Wavelet and artificial intelligence application to automated fault analysis /
A normal operation of an interconnected power system is very important to the whole society. When a power system fault occurs, it must be corrected in a timely manner to ensure the continual supply of power to industries and residential customers. Detection and classification of the type of power sy...
| Main Author: | |
|---|---|
| Format: | Thesis eBook |
| Language: | English |
| Published: |
[Place of publication not identified] :
[publisher not identified] ;
2001.
|
| Subjects: | |
| Online Access: | Link to OAKTrust copy |
| Summary: | A normal operation of an interconnected power system is very important to the whole society. When a power system fault occurs, it must be corrected in a timely manner to ensure the continual supply of power to industries and residential customers. Detection and classification of the type of power system fault are two main functions of a power system fault analysis. Previously, a solution framework utilizing Fourier transform and well-defined mathematical models for the power system has been adopted for the analysis of power system. Though this solution framework has been successfully employed for several years, it has some drawbacks that need to be addressed. This thesis investigates two alternative techniques: the wavelet transform and the fuzzy-neuro system for improvements in detection and classification of power system faults. Wavelet transform separates a signal into components like Fourier transform. However, it employs analyzing functions that are localized both in time and frequency domains. The ability of wavelet functions to focus on short time intervals for high frequency components and long time intervals for low frequency components improves the analysis of signals such as power system transients with localized oscillations and impulses, particularly in the presence of base frequency and low order harmonics. A detailed mathematical model of the interested power system is essential in the traditional power system fault analysis approach. With the increased complexity of modern power system, this model can not be always obtained. Also the variety of operating conditions increase the uncertainties of the topology of power network, making finding a solution more difficult. The neural network is well suited for the classification problem with nonlinearity. It does not require a precise model in the problem domain. Fuzzy logic represents knowledge with ambiguity. It can handle the uncertainties well. The hybrid intelligent system based on the fuzzy logic and neural network utilize the virtue of the individual techniques while overcoming their respective drawbacks. It is used as the technique for the classifier proposed in this thesis. In this thesis, the proposed classifier is implemented in Matlab. The simulation data is obtained with ATP (Alternative Transients Program). The results and analysis of the classifier are also presented. |
|---|---|
| Item Description: | "Major subject: Electrical Engineering". Vita. |
| Physical Description: | x, 65 leaves : illustrations ; 28 cm. Also available online. Issued also on microfiche from Lange Micrographics. |
| Bibliography: | Includes bibliographical references (leaves 61-64). |