Application of Hybrid Algorithm Based on LSTM in Anomaly Detection of User Electricity Data

Chen Min

Integrated Circuits and Embedded Systems ›› 2022, Vol. 22 ›› Issue (10) : 21-24.

PDF(999 KB)
PDF(999 KB)
Integrated Circuits and Embedded Systems ›› 2022, Vol. 22 ›› Issue (10) : 21-24.
TOPICAL DISCUSS

Application of Hybrid Algorithm Based on LSTM in Anomaly Detection of User Electricity Data

  • Chen Min
Author information +
History +

Abstract

In the paper,an anomaly detection model of user power data based on LSTM hybrid algorithm is proposed based on the study of CNN and LSTM.Firstly,a detector for feature extraction of power data is proposed.The detector extracts abstract features from samples,reconstructs the input,and compares the reconstruction error with the threshold to realize the high-dimensional feature representation of power users.Secondly,this paper proposes a hybrid cnn-lstm power data anomaly analysis network,which uses CNN superposition feature representation,and captures the context of power data based on double-layer LSTM,so as to effectively improve the classification and regression ability of the model.In the experimental stage,taking the power data provided by a power company as an example,the proposed model is verified.The results show that the performance of the proposed model is the best,and the accuracy and recall are 89.3% and 69% respectively.

Key words

power system / anomaly detection / deep learning / long and short term memory / convolutional neural network

Cite this article

Download Citations
Chen Min. Application of Hybrid Algorithm Based on LSTM in Anomaly Detection of User Electricity Data[J]. Integrated Circuits and Embedded Systems, 2022, 22(10): 21-24

References

[1] 毛龙灿,杨南.基于大数据背景的皮革人才培养优化研究[J].中国皮革,2021,50(9):3841.
[2] 杨涛.互联网时代下皮革行业电商物流体系研究分析[J].中国皮革,2021,50(8):8285.
[3] 杨森,马文,孙占功,等.一种基于大数据的低压单相用户窃电分析方法[J].电力信息与通信技术,2022,20(2):3441.
[4] 谢连芳.大数据架构下的用电评价与反窃电技术探讨[J].单片机与嵌入式系统应用,2022,22(2):2226.
[5] 梁捷,梁广明,黄水莲.基于LSSVM的用户异常用电行为辨识研究[J].青海电力,2021,40(2):2731,65.
[6] 王安军,韩丽,周亚静.基于用电数据和机器学习的窃电行为识别方法[J].信息技术,2021(5):116121.
[7] 吴少雷,肖坚红,张波,等.基于阻抗分析的居民用户异常用电研判模型研究[J].电器与能效管理技术,2019(3):3034,67.
[8] 蔡嘉辉,王琨,董康,等.基于DenseNet和随机森林的电力用户窃电检测[J].计算机应用,2021,41(S1):7580.
[9] 王楚,王忠锋,徐志远,等.电力营业厅智能服务机器人个性化情感分析技术研究[J].供用电,2020,37(6):2126.
[10] 招景明,唐捷,潘峰,等.基于SDAE和双模型联合训练的低压用户窃电检测方法[J].电测与仪表,2021,58(12):161168.
[11] 马吉科,祝永晋,许杰雄.基于卷积神经网络的计量装置运行状态异常诊断方法[J].电工电气,2020(12):5660.
[12] 万磊,陈成,黄文杰,等.基于BRB和LSTM网络的电力大数据用电异常检测方法[J].电力建设,2021,42(8):3845.
PDF(999 KB)

47

Accesses

0

Citation

Detail

Sections
Recommended

/