基于神经网络技术的水轮发电机组轴心轨迹识别与故障判定研究  

Research on Fault Determination and the Identification of the Axis Orbit of Hydro-generator Units Based on Neural Network Technology

作　　者：王启发 刘万江[1] 白洁[1] 刘忠仁[1] 王润鹏 WANG Qifa;LIU Wanjiang;BAI Jie;LIU Zhongren;WANG Runpeng(Harbin Electric Machinery Company Limited,Harbin 150040,China)

机构地区：[1]哈尔滨电机厂有限责任公司,哈尔滨150040

出　　处：《大电机技术》2020年第2期44-50,共7页Large Electric Machine and Hydraulic Turbine

基　　金：智能制造综合标准化与新模式应用项目:水力发电设备智能远程运维新模式。

摘　　要：轴心轨迹图形包含了丰富的机组运行状态信息,轴心轨迹识别是判断水轮发电机组轴系故障的重要技术手段;神经网络技术是一种通过计算机程序模拟人类大脑结构功能的智能化技术思想,具有自学习的优点。本文首次提出采用神经网络技术通过仿射不变矩算法计算出特征向量,以特征向量作为数据分类依据识别轴心轨迹图形,从而达到故障判定的目标。数值模拟测试结果表明,该技术的故障识别率能达到95%以上,证明基于神经网络技术的水轮发电机组轴心轨迹识别算法对于机组故障诊断具有重要价值。Axis track graph contains abundant information of unit operation state,and axis track recognition is an important technical means to judge the fault of hydro-generator shaft system.The neural network technology is an intelligent technical idea to simulate human brain structure and function through computer program,which has the advantage of self-learning.In this paper,the neural network technology is used for the first time to calculate the eigenvector through the affine moment invariant algorithm.The eigenvector is used as the data classification basis to draw the axis track graph,so as to achieve the goal of fault determination.The test results of numerical simulation show that the fault recognition rate of this technology can reach more than 95%.It is proved that the algorithm based on neural network technology is of great value for the fault diagnosis of hydro-generator units.

关 键 词：神经网络 水轮发电机组 轴心轨迹 故障判定 

分 类 号：TM312[电气工程—电机]