Distribution Grid Line Outage Identification with Unknown Pattern and Performance Guarantee

Line outage identification in distribution grids is essential for sustainable grid operation. In this work, we propose a practical yet robust detection approach that utilizes only readily available voltage magnitudes, eliminating the need costly phase angles or power flow data. Given sensor data, many existing methods based on change-point require prior knowledge of patterns, which are unknown real-world scenarios. To remove impractical requirement, data-driven method to learn parameters post-outage through gradient descent. However, directly using descent presents feasibility issues. address this, modify our by adding Bregman divergence constraint control trajectory parameter updates, eliminates problems. As timely operation key nowadays, prove optimal can be learned with convergence guarantees via leveraging statistical and physical properties We evaluate representative real load profiles 17 configurations. The results show detect localize manner magnitudes without assuming patterns.