Machine Learning-Driven Three-Phase Current Relay Protection System for Small Transient Periods in Sustainable Power Systems

This study focuses on improving the effectiveness of three-phase current relay protection systems, which is a significant problem. It is achieved through the use of machine learning techniques. The primary objective is to enhance defect detection capabilities by utilizing artificial neural networks (ANNs). Especially during brief, intermittent durations. The process encompasses ameticulous design of the relay protection system, delineating crucial elements such as relays, sensors, and communication infrastructure. Artificial neural networks (ANNs) are vital to machine learning methodologies since they are employed to represent intricate connections within present patterns. This gets around the troubles that emerge. Encompassing conventional relay protection techniques. The dataset employed for training and testing comprises numerous transitory circumstances. It facilitates the development of strong and resilient model training. The results exhibit amazing performance increases. With accuracy surpassing 95%. The proposed technique is demonstrated to be superior by conducting a comparative study against traditional relay protection systems. The system displays resilience and generalization. Including sensitivity to parameter adjustments. It affirms its efficacy in varied operational situations. Future work requires further development of the machine learning model. Dataset expansion. Include in the research of new technologies for real-time application situations. This research contributes to enhancing relay protection systems and supporting grid stability. Including dependability in contemporary power networks.