The Optimal Classification of Partial Discharge Defects within XLPE Cable by Using ANN and Statistical Techniques
The classification of medium voltage cable defects is the most important tool to avoid the inaccurate Partial Discharge (PD) measurements. This paper presents a proposed methodology based on pattern recognition technique to classify the PD occurred in cables into two distinct groups: internal and ex...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Port Said University
2014-09-01
|
| Series: | Port Said Engineering Research Journal |
| Subjects: | |
| Online Access: | https://pserj.journals.ekb.eg/article_45254_c19f46cf4150d5a9c58c37e8346e454e.pdf |
| Summary: | The classification of medium voltage cable defects is the most important tool to avoid the inaccurate Partial Discharge (PD) measurements. This paper presents a proposed methodology based on pattern recognition technique to classify the PD occurred in cables into two distinct groups: internal and external. The Artificial Neural Networks (ANNs) with different input schemes have been built to obtain the optimal classification. Many statistical features, which extracted by different techniques from measurements, have established the input schemes 3D-pattern PD to improve the performance and classification speed of ANN. In order to obtain the effective statistical features, the study and comparison between all ANN has been finished by evaluating the classification through two parameters: the mean square error (MSE) and the accuracy of neural network. As a result, the proposed approach provides high recognition rate of classification between internal and external PD within Cross-Linked Polyethylene (XLPE) insulated medium voltage cable |
|---|---|
| ISSN: | 1110-6603 2536-9377 |