混沌优化决策树下水轮发电机组局部放电检测  

Partial Discharge Detection of Hydroelectric Generator Units under Chaotic Optimization Decision Tree

作　　者：田云 王荣 赵娅 邹佳成 TIAN Yun;WANG Rong;ZHAO Ya;ZOU Jia-cheng(Spic Guizhou Jinyuan Zunyi Hydropower Development CO.,LTD,Zunyi Guizhou 563000,China)

机构地区：[1]国家电投集团贵州金元遵义水电开发有限公司,贵州遵义563000

出　　处：《计算机仿真》2025年第1期96-100,共5页Computer Simulation

基　　金：国家电投集团贵州金元股份有限公司科技项目(KY-C-2021-JY-SD02)。

摘　　要：水轮发电机组A、B、C相的局放量差异较大,局放频次不同,导致局部放电准确检测难度较大。为此,提出混沌优化决策树下水轮发电机组局部放电检测方法。提取水轮发电机组脉冲信号脉冲等效长度与等效速度特征,融合时域特征与频域特征;建立决策树分割水轮发电机组脉冲信号特征集;引入混沌海鸥优化算法优化决策树,结合混沌映射与鲸鱼算法展开决策树参数优化,将机组时频特征输入到优化后的决策树中,实现水轮发电机组局部放电检测。仿真结果表明,所提方法应用后,可准确地完成局放量检测与局放频次检测,具有较优的检测效果。The difference in partial discharge volume and frequency among phases A,B,and C of the hydroelectric generator set makes accurate detection of partial discharge dificult.For this reason,a chaotic optimization decision tree based partial discharge detection method for hydroelectric generators is proposed.Extract the equivalent length and equivalent velocity characteristics of the pulse signal of the hydroelectric generator unit,and fuse the time-domain and frequency-domain characteristics;Establish a decision tree to segment the feature set of pulse signals for hydroelectric generators;Introducing the chaotic seagull optimization algorithm to optimize the decision tree,combining chaotic mapping and whale algorithm to optimize the decision tree parameters,and inputting the time-frequency characteristics of the unit into the optimized decision tree to achieve partial discharge detection of the hydroelectric generator unit.The simulation results show that the proposed method can accurately complete partial discharge detection and partial discharge frequency detection,and has better detection effect.

关 键 词：时频信号 小波变换方法 决策树算法 局部放电检测 混沌优化 

分 类 号：TM835[电气工程—高电压与绝缘技术]