Renewable energies represent one of the most effective responses to the environmental and energy challenges of our time. In a world increasingly focused on sustainability and reducing CO2 emissions, the transition to clean energy sources has become a necessity. Among these, renewable energies stand out for their ability to harness inexhaustible natural resources, such as the sun, wind, Earth’s heat, and water. Among the oldest and most efficient renewable sources is hydroelectric energy, which utilizes the power of water to generate electricity. This technology, based on the use of rivers and water reservoirs, has enabled energy production in a reliable and sustainable way for centuries. Today, it is one of the leading resources for global energy production, accounting for more than 15% of the world’s electricity. In recent years, the power grid has been increasingly supplied with energy generated from renewable sources. Hydropower stands out as one of the most reliable and highquality energy sources. Water reservoirs associated with dams can be conceptually likened to large-scale energy storage systems, where energy is released on demand through the controlled discharge of water masses across hydroelectric turbines. Hydroelectric turbines are engineered to operate within specific design parameters that ensure optimal efficiency and structural integrity. However, deviations from these optimal conditions known as off design conditions can lead to significant challenges. Such conditions often result in mechanical vibrations and complex fluid dynamic phenomena, including turbulence and vortex formation, which adversely affect turbine performance and longevity. These instabilities cause pressure fluctuations and unsteady forces on turbine components, contributing to mechanical vibrations and potential structural damage. The objective of this study was to analyze the off-design operating conditions of a hydraulic turbine, particularly during low energy demand periods when full-load operation is not required. In such conditions, the turbine is forced to work outside its optimal design range, leading to a decrease in efficiency. To address this challenge, the study leverages advanced data-driven techniques, including machine learning and artificial intelligence, to develop a virtual sensor capable of detecting early signs of potentially harmful dynamic behavior. This sensor enables a preventive approach to turbine monitoring by identifying patterns and anomalies associated with off-design operation, even before they evolve into critical failures.

Incoming