第6期 尹诗，等：基于AC-GAN数据重构的风电机组主轴承温度监测方法 ·1115· 进行重构，解决了人为设定阈值的相关问题，提 [7]QIU Yingning,FENG Yanhui,INFIELD D.Fault dia- 升了主轴承异常和正常数据标签标注的准确率。 gnosis of wind turbine with SCADA alarms based multi- 3)在同等条件下，NGBoost算法在风电机组 dimensional information processing method[J].Renew- 主轴承状态决策模型中优于XGBoost算法和随机 able energy,2020,145:1923-1931. [8]LIANG Tao,QIAN Siqi,MENG Zhaochao,et al.Early 森林算法。而且，NGBoost、XGBoost和随机森林 fault warning of wind turbine based on BRNN and large 3种算法经过残差重构的状态决策模型的准确度 sliding window[J].Journal of intelligent&fuzzy systems, 分别提高了21.5%、19.2%、10.8%，选择6台机组 2020,38(3):3389-3401. 进行测试分析(3台主轴承正常机组，3台主轴承 [9]尹诗，侯国莲，于晓东，等.基于Bi-RNN的风电机组主 异常机组)，均能够判断正确。基于NGBoost的状 轴承温度预警方法研究[].郑州大学学报（工学版）， 态决策模型的平均准确率从60.5%（无残差序列重 2019,40(5):4450 构)提升至72.3%（利用残差数据重构）。 YIN Shi,HOU Guolian,YU Xiaodong,et al.Research on temperature prediction method for main bearing of wind 参考文献： turbine based on Bi-RNN[J].Journal of Zhengzhou Uni- versity (engineering science edition),2019,40(5):44-50. [1]曾军，陈艳峰，杨苹，等.大型风力发电机组故障诊断综 [10]陈雪峰，郭艳婕，许才彬，等.风电装备故障诊断与健 述.电网技术，2018,42(3)：849-860 康监测研究综述J.中国机械工程，2020,31(2)： ZENG Jun,CHEN Yanfeng,YANG Ping,et al.Review 175-189 of fault diagnosis methods of large-scale wind turbines[J] CHEN Xuefeng,GUO Yanjie,XU Caibin,et al.Review Power system technology,2018,42(3):849-860. of fault diagnosis and health monitoring for wind power [2]金晓航，孙毅，单继宏，等.风力发电机组故障诊断与预 equipment[J].China mechanical engineering,2020 测技术研究综述[J].仪器仪表学报，2017,38(5)： 31(2):175-189. 1041-1053 [11]沈小军，付雪姣，周冲成，等.风电机组风速-功率异常 JIN Xiaohang,SUN Yi,SHAN Jihong,et al.Fault dia- 运行数据特征及清洗方法[J].电工技术学报，2018， gnosis and prognosis for wind turbines:an overview[J]. 33(14):3353-3361. Chinese journal of scientific instrument,2017,38(5): SHEN Xiaojun,FU Xuejiao,ZHOU Chongcheng,et al. 1041-1053 Characteristics of outliers in wind speed-power opera- [3]刘帅.基于实时监测数据挖掘的风电机组故障预警方 tion data of wind turbines and its cleaning method[]. 法研究D].北京：华北电力大学（北京），2019. Transactions of China electrotechnical society,2018. LIU Shuai.Research on fault warning method of wind 33(14):3353-3361. turbine based on real-time monitoring data mining[D]. [12]贾科，杨哲，魏超，等.基于斯皮尔曼等级相关系数的 Beijing:North China Electric Power University (Beijing) 新能源送出线路纵联保护円.电力系统自动化，2020 2019. 4415):103-111. [4]李俊卿，李斯璇，陈雅婷，等.同步发电机定子故障预警 JIA Ke,YANG Zhe,WEI Chao,et al.Pilot protection 模型[).电力科学与工程，2020,36(5)：7-14 based on spearman rank correlation coefficient for trans- LI Junqing,LI Sixuan,CHEN Yating,et al.Synchronous mission line connected to renewable energy source[J]. generator stator fault prediction model[J].Electric power Automation of electric power systems,2020,44(15): science and engineering,2020,36(5):7-14. 103-111. [5]王梓齐，刘长良.基于Box-Cox变换和相对嫡残差分析 [13]CHEN Tianqi,GUESTRIN C.XGBoost:a scalable tree 的风电机组齿轮箱状态监测].中国电机工程学报 boosting system[C]//22nd ACM SIGKDD International 2020,40(13:4210-4218 Conference on Knowledge Discovery and Data Mining. WANG Ziqi,LIU Changliang.Wind turbine gearbox con- San Francisco,USA,2016:785-794. dition monitoring based on Box-Cox transformation and [14]KE Guolin,MENG Qi,FINLEY T,et al.LightGBM:a relative entropy residual analysis[].Proceedings of the highly efficient gradient boosting decision tree[Cl//Pro- CSEE,2020,40(13):4210-4218. ceedings of the 31st International Conference on Neural [6]刘帅，刘长良，甄成刚.基于数据分类重建的风电机组 Information Processing Systems.Long Beach,USA 故障预警方法.仪器仪表学报，2019,40(8)：1-11. 2017:3149-3157. LIU Shuai,LIU Changliang,ZHEN Chenggang.Fault [15]黄伟，李阳.基于MCS-MIFS与LightGBM的燃气轮 warning method for wind turbine based on classified data 机功率预测方法[.电力科学与工程，2020,36(5)： reconstruction[J].Chinese journal of scientific instrument, 23-31. 2019,40(8):1-11. HUANG Wei,LI Yang.Gas turbine power forecasting进行重构，解决了人为设定阈值的相关问题，提 升了主轴承异常和正常数据标签标注的准确率。 3) 在同等条件下，NGBoost 算法在风电机组 主轴承状态决策模型中优于 XGBoost 算法和随机 森林算法。而且，NGBoost、XGBoost 和随机森林 3 种算法经过残差重构的状态决策模型的准确度 分别提高了 21.5%、19.2%、10.8%，选择 6 台机组 进行测试分析 (3 台主轴承正常机组，3 台主轴承 异常机组)，均能够判断正确。基于 NGBoost 的状 态决策模型的平均准确率从 60.5%(无残差序列重 构) 提升至 72.3%(利用残差数据重构)。 参考文献： 曾军, 陈艳峰, 杨苹, 等. 大型风力发电机组故障诊断综 述 [J]. 电网技术, 2018, 42(3): 849–860. ZENG Jun, CHEN Yanfeng, YANG Ping, et al. Review of fault diagnosis methods of large-scale wind turbines[J]. Power system technology, 2018, 42(3): 849–860. [1] 金晓航, 孙毅, 单继宏, 等. 风力发电机组故障诊断与预 测技术研究综述 [J]. 仪器仪表学报, 2017, 38(5): 1041–1053. JIN Xiaohang, SUN Yi, SHAN Jihong, et al. Fault dia￾gnosis and prognosis for wind turbines: an overview[J]. Chinese journal of scientific instrument, 2017, 38(5): 1041–1053. [2] 刘帅. 基于实时监测数据挖掘的风电机组故障预警方 法研究 [D]. 北京: 华北电力大学 (北京), 2019. LIU Shuai. Research on fault warning method of wind turbine based on real-time monitoring data mining[D]. Beijing: North China Electric Power University (Beijing), 2019. [3] 李俊卿, 李斯璇, 陈雅婷, 等. 同步发电机定子故障预警 模型 [J]. 电力科学与工程, 2020, 36(5): 7–14. LI Junqing, LI Sixuan, CHEN Yating, et al. Synchronous generator stator fault prediction model[J]. Electric power science and engineering, 2020, 36(5): 7–14. [4] 王梓齐, 刘长良. 基于 Box-Cox 变换和相对熵残差分析 的风电机组齿轮箱状态监测 [J]. 中国电机工程学报, 2020, 40(13): 4210–4218. WANG Ziqi, LIU Changliang. Wind turbine gearbox con￾dition monitoring based on Box-Cox transformation and relative entropy residual analysis[J]. Proceedings of the CSEE, 2020, 40(13): 4210–4218. [5] 刘帅, 刘长良, 甄成刚. 基于数据分类重建的风电机组 故障预警方法 [J]. 仪器仪表学报, 2019, 40(8): 1–11. LIU Shuai, LIU Changliang, ZHEN Chenggang. Fault warning method for wind turbine based on classified data reconstruction[J]. Chinese journal of scientific instrument, 2019, 40(8): 1–11. [6] QIU Yingning, FENG Yanhui, INFIELD D. Fault dia￾gnosis of wind turbine with SCADA alarms based multi￾dimensional information processing method[J]. Renew￾able energy, 2020, 145: 1923–1931. [7] LIANG Tao, QIAN Siqi, MENG Zhaochao, et al. Early fault warning of wind turbine based on BRNN and large sliding window[J]. Journal of intelligent & fuzzy systems, 2020, 38(3): 3389–3401. [8] 尹诗, 侯国莲, 于晓东, 等. 基于 Bi-RNN 的风电机组主 轴承温度预警方法研究 [J]. 郑州大学学报(工学版), 2019, 40(5): 44–50. YIN Shi, HOU Guolian, YU Xiaodong, et al. Research on temperature prediction method for main bearing of wind turbine based on Bi-RNN[J]. Journal of Zhengzhou Uni￾versity (engineering science edition), 2019, 40(5): 44–50. [9] 陈雪峰, 郭艳婕, 许才彬, 等. 风电装备故障诊断与健 康监测研究综述 [J]. 中国机械工程, 2020, 31(2): 175–189. CHEN Xuefeng, GUO Yanjie, XU Caibin, et al. Review of fault diagnosis and health monitoring for wind power equipment[J]. China mechanical engineering, 2020, 31(2): 175–189. [10] 沈小军, 付雪姣, 周冲成, 等. 风电机组风速−功率异常 运行数据特征及清洗方法 [J]. 电工技术学报, 2018, 33(14): 3353–3361. SHEN Xiaojun, FU Xuejiao, ZHOU Chongcheng, et al. Characteristics of outliers in wind speed-power opera￾tion data of wind turbines and its cleaning method[J]. Transactions of China electrotechnical society, 2018, 33(14): 3353–3361. [11] 贾科, 杨哲, 魏超, 等. 基于斯皮尔曼等级相关系数的 新能源送出线路纵联保护 [J]. 电力系统自动化, 2020, 44(15): 103–111. JIA Ke, YANG Zhe, WEI Chao, et al. Pilot protection based on spearman rank correlation coefficient for trans￾mission line connected to renewable energy source[J]. Automation of electric power systems, 2020, 44(15): 103–111. [12] CHEN Tianqi, GUESTRIN C. XGBoost: a scalable tree boosting system[C]//22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Francisco, USA, 2016: 785−794. [13] KE Guolin, MENG Qi, FINLEY T, et al. LightGBM: a highly efficient gradient boosting decision tree[C]//Pro￾ceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach, USA, 2017: 3149−3157. [14] 黄伟, 李阳. 基于 MCS-MIFS 与 LightGBM 的燃气轮 机功率预测方法 [J]. 电力科学与工程, 2020, 36(5): 23–31. HUANG Wei, LI Yang. Gas turbine power forecasting [15] 第 6 期 尹诗，等：基于 AC-GAN 数据重构的风电机组主轴承温度监测方法 ·1115·