基于GRU-LightGBM的风电机组发电机滚动轴承故障检测  

Fault Detection of Rolling Bearing of Wind Turbine Generator Based on GRU-LightGBM

作　　者：贾洪岩 刘玉龙 魏宏杰 吴劲芳 闫志强 JIA Hongyan;LIU Yulong;WEI Hongjie;WU Jinfang;YAN Zhiqiang(State Grid Jibei Zhangjiakou Wind and Solar Energy Storage and Transportation New Energy Co.,LTD.,Zhangjiakou Hebei 075000,China;Dongfang Electric New Energy Technology(Chengdu)Co.,LTD.,Chengdu 610036,China)

机构地区：[1]国网冀北张家口风光储输新能源有限公司,河北张家口075000 [2]东方电气新能科技(成都)有限公司,成都610036

出　　处：《微电机》2024年第12期60-64,共5页Micromotors

摘　　要：在复杂作业环境、冗余负荷条件等多种因素的影响下,滚动轴承故障发生频率居高不下,威胁着风电机组发电机运行的安全性与可靠性,故提出基于GRU-LightGBM的风电机组发电机滚动轴承故障检测方法研究。有效融合GRU模型与LightGBM算法,建立风电机组发电机滚动故障检测架构,基于GRU模型(更新门和重置门)处理滚动轴承振动数据和电气数据特征,通过LightGBM算法中的弱学习器与强学习器选择并合并转速数据,获取载荷数据特征融合结果,加载滚动轴承故障特征集合,计算两者之间的相关系数,以此为基础,获取滚动轴承故障检测结果。实验结果显示:应用提出方法获得的滚动轴承转速数据并行处理效率最大值达到了96 MB/min,载荷数据特征融合与实际结果趋于一致,滚动轴承故障检测结果与实际结果相同,充分证实提出方法滚动轴承故障检测性能更佳。Under the influence of various factors such as complex work environments and redundant load conditions,the frequency of rolling bearing failures remains high,posing a threat to the safety and reliability of wind turbine generator operation.Therefore,a fault detection method for wind turbine generator rolling bearings based on GRU-LightGBM was proposed.Effectively integrating the GRU model with the LightGBM algorithm,a rolling fault detection architecture for wind turbine generators was established.Based on the GRU model(update gate and reset gate),rolling bearing vibration data and electrical data features were processed.The speed data was selected and merged using weak and strong learners in the LightGBM algorithm to obtain load data feature fusion results.The rolling bearing fault feature set was loaded,and the correlation coefficient between the two was calculated.Based on this,the rolling bearing fault detection results were obtained.The experimental results showed that the maximum parallel processing efficiency of the rolling bearing speed data obtained by the proposed method reached 96MB/min,and the fusion of load data features tended to be consistent with the actual results.The fault detection results of rolling bearings were consistent with the actual results,fully confirming that the proposed method has better fault detection performance for rolling bearings.

关 键 词：风电机组 滚动轴承 GRU-LightGBM 发电机 故障检测 振动数据和电气数据特征提取 

分 类 号：TM315[电气工程—电机]