Deep learning framework for low-observable distribution system state estimation with multitimescale measurements

Zhang, Xihai, Ge, Shaoyun, Zhou, Yue ORCID: https://orcid.org/0000-0002-6698-4714 and Liu, Hong 2024. Deep learning framework for low-observable distribution system state estimation with multitimescale measurements. IEEE Transactions on Industrial Informatics 20 (11) , pp. 13273-13283. 10.1109/TII.2024.3435430

Preview

PDF - Accepted Post-Print Version Download (1MB) | Preview

Abstract

The deployment of microphasor measurement units, supervisory control and data acquisition systems, and smart meters has revolutionized distribution systems, moving toward more intelligence and facilitating state estimation. However, the asynchronous nature and varying sample rates of the measurements raise a significant challenge to state estimation in low-observable distribution systems. This article proposes a novel data-driven framework for state estimation in low-observable distribution systems. The framework incorporates super-resolution imputation techniques that are capable of handling multitime scale measurements. Specifically, the proposed framework utilizes Wasserstein divergence conditional adversarial networks to implement multivariate super-resolution imputation. This involves employing a gated recurrent unit generator with additive attention to impute super-resolution measurements, as well as a discriminator with self-attention to approximate the Wasserstein divergence between the imputed values and the ground truth. Furthermore, a tailor-made physical-guided bilinear neural network is employed to estimate the operational state of low-observable distribution systems by leveraging matrix factorization techniques to capture the low-rank characteristics. Simulation results illustrate the superiority of the proposed approach from the standpoint of multitimescale measurements imputation and state estimation in low-observable distribution systems.

Item Type:	Article
Date Type:	Publication
Status:	Published
Schools:	Engineering
Publisher:	Institute of Electrical and Electronics Engineers
ISSN:	1551-3203
Date of First Compliant Deposit:	23 September 2024
Date of Acceptance:	17 July 2024
Last Modified:	17 Dec 2024 13:30
URI:	https://orca.cardiff.ac.uk/id/eprint/171887

Actions (repository staff only)

Edit Item