The Galileo Institutional Repository as an Information Stock in the Training of Masters and Doctorates Focused on Industry 4.0 in the Amazon

Abstract

Transformers are crucial components in contemporary power systems, ensuring efficient distribution and transmission of electrical energy. They pose a risk of internal faults, such as inter-winding short circuits, which are difficult to identify in real time with conventional techniques like thermal monitoring and gas dissolved analysis. The detected anomalies can severely impair transformer efficiency and result in expensive operational failures. This study introduces a technique for identifying winding short-circuit faults via vibration analysis, employing artificial neural networks (ANN) in conjunction with the Fast Fourier Transform (FFT). The method examines variations in vibration frequency as signs of potential failure and utilizes ANN to accurately classify various situations. Experimental results demonstrate that the proposed technique successfully differentiates between normal and defective conditions across various load scenarios, enabling rapid and accurate fault detection. The system's ability to continuously assess transformers without interruptions enhances operational efficiency, lowers maintenance expenses, and increases the overall precision of the power grid.