Please use this identifier to cite or link to this item:
|Title:||Machine learning based resolutions for partial discharge detection||Authors:||Xu, Ning||Keywords:||Engineering::Electrical and electronic engineering||Issue Date:||2022||Publisher:||Nanyang Technological University||Source:||Xu, N. (2022). Machine learning based resolutions for partial discharge detection. Master's thesis, Nanyang Technological University, Singapore. https://hdl.handle.net/10356/162346||Abstract:||Partial discharge (PD) is an important inducement of power failures such as insulation degradation. Effective, timely, and economical detection of PD is the basis for maintaining power system stability. This thesis presents novel resolutions for PD detection of medium voltage (MV) overhead power lines. The research is based on the large PD-related data set shared by Technical University of Ostrava (VSB). The data-driven machine learning-based approaches are employed for the complicated, nonlinear PD detection tasks. The first area of interest in this thesis is to propose a long short-term memory (LSTM) neural network based classifier with optimal feature extraction schemes for PD detection. It establishes a loop optimization process to greatly coordinate the data processing and the data-driven based machine learning algorithm. The combination of the different methods has combined advantages and can overcome each individual’s drawback. The second area of interest in this thesis is to propose a novel transformer-based multilevel filtering framework for PD detection. The proposed framework demonstrates superior performance with high robustness and less manual intervention.||URI:||https://hdl.handle.net/10356/162346||DOI:||10.32657/10356/162346||Rights:||This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0).||Fulltext Permission:||embargo_20231011||Fulltext Availability:||With Fulltext|
|Appears in Collections:||EEE Theses|
Files in This Item:
|4.72 MB||Adobe PDF||Under embargo until Oct 11, 2023|
Updated on Nov 25, 2022
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.