Rapid Assessment of Distribution Line Damage via Autonomous UAV

Two-thirds of all electric outage minutes are weather-related, mostly the result of damage to above ground distribution lines. These outages cost the U.S. economy between $18 and $33 billion per year. We propose to rapidly assess line damage immediately following an extreme weather event using specially equipped unmanned aerial vehicles (UAVs). Efforts to restore power start with the utility’s spotters driving through flooded, blocked and unlit roads to document the extent of storm damage to distribution lines. In order to assess the inaccessible spans, the spotter can launch an autonomous UAV from their vehicle that will use stereoscopic 3D depth perception to sense and avoid obstacles as it maintains a fixed distance from the overhead lines, collecting images as it travels. Onboard algorithms will send a curated set of geo-tagged imagery to the spotter and the Incident Command Center. Other first responders will also be alerted to unsafe conditions such as downed wires and flooding. In Phase I, with the assistance of a major utility partner, we developed and field-tested pattern recognition algorithms to detect poles and wires. We flagged anomalies such as missing lines, downed poles and encroaching vegetation. We also developed an app to present this information to the spotter. In Phase II, we will detect more elements of interest, such as damaged insulators, broken cross arms and blown fuses. We will also deploy our technology on an autonomous UAV that can populate the utility’s asset management system with time-stamped aerial images of the damage and repairs. Ultimately, we will reduce outage minutes by helping utilities to prioritize work orders and deploy suitably equipped restoration crews.