水轮发电机转子磁极主引线在线监测技术  

Online monitoring technology of hydro-generator rotor magnet poles

作　　者：杨晖 赵聚平 万正喜 王俏 YANG Hui;ZHAO Juping;WAN Zhengxi;WANG Qiao(Hunan Heimifeng Pumped Storage Co.,Ltd.,Changsha 410200,Hunan,China)

机构地区：[1]湖南黑麋峰抽水蓄能有限公司,湖南长沙410200

出　　处：《金属功能材料》2023年第3期77-82,共6页Metallic Functional Materials

基　　金：国网新源公司科技项目资助项目(SGXYKJ-2022-031)。

摘　　要：水轮发电机转子磁场不稳定、磁拉力不平衡,容易导致主引线状态异常,影响到整个水力发电系统运行安全,研究水轮发电机转子磁极主引线在线监测技术。确定水轮发电机不稳定磁场下,不平衡磁拉力是主引线故障状态的主要指标。采集磁拉力的数据,采用波阈值去噪算法进行数据去噪处理。结合模糊神经网络建立主引线在线监测模型,获取模型的概率密度分布,结合梯度下降法设定发电机转子磁极主引线异常阈值门限,从而实现水轮发电机转子磁极主引线在线监测。实验结果表明,该方法的查全率、查准率、平均监测精度较高,数据去噪处理效果好,主引线异常瞬时功率曲线与实际曲线非常接近,监测效果好,可以在主引线在线监测领域得到进一步应用。The unstable magnetic field and unbalanced magnetic tension of the hydro-generator rotor will easily lead to the abnormal state of the main lead,which will affect the operation safety of the entire hydro power generation system.The online monitoring technology of the main lead of the hydro-generator rotor pole is studied.The determination of the unstable magnetic field and unbalanced magnetic pull of the hydro-generator is the main indicator of the fault state of the main lead.Magnetic field and magnetic pull data were collected,and the wave threshold denoising algorithm was used for data denoising.The fuzzy neural network was used to establish the online monitoring model of the main lead,obtained the probability density distribution of the model,and the gradient descent method was used to set the threshold value of the abnormal main lead of the generator rotor pole,so as to realize the online monitoring of the main lead of the hydro-generator rotor pole.The experimental results show that the recall,precision and average monitoring accuracy of this method are high,the data denoising effect is good,the abnormal instantaneous power curve of the main lead is very close to the actual curve,and the monitoring effect is good,which can be further applied in the field of online monitoring of the main lead..

关 键 词：水轮发电机 转子磁极主引线 在线监测技术 数据去噪 模糊神经网络 

分 类 号：TV734.2[水利工程—水利水电工程]