电信科学 ›› 2020, Vol. 36 ›› Issue (7): 136-145.doi: 10.11959/j.issn.1000-0801.2020188

• 运营技术广角 • 上一篇    下一篇

一种基于LSTM自动编码机的工业系统异常检测方法

沈潇军1,葛亚男2,沈志豪1,倪阳旦1,吕明琪2,翁正秋3()   

  1. 1 国网浙江省电力有限公司,浙江 杭州 310007
    2 浙江工业大学,浙江 杭州 310023
    3 温州职业技术学院,浙江 温州 325035
  • 修回日期:2020-06-23 出版日期:2020-07-20 发布日期:2020-07-28
  • 作者简介:沈潇军(1975- ),男,国网浙江省电力有限公司高级工程师,主要研究方向为电力信息化技术及信息管理|葛亚男(1994- ),女,浙江工业大学研究生,主要研究方向为网络安全|沈志豪(1986- ),男,国网浙江省电力有限公司高级工程师,主要研究方向为电力系统信息化运维|倪阳旦(1986- ),男,国网浙江省电力有限公司高级工程师,主要研究方向为电力信息技术|吕明琪(1981- ),男,浙江工业大学副教授,主要研究方向为网络安全、移动安全、系统安全|翁正秋(1981- ),女,温州职业技术学院副教授,主要研究方向为数据安全与大数据技术
  • 基金资助:
    国家自然科学基金资助项目(U1936215);浙江省自然科学基金资助项目(LY18F020033);工业和信息化部2019年工业互联网创新发展工程(TC190H3WN);温州市551人才工程科研项目(温人社发〔2020〕61号(5);温人社发〔2019〕55号(17);温州职业技术学院重大科研课题(WZY2020001)

An LSTM auto-encoder based anomaly detection for industrial system

Xiaojun SHEN1,Yanan GE2,Zhihao SHEN1,Yangdan NI1,Mingqi LV2,Zhengqiu WENG3()   

  1. 1 State Grid Zhejiang Electric Power Company,Hangzhou 310007,China
    2 Zhejiang University of Technology,Hangzhou 310023,China
    3 Wenzhou Polytechnics,Wenzhou 325035,China
  • Revised:2020-06-23 Online:2020-07-20 Published:2020-07-28
  • Supported by:
    The National Natural Science Foundation of China(U1936215);The Natural Science Foundation of Zhejiang Province of China(LY18F020033);2019 Industrial Internet Innovation and Development Project of the Ministry of Industry and Information Technology(TC190H3WN);Wenzhou Scientific Research Projects for 551 Talent Engineering Project (WenRenSheFa[2020]61(5);WenRenSheFa[2019]55(17);Major Scientific Research Projects of Wenzhou Polytechnics(WZY2020001)

摘要:

在工业互联网的环境下,自动有效的异常检测方法对工业系统的安全、稳定生产具有重要的意义。传统的异常检测方法存在需要大量标注样本、不适应高维度时序数据等不足,提出一种基于LSTM自动编码机的工业系统异常检测方法。为克服现有方法依赖标注样本的不足,提出采用自动编码机,通过无监督的方式学习大量正常样本的特征和模式,在此基础上通过对样本进行重构和计算重构误差的方式进行异常检测。其次,为克服现有方法不适应高维度时序数据的不足,提出采用双向LSTM作为编码器,进而挖掘多维时序数据的潜在特征。基于一个真实造纸工业的数据集的实验表明,所提方法在各项指标上都对现有无监督异常检测方法有一定的提升,检测的总体精度达到了93.4%。

关键词: 异常检测, 工业互联网, 自动编码机, LSTM

Abstract:

In the context of the industrial internet,automatic and effective anomaly detection methods are of great significance to the safe and stable production of industrial systems.Traditional anomaly detection methods have the disadvantages of requiring a large number of labeled samples,and not adapting to high-dimensional time series data.Aiming at these limitations,an industrial system anomaly detection method based on LSTM (long short-term memory)auto-encoder was proposed.Firstly,to address the limitation of relying on labeled samples,an encoder used to learn the features and patterns of a large number of normal samples in an unsupervised manner.Then,anomaly detection was performed via reconstructing samples and calculating the reconstruction error.Secondly,to adapt to high-dimensional time series data,a BiLSTM (bidirectional LSTM) was used as an encoder,and then the potential characteristics of multi-dimensional time series data were mined.Experiments based on a real paper industry data set which demonstrate this method has improved the existing unsupervised anomaly detection methods in various indicators,and the overall accuracy of the detection has reached 93.4%.

Key words: abnormal detection, industrial internet, auto-encoder, LSTM

中图分类号: 

No Suggested Reading articles found!