Abstract:
This paper presents a methodology for constructing a digital twin model of a vertical turbine used in hydroelectric power generation. The objective is to implement displacement virtual sensors in areas where physical measurements are impractical or unavailable, such as the clearance between the turbine runner and the discharge ring (water-gap). This is achieved by simulating the displacements and deformations of the shaft and support structure, based on on-line measured vibrations. The method consists in constructing a finite elements model which accurately represents the dynamic behavior of the hydrogenerator’s main components, including the shaft, runner, generator and support structure. During turbine operation, bearing displacements are measured and the model calculates rotor displacements at specified virtual sensor locations to assess potential violations of clearances between stationary and rotating parts. A case study is presented where the validity of the method is evaluated by comparing virtual sensors results with actual measurements during a generator trip event of a 92 MW Kaplan turbine, demonstrating excellent agreement. Notably, the digital twin detected excessive radial displacement of the runner during the event, indicating a potential violation of the water-gap. This result was corroborated by on-line physical water-gap measurements and visual inspection during maintenance, which revealed evidence of contact between the runner blades and the draft tube. This paper highlights the practical applicability of utilizing a digital twin for rotor dynamic behavior analysis of hydroelectric turbines, ultimately contributing to performance optimization, improved maintenance strategies, enhancing reliability and safety. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
