Topology and FDIA identification in distribution system state estimation using a data-driven approach

In general, the solution of distribution system state estimation (DSSE) is highly dependent on the accuracy of measurement data and accurate topology data. Since the measurement devices are more vulnerable to different types of cyberattacks such as denial-of-service (DoS) attacks, outliers, and stru...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Measurement : journal of the International Measurement Confederation Ročník 253; s. 117741
Hlavní autori: Raghuvamsi, Y., Batchu, Sreenadh, Teeparthi, Kiran
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Elsevier Ltd 01.09.2025
Predmet:
Denoising autoencoder
Distribution system state estimation
False data injection attacks
Graph-based temporal convolutional network
Line current sensors
Topology identification
Distribution system state estimation
Graph-based temporal convolutional network
Denoising autoencoder
Line current sensors
False data injection attacks
Topology identification
ISSN:0263-2241
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:In general, the solution of distribution system state estimation (DSSE) is highly dependent on the accuracy of measurement data and accurate topology data. Since the measurement devices are more vulnerable to different types of cyberattacks such as denial-of-service (DoS) attacks, outliers, and structured false data injection attacks (FDIAs), the accuracy of DSSE is significantly affected. In addition, the topology identification is also affected when the measurements are used for the task of topology detection. To address these issues, a novel denoising autoencoder (DAE) is developed with the use of graph-based temporal convolutional layers in the encoder and decoder stages. The graph-based temporal convolutional DAE (G-TCDAE) utilizes the graph knowledge and temporal convolutional structure to understand the spatial correlations and temporal correlations among the input measurements respectively, which enhances the accuracy of reconstructed data against cyberattacks, and classification tasks of topology and FDIA locations. A DAE model can reconstruct the measurement data when input data are having cyberattacks which are considered as noise. In addition, the same model can identify the correct topology as well as measurements having FDIAs. The G-TCDAE model’s performance is assessed and compared with other types of DAE models by performing simulation works on modified IEEE 13-node and IEEE 37-node distribution test systems. The robustness of the model’s capability is tested at different percentages of cyberattacks and results demonstrated that the G-TCDAE model is highly skillful in achieving efficient performance. Also, the model performance is not significantly affected by the scalability which is tested with the IEEE 123-node distribution system. •A novel graph-based temporal convolutional DAE is developed for cyberattacks.•The proposed model is utilized for the detection of stealthy FDIAs.•Also, the same model is leveraged for topology identification.•The model’s output is used for the reconstruction of FDIA and missing data.•The model performance is compared with other DAE-based models.
ISSN:0263-2241
DOI:10.1016/j.measurement.2025.117741