Research on Recognizing Power Cable Fault Based on the ACCLN

Author

Xuebin Qin ; Mei Wang ; Xiaowei Li ; Jzau-Sheng Lin

Author_Institution

Control Eng. Dept., Xi´an Univ. of Sci. & Technol., Xi´an, China

fYear

2014

fDate

10-12 June 2014

Firstpage

219

Lastpage

222

Abstract

Power cable problem under the normal operating condition is very important in electric power systems. In practical operation, Various cable faults will happen. Recognizing the cable faults correctly and timely is very crucial. A method is proposed that an annealed chaotic competitive learning network (ACCLN) recognizes power cable types in this paper. The result shows a better performance than SVM and IPSO-SVM method. The result shows that the fault recognition accuracy reached 96.2% by 54 data test sample data. The network training time is about 0.032 second. The proposed method is applied for the cable fault classification effectively.

Keywords

chaos; fault diagnosis; learning (artificial intelligence); power cables; power engineering computing; simulated annealing; ACCLN; annealed chaotic competitive learning network; cable fault classification; chaotic simulated annealing; electric power systems; network training time; normal operating condition; power cable fault recognition; power cable problem; Annealing; Circuit faults; Entropy; Neurons; Power cables; Support vector machines; Training; ACCLN; IPSO-SVM; cable faults; power cable; recognition;

fLanguage

English

Publisher

ieee

Conference_Titel

Computer, Consumer and Control (IS3C), 2014 International Symposium on

Conference_Location

Taichung

Type

conf

DOI

10.1109/IS3C.2014.67

Filename

6845858