DC Arc Fault Signatures Analysis Based on Improved Bilinear Time-Frequency Representation

Series arc fault is widely concerned in DC distribution system, as they could introduce transient electromagnetic interference to nearby equipment and cause serious system failure to electrical system and signalling system. Extracting the signatures of series arc fault has an important role in arc-fault detection algorithms and protection techniques for sensitive equipment. Current studies show that characteristic analysis methods based on frequency domain are deficient in identifying arc fault from other transient emission [1] - [3] , therefore the time-frequency representation is considered as an efficient tool in extracting the arc fault signatures.