الفهرس يوجد فقط 14 صفحة متاحة للعرض العام
 |  | from | 94 |  |  |
| | | from | 94 | | |
المستخلص
ABSTRACT
Fault location algorithms aims to accunstely localize the fault by analyzing the data available after the inception of. disturt.nce. 1’bey also give infonnation ahout the location. possible cause and phases of the line involved in the fault. In this thesis OJ new fault location technique for transmission lines is dcscrihc..’C1. The technique is based on signal analysis using Prony method. and artificial neural networks.
The power system n\Odd used in this thesis consists of two sources: scnding-end source. recciving-end source. and 220 KV transmission line. The short circuit capacity of the equivalent thevenin’s sending and R.’CCiving end sources is considered to be 12 GV A and 4.5 GV A. respectively.
Prony method is used for analyzing the fault transient current signal. and extracting useful iaformation which can be used for training a neural netWork to locate the faults. The output of Prony analysis reprdine dominant fiequencies showed that there is • relation between 1bac fiequcncies and the fault location. The dominant ~uencies dre chosen based on criteria accounting for the damping factors and amplitudes.
1”he proposed neural networt is a feed forward neural network with two hidden layers. Dominant modal frequencies are used as the input patterns for the neural network. The training data is prepar.:d by simulating a large number of faults occurring at different &oc.tions, with various fault resist~ and inception angles. faults are simulated using PSCADIEMTDC program. The result of tl.!sting the neural net\’lork gives a percentage error less tbM 1% en•Of.
The proposed fault location techniql.&C. Prony analysis, neural network structure, training. and the results of testing the perfonnance of the neural network are given in this thesis.