面向电力设备异常检测的深度自编码支持向量数据描述模型研究

doi:10.3969/j.issn.1008-0198.2024.01.017

湖南电力 ›› 2024, Vol. 44 ›› Issue (1): 119-127.doi: 10.3969/j.issn.1008-0198.2024.01.017

面向电力设备异常检测的深度自编码支持向量数据描述模型研究

耿波¹, 潘曙辉², 董晓旭³

1.核动力运行研究所,湖北武汉 430223;
2.华中科技大学计算机科学与技术学院,湖北武汉 430074;
3.华北理工大学冶金与能源学院,河北唐山 063210

收稿日期:2023-09-07 修回日期:2023-10-09 出版日期:2024-02-25 发布日期:2024-03-11
通信作者: 董晓旭(2000),女,硕士,研究方向为热力系统模拟优化。
作者简介:耿波(1974),男,硕士,高级工程师,研究方向为设备监测与故障诊断。
基金资助:
国家自然科学基金青年基金资助项目(52004094);河北省自然科学基金项目(E2021209037);中央引导地方科技发展基金项目(236Z1017G)

Research on Deep Auto Encoding Support Vector Data Description Model for Power Equipment Anomaly Detection

GENG Bo¹, PAN Shuhui², DONG Xiaoxu³

1. Research Institute of Nuclear Power Operation, Wuhan 430223, China;
2. School of Computer Science and Technology, Huazhong University of Science and Technology, Wuhan 430074, China;
3. School of Metallurgy and Energy, North China University of Science and Technology, Tangshan 063210, China

Received:2023-09-07 Revised:2023-10-09 Online:2024-02-25 Published:2024-03-11

摘要/Abstract

摘要： 针对深度自编码支持向量数据描述模型对电力设备部分异常区分能力不足的问题,提出自监督混合专家增强的深度自编码支持向量数据描述模型,构造多种自监督变换数据集模拟潜在未知异常,引入自监督分类和掩码重构任务以学习更具区分性的表示。此外,将编码器部分改造为混合专家模型结构,将数据分配给不同专家子模块进行专业化的学习,使异常决策边界更清晰。在4个公开数据集和3个电厂设备数据集上的实验结果证实了自监督学习和混合专家模型的有效性。

关键词: 异常检测, 深度自编码支持向量数据描述, 自监督学习, 混合专家模型

Abstract: In response to the problem of insufficient ability of deep auto encoding support vector data description model to distinguish some anomalies of power equipment, this paper proposes deep auto encoding support vector data description model with self-supervised and mixture-of-experts.The multiple self-supervised transformation datasets is constructed to simulate potential unknown anomalies, and the self-supervised classification and mask reconstruction tasks is introduced to learn more discriminative representations. In addition, the encoder is transformed into a mixture-of-experts structure, and the data is allocated to different expert sub-modules for professional learning, making the abnormal decision boundary clearer. Experimental results from four public data sets and three power plant equipment data sets prove the effectiveness of self-supervised learning and mixture-of-experts.

Key words: anomaly detection, deep auto encoding support vector data description, self-supervised learning, mixture-of-experts

中图分类号:

TM762

耿波, 潘曙辉, 董晓旭. 面向电力设备异常检测的深度自编码支持向量数据描述模型研究[J]. 湖南电力, 2024, 44(1): 119-127.

GENG Bo, PAN Shuhui, DONG Xiaoxu. Research on Deep Auto Encoding Support Vector Data Description Model for Power Equipment Anomaly Detection[J]. Hunan Electric Power, 2024, 44(1): 119-127.

参考文献

[1] LENNOX B, MONTAGUE G A, HIDEN H G, et al. Process monitoring of an industrial fed-batch fermentation[J]. Biotechnology and Bioengineering, 2001, 74(2):125-135.
[2] 本刊综合.党的十八大以来工业和信息化发展成就[J]. 中国信息化, 2022(6):17-23.
[3] CHEN J H, SATHE S, AGGARWAL C, et al. Outlier detection with autoencoder ensembles[M]//Proceedings of the 2017 SIAM International Conference on Data Mining.philadelphia,PA:Society for Industrial and Applied Mathematics,2017:90-98.
[4] NGUYEN M N, VIEN N A. Scalable and interpretable one-class SVMS with deep learning and random fourier features[C]//Joint European Conference on Machine Learning and Knowledge Discovery in Databases.cham:Springer,2019:157-172.
[5] SCHÖLKOPF B, PLATT J C, SHAWE-TAYLOR J, et al. Estimating the support of a high-dimensional distribution[J]. Neural Computation, 2001, 13(7):1443-1471.
[6] RUFF L, VANDERMEULEN R, GOERNITZ N, et al. Deep one-class classification[C]//International Conference on Machine Learning. Stockholm, 2018:4393-4402.
[7] ZONG B, SONG Q, MIN M R, et al. Deep autoencoding Gaussian mixture model for unsupervised anomaly detection[C]//International Conference on Learning Representations. BC, 2018:1-19.
[8] CHEN Y, ASHIZAWA N, YEO C K, et al. Multi-scale self-organizing map assisted deep autoencoding gaussian mixture model for unsupervised intrusion detection[J]. Knowledge-Based Systems, 2021, 224(1):107086.
[9] WANG H, PANG G, SHEN C, et al. Unsupervised representation learning by predicting random distances[C]//Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence.Yokohama,Japan.California:International Joint Conferences on Artificial Intelligence Organization,2020:2950-2956.
[10] PANG G S, CAO L B, CHEN L, et al. Learning representations of ultrahigh-dimensional data for random distance-based outlier detection[C]//Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. London,United Kindom New York:ACM,2018:2041-2050.
[11] LIU Y Z, LI Z, ZHOU C, et al. Generative adversarial active learning for unsupervised outlier detection[J]. IEEE Transactions on Knowledge and Data Engineering, 2020, 32(8):1517-1528.
[12] NGO P C, WINARTO A A, KOU C K L, et al. Fence GAN: towards better anomaly detection[C]//2019 IEEE 31st international conference on tools with artificial intelligence (ICTAI).Portland,OR,USA,IEEE,2020:141-148.
[13] HOJJATI H, ARMANFARD N. Dasvdd: deep autoencoding support vector data descriptor for anomaly detection[J].IEEE Transactions on Knowledge and Data Engineering,2023(10):1-12.
[14] ZHANG Z, DENG X G. Anomaly detection using improved deep SVDD model with data structure preservation[J]. Pattern Recognition Letters, 2021, 148:1-6.
[15] HOJJATI H, HO T K K, ARMANFARD N. Self-Supervised anomaly detection in computer vision and beyond: A survey and outlook[J]. Neural Networks,2024(172):106106.
[16] ZHAO Y E, ZHENG G Q, MUKHERJEE S, et al. Admoe: Anomaly detection with mixture-of-experts from noisy labels[J]//Proceedings of the AAAI Conference on Artificial Intelligence. Washington, 2023,37(4):4937-4945.
[17] SCHULZE J-P, SPERL P, BÖTTINGER K. Anomaly detection by recombining gated unsupervised experts[C]//2022 International Joint Conference on Neural Networks (IJCNN). Padua, Italy.IEEE,2022:1-8.
[18] LI Z, ZHAO Y, BOTTA N, et al. COPOD: copula-based outlier detection[C]//2020 IEEE International Conference on Data Mining (ICDM). Sorrento,Italy.IEEE,2021:1118-1123.
[19] HAN S Q, HU X Y, HUANG H L, et al. Adbench: Anomaly detection benchmark[J]. SSRN Electronics Journal,2022, 35:32142-32159.
[20] ASUNCION A, NEWMAN D. UCI machine learning repository[EB/OL] (2007-01-01)[2023-09-04].http://www.ics.uci.edu/~mlearn/MLRepository.html.

面向电力设备异常检测的深度自编码支持向量数据描述模型研究

Research on Deep Auto Encoding Support Vector Data Description Model for Power Equipment Anomaly Detection

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 2

编辑推荐

Metrics

本文评价

[1]	尹超勇, 李刚, 李辉, 徐彪, 刘伟良, 熊尚峰. 基于5G通信方式的变电站远方备自投原理及实现方法[J]. 湖南电力, 2023, 43(1): 58-62.
[2]	龙雪梅, 许立强, 徐彪, 欧阳帆, 李辉. 标准化安全稳定控制系统的策略及分配研究[J]. 湖南电力, 2022, 42(6): 104-109.