Primjena neuronskih mreža u distantnoj zaštiti visokonaponskih vodova

scientific master's theses

• Title (english): Application of artificial neural networks to distance protection on HV lines
• Author: Džebel Kulenović
• Mentor: Ante Marušić (mentor)
• Committee member: Ante Marušić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2000, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering; Power Engineering
• Universal decimal classification (UDC): 621.3 - Electrical engineering

Abstract

Primjena neuronskih mreža u elektroenergetici, a jednako tako i u dostantnoj zaštiti, zadnjih je godina u stalnom porastu. Da bi se smanjio negativan utjecaj otpora na mjestu kvara u distantnoj zaštiti visokonaponskih vodova, u ovom je radu upotrebljen višeslojni perceptron, koji je učen Levenberg-Marquardt algoritmom. Poglavlja 2 i 3 obuhvaćaju poznata saznanja o razmatranoj problematici. U poglavlju 4 se opisuje razmatrana mreža i matematički model zasnovan na teoriji simetričnih komponenti. Poglavlje 5 objašnjava matematičke principe učenja pomoću algoritma sa povratnim rasprostiranjem pogreške i Levenberg-Marquardt algoritma. Poglavlja 6 do 8 prikazuju dobivene rezultate, analizu rezultata te zaključak. Višeslojni perceptron je ispitan simuliranjem pojave jednopolnog kratkog spoja uzduž dvostrano napajanog visokonaponskog dalekovoda, za različite iznose otpora na mjestu kvara te različite udaljenosti do mjesta kvara. Rezultati istraživanja su pokazali primjenjivost neuronskih mreža u distantnoj zaštiti i lokatoru kvara.

Abstract (english)

In the last few years use of artificial neural networks in power systems, and also in distance protection, is rapidly increasing. Multilayer perceptron with Levenberg-Marquardt algorithm are used to decrease negative influence of fault resistance in distance protection. Chapters 2 and 3 give the present knowledge concerning analysed topics. Chapter 4 explains the analysed network and the coresponding mathematical model based on symmetrical components. Chapter 5 explains mathematical background of multilayer perceptron. Basic and faster back-propagation lerning algorithm are also presented. Chapters 6 to 8 show the testing results, analisys of this results and conclusion. Multilayer perceptron with Levenberg-Marquardt algorithm are tested by simulating the occurrence of single phase-to-earth fault on different fault points along the HV line, varying the fault resistance. The testing results prove that the neural networks are useful to implement in distance protection and fault locator algorithm.

• Keywords: Neuronske mreže; višeslojni perceptron; algoritam sa povratnim rasprostiranjem pogreške; Levenberg-Marquardt algoritam; digitalna distantna zaštita; digitalni lokator kvara
• Keywords (english): Neural network; multilayer perceptron; back-propagation algorithm; Levenberg-Marquardt algorithm; digital distance protection; digital fault locator
• Language: croatian
• URN:NBN: urn:nbn:hr:168:937716
• Study programme: Title: Postgraduate master programme in electrical engineering; Study programme type: university; Study level: postgraduate; Academic / professional title: Magistar znanosti elektrotehnike (Magistar znanosti elektrotehnike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-07-22 07:50:03