Analysis and Modelling of Faults in Solar Power Plant

Authors

  • Emilija Bareikaitė Vilnius University, Faculty of Mathematics and Informatics
  • Aušrinė Budzevičiūtė Vilnius University, Faculty of Mathematics and Informatics
  • Rūta Levulienė Vilnius University, Faculty of Mathematics and Informatics

DOI:

https://doi.org/10.31449/upinf.275

Keywords:

machine learning, modelling, solar power plants, statistical analysis

Abstract

Considering the rapid increase in electricity consumption and the necessity to transition to sustainable energy sources, the reliable operation of solar power plants has become particularly crucial. This study analyses time-series data from an active solar power plant, consisting of electrical parameters recorded every 5 minutes and corresponding fault labels. The primary focus is on the investigation of anomalies and the development of significant covariates that contribute to power plant failures. Subsequently, these variables are utilized in logistic regression models to predict operational disruptions, with the best model achieving an accuracy of 96.5%. The study is further enhanced by the application of a machine learning technique – random forests – which attained an accuracy of 98.4%. This research contributes to the more effective management of solar power plants, and aids in ensuring their reliable operation over the long term.

Author Biographies

  • Emilija Bareikaitė, Vilnius University, Faculty of Mathematics and Informatics
    Emilija Bareikaitė graduated magna cum laude with a Bachelor's degree in Data Science from Vilnius University, Lithuania, where she will continue to pursue a Master's degree in Data Science. She is an ML/AI Researcher at AAI-Labs. Her research interests include natural language processing, machine learning, statistical modeling, and AI applications. She has authored and co-authored scientific publications and presented her research at national and international conferences.
  • Aušrinė Budzevičiūtė, Vilnius University, Faculty of Mathematics and Informatics
    Aušrinė Budzevičiūtė received a Bachelor's degree in Data Science from the Faculty of Mathematics and Informatics at Vilnius University, Lithuania. Her research interests include machine learning, statistical modeling, and artificial intelligence. She has worked on the application of machine learning and statistical methods to solar power plant data analysis, including fault analysis and predictive modeling.
  • Rūta Levulienė, Vilnius University, Faculty of Mathematics and Informatics
    Rūta Levulienė is an associate professor at the Faculty of Mathematics and Informatics of Vilnius University in Lithuania. Her areas of research interest are mathematical statistics and its applications. She has over 20 years of teaching experience in various statistics courses. Recently, she participated in the AI2SEP project, a scientific project related to the COVID-19 pandemic, where statistical and machine learning methods were applied to forecasting and modeling. She is a member of the Council of the Lithuanian Mathematical Society and the Lithuanian Statistical Society.

Downloads

Published

2026-10-13

Issue

Online-first

Section

Scientific articles

License

Copyright (c) 1970 Applied Informatics

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

How to Cite

[1]
2026. Analysis and Modelling of Faults in Solar Power Plant. Applied Informatics. (Oct. 2026). DOI:https://doi.org/10.31449/upinf.275.