سال انتشار: ۱۳۸۹

محل انتشار: پنجمین کنفرانس پایش وضعیت و عیب یابی

تعداد صفحات: ۶

نویسنده(ها):

Javad haddadnia – Associate Professor Department of Engineering, Sabzevar Tarbiat Moallem University, Sabzevar,Iran
Arash Khakzadian – Msc Student
Hassan Moslemi – Msc Student

چکیده:

This paper presents a new intelligent method for electrical equipment fault detection in power distribution networks based on thermo images by using of Legendre Moment (LM) for feature extraction and neural network for classifier. There are several types of faults in substations which transformer bushing breakdown, loose connection between conductor and section insulator, fuse deficiency and destruction of cable bug are the most commonly faults in substations that have been selected in this paper. This method has been conducted on practical thermo images from the electrical distribution network of Tehran. Simulation results indicate the validity of approach method with accuracy of 97.2 percentages using of MLP neural network