基于SCADA参量耦合网络变分图自编码的风电机组异常检测方法

刘小峰; 李俊锋; 柏林

doi:10.19912/j.0254-0096.tynxb.2024-0138

PDF(2009 KB)

太阳能学报 ›› 2025, Vol. 46 ›› Issue (5) : 567-576. DOI: 10.19912/j.0254-0096.tynxb.2024-0138

基于SCADA参量耦合网络变分图自编码的风电机组异常检测方法

刘小峰, 李俊锋, 柏林

作者信息 +

WIND TURBINE ANOMALY DETECTION BASED ON VARIATIONAL GRAPH AUTO-ENCODING OF SCADA PARAMETER COUPLING NETWORK

Liu Xiaofeng, Li Junfeng, Bo Lin

Author information +

文章历史 +

摘要

利用风电机组数据采集与监控系统（SCADA）数据监测参量本身数值信息及其相互间的耦合关联性,提出基于多参量耦合关系变分图自编码的风电机组异常检测方法。该方法利用时间序列自适应符号传递熵构建SCADA数据的参量耦合关系网络,设计变分图自编码再编码模型对参量耦合关系网络进行编码重构。结合SCADA参量耦合关系网络的编码重构误差构建风电机组的健康状态评估指标,采用支持向量回归的迭代更新法,对机组实时健康阈值进行自适应设置。两个风场的风力发电机组SCADA数据分析结果表明：该文方法充分利用了SCADA数据本身的数值信息及耦合关系结构信息,有效提高了风电机组异常状态检测的准确性及对环境工况的鲁棒性。

Abstract

By utilizing the numerical information of the wind turbine SCADA data monitoring parameters and their mutual coupling correlation, a wind turbine anomaly detection method based on variational graph auto-coding of multi-parameter coupling relationship is proposed. In this method, the adaptive symbolic transfer entropy of time series is used to construct a parameter coupling correlation network for SCADA data. A variational graph auto-encoding-recoding model is established to encode and reconstruct the parameter coupling correlation network. Combined with the coding reconstruction error of the SCADA parameter coupling correlation network, the health state assessment index of the wind turbine is constructed, and the support vector regression iterative updating method is applied to adaptively set the real-time health threshold of the wind turbine. The proposed method is verified using the SCADA monitoring data of wind turbines in two wind farms. The analysis results showed that this method can effectively extract the numerical information and the internal coupling correlation information from SCADA multivariable time series, which can effectively improve the accuracy of wind turbine anomaly detection and has good robustness to environmental interferences.

导出引用

刘小峰, 李俊锋, 柏林. 基于SCADA参量耦合网络变分图自编码的风电机组异常检测方法[J]. 太阳能学报. 2025, 46(5): 567-576 https://doi.org/10.19912/j.0254-0096.tynxb.2024-0138

Liu Xiaofeng, Li Junfeng, Bo Lin. WIND TURBINE ANOMALY DETECTION BASED ON VARIATIONAL GRAPH AUTO-ENCODING OF SCADA PARAMETER COUPLING NETWORK[J]. Acta Energiae Solaris Sinica. 2025, 46(5): 567-576 https://doi.org/10.19912/j.0254-0096.tynxb.2024-0138

中图分类号： TK83

参考文献

[1] 江国乾, 周俊超, 武鑫, 等. 基于空洞因果卷积网络的风电机组异常检测[J]. 太阳能学报, 2023, 44(5): 368-375.
JIANG G Q, ZHOU J C, WU X, et al.Wind turbine anomaly detection based on dilated causal convolution network[J]. Acta energiae solaris sinica, 2023, 44(5): 368-375.
[2] MARTI-PUIG P, NÚÑEZ-VILAPLANA C. Dynamic clustering of wind turbines using SCADA signal analysis[J]. Energies, 2024, 17(11): 2514.
[3] 马良玉, 袁乃正. 基于CFSFDP与LightGBM的风电机组异常状态预警研究[J]. 太阳能学报, 2023, 44(5): 401-406.
MA L Y, YUAN N Z.Research on abnormal condition early warning for wind turbine based on CFSFDP and LightGBM[J]. Acta energiae solaris sinica, 2023, 44(5): 401-406.
[4] 王红. 基于数据驱动的风电机组健康状态监测方法研究[D]. 秦皇岛: 燕山大学, 2020.
WANG H.Research on health condition monitoring method of wind turbine based on data-driven[D]. Qinhuangdao: Yanshan University, 2020.
[5] 金晓航, 秦治伟, 郭远晶, 等. 基于改进劣化度模型的风电机组日常运行状态评估[J]. 太阳能学报, 2023, 44(1): 239-246.
JIN X H, QIN Z W, GUO Y J, et al.Evaluation of wind turbines daily operation conditions based on improved degradation model[J]. Acta energiae solaris sinica, 2023, 44(1): 239-246.
[6] 马东, 孔德同, 郭鹏, 等. 基于SCADA运行数据的风电机组发电性能劣化监测研究[J]. 可再生能源, 2021, 39(1): 45-49.
MA D, KONG D T, GUO P, et al.Wind turbine performance degradation monitoring with SCADA data[J]. Renewable energy resources, 2021, 39(1): 45-49.
[7] ENCALADA-DÁVILA Á, PURUNCAJAS B, TUTIVÉN C, et al. Wind turbine main bearing fault prognosis based solely on SCADA data[J]. Sensors, 2021, 21(6): 2228.
[8] XIANG L, WANG P H, YANG X, et al.Fault detection of wind turbine based on SCADA data analysis using CNN and LSTM with attention mechanism[J]. Measurement, 2021, 175: 109094.
[9] LIU J Y, WANG X S, XIE F Q, et al.Condition monitoring of wind turbines with the implementation of spatio-temporal graph neural network[J]. Engineering applications of artificial intelligence, 2023, 121: 106000.
[10] LYU Q C, LIU J Y, HE Y W, et al.Condition monitoring of wind turbines with implementation of interactive spatio-temporal deep learning networks[J]. IEEE transactions on instrumentation and measurement, 2023, 72: 3520811.
[11] 朱渊, 何瑞瑞, 刘源, 等. DeepCKI: 一个基于变分图自编码器预测细胞-细胞因子相互作用的生物信息学模型[J]. 中国生物化学与分子生物学报, 2022, 38(8): 1033-1042.
ZHU Y, HE R R, LIU Y, et al.DeepCKI, a bioinformatics model for predicting cell-cytokine interactions based on variational graph auto-encoder[J]. Chinese journal of biochemistry and molecular biology, 2022, 38(8): 1033-1042.
[12] LUO X, WANG L W, HU P W, et al.Predicting protein-protein interactions using sequence and network information via variational graph autoencoder[J]. IEEE/ACM transactions on computational biology and bioinformatics, 2023, 20(5): 3182-3194.
[13] 严盛辉, 陈志德. 基于图自编码器的无监督多变量时间序列异常检测[J]. 计算机系统应用, 2023, 32(5): 308-315.
YAN S H, CHEN Z D.GAE-based unsupervised anomaly detection of multivariable time series[J]. Computer systems and applications, 2023, 32(5): 308-315.
[14] SCHREIBER T.Measuring information transfer[J]. Physical review letters, 2000, 85(2): 461-464.
[15] KAISER A, SCHREIBER T.Information transfer in continuous processes[J]. Physica D: nonlinear phenomena, 2002, 166(1/2): 43-62.
[16] CAO L Y.Practical method for determining the minimum embedding dimension of a scalar time series[J]. Physica D: nonlinear phenomena, 1997, 110(1/2): 43-50.
[17] WESSEL N, ZIEHMANN C, KURTHS J, et al.Short-term forecasting of life-threatening cardiac arrhythmias based on symbolic dynamics and finite-time growth rates[J]. Physical review E, statistical physics, plasmas, fluids, and related interdisciplinary topics, 2000, 61(1): 733-739.
[18] KIPF T N, WELLING M. Variational graph auto-encoders[EB/OL].2016: 1611.07308. https://arxiv.org/abs/1611.07308v1.
[19] VAPNIK V N.The nature of statistical learning theory[M]. New York: Springer New York, 2000.
[20] CHABCHOUB Y, TOGBE M U, BOLY A, et al.An in-depth study and improvement of isolation forest[J]. IEEE access, 2022, 10: 10219-10237.
[21] WANG G, CHEN Y F.Robust feature matching using guided local outlier factor[J]. Pattern recognition, 2021, 117: 107986.
[22] RENSTRÖM N, BANGALORE P, HIGHCOCK E. System-wide anomaly detection in wind turbines using deep autoencoders[J]. Renewable energy, 2020, 157: 647-659.
[23] MAO J X, WANG H, SPENCER B F Jr. Toward data anomaly detection for automated structural health monitoring: exploiting generative adversarial nets and autoencoders[J]. Structural health monitoring, 2021, 20(4): 1609-1626.
[24] LEAHY K, GALLAGHER C, O’DONOVAN P, et al. A robust prescriptive framework and performance metric for diagnosing and predicting wind turbine faults based on SCADA and alarms data with case study[J]. Energies, 2018, 11(7): 1738.