COUPLED FLUID STRUCTURE SIMULATION METHOD IN THE FREQUENCY DOMAIN FOR TURBOMACHINERY APPLICATIONS

Abstract

Turbomachinery components are exposed to unsteady aero- dynamic loads which must be considered during the design pro- cess to ensure the structural mechanical integrity. There are two primary mechanisms which cause structural vibrations and can lead to high-cycle fatigue due to high dynamic stresses: flutter (self-excited vibrations) and forced response (forced excitation, e.g. wakes from upstream blade rows). In this work an emerging numerical frequency-domain method which is designed to effi- ciently simulate coupled fluid-structure interaction (FSI) prob- lems considering nonlinearities in the flow and structure is mod- ified and applied to an academic and a realistic test case. Fur- thermore complex structural eigenmodes are considered instead of purely real modes as was demonstrated in the literature so far. This method is able to predict limit cycle oscillations and forced response amplitudes. The coupled solver uses the Har- monic Balance (HB) method with an alternating frequency time approach to model periodically unsteady flows and structure dy- namics. The resulting nonlinear HB equations of the flow are solved with a pseudo-time stepping method while the nonlinear HB equations of the structure are solved with a Newton method. The dynamics of the involved structure are further simplified by considering only one relevant eigenmode of the structure. The method is applied to a 3D axial turbine configuration with a mod- ified Youngs modulus for the material of the blisk. The standard flutter curve of the blade row shows that at least one eigenmode is aerodynamically unstable at certain nodal diameters. As a first model test case for the harmonic balance solver, the non- linear structural damping is defined as a cubic modal damping term. The results of the frequency-domain method are compared to coupled FSI simulations in the time domain. The analysis shows that the frequency-domain method is very promising in terms of both computational efficiency and accuracy.