Fault Detection in Microgrid Using Artificial Intelligence

Abstract

newline Abstract newlineMicrogrids, composed of distributed energy resources (DERs) like solar, wind, and battery newlinestorage, enhance energy reliability, efficiency, and sustainability. They operate both in grid newlineconnected and islanded modes, ensuring power availability during main grid failures. However, newlinefault detection in microgrids is complex due to bidirectional power flow, inverter-based generation, newlineand fluctuating renewable energy sources. Traditional fault detection methods often lack the speed, newlineaccuracy, and adaptability required for modern microgrid operations, necessitating advanced AI newlinedriven solutions. newlineThis thesis investigates the application of Artificial Intelligence (AI) for fault detection in newlinemicrogrids, leveraging machine learning techniques such as artificial neural networks, decision newlinetrees, and random forests. The approach involves analysing phasor current signals and their newlinefundamental components under normal and abnormal conditions using Empirical Mode newlineDecomposition (EMD). By extracting key features from intrinsic mode functions (IMFs) and newlineapplying the Hilbert Transform, AI models are trained and tested on various fault scenarios. newlineSimulation results demonstrate that AI-based fault detection systems significantly enhance newlineaccuracy and response time compared to conventional techniques. The models effectively classify newlinefaults, improving the reliability and resilience of microgrid operations. This research highlights the newlinepotential of AI in microgrid protection and offers insights into integrating AI-driven fault detection newlineinto existing infrastructure. Future work should focus on real-time deployment, adaptive learning newlinetechniques, and optimizing AI models for large-scale microgrid networks, ensuring robust and newlineintelligent fault management systems. newlinevi

Description

Keywords

Citation

item.page.endorsement

item.page.review

item.page.supplemented

item.page.referenced