Document Type

Theses, Ph.D

Rights

Available under a Creative Commons Attribution Non-Commercial Share Alike 4.0 International Licence

Disciplines

Electrical and electronic engineering

Publication Details

Thesis submitted for the Degree of Doctor of Philosophy to the School of Electrical and Electronic Engineering, Technological University Dublin, December 2015.

Abstract

The detection and fast clearance of faults are important for the safe and optimal operation of HVDC systems. In HVDC systems, various types of AC faults (rectifier & inverter side) and DC faults can occur. It is therefore necessary to detect the faults and classify them for better protection and diagnostics purposes. Various techniques for fault detection and classification in HVDC systems using signal processing techniques are presented and investigated in this research work. In this research work, it is shown that the wavelet transformation can effectively detect abrupt changes in system signals which are indicative of a fault. This research has focused on DC faults at various distances along the lines and AC faults on the converter side. The DC line current is chosen as the input to the wavelet transform. The 5th level coefficients have been used to identify the various faults in the LCC-HVDC system. Moreover, the value of these coefficients has been used for the classification of the different faults. For more accurate classification of faults, the wavelet entropy principle is proposed. In LCC-HVDC systems, a different approach for fault identification and classification is proposed. In this investigation an algorithm is developed that provides the trade-off between large input data size and minimal number of neurons in the hidden layer, without compromising the accuracy. The claim is confirmed by the results provided from the investigation for various fault conditions and its corresponding ANN output which confirms the specific fault detection and its classification. A fault identification and classification strategy based on fuzzy logic for VSC–HVDC systems is proposed. Initially, the developed Fuzzy Inference Engine (FIE) detects AC faults occurring in the rectifier side and DC faults on the cable successfully. However, it could not identify the line on which the fault has occurred. Hence, to classify the faults occurring in either AC section or DC section of the HVDC system, the FIE has to be restructured with appropriate data input. Therefore, a FIE which identifies different types of fault and the corresponding line where the fault occurs anywhere in the HVDC system was developed. Initially the developed FIE with three input and seven output parameters results in an accuracy level of 99.47% being achieved. After a modified FIE was developed with five inputs and seven output parameters, 21 types of faults in the VSC HVDC system were successfully classified with 100% accuracy. The FIE was further developed to successfully classify with 100% accuracy faults in Multi-Terminal HVDC systems.

DOI

https://doi.org/10.21427/D7MW3W


Share

COinS