有效诊断输电系统故障对于精确确定设备的运行状态具有重要意义。提出基于深度学习网络的输变电系统故障诊断方法;面向非结构化数据处理问题,提出非结构化数据特征提取、深度神经网络构建、深度神经网络训练、故障诊断等一系列深度学习处理方式;构建并改进循环神经网络RNN,将长短时记忆模型(LSTM)添加到神经网络中的记忆单元,并提出相应的神经网络训练策略。以南方电网故障巡检报告作为数据源,仿真分析结果验证了该方法的有效性。
[Abstract]:It is very important to extract information from unstructured data and to diagnose the fault of transmission system accurately.This paper presents a fault diagnosis method of transmission and transformer system based on deep learning network, and presents the feature extraction of unstructured data, the construction of depth neural network, the training of depth neural network, and the problem of unstructured data processing.A series of advanced learning methods such as fault diagnosis are used to construct and improve the cyclic neural network (RNNN) and to add LSTM (long and short time memory model) to the memory unit of the neural network and to put forward the corresponding training strategy of the neural network.The fault inspection report of Southern Power Grid is used as the data source, and the simulation results verify the effectiveness of the method.
【作者单位】：贵州电网有限责任公司电力科学研究院;武汉大学电气工程学院;
【基金】：中国南方电网公司重点科技项目(GZ2014-2-0049)
【分类号】：TM711

【相似文献】