A REDUCED ORDER MODEL BASED ON AERODYNAMIC INFLUENCE COEFFICIENTS FOR AEROELASTIC COMPUTATIONS IN TRANSONIC COMPRESSORS

Reliable and computationally affordable aeroelastic simulations of compressor blades are of capital relevance today, as modern turbomachinery for aircraft propulsion are expected to meet stricter goals on weight saving and fuel efficiency through slender and highly-loaded geometries with a reduced number of blades, that may ultimately lead to a higher flutter susceptibility. In this paper, an aeroelastic reduced order model (ROM) based on the aerodynamic influence coefficient (AIC) technique for flutter simulation in transonic compressor blades is adopted to compute the aerodynamic damping for two resonance conditions. The ROM is capable of performing flutter simulations in both an uncoupled and a coupled way. In the former, the motion of blades is imposed as a travelling wave vibration pattern, with constant amplitude, frequency and phase, and the exchanged work between the fluid and the blade is used to assess stability; in the latter, a structural and an aerodynamic subsystem are built for each blade, exchanging displacements and forces respectively at the blade surface in a time-marching model. The key point in model order reduction is the accurate representation of unsteady aerodynamic forces. To this purpose a training simulation is setup, which relates the displacement of a blade to the induced aerodynamic loads on the neighbouring ones, called aerodynamic influence coefficients. In the uncoupled model the unsteady pressure field on the blade surface is adopted as AIC, while in the coupled version the AIC are the modal aerodynamic forces, related to the motion of blades with system identification techniques. The ROM is applied to the TU-Darmstadt Open Test Rotor geometry, a representative case of modern transonic compressor designs. Two resonance conditions are selected, where the crossing between a natural frequency and the rotational speed allows the experimental measurement of aerodynamic damping of the selected mode at several mass flows through a blade tip-timing system. Validation of the CFD model is performed by computing the compressor map in steady conditions and comparing the data against experimental measurements. After that, the uncoupled ROM accuracy is assessed, by comparing the aerodynamic damping for the two resonances at different mass flows and IBPA to full order flutter simulations adopting Fourier transformation and harmonic balance methods. The coupled ROM is then setup and compared to experimental data on damping ratio at the same conditions. The results prove the accuracy of the developed ROM, as well as a consistent reduction in computational cost compared to full order calculations.