Abstract
In electrical fires, arc fault is one of the important reasons. In virtue of cross talk, randomness and weakness of arc faults in low-voltage circuits, very few of techniques have been well used to protect loads from all arc faults. Thus, a novel detection method based on BP neural network was developed in this paper. When arc faults occur in circuits, current integrations of cycles were variable and erratic. However, current integrations of cycles would also vary while the working conditions of circuits change. To better discriminate the current integrations, four characteristics were extracted to represent their differences through chain code. Based on these characteristics, BP neural network was used to distinguish arc faults from normal operations. The validity of the developed method was verified via an experimental platform set up. The results show that arc faults are well detected based on the developed method.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call