NUMERICAL INVESTIGATION OF POTENTIAL FLOW INDUCED VIBRATIONS OF STEAM TURBINE LAST STAGE ROTOR AT LOW LOAD OPERATION - PART 2: ROTATING INSTABILITIES DETECTION

Steam turbines play an important role in global power generation since they are widely used, as thermal engines, in fossil-fueled, nuclear, and concentrated solar power plants. Therefore, recent trends in steam turbine design practices are closely related to the development of the energy market, which is especially focused on expansion of fast renewable energy. As a direct consequence, due to intrinsic variability of the green-energy resources, the steam turbines address the need to increase their flexibility to ensure the stable functioning of the power grid. Greater flexibility is linked to even increasing Low Volume Flow (LVF) operating conditions which could trigger dangerous non-synchronous aerodynamic excitations of the last stage bucket (LSB). In order to discover the source of such excitations, an extensive numerical study, presented in a two-part publication, has been carried out to investigate two different mechanisms potentially accountable for flow induced vibrations. In part one, the focus is on the flutter stability, while the present part deals with the detection of rotating instability phenomena that might arise in the last stage during LVF conditions. Such aerodynamic instabilities are investigated using CFD simulations by performing 3D, Unsteady Reynolds Averaged Navier Stokes (URANS) of the low pressure, last turbine stage coupled with an axial exhaust hood, with structural struts. The full annulus mesh of both the last stage and diffuser is considered with the transient stator-rotor interface to properly account for unsteady interaction effects. The influence of the operating conditions on the fluid dynamic behavior is assessed by considering six different operating conditions, starting from the design condition and gradually decreasing the mass flow rate. The presence of rotating instabilities is demonstrated by monitoring the fluid dynamic variables during the simulation and by using advanced post-processing techniques, such as Proper Orthogonal Decomposition (POD).