MSDG: Multi-Scale Dynamic Graph Neural Network for Industrial Time Series Anomaly Detection
A large number of sensors are typically installed in industrial plants to collect real-time operational data. These sensors monitor data with time series correlation and spatial correlation over time. In previous studies, GNN has built many successful models to deal with time series data, but most o...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-11-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/24/22/7218 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850227718570901504 |
|---|---|
| author | Zhilei Zhao Zhao Xiao Jie Tao |
| author_facet | Zhilei Zhao Zhao Xiao Jie Tao |
| author_sort | Zhilei Zhao |
| collection | DOAJ |
| description | A large number of sensors are typically installed in industrial plants to collect real-time operational data. These sensors monitor data with time series correlation and spatial correlation over time. In previous studies, GNN has built many successful models to deal with time series data, but most of these models have fixed perspectives and struggle to capture the dynamic correlations in time and space simultaneously. Therefore, this paper constructs a multi-scale dynamic graph neural network (MSDG) for anomaly detection in industrial sensor data. First, a multi-scale sliding window mechanism is proposed to input different scale sensor data into the corresponding network. Then, a dynamic graph neural network is constructed to capture the spatial–temporal dependencies of multivariate sensor data. Finally, the model comprehensively considers the extracted features for sequence reconstruction and utilizes the reconstruction errors for anomaly detection. Experiments have been conducted on three real public datasets, and the results show that the proposed method outperforms the mainstream methods. |
| format | Article |
| id | doaj-art-005b5dfa9fa04a39abebed6a27687e3d |
| institution | OA Journals |
| issn | 1424-8220 |
| language | English |
| publishDate | 2024-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj-art-005b5dfa9fa04a39abebed6a27687e3d2025-08-20T02:04:44ZengMDPI AGSensors1424-82202024-11-012422721810.3390/s24227218MSDG: Multi-Scale Dynamic Graph Neural Network for Industrial Time Series Anomaly DetectionZhilei Zhao0Zhao Xiao1Jie Tao2School of Computer Science and Engineering, Hunan University of Science and Technology, Xiangtan 411201, ChinaSchool of Mechanical Engineering, Hunan University of Science and Technology, Xiangtan 411201, ChinaSchool of Computer Science and Engineering, Hunan University of Science and Technology, Xiangtan 411201, ChinaA large number of sensors are typically installed in industrial plants to collect real-time operational data. These sensors monitor data with time series correlation and spatial correlation over time. In previous studies, GNN has built many successful models to deal with time series data, but most of these models have fixed perspectives and struggle to capture the dynamic correlations in time and space simultaneously. Therefore, this paper constructs a multi-scale dynamic graph neural network (MSDG) for anomaly detection in industrial sensor data. First, a multi-scale sliding window mechanism is proposed to input different scale sensor data into the corresponding network. Then, a dynamic graph neural network is constructed to capture the spatial–temporal dependencies of multivariate sensor data. Finally, the model comprehensively considers the extracted features for sequence reconstruction and utilizes the reconstruction errors for anomaly detection. Experiments have been conducted on three real public datasets, and the results show that the proposed method outperforms the mainstream methods.https://www.mdpi.com/1424-8220/24/22/7218multi-scale sliding window mechanismgraph neural networklong short-term memorymultivariate sensor monitoring dataindustrial equipmentspatial–temporal correlations |
| spellingShingle | Zhilei Zhao Zhao Xiao Jie Tao MSDG: Multi-Scale Dynamic Graph Neural Network for Industrial Time Series Anomaly Detection Sensors multi-scale sliding window mechanism graph neural network long short-term memory multivariate sensor monitoring data industrial equipment spatial–temporal correlations |
| title | MSDG: Multi-Scale Dynamic Graph Neural Network for Industrial Time Series Anomaly Detection |
| title_full | MSDG: Multi-Scale Dynamic Graph Neural Network for Industrial Time Series Anomaly Detection |
| title_fullStr | MSDG: Multi-Scale Dynamic Graph Neural Network for Industrial Time Series Anomaly Detection |
| title_full_unstemmed | MSDG: Multi-Scale Dynamic Graph Neural Network for Industrial Time Series Anomaly Detection |
| title_short | MSDG: Multi-Scale Dynamic Graph Neural Network for Industrial Time Series Anomaly Detection |
| title_sort | msdg multi scale dynamic graph neural network for industrial time series anomaly detection |
| topic | multi-scale sliding window mechanism graph neural network long short-term memory multivariate sensor monitoring data industrial equipment spatial–temporal correlations |
| url | https://www.mdpi.com/1424-8220/24/22/7218 |
| work_keys_str_mv | AT zhileizhao msdgmultiscaledynamicgraphneuralnetworkforindustrialtimeseriesanomalydetection AT zhaoxiao msdgmultiscaledynamicgraphneuralnetworkforindustrialtimeseriesanomalydetection AT jietao msdgmultiscaledynamicgraphneuralnetworkforindustrialtimeseriesanomalydetection |