Istraživanje materijala primjenom strojnog učenja

master's thesis

• Title (english): Study of materials using machine learning
• Author: Mia Sitar
• Mentor: Goranka Bilalbegović (mentor)
• Committee member: Goranka Bilalbegović (predsjednik povjerenstva)
• Committee member: Mario Basletić (član povjerenstva)
• Committee member: Ivan Kupčić (član povjerenstva)
• Committee member: Mihael Srđan Grbić (član povjerenstva)
• Committee member: Danko Radić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Science; (Department of Physics); Zagreb
• Defense date and country: 2019-09-23, Croatia
• Scientific / art field, discipline and subdiscipline: NATURAL SCIENCES; Physics

Abstract

Iskorištavanje sunčeve energije predstavlja jedan od ključnih elemenata prelaska na čistu energiju. Perovskitni spojevi su medju najistraživanijim i najperspektivnijim materijalima za izradu solarnih ćelija. U ovom radu su algoritmima nadziranog strojnog učenja predvideni energijski procjep, standardna toplina stvaranja i najstabilnija faza kristalne rešetke perovskita. Korišteni su algoritmi L2 regularizirane linearne regresije i stabla odlučivanja. Pronađeni su perovskiti koji bi mogli biti pogodni kao gradivni materijal za solarne ćelije te su uspoređeni s onima koji se trenutno najčešće koriste u tu svrhu. U metodičkom dijelu rada opisano je nekoliko pokusa sa solarnim ćelijama koji bi se mogli izvesti u nastavi s ciljem naglašavanja važnosti solarne energije i informiranja učenika o temeljnim značajkama solarnih ćelija.

Abstract (english)

Exploitation of solar energy is one of the key elements for a clean energy transition. Perovskite compounds are among the most researched and promising materials for the solar cell production. In this diploma thesis supervised machine learning algorithms have been used to predict the band gap, heat of formation and the most stable phase of the crystal lattice for perovskites. Predictions were made with the L2 regularized linear regression and decision tree algorithms. Perovskites which might be suitable for the production of solar cells were found. They have been compared with ones currently used for this purpose. In the methodical section of this thesis a few solar cell experiments have been decribed, with the goal of emphasizing the importance of solar energy and introducing students to basic features of solar cells.

• Keywords: solarne ćelije; perovskiti; strojno učenje; linearna regresija; stabla odlučivanja
• Keywords (english): solar cells; perovskites; machine learning; linear regression; decision trees
• Language: croatian
• URN:NBN: urn:nbn:hr:217:559686
• Study programme: Title: Physics and Computer Science; specializations in: Education; Course: Education; Study programme type: university; Study level: integrated undergraduate and graduate; Academic / professional title: magistar/magistra fizike i informatike (magistar/magistra fizike i informatike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2020-01-07 09:44:51