Model sudjelovanja solarne elektrane i vodikova postrojenja na tržištu električne energije

master's thesis

• Title (english): Model for joint electricity market participation of solar power plan and hydrogen facility
• Author: Filip Grebenar
• Mentor: Tomislav Capuder (mentor)
• Committee member: Tomislav Capuder (predsjednik povjerenstva)
• Committee member: Marko Delimar (član povjerenstva)
• Committee member: Mirna Gržanić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2023-07-07, Croatia

Abstract

Kontinuirano povećanje novoinstaliranog kapaciteta fotonaponskih elektrana (FNE) osim niza prednosti donosi različite tehničke i operativne izazove za postojeći elektroenergetski sustav budući da su varijabilnost i nepredvidivost inherentna obilježja solarne energije. Kompenzacija navedenih obilježja može se donekle ostvariti razvojem efikasnih modela predviđanja proizvodnje električne energije (EE) iz FNE, ali i korištenjem fleksibilnih jedinica kao što je vodikovo postrojenje u sklopu kojeg se nalaze elektrolizator, stog gorivnih članaka i spremnik vodika.

U ovome radu opisana je matematička formulacija odabranih modela strojnog učenja koji su zatim, koristeći različite metodologije predviđanja EE iz FNE, implementirani i evaluirani na stvarnim podacima na kojima je prethodno izvršena eksploratorna analiza. Rezultati predviđanja upotrebljeni su kao ulazni podaci u dvorazinski optimizacijski model s ciljem određivanja optimalnog nastupa bilančne grupe FNE i vodikova postrojenja na dan unaprijed tržištu minimizirajući pritom odstupanja u stvarnom vremenu koristeći Monte Carlo scenarijsku analizu ostvarenja proizvodnje EE iz FNE.

Abstract (english)

The capacity of newly installed solar photovoltaic (PV) power plants constantly increases worldwide, which, aside from certain environmental benefits, brings various technical and operational challenges for power system operators and market participants. The main driver of these challenges is the variable and intermittent nature of solar energy, which can be partially compensated by developing efficient and dedicated PV power forecasting techniques or through the addition of flexible units such as hydrogen facility composed of electrolyzer, fuel cell stack and hydrogen storage tank.

This paper describes and elaborates on the mathematical formulation of proposed machine learning techniques that were later implemented and evaluated on historical data using different methodologies. The obtained forecasting results are then fed into the input of a two-stage optimization model with the aim of determining the optimal day-ahead market scheduling of the balance group (PV power plant and hydrogen facility) while minimizing its real-time imbalances. Scenario analysis based on Monte Carlo method is adopted to model the uncertainty in PV power plant production.

• Keywords: Vodikovo postrojenje; Solarna elektrana; Modeli predviđanja; Umjetne neuronske mreže; Stroj potpornih vektora; Optimizacija; Dan unaprijed tržište; Financijsko poravnanje odstupanja; Analiza scenarija
• Keywords (english): Hydrogen facility; Solar power plant; Forecasting methods; Artificial neural networks; Support vector machine; Optimization; Day-ahead market; Imbalances settlement; Scenario analysis
• Language: croatian
• URN:NBN: urn:nbn:hr:168:842024
• 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: 2023-10-11 08:56:46