Fault Detection and Classification of VSC-HVDC Transmission Lines Using a Deep Intelligent Algorithm | ||
| International Journal of Research and Technology in Electrical Industry | ||
| مقاله 1، دوره 2، شماره 2، 2023، صفحه 137-146 اصل مقاله (1.31 M) | ||
| نوع مقاله: Research Article | ||
| شناسه دیجیتال (DOI): 10.48308/ijrtei.2023.103583 | ||
| نویسندگان | ||
| Amir Inanloo salehi؛ Navid Ghaffarzadeh* | ||
| Electrical Engineering Department, Faculty of Engineering and Technology, Imam Khomeini International University, Qazvin, Iran | ||
| چکیده | ||
| Considering the sensitivity of high voltage direct current transmission system protection and the difficulty in identifying external DC faults, this paper presents two methods for fault detection and classification in VSC-HVDC transmission lines. These studied methods are evaluated in terms of efficiency and accuracy. This research is focused on DC faults at different distances along the lines. DC line current and voltage are selected as input to wavelet transform, and in the next step, unique and valuable features of each signal are extracted with modern signal processing methods, and then these features are used as input data for algorithms to detection and classification faults. Deep Neural Network (DNN) and Support Vector Machine (SVM) have been investigated to detect and classify faults. In the next step, the efficiency of these algorithms was investigated and analyzed in noise conditions. The innovation of this research is replacing a new method of extracting features from the fractal dimension, which has been used to study more prominent features, and improve performance with a small number of study data and considering different conditions, and using the new feature extraction method and improving the performance of the algorithms, 96% accuracy has been achieved. | ||
| کلیدواژهها | ||
| Fault detection؛ HVDC؛ Machine learning؛ Deep learning؛ Fault classification؛ Transmission line protection | ||
|
آمار تعداد مشاهده مقاله: 215 تعداد دریافت فایل اصل مقاله: 356 |
||