Abstract:To better monitor and control distribution grids, the exact knowledge of system topology and parameters is a fundamental requirement. However, topology information is usually incomplete due to limited sensors in the grid. Therefore, estimating the system parameters using partial data is a critical topic for distribution systems. Due to the high nonlinearity of unobservable system quantities and noises, Deep Neural Networks (DNNs) are widely utilized for accurate estimation. While traditional approaches either treat DNNs as a black box or embed little physical knowledge into DNNs, they cannot guarantee that the DNN model is consistent with physical equations and hence lack accuracy and interpretability. Therefore, we propose a Deep-Shallow neural Network (DSN) for distribution system estimation. The key is to create virtual nodes to represent nodes without sensors in the system, and denote a DNN to approximate missing quantities at virtual nodes. Then, the Power Flow (PF) equations can be estimated via a shallow neural network, achieving physical consistency. Isolated by virtual nodes, the whole system is decomposed into a set of reduced graphs with approximate PF equations. Likewise, we introduce a Reinforcement Learning-based search algorithm to connect the reduced grids into one connected system. Correspondingly, the DSN is fine-tuned to achieve consistency with the PF equations of the connected graph. Finally, the paper illustrates the superiority of the proposed DSN due to its physical consistency. Specifically, comprehensive experiments demonstrate the high performances of our DSN over other methods in distribution grids.

Real-Time Topology Estimation for Active Distribution System Using Graph-Bank Tracking Bayesian Networks

Forecasting-aided Joint Topology/State Estimation for Distribution Systems via Bayesian Non-parametric Modeling

Full-Scale Distribution System Topology Identification Using Markov Random Field.

A Data-Driven Topology and Parameter Joint Estimation Method in non-PMU Distribution Networks

Fault Location in Active Distribution Networks Using Multiple Measurement-Based Bayesian Learning

Topology Identification and Line Parameter Estimation for Non-PMU Distribution Network: A Numerical Method.

Topology Identification for Distribution Network Driven by Diverse Measurement Data

Low-Voltage Distribution Grid Topology Identification With Latent Tree Model

Data-driven Distribution Network Topology Identification Considering Correlated Generation Power of Distributed Energy Resource

Distribution Network Topology Identification Using Smart Meter Data and Considering the Same-Bus-Different-Feeder Condition

Block-Sparse Bayesian Learning Method for Fault Location in Active Distribution Networks with Limited Synchronized Measurements

Distribution Grid Topology and Parameter Estimation Using Deep-Shallow Neural Network with Physical Consistency

Real-Time Distribution System Topology Monitoring with Limited Communication

Topology Estimation in Bulk Power Grids: Guarantees on Exact Recovery

Research on Real-time Automatic Acquisition Technology of Topology Data in Distribution Network

A Distributed Maximum-Likelihood-Based State Estimation Approach for Power Systems.

Robust Hidden Topology Identification in Distribution Systems

Fault Location in Active Distribution Networks Using Block Sparse Bayesian Learning

A Fully Distributed Topology Identification Approach for Active Distribution Network Based on Multi-Agent Framework

Real-time Topology Analysis Method for Power Network Based on Sparse PMU Placement

Data-Driven Approach for Distribution Network Topology Detection