Series Arc Fault Identification in DC Distribution Based on Random Forest Predicted Probability
Vu Le, Chad Miller, Bang‐Hung Tsao, Xiu Yao
Abstract
Series dc arc fault generates arcing noise crosstalk to adjacent electrical loads, making it challenging to identify the arc fault location, which creates a fire hazard that endangers the system’s safety. This study proposes a series arc fault identification method in a dc zonal electrical distribution (ZED) using random forest (RF)-based local detectors (LDs) to monitor constant power loads (CPLs), and output predicted nominal and arc fault probabilities. The predicted probabilities are then sent to a centralized master detector (MD) to obtain the final decision. With full communication capability among all LDs, the MD makes the final decision using RF algorithms. If there is any disconnection in the communication links, the LD can operate independently using the predicted arc fault probability to output the flag signal autonomously, while the MD continues to operate with other LDs. The proposed fault identification method is experimentally verified with a ZED testbed that comprises three CPLs.