Elevator multivariate time series anomaly detection method based on dynamic spatial-temporal encoder

doi:10.11772/j.issn.1001-9081.2025060800

Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (7): 2364-2372.DOI: 10.11772/j.issn.1001-9081.2025060800

• Frontier and comprehensive applications • Previous Articles

Elevator multivariate time series anomaly detection method based on dynamic spatial-temporal encoder

Jianwei CHEN, Jiawei LU(), Qibing WANG, Mengke ZHAO

School of Mechanical and Electrical Engineering，China Jiliang University，Hangzhou Zhejiang 310018，China

Received:2025-07-18 Revised:2025-09-16 Accepted:2025-09-25 Online:2025-10-30 Published:2026-07-10
Contact: Jiawei LU
About author:CHEN Jianwei， born in 2001， M. S. candidate. His research interests include time series， anomaly detection.
WANG Qibing， born in 1978， Ph. D.， professor. His research interests include fault diagnosis， intelligentization for special equipment.
ZHAO Mengke， born in 2002， M. S. candidate. Her research interests include knowledge graph， large language model.
Supported by:
“Pioneer” and “Leading Goose” R&D Program of Zhejiang Province, China(2025C01022);Huzhou City Key Research and Development Science and Technology Program(2022ZD2019)

基于动态时空编码器的电梯多元时间序列异常检测方法

陈坚伟, 陆佳炜(), 王琪冰, 赵梦珂

中国计量大学机电工程学院，杭州 310018

通讯作者: 陆佳炜
作者简介:陈坚伟（2001—），男，浙江台州人，硕士研究生，主要研究方向：时间序列、异常检测
王琪冰（1978—），男，浙江绍兴人，教授，博士，主要研究方向：故障诊断、特种设备智能化
赵梦珂（2002—），女，河南新乡人，硕士研究生，主要研究方向：知识图谱、大语言模型。
基金资助:
浙江省“尖兵领雁”研发攻关计划项目(2025C01022);湖州市重点研发科技计划项目(2022ZD2019)

Abstract

Abstract:

The multivariate time series data generated during elevator operation are crucial for building an elevator safety monitoring system. To address the issue that existing methods ignore the correlation between series， an elevator multivariate time series anomaly detection method based on Dynamic Spatial-Temporal Encoder （DSTE） was proposed. First， a dynamic graph aggregation module was introduced to compute the global node self-attention through graph convolution and attention mechanism for spatial feature aggregation. Meanwhile， a graph pooling method was employed to dynamically update the graph structure and adjust the degree of correlation between series. Then， a temporal feature aggregation module was constructed by incorporating gated convolution modules into the traditional Temporal Convolutional Network （TCN）， and the temporal feature propagation was calculated utilizing neural ordinary differential equations， thereby realizing precise spatial-temporal feature modeling. Finally， multi-layer convolutions were used for decoding and prediction， so as to identify anomalies. Experimental results on a real-world elevator operation dataset show that compared to the suboptimal model DTAAD （Dual TCN-Attention networks for Anomaly Detection）， DSTE achieves a 0.34% improvement in precision， a 10.85% increase in recall， a 5.39% enhancement in F1-Score， and a 56.78% reduction in False Negative Rate （FNR）， while balancing the FNR and False Positive Rate （FPR）， verifying that the proposed method can identify elevator anomalies effectively and reduce the missed detection of abnormal events.

Key words: elevator, multivariate time series, anomaly detection, dynamic graph aggregation, neural ordinary differential equation

摘要：

电梯在运行过程中所产生的多元时间序列数据对于构建电梯安全监测系统至关重要。针对现有方法忽略了序列之间相关性影响的问题，提出一种基于动态时空编码器的电梯多元时间序列异常检测方法——DSTE （Dynamic Spatial-Temporal Encoder）。首先，引入动态图聚合模块，通过图卷积与注意力机制计算全局节点的自注意力，并用得到的结果进行空间特征聚合；同时，采用图池化的方法进行动态图结构更新，动态地调整序列间的关联程度；其次，构建时域特征聚合模块，向传统的时间卷积神经网络（TCN）添加门控卷积模块，并通过神经常微分方程进行时间特征传播的计算，从而实现精准的时空特征建模；最后，结合多层卷积进行解码预测，并判断电梯异常。在真实电梯运行数据集上的实验结果表明，与次优模型DTAAD （Dual TCN-Attention networks for Anomaly Detection）相比， DSTE进行异常检测的精确率提升了0.34%，召回率提升了10.85%， F1-Score提升5.39%，漏报率（FNR）降低了56.78%，同时实现了FNR与误报率（FPR）之间的平衡。本文方法能有效识别电梯异常，并降低异常事件的漏检概率。

关键词: 电梯, 多元时间序列, 异常检测, 动态图聚合, 神经常微分方程

CLC Number:

TP391.9

Jianwei CHEN, Jiawei LU, Qibing WANG, Mengke ZHAO. Elevator multivariate time series anomaly detection method based on dynamic spatial-temporal encoder[J]. Journal of Computer Applications, 2026, 46(7): 2364-2372.

陈坚伟, 陆佳炜, 王琪冰, 赵梦珂. 基于动态时空编码器的电梯多元时间序列异常检测方法[J]. 《计算机应用》唯一官方网站, 2026, 46(7): 2364-2372.

Figures/Tables 9

Fig. 1 Framework of DSTE

Tab. 1 Symbol description

符号	说明
$A 0$	初始图邻接矩阵
$A o u t$	经过图池化的邻接矩阵
$X ∈ R N × D × S$	历史多元时间序列数据
$X s ∈ R N × D$	时间点s下的时序数据
$X n ∈ R S × D$	第n个传感器的时序数据
$X t + 1 : t + T ∈ R N × D × T$	长度为T的输入时序特征矩阵
$X' ∈ R N × D × T$	经动态图聚合更新后的时序特征矩阵
$H T ∈ R N × D' × Q$	时域特征聚合模块输入特征
$H T (t *) ∈ R N × D' × Q$	积分时间 $t *$ 时刻隐藏层状态
$H o u t T ∈ R N × D'$	时域特征聚合模块输出
$Y ∈ R N × D × τ$	解码器预测结果

Tab. 1 Symbol description

符号	说明
$A 0$	初始图邻接矩阵
$A o u t$	经过图池化的邻接矩阵
$X ∈ R N × D × S$	历史多元时间序列数据
$X s ∈ R N × D$	时间点s下的时序数据
$X n ∈ R S × D$	第n个传感器的时序数据
$X t + 1 : t + T ∈ R N × D × T$	长度为T的输入时序特征矩阵
$X' ∈ R N × D × T$	经动态图聚合更新后的时序特征矩阵
$H T ∈ R N × D' × Q$	时域特征聚合模块输入特征
$H T (t *) ∈ R N × D' × Q$	积分时间 $t *$ 时刻隐藏层状态
$H o u t T ∈ R N × D'$	时域特征聚合模块输出
$Y ∈ R N × D × τ$	解码器预测结果

Fig. 2 Self-attention graph pooling

Tab. 2 Elevator data types （partial）

类别	数据名称	数据类型
电梯控制柜数据（部分）	曳引机额定电流	Float
	曳引机额定转矩	Float
	曳引机额定转速	Float
	曳引机实际电流	Float
	曳引机实际转速	Float
	曳引机输出电压	Float
传感器采集数据（部分）	电梯速度	Varchar（max）
	X轴振动加速度	Varchar（max）
	Y轴振动加速度	Varchar（max）
	Z轴振动加速度	Varchar（max）
	轿厢温度	Float

Tab. 3 Comparison experimental results of anomaly detection

方法	Precision	Recall	F1-Score
LSTM-VAE	0.572	0.681	0.622
OmniAnomaly	0.597	0.623	0.610
MAD-GAN	0.644	0.604	0.623
DAGMM	0.643	0.658	0.650
MSCRED	0.778	0.662	0.715
MTAD-GAT	0.803	0.764	0.783
Xu方法^［30］	0.851	0.736	0.789
DTAAD	0.868	0.838	0.853
DSTE	0.871	0.929	0.899

Tab. 4 Comparison of FPR and FNR

方法	FPR	FNR
LSTM-VAE	0.074	0.319
OmniAnomaly	0.062	0.377
MAD-GAN	0.049	0.395
DAGMM	0.054	0.342
MSCRED	0.025	0.282
MTAD-GAT	0.027	0.236
Xu方法^［10］	0.019	0.264
DTAAD	0.018	0.162
DSTE	0.020	0.070

Fig. 3 Comparison of different time windows and prediction horizons

Fig. 4 Results of ablation experiments

Fig. 5 Visualization of anomaly detection

References 31

[1]	Wang Q， Leng Y， Li D， et al. MCU system-based intelligent high-speed elevator door operator fault analysis and research ［J］. IOP Conference Series： Materials Science and Engineering， 2018， 428： No.012028.
[2]	Xie P， Zhang L， Li M， et al. An elevator door anomaly detection method based on improved deep multi-sphere support vector data description ［J］. Computers and Electrical Engineering， 2024， 120（Pt B）： No.109660.
[3]	Ishimtsev V， Bernstein A， Burnaev E， et al. Conformal k-NN anomaly detector for univariate data streams ［C］// COPA 2017. New York： JMLR.org， 2017： 213-227.
[4]	Klein P， Malburg L， Bergmann R. Combining informed data-driven anomaly detection with knowledge graphs for root cause analysis in predictive maintenance ［J］. Engineering Applications of Artificial Intelligence， 2025， 145： No.110152.
[5]	Munir M， Siddiqui S A， Dengel A， et al. DeepAnT： a deep learning approach for unsupervised anomaly detection in time series ［J］. IEEE Access， 2019， 7： 1991-2005.
[6]	Luo Q， Chen J， Zi Y， et al. Multi-mode non-Gaussian variational autoencoder network with missing sources for anomaly detection of complex electromechanical equipment ［J］. ISA Transactions， 2023， 134： 144-158.
[7]	袁安妮，邹春明，王勇，等.基于改进TCN的多元时间序列异常检测算法［J］.计算机工程与设计， 2025， 46（2）： 416-422.
	Yuan Aanni， Zou Chunming， Wang Yong， et al. Multivariate time series anomaly detection algorithm based on improved TCN ［J］. Computer Engineering and Design， 2025， 46（2）： 416-422.
[8]	侯立，王健.改进密度峰值聚类的多源数据异常值检测算法［J］.计算机仿真， 2024， 41（6）： 565-569.
	Hou Li， Wang Jian. Outlier detection algorithm for multi-source data based on improved density peak clustering ［J］. Computer Simulation， 2024， 41（6）： 565-569.
[9]	尹春勇，张不凡.基于多尺度的多变量时间序列异常检测模型［J］.计算机应用， 2026， 46（3）： 790-797.
	Yin Chunyong， Zhang Bufan. Multi-scale based multivariate time series anomaly detection model ［J］. Journal of Computer Applications， 2026， 46（3）： 790-797.
[10]	Zhang C， Song D， Chen Y， et al. A deep neural network for unsupervised anomaly detection and diagnosis in multivariate time series data ［C］// AAAI 2019. Palo Alto： AAAI Press， 2019： 1409-1416.
[11]	Zhao H， Wang Y， Duan J， et al. Multivariate time-series anomaly detection via graph attention network ［C］// ICDM 2020. Piscataway： IEEE， 2020： 841-850.
[12]	Wang S， Guo Y， Huang S， et al. A deep graph neural network-based link prediction model for proactive anomaly detection in discrete manufacturing workshop ［J］. Journal of Manufacturing Systems， 2025， 79： 301-317.
[13]	Chen W， Tian L， Chen B， et al. Deep variational graph convolutional recurrent network for multivariate time series anomaly detection ［C］// ICML 2022. New York： JMLR.org， 2022： 3621-3633.
[14]	高雄飞.基于振动信号的电梯运行状态监测与预测技术研究［D］.沈阳：东北大学， 2022： 17-20.
	Gao Xiongfei. Research on elevator running state monitoring and prediction technology based on vibration signal ［D］. Shenyang： Northeastern University， 2022： 17-20.
[15]	周宇偲，单志勇，潘峰.基于LSTM-VAE的电梯异常检测［J］.自动化与仪器仪表， 2022（4）： 6-10.
	Zhou Yusi， Shan Zhiyong， Pan Feng. Elevator anomaly detection based on LSTM-VAE model ［J］. Automation and Instrumentation， 2022（4）： 6-10.
[16]	范玉川，范波，陈卓，等.电梯门机异常检测和故障定位方法［J］.东北大学学报（自然科学版）， 2024， 45（6）： 761-768.
	Fan Yuchuan， Fan Bo， Chen Zhuo， et al. A method for anomaly detection and fault diagnosis of elevator door machine ［J］. Journal of Northeastern University （Natural Science）， 2024， 45（6）： 761-768.
[17]	Park J， Kim Y， Na K， et al. Variance of Energy Residual （VER）： an efficient method for planetary gear fault detection under variable-speed conditions ［J］. Journal of Sound and Vibration， 2019， 453： 253-267.
[18]	Wang Y， Perry M， Whitlock D， et al. Detecting anomalies in time series data from a manufacturing system using recurrent neural networks ［J］. Journal of Manufacturing Systems， 2022， 62： 823-834.
[19]	Song H， Sun Y， Gao H， et al. Residual feature-based semi-supervised anomaly detection for Gearbox under varying rotational speeds ［J］. Measurement， 2025， 256（Pt B）： No.118236.
[20]	Haider Z A， Zeb A， Rahman T， et al. A survey on anomaly detection in IoT： techniques， challenges， and opportunities with the integration of 6G ［J］. Computer Networks， 2025， 270： No.111484.
[21]	Li G， Ge M， Wan J， et al. MemMambaAD： memory-augmented state space model for multivariate time series anomaly detection ［J］. Engineering Applications of Artificial Intelligence， 2025， 158（Pt A）： No.111308.
[22]	Wang K， Kong J， Zhang M， et al. MAD-DGTD： multivariate time series anomaly detection based on dynamic graph structure learning with time delay ［J］. Neurocomputing， 2025， 635： No.129887.
[23]	Zhao M， Peng H， Li L， et al. Multivariate time series anomaly detection based on spatial-temporal network and Transformer in industrial internet of things ［J］. Computers， Materials and Continua， 2024， 80（2）： 2815-2837.
[24]	Jia X， Xun P， Peng W， et al. Deep anomaly detection for time series： a survey ［J］. Computer Science Review， 2025， 58： No.100787.
[25]	Siffer A， Fouque P A， Termier A， et al. Anomaly detection in streams with extreme value theory ［C］// KDD 2017. New York： ACM， 2017： 1067-1075.
[26]	Park D， Hoshi Y， Kemp C C. A multimodal anomaly detector for robot-assisted feeding using an LSTM-based variational autoencoder ［J］. IEEE Robotics and Automation Letters， 2018， 3（3）： 1544-1551.
[27]	Su Y， Zhao Y， Niu C， et al. Robust anomaly detection for multivariate time series through stochastic recurrent neural network ［C］// KDD 2019. New York： ACM， 2019： 2828-2837.
[28]	Li D， Chen D， Jin B， et al. MAD-GAN： multivariate anomaly detection for time series data with generative adversarial networks ［C］// ICANN 2019. Cham： Springer， 2019： 703-716.
[29]	Zong B， Song Q， Min M R， et al. Deep autoencoding Gaussian mixture model for unsupervised anomaly detection ［EB/OL］. ［2025-09-09］. .
[30]	Xu J， Wu H X， Wang J， et al. Anomaly Transformer： time series anomaly detection with association discrepancy ［EB/OL］. ［2025-09-09］. .
[31]	Yu L R， Lu Q H， Xue Y. DTAAD： dual TCN-attention networks for anomaly detection in multivariate time series data ［J］. Knowledge-Based Systems， 2024， 295： No.111849.

Elevator multivariate time series anomaly detection method based on dynamic spatial-temporal encoder

基于动态时空编码器的电梯多元时间序列异常检测方法

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 9

References 31

Related Articles 15

Recommended Articles

Metrics

[1]	Zhiqiang REN, Xuebin CHEN, Changsheng QU. Active defense framework against Byzantine attacks in federated learning [J]. Journal of Computer Applications, 2026, 46(7): 2196-2207.
[2]	Binhong XIE, Erdan ZHU, Rui ZHANG. Appearance-motion collaborative modeling for video anomaly detection [J]. Journal of Computer Applications, 2026, 46(5): 1551-1559.
[3]	Yifan SUO, Songhua LIU, Qiuzhi HAO. Time series anomaly detection method based on high-order feature aggregation [J]. Journal of Computer Applications, 2026, 46(4): 1131-1138.
[4]	Chunyong YIN, Bufan ZHANG. Multi-scale based multivariate time series anomaly detection model [J]. Journal of Computer Applications, 2026, 46(3): 790-797.
[5]	Limei DONG, Yanzi LI, Jiayin LI, Li XU. Neighborhood-enhanced unsupervised graph anomaly detection [J]. Journal of Computer Applications, 2026, 46(2): 458-466.
[6]	Feng HAN, Yongfeng BU, Haoxiang LIANG, Shuwen HUANG, Zhaoyang ZHANG, Shijie SUN. Vehicle trajectory anomaly detection based on multi-level spatio-temporal interaction dependency [J]. Journal of Computer Applications, 2026, 46(2): 604-612.
[7]	Qi ZHONG, Shufen ZHANG, Zhenbo ZHANG, Yinlong JIAN, Zhongrui JING. Detection and defense mechanism for poisoning attacks to federated learning [J]. Journal of Computer Applications, 2026, 46(2): 445-457.
[8]	Xiang WANG, Zhixiang CHEN, Guojun MAO. Multivariate time series prediction method combining local and global correlation [J]. Journal of Computer Applications, 2025, 45(9): 2806-2816.
[9]	Longbo YAN, Wentao MAO, Zhihong ZHONG, Lilin FAN. Robust unsupervised multi-task anomaly detection method for defect diagnosis of urban drainage pipe network [J]. Journal of Computer Applications, 2025, 45(6): 1833-1840.
[10]	Wenpeng WANG, Yinchang QIN, Wenxuan SHI. Review of unsupervised deep learning methods for industrial defect detection [J]. Journal of Computer Applications, 2025, 45(5): 1658-1670.
[11]	Zihe CHEN, Bin CHEN. Unsupervised point cloud anomaly detection based on multi-representation fusion [J]. Journal of Computer Applications, 2025, 45(5): 1677-1685.
[12]	Lihu PAN, Shouxin PENG, Rui ZHANG, Zhiyang XUE, Xuzhen MAO. Video anomaly detection for moving foreground regions [J]. Journal of Computer Applications, 2025, 45(4): 1300-1309.
[13]	Zhenhua XUE, Qiang LI, Chao HUANG. Vision foundation model-driven pixel-level image anomaly detection method [J]. Journal of Computer Applications, 2025, 45(3): 823-831.
[14]	Dixin WANG, Jiahao WANG, Min LI, Hao CHEN, Guangyao HU, Yu GONG. Abnormal attack detection for underwater acoustic communication network [J]. Journal of Computer Applications, 2025, 45(2): 526-533.
[15]	Miaolei DENG, Yupei KAN, Chuanchuan SUN, Haihang XU, Shaojun FAN, Xin ZHOU. Summary of network intrusion detection systems based on deep learning [J]. Journal of Computer Applications, 2025, 45(2): 453-466.