Centralized reinforcement learning-based voltage regulation in distribution networks is becoming increasingly difficult due to the growing penetration of distributed energy resources, high computational burden, repeated power flow calculations, and increasing privacy concerns. This paper proposes a physics-informed fully distributed reinforcement learning framework that enables autonomous voltage regulation using only local smart meter data. A Thevenin-equivalent-based local voltage estimation model and a hybrid correction mechanism are developed to support accurate local decision-making without synchronized global measurements or centralized power flow solvers. A lightweight coordination mechanism is further introduced to refine the actions of independently trained local agents. Case studies show that the proposed framework reduces voltage violations by approximately 80%, achieves performance close to that of power flow-based training environments, and achieves a training speedup of about 6×[jls-end-space/]. The results also indicate that the relaxation factors in the reward function and the coordination scaler are critical to voltage regulation efficiency, whereas the discount factor has a smaller impact. These findings demonstrate the practicality of the proposed framework for privacy-aware fully distributed voltage regulation.

To correct outdated and incomplete topologies in low-voltage distribution networks (LVDNs) using only voltage magnitude measurements, a data-driven approach is developed by integrating machine learning algorithms with correlation analysis. Similar to existing data-driven topology identification and correction methods, the proposed approach exploits smart meter data to infer topology information. However, unlike many conventional approaches that require repeated preprocessing, multiple data sources, or separate procedures for different topology elements, it provides a unified framework that consistently uses the same up-to-date voltage magnitude dataset across all processing stages. Specifically, switch states are identified via supervised learning, while user–feeder connections and customer phase labels are refined using a modified hierarchical clustering algorithm. To address the similarity among smart meter data induced by distributed photovoltaic (PV) systems, a time-based data selection strategy is incorporated into the correlation analysis. The feasibility and robustness of the proposed approach are validated using modified real-world LVDNs and multiple incomplete smart meter datasets collected from customers in the Netherlands. The results demonstrate that the proposed approach can effectively mitigate the impact of PV-induced similarity on phase identification and improve topology correction performance. Although the approach is designed for topology correction rather than full topology reconstruction, the corrected topology improves network observability and supports distribution system operators in load balancing and PV integration.

With the increasing availability of smart meter (SM) data and the frequent lack of accurate network topology information, model-free power flow (PF) calculation has gained traction, often leveraging artificial neural networks (ANNs). However, training such models typically requires large volumes of SM data, raising significant privacy concerns for households in distribution networks. To address this challenge, we propose a privacy-preserving PF calculation framework that incorporates two local privacy-enhancing mechanisms: a Local Randomisation Strategy (LRS) and a Zero-Knowledge Proof (ZKP)-based data collection strategy. The LRS provides irreversible transformation of power data, ensuring strong privacy protection while preserving data utility. In parallel, the ZKP-based strategy enables secure and trustworthy voltage data collection, allowing smart meters to interact with distribution system operators without disclosing actual voltage magnitudes. To address performance degradation caused by seasonal variations in load profiles, we further integrate an incremental learning strategy into the online application. Extensive evaluations across three datasets demonstrate that the proposed framework can efficiently collect one month of SM data within one hour while maintaining most voltage magnitude estimation errors lower than 0.01 p.u. under varying measurement noise and seasonal conditions.

The topology of low-voltage distribution networks (LVDNs) is crucial for system analysis, e.g., distributed energy resources (DERs) integration, network hosting capacity analysis, state estimation, and electric vehicle charging management. However, it is frequently unavailable or incomplete. This paper develops a data-driven topology identification approach for LVDNs with a high proportion of underground cables. The proposed approach exploits the fact that underground cables usually follow the street pattern, thus relying on open street map (OSM) and smart meter (SM) data. Three stages compose the proposed approach: In the first stage, a hierarchical minimum spanning tree algorithm is proposed to generate the initial topology with an accurate number of sub-branches from the pre-processed OSM data and peak demand. In the second stage, based on the limited SM data, the location of breakpoints in mesh topology caused by circle roads is verified and reconstructed to guarantee the radial structure of LVDNs. Finally, given multiple incomplete SM datasets, three data-driven optimization models based on a state estimation model are constructed to mitigate the error of cable length induced by OSM data. The feasibility of the proposed topology identification approach is verified on three actual LVDNs in The Netherlands and multiple incomplete SM datasets. Furthermore, the minimal amount of SM data needed to minimize the error of cable length is analyzed.

Low-voltage distribution networks (LVDNs) topology is significant for distributed energy resources (DERs) integration, and network operation management, among others. However, topology identification is a difficult task due to the outdated recordings of networks, the uncertainty of DERs and data privacy. To address this issue, a data-driven topology generation approach is proposed based on open GIS and voltage magnitude data. The proposed approach aims to generate a topology with an accurate number of main feeders and sub-branches for adjacent substations. The boundaries between adjacent substations are first identified by using hierarchical clustering (HC) to cluster normalized voltage magnitude. Given the boundaries and the location of LV transformers, a hierarchical minimum spanning tree algorithm (HMST) is adopted to generate graph topologies using GIS data, which simultaneously verifies the number of cables under the streets. Finally, the endpoints of each feeder are estimated by clustering the transformed Pearson correlation coefficient of voltage magnitude. The feasibility of the proposed approach is evaluated on two real LVDNs in the Netherlands.