The intensity and frequency of wildfires have increased rapidly in recent years in the US, particularly the Western states and other parts of the world. There is an increasing need for addressing the risk of wildfires initiated by faults in electrical equipment used in the power distribution system. Arcing faults can potentially create dangerous conditions such as sparks or manifest deteriorating equipment condition that potentially cause larger issues such as outages to the customers. Arcing faults are not well understood and consequently are difficult to identify with existing equipment.
High-resolution, high-fidelity sensor measurements can be used to detect unique signatures of electric power grid equipment malfunction and anomalies such as arcing faults that can potentially cause outages and wildfires. This LLNL invention proposes an unsupervised approach to detect arcing fault anomalies using a reduced dimensional gradient-based filtering algorithm followed by a similarity-based clustering algorithm. This approach would help identify events that are not related to arcing, such as normal voltage regulation events. Furthermore, the LLNL researchers propose a classifier-based approach to identify and validate an arcing event from the pool of potential arcing events.
I. Chakraborty and J. Joo, "Data-Driven Detection Of Low-Current Arcing Events In Power Distribution Systems," 2022 IEEE/PES Transmission and Distribution Conference and Exposition (T&D), New Orleans, LA, USA, 2022, pp. 01-05, doi: 10.1109/TD43745.2022.9816885 (IEEE Conference Publication)
Current stage of technology development: TRL ☒ 0-2 ☐ 3-5 ☐ 5-9
LLNL has filed for patent protection on this invention.
U.S. Patent Application No. 20230059561 IDENTIFICATION OF ARCING HAZARDS IN POWER DISTRIBUTION SYSTEMS published 2/23/2023
- Can identify anomalies from high-resolution sub-cycle measurements to detect arcing signatures in power distribution systems.
- Measurements can be obtained from a distribution substation using two types of devices: phasor and waveform, or point-on-wave (POW) sensors.
- Does not require defining detection threshold parameters as required by current anomaly detection approaches.
electric power delivery systems offered by electric utilities and their equipment suppliers
