Outlier Detection and Data Reconstruction of Dissolved Gas in Oil for Power Transformers
Jiang Jun1, Zhang Wenqian1, Li Bo2, Li Xiaohan3, Fan Lidong4
1. Jiangsu Key Laboratory of New Energy Generation and Power Conversion Nanjing University of Aeronautics and Astronautics Nanjing 211106 China; 2. State Grid Anyang Power Supply Company Anyang 455000 China; 3. State Grid Jiangsu Electric Power Co. Ltd Research Institute Nanjing 211103 China; 4. Hangzhou Qianjiang Electric Group Co. Ltd Hangzhou 311243 China
Abstract Transformer plays a vital role of power transmission and voltage conversion in power system, and its health is of great significance to the safe operation and reliability of the whole power grid. Dissolved gas analysis (DGA) in oil is an important indicator for transformer condition evaluation and health evaluation, especially in the process of power equipment condition maintenance and the promotion of digital power grid. Its data quality and data accumulation are of great value for transformer health index analysis, equipment management and even smart grid development. However, due to the degradation of performance, transmission interruption or other interference factor during the operation of the on-line sensors, the monitoring data occur errors and outliers, which greatly affects the use of data, and results in a waste of resources. Considering that Copula based outlier detection (COPOD) algorithm could handle multidimensional, high and large-scale data sets, and isolation forest (IForest) algorithm had a better detection effect on dense data sets, the two algorithms were combined to mark more outliers by setting a higher detection threshold as preliminary test. To reduce the influence of both ends of data on the detection algorithm caused by the trend of data, the method of sliding window was used to detect outliers by the two algorithms respectively, and the intersection of the results was taken as the outlier data set. Due to the high threshold, the suspect outlier data set were mixed with some normal DGA data. Further, Grubbs test with high sensitivity, wide applicability and suitable for single detection was introduced, and its performance is better in small samples. Then the suspect outlier data was tested twice through Grubbs method for improving the accuracy of outlier identification, and the misidentified normal data was excluded from the detection results. After the outliers are removed, Transformer neural network is used to fill in the outliers. Mask filling and observation reconstruction were adopted to optimize the loss function and train the network structure. Finally, the predicted values using neural network were used to fill in the data to construct DGA data. Simulation data and actual online monitoring data of 11 500 kV oil-immersed transformers in operation were used to verify the performance of the algorithm. Results show that, the number of correct identification points, the number of correct identification outlier points and the average area of the receiver operating characteristic (ROC) curve of the proposed algorithm are improved by 3.5%, 29.4% and 5.0% on the same gas series compared with the traditional K-nearest neighbors (KNN). For data filling, compared with the bidirectional scaling (BiScaler) algorithm, the mean of root mean square error (RMSE) is 7.29 μL/L, and the mean of mean absolute error (MAE) is 2.66 μL/L. The performance is improved by 9.7% and 9.2% respectively, effectively improving the quality and utilization of data. The following conclusions can be drawn from the simulation analysis: (1) The combined algorithm of COPOD, IForest and Grubbs, which is used to identify outliers in transformer DGA online data, is more accurate than the traditional algorithm KNN, the misidentification of outliers is also reduced. (2) By optimizing the loss function using mask filling and observation reconstruction, the Transformer neural network is trained to improve the prediction performance and effectively complete the data filling task. (3) After the transformer DGA data is filled, the data is more concentrated and the quality is significantly improved without changing the overall data distribution. This method has moderate complexity and high reliability,and it is suitable for large-scale data preprocessing in practical applications.
Jiang Jun,Zhang Wenqian,Li Bo等. Outlier Detection and Data Reconstruction of Dissolved Gas in Oil for Power Transformers[J]. Transactions of China Electrotechnical Society, 2024, 39(17): 5521-5533.
[1] 杨童亮, 胡东, 唐超, 等. 基于SMA-VMD-GRU模型的变压器油中溶解气体含量预测[J]. 电工技术学报, 2023, 38(1): 117-130. Yang Tongliang, Hu Dong, Tang Chao, et al.Prediction of dissolved gas content in transformer oil based on SMA-VMD-GRU model[J]. Transactions of China Electrotechnical Society, 2023, 38(1): 117-130. [2] Walker C, Al Rashdan A, Agarwal V.Transformer health monitoring using dissolved gas analysis: a technical brief[J]. International Journal of Prognostics and Health Management, 2022, 13(2): 1-8. [3] 谢庆, 蔡扬, 谢军, 等. 基于ALBERT的电力变压器运维知识图谱构建方法与应用研究[J]. 电工技术学报, 2023, 38(1): 95-106. Xie Qing, Cai Yang, Xie Jun, et al.Research on construction method and application of knowledge graph for power transformer operation and maintenance based on ALBERT[J]. Transactions of China Electrotechnical Society, 2023, 38(1): 95-106. [4] 池源, 郭莹霏, 孟庆昊, 等. 主动配电网储能应用新形态:虚拟配电变压器与馈线[J]. 电力系统自动化, 2023, 47(12): 163-175. Chi Yuan, Guo Yingfei, Meng Qinghao, et al.New form of energy storage application in active distribution network: virtual distribution transformers and feeders[J]. Automation of Electric Power Systems, 2023, 47(12): 163-175. [5] 吴瞻宇, 董明, 王健一, 等. 基于模糊关联规则挖掘的电力变压器故障诊断方法[J]. 高压电器, 2019, 55(8): 157-163. Wu Zhanyu, Dong Ming, Wang Jianyi, et al.Fault diagnosis of power transformer based on fuzzy association rules mining[J]. High Voltage Apparatus, 2019, 55(8): 157-163. [6] 刘文君, 贺馨仪, 王彬, 等. 基于异常检测集成算法的油色谱在线监测数据质量评价体系[J]. 电网与清洁能源, 2022, 38(8): 16-23. Liu Wenjun, He Xinyi, Wang Bin, et al.Online monitoring data quality evaluation system for oil chromatography based on anomaly detection integrated algorithm[J]. Power System and Clean Energy, 2022, 38(8): 16-23. [7] 王文森, 杨晓西, 刘阳, 等. 基于层次聚类分析的变压器油中溶解气体在线监测数据异常检测[J]. 高压电器, 2023, 59(1): 142-147. Wang Wensen, Yang Xiaoxi, Liu Yang, et al.Anomaly detection of online monitoring data of dissolved gases in transformer oil based on hierarchical cluster analysis[J]. High Voltage Apparatus, 2023, 59(1): 142-147. [8] 吴睿涵, 何亚倩, 江军, 等. 变压器油中溶解乙炔光热干涉检测系统的温度和压强特性[J]. 电力工程技术, 2023, 42(5): 30-36. Wu Ruihan, He Yaqian, Jiang Jun, et al.Temperature and pressure characteristics of dissolved acetylene in transformer oil based on photothermal interference detection system[J]. Electric Power Engineering Technology, 2023, 42(5): 30-36. [9] Adikaram K K L B, Hussein M A, Effenberger M, et al. Data transformation technique to improve the outlier detection power of Grubbs' test for data expected to follow linear relation[J]. Journal of Applied Mathematics, 2015, 2015: 1-9. [10] 韩崇, 袁颖珊, 梅焘, 等. 基于K-means的数据流离群点检测算法[J]. 计算机工程与应用, 2017, 53(3): 58-63. Han Chong, Yuan Yingshan, Mei Tao, et al.Data stream outlier detection algorithm based on K-means[J]. Computer Engineering and Applications, 2017, 53(3): 58-63. [11] Liu Hang, Wang Youyuan, Chen Weigen.Anomaly detection for condition monitoring data using auxiliary feature vector and density‐based clustering[J]. IET Generation Transmission & Distribution, 2019, 14(1): 108-118. [12] 严英杰, 盛戈皞, 陈玉峰, 等. 基于时间序列分析的输变电设备状态大数据清洗方法[J]. 电力系统自动化, 2015, 39(7): 138-144. Yan Yingjie, Sheng Gehao, Chen Yufeng, et al.Cleaning method for big data of power transmission and transformation equipment state based on time sequence analysis[J]. Automation of Electric Power Systems, 2015, 39(7): 138-144. [13] 顾仲翔, 马宏忠, 张勇, 等. 基于VMD与优化SVM的变压器绕组松动缺陷振动信号诊断方法[J]. 高压电器, 2023, 59(1): 117-125. Gu Zhongxiang, Ma Hongzhong, Zhang Yong, et al.Vibration signal diagnosis method of transformer winding looseness defect based on VMD and optimized SVM[J]. High Voltage Apparatus, 2023, 59(1): 117-125. [14] 朱倩雯, 叶林, 赵永宁, 等. 风电场输出功率异常数据识别与重构方法研究[J]. 电力系统保护与控制, 2015, 43(3): 38-45. Zhu Qianwen, Ye Lin, Zhao Yongning, et al.Methods for elimination and reconstruction of abnormal power data in wind farms[J]. Power System Protection and Control, 2015, 43(3): 38-45. [15] 赵厚翔, 沈晓东, 吕林, 等. 基于GAN的负荷数据修复及其在EV短期负荷预测中的应用[J]. 电力系统自动化, 2021, 45(16): 143-151. Zhao Houxiang, Shen Xiaodong, Lü Lin, et al.Load data restoration based on generative adversarial network and its application in short-term load forecasting of electric vehicle[J]. Automation of Electric Power Systems, 2021, 45(16): 143-151. [16] Khoury R, Harder D W.Numerical Methods and Modelling for Engineering[M]. Cham: Springer, 2016. [17] 张若愚, 齐波, 张鹏, 等. 面向电力变压器状态评价的油中溶解气体监测数据补全方法[J]. 电力自动化设备, 2019, 39(11): 181-187. Zhang Ruoyu, Qi Bo, Zhang Peng, et al.Method for interpolating monitoring data of dissolved gas in oil for power transformer state assessment[J]. Electric Power Automation Equipment, 2019, 39(11): 181-187. [18] Che Zhengping, Purushotham S, Cho K, et al.Recurrent neural networks for multivariate time series with missing values[J]. Scientific Reports, 2018, 8: 6085. [19] Luo Yonghong, Cai Xiangrui, Zhang Ying, et al.Multivariate time series imputation with generative adversarial networks[C]//Proceedings of the 32nd International Conference on Neural Information Processing Systems, Montréal, Canada, 2018: 1603-1614. [20] 周远翔, 林孟龙, 陈健宁, 等. 基于自注意力生成对抗网络的电力设备在线监测缺失数据填补[J]. 高电压技术, 2023, 49(5): 1795-1809. Zhou Yuanxiang, Lin Menglong, Chen Jianning, et al.Missing data imputation for online monitoring of power equipment based on self-attention generative adversarial networks[J]. High Voltage Engineering, 2023, 49(5): 1795-1809. [21] 李阳, 沈小军, 张扬帆, 等. 基于速度-关联约束的风电机组风速感知异常数据识别方法[J]. 电工技术学报, 2023, 38(7): 1793-1807. Li Yang, Shen Xiaojun, Zhang Yangfan, et al.Cleaning method of wind speed outliers for wind turbines based on velocity and correlation constraints[J]. Transactions of China Electrotechnical Society, 2023, 38(7): 1793-1807. [22] 郭慧军, 李永亭, 齐咏生, 等. 两阶段CP-Copula的风电机组异常数据清洗算法[J]. 计算机仿真, 2022, 39(11): 85-91. Guo Huijun, Li Yongting, Qi Yongsheng, et al.A two-stage CP-Copula algorithm for clearing abnormal data of wind turbine[J]. Computer Simulation, 2022, 39(11): 85-91. [23] 陈雨鸽, 陈昌铭, 张思, 等. 考虑时空耦合的小水电富集型虚拟电厂优化调度策略[J]. 电力系统自动化, 2022, 46(18): 90-98. Chen Yuge, Chen Changming, Zhang Si, et al.Optimal dispatching strategy of small hydropower enriched virtual power plant considering temporal-spatial coupling[J]. Automation of Electric Power Systems, 2022, 46(18): 90-98. [24] 唐立, 郝鹏, 任沛阁, 等. 基于改进孤立森林算法的无人机异常行为检测[J]. 航空学报, 2022, 43(8): 326789. Tang Li, Hao Peng, Ren Peige, et al.UAV abnormal behavior detection based on improved IForest algorithm[J]. Acta Aeronautica et Astronautica Sinica, 2022, 43(8): 326789. [25] 徐东, 王岩俊, 孟宇龙, 等. 基于Isolation Forest改进的数据异常检测方法[J]. 计算机科学, 2018, 45(10): 155-159. Xu Dong, Wang Yanjun, Meng Yulong, et al.Improved data anomaly detection method based on isolation forest[J]. Computer Science, 2018, 45(10): 155-159. [26] 李新鹏, 高欣, 阎博, 等. 基于孤立森林算法的电力调度流数据异常检测方法[J]. 电网技术, 2019, 43(4): 1447-1456. Li Xinpeng, Gao Xin, Yan Bo, et al.An approach of data anomaly detection in power dispatching streaming data based on isolation forest algorithm[J]. Power System Technology, 2019, 43(4): 1447-1456. [27] Ding Kun, Zhang Jingwei, Ding Hanxiang, et al.Fault detection of photovoltaic array based on Grubbs criterion and local outlier factor[J]. IET Renewable Power Generation, 2020, 14(4): 551-559. [28] Vaswani A, Shazeer N, Parmar N, et al.Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, California, USA, 2017: 6000-6010. [29] 黄飞虎, 赵红磊, 弋沛玉, 等. 一种改进Transformer的电力负荷预测方法[J]. 现代电力, 2023, 40(1): 50-58. Huang Feihu, Zhao Honglei, Yi Peiyu, et al.An improved power load forecasting method based on Transformer[J]. Modern Electric Power, 2023, 40(1): 50-58. [30] 叶林, 李奕霖, 裴铭, 等. 寒潮天气小样本条件下的短期风电功率组合预测[J]. 中国电机工程学报, 2023, 43(2): 543-555. Ye Lin, Li Yilin, Pei Ming, et al.Combined approach for short-term wind power forecasting under cold weather with small sample[J]. Proceedings of the CSEE, 2023, 43(2): 543-555. [31] Du Wenjie, Côté D, Liu Yan.SAITS: self-attention-based imputation for time series[J]. Expert Systems With Applications, 2023, 219: 119619. [32] 张鹏, 齐波, 刘娟, 等. 电力变压器油中溶解气体数据的分布特征参数快速计算方法[J]. 中国电机工程学报, 2022, 42(5): 2001-2012. Zhang Peng, Qi Bo, Liu Juan, et al.Fast calculation method for distribution characteristic parameters of dissolved gas data in power transformer oil[J]. Proceedings of the CSEE, 2022, 42(5): 2001-2012. [33] Qi Bo, Zhang Peng, Rong Zhihai, et al.Differentiated warning rule of power transformer health status based on big data mining[J]. International Journal of Electrical Power & Energy Systems, 2020, 121: 106150. [34] Hastie T, Mazumder R, Lee J D, et al.Matrix completion and low-rank SVD via fast alternating least squares[J]. Journal of Machine Learning Research, 2015, 16: 3367-3402. [35] Cao Wei, Wang Dong, Li Jian, et al.BRITS: bidirectional recurrent imputation for time series[C]// Proceedings of the 32nd International Conference on Neural Information Processing Systems, Montréal, Canada, 2018: 6776-6786.
2024, Vol. 39 
