Prediction, Monitoring, and Fault Diagnosis of Hydropower Plants Using Digital Twin Model and Differential Evolution Algorithm

The maintenance of rotating machinery is critical to ensuring operational efficiency and safety in industrial systems. Among the main challenges is the early identification of faults, such as mechanical unbalance and misalignment. This work presents a fault detection methodology based on an inverse technique, combining a representative digital twin model with the Differential Evolution optimization algorithm. Numerical vibration responses are compared with experimental measurements, and faults are iteratively introduced into the digital twin model until both signals converge. The proposed methodology was applied to the digital twin of a generating unit at the Foz do Chapecó hydroelectric plant, demonstrating its ´ effectiveness in fault diagnosis.