Research on Low-Voltage Series Arc Fault Detection Based on Higher-Order Cumulants

2014 ◽  
Vol 889-890 ◽  
pp. 741-744 ◽  
Author(s):  
Kai Yang ◽  
Ren Cheng Zhang ◽  
Jian Hong Yang ◽  
Xiao Mei Wu

Arc faults are one of the main reasons of electrical fires. For the difficulty of series arc fault detection, very few of techniques have successfully protected loads from arc faults in low-voltage circuits, especially in China. When series arc faults occur, shoulders will appear in load currents. The shoulder widths of non-arc faults such as normal load arcs are stable, while those of arc faults are variable because the appearance of arc faults is erratic. Therefore, a novel detection method based on shoulder characteristics was proposed. To better capture shoulder widths, original currents were firstly converted into pulses. Then, the pulse widths were used to detect arcs and the fourth-order cumulants of their differential were used to distinguish arc faults from normal operations. Finally, an arc fault detection device (AFDD) prototype was developed for test. The results show this prototype can discriminate arc faults effectively from normal operations.

IEEE Access ◽  
2019 ◽  
Vol 7 ◽  
pp. 4586-4597 ◽  
Author(s):  
Guanghai Bao ◽  
Run Jiang ◽  
Dejun Liu

2014 ◽  
Vol 651-653 ◽  
pp. 499-502 ◽  
Author(s):  
Ren Cheng Zhang ◽  
Kai Yang ◽  
Qi Yong Wu ◽  
Jian Hong Yang

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.


Author(s):  
Weihai Sun ◽  
Lemei Han

Machine fault detection has great practical significance. Compared with the detection method that requires external sensors, the detection of machine fault by sound signal does not need to destroy its structure. The current popular audio-based fault detection often needs a lot of learning data and complex learning process, and needs the support of known fault database. The fault detection method based on audio proposed in this paper only needs to ensure that the machine works normally in the first second. Through the correlation coefficient calculation, energy analysis, EMD and other methods to carry out time-frequency analysis of the subsequent collected sound signals, we can detect whether the machine has fault.


2016 ◽  
Vol 136 (11) ◽  
pp. 878-883 ◽  
Author(s):  
Kazunori Nishimura ◽  
Yusaku Marui ◽  
Satonori Nishimura ◽  
Wataru Sunayama

Sign in / Sign up

Export Citation Format

Share Document