Usage of Machine Learning Methods to Locate Short Circuits in an Onshore Wind Farm

• Grant number: 24/23217-0
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: June 01, 2025
• End date: May 31, 2026
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Mário Oleskovicz
• Grantee: Miguel Rodrigues Fonseca
• Host Institution: Escola de Engenharia de São Carlos (EESC). Universidade de São Paulo (USP). São Carlos , SP, Brazil

Abstract

With the increase in the use of Inverter-Based Resources (IBRs) in the power grid and in the context of renewable energies such as wind and solar, new challenges have been added to the traditionally used fault location methods (short circuits), since the characteristics of the IBRs contributions to faults are influenced by the inverters used to connect these generators to the power grid. From this perspective, studies analyzing the impact of IBRs on these fault location methods are still scarce in the literature, as well as new proposals that take these new impacts into account. In this context, aiming at the analysis and better handling of these problems, the increased availability of data obtained from transmission systems becomes a potential ally in the use of Machine Learning (ML) methods, as such methods rely on massive databases for proper training. Thus, with this project, it is intended to explore a methodology that encompasses ML methods for fault location in a real onshore wind farm, evaluating the performance of different ML methods available in the Scikit-Learn library developed in Python for this purpose. For this, a pre-existing database will be used, which includes waveform data from several fault scenarios obtained through computational simulations using the PSCAD software. The performances of different ML methods will be compared in order to determine which ones will be the most efficient and appropriate for developing a future methodology for electrical fault location.