Flutter and Forced Response Computations of a Compressor Front Stage

Abstract

Nowadays, the trend towards designing lighter, more compact and highly efficient engines has become a major issue within the gas turbines industry. The compromise between size and/or weight becomes more significant for those engines destined for aviation purposes. Industrial gas turbines target is to continuously produce engines with longer operation time and increased efficiency. In general, these improvements have been proved to be achieved by introducing a series of changes on the early designs such as reduction of number of stages, different spool configuration or reduction of the axial gap between blade rows. However, some of the measures above presented introduce new challenges when designing a new engine. The reduction of the number of stages, resulting in more compact engines, leads to a significantly higher compression rate in the front part of the engine. A direct consequence of this modification is a more abrupt change in flow velocity and pressure distribution through the compressor. Gas turbines are then subjected to higher loading forces which compromise the operation and lifespan of their structural parts. Computational fluid dynamics provides a high fidelity approach for predicting the aeroelastic behaviour of turbomachinery components. The main issue investigated in this thesis covers the study validation of a high-fidelity CFD tool for flutter predictions on the first stage and a half of a compressor from a Siemens industrial gas turbine. Besides, forced response due to blade row interaction is further analyzed. The 1 1/2 stages is modeled with the commercial software STAR CCM+. The steady state solution is first obtained for different domain sizes. A comparison between different turbulence models is presented. Once the steady solution is reached, the most suitable mesh is selected for the later unsteady simulations. STAR CCM+ features a Harmonic Balance solver instead of the commonly used time-marching approaches, which allows to significantly reduce the computational time. Flutter motion for different nodal diameters is simulated with the Harmonic Balance solver and compared to previous results from different solvers for validation purposes. The novelty of the present work is found on the multistage configuration for flutter prediction opposed to traditional single row calculations. Additionally, blade row interaction effects are included in the flutter model to further investigate the aeroelastic response of the compressor blades subjected to downstream excitations.