FFT based anomaly detection in railway systems

The exposure to various environmental influences and different type of human impact carries a wide variety of harmful effects to many infrastructures. An efficient and precise technology to identify structural anomalies based in the vibrations of these infrastructures becomes essential to provide a safe environment. There is precisely where Predictive-Cognitive Maintenance (PCM)
(Araquistain (2024)) gains significance, specially in Advanced Integrated Railway Management as a transformative approach for ensuring the safety, reliability, and efficiency of these systems. To save computational effort and improve time efficiency effective Structural Health Monitoring (SHM) using low-cost sensor devices is required. At the same time, an efficient and cheap algorithm
in terms of computational resources would be beneficial to avoid excessive loads and costs. In that sense, in this work we propose a simplistic solution to
detect anomalies of the vibration data of a railway. We developed an algorithm based in the dissimilarities between the average frequencies of the main modes and the frequencies of the modes of a given vibration data. Experiments have shown that it is a useful solution without the need of complex algorithms involving heavy and time consuming Machine Learning (ML) tasks.