Predviđanje gubitaka djelatne snage u niskonaponskoj distribucijskoj mreži metodama strojnog učenja

master's thesis

• Title (english): Forecasting active power losses in low voltage distribution networks by machine learning methods
• Author: Terezija Matijašević
• Mentor: Tomislav Capuder (mentor)
• Committee member: Tomislav Capuder (predsjednik povjerenstva)
• Committee member: Juraj Havelka (član povjerenstva)
• Committee member: Božidar Filipović-Grčić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-07-05, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

Brojne klimatske promjene koje su pogodile Europu utjecale su na donošenje novih zakona i politika kojima je cilj smanjenje stakleničkih plinova i povećanje energetske efikasnosti. Sve to utječe na tranziciju koja je u značajnijoj mjeri orijentirana prema distribucijskoj mreži i krajnjim korisnicima. Distribuirana proizvodnja, električna vozila i aktivniji korisnici imaju značajniji utjecaj na ovu mrežu koja tradicionalno prepoznaje jednosmjerne tokove snaga i pasivne potrošače. Takve niskonaponske mreže slabo su osmotrive te postoji mali broj alata i analiza za modeliranje tehničkih uvjeta i karakteristika u njima, a taj problem raste s većom integracijom novih sudionika u mrežu bližu krajnjim korisnicima. Kako bi se smanjila takva nepogodnost te omogućila energetska efikasnost i isporuka električne energije zadovoljavajuće kvalitete, potrebno je implementirati alate bazirane na metodama strojnog učenja.
U ovom diplomskom radu izrađen je model za simulaciju i analizu niskonaponskih distribucijskih mreža s naglaskom na trofazne tokove snaga i analize gubitaka djelatne snage. Paralelno s modelom tokova snaga izrađen je model za predviđanje opterećenja, koji je baziran na odabranoj SVR metodi. Oba modela korištena su kao ulazni parametri u novo razvijeni alat za predviđanje gubitaka djelatne snage u niskonaponskim distribucijskim mrežama.

Abstract (english)

Numerous climate changes that have affected Europe have influenced the adoption of new laws and policies aimed at reducing greenhouse gases and increasing energy efficiency. All this affects the transition, which is more oriented towards the distribution network and end-users. Distributed generation, electric vehicles, and more active users have a significant impact on this network, which traditionally recognizes unidirectional power flows and passive consumers. Such low-voltage networks are poorly observable and there are a small number of tools and analyzes for modeling technical conditions and characteristics in them, and this problem grows with greater integration of new participants into the network closer to end-users. In order to reduce such inconvenience and enable energy efficiency and delivery of electricity of satisfactory quality, it is necessary to implement tools based on machine learning methods.
In this thesis, a model for simulation and analysis of low voltage distribution networks with emphasis on three-phase power flows and analysis of active power losses was developed. In parallel with the power flow model, a load forecasting model was created, which is based on the selected SVR method. Both models were used as input parameters in a newly developed tool for predicting active power losses in low-voltage distribution networks.

• Keywords: tranzicija; novi sudionici; distribucijska mreža; strojno učenje; trofazni tokovi snaga; gubici djelatne snage
• Keywords (english): transition; new participants; distribution network; machine learning; three-phase power flows; active power losses
• Language: english
• URN:NBN: urn:nbn:hr:168:515315
• Study programme: Title: Electrical Engineering and Information Technology; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka elektrotehnike i informacijske tehnologije (magistar/magistra inženjer/inženjerka elektrotehnike i informacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-13 15:34:40