Detecting Unobservable Contingencies in Active Distribution Systems Using a Stochastic Hybrid Systems Approach

TL;DR

The paper addresses the challenge of detecting unobservable contingencies in active distribution systems with limited measurements by formulating a distributed stochastic hybrid system (SHS) model that couples PV-BESS dynamics, network topology, and loads. It introduces a segmentation approach with auxiliary buses to enable distributed monitoring and develops MaMI (Magnitude-Modulated Input) probing to overcome indistinguishability across contingencies. A contingency-detection strategy estimates unknown initial conditions and uses a switching-state comparison to identify the most likely contingency from real-time measurements. Simulations on a PV-BESS-equipped distribution network demonstrate that MaMI can reliably distinguish and track switching contingencies, offering a scalable, practical tool for enhancing resilience in low-inertia grids.

Abstract

This paper introduces a distributed contingency detection algorithm for detecting unobservable contingencies in power distribution systems using stochastic hybrid system (SHS) models. We aim to tackle the challenge of limited measurement capabilities in distribution networks that restrict the ability to detect contingencies promptly. We incorporate the dynamics of distribution network connections, load feeders, PV, and battery energy storage system (BESS) hybrid resources into a fully correlated SHS model representing the distribution system as a randomly switching system between different structures during contingency occurrence. We show that jumps in the SHS model correspond to contingencies in the physical power grid. We propose a probing approach based on magnitude-modulation inputs (MaMI) to make contingencies detectable. The effectiveness of the proposed approach is validated through simulations on a sample distribution system.

Figures (6)

• Figure 1: Architecture of the active distribution system.
• Figure 2: A generic representation of each segment of the distribution system.
• Figure 3: The PV-B Integrated Distribution System.
• Figure 4: Eigenvalue analysis under line 1-4 variations for normal, short circuit, single line outage, and line disconnection.
• Figure 5: System responses for probing input ($y_p$) and random initial conditions for $\alpha_0$ to $\alpha_3$ ((a)-(d)), and System zero initial responses to probing input for $\alpha \in \mathcal{S}$ (e).