An Integrated DC Series Arc Fault Detection Method for Different Operating Conditions

Series arc fault (SAF) has severe impacts on the safety of dc power supply systems. Timely and accurate SAF detection under different operating conditions is an open challenging problem. To address this problem, article proposes integrated method for conditions. In proposed method, dual-tree complex wavelet transform (DT-CWT) employed to obtain current signal decomposition. The singular values each component are then extracted by using improved matrix construction which can effectively reduce computational cost constructing high-dimension features. Finally, kernel extreme learning machine (KELM) applied fuse feature information detection. A series experiments presented demonstrate effectiveness method. results offline experiment show that accuracy higher than six state-of-the-art methods article, embedded into hardware experimental platform online in-service implementation. achieves fast and, at same time, offers outstanding reliability stability in system dynamic transients.