A rotor blisk of a supersonic space turbine has previously been designed to allow for free flutter to occur in an air test rig [1]. Flutter occurred at several operating conditions and the flutter boundary for the test turbine was established. In this paper the rotor blisk is redesigned in order to inhibit flutter. The design strategy chosen is to introduce a mistuning concept. Based on aeroelastic analyses using a reduced order model (ROM) a criterion for the required level of mistuning is established in order to stabilize the lower system modes. Proposals in literature suggest and analyze mistuning by varying blade mode frequencies in random patterns or by modifying blades in an odd-even pattern. Here a modification of sectors of the blisk is introduced in order to bring a sufficient split of the system mode frequencies. To verify that the redesigned blisk efficiently could inhibit flutter an experiment similar to that in [1] is performed with the mistuned rotor blisk. By running the redesigned blisk at operating conditions deep in to the unstable region of the tuned blisk it is demonstrated that a relative low level of mistuning is sufficient to eliminate rotor flutter.

A finite volume method for the computation of rotor/stator interaction for stages with arbitrary rotor/stator pitch ratios is presented and partly validated in this paper. The method which solves the unsteady three-dimensional Euler equations is formulated in the four-dimensional time-space domain. The method of time inclination is utilized to account for unequal pitchwise periodicity by distributing time co-ordinates at the grid nodes such that phase lagged boundary conditions can be employed. Calculated results show excellent agreement with the results of a reference solver for the validation test case. Furthermore the method was applied to the simulation of the unsteady flow field in a transonic test turbine stage with a stator/rotor pitch ratio of 1.875. The results were compared with measurements of the unsteady rotor blade pressure and a reference solver calculation where an approximate pitch ratio of 2.0 with a 6.7% scaled rotor geometry was employed. Both computational cases show satisfactory agreement with the experiments for both time averaged pressure distributions and pressure perturbation amplitudes.

In the design of modern compressor blades of wide chord (low aspect ratio) type it is often hard to avoid having modes that are close to each other in frequency. Modes which are closely spaced can interact dynamically. Mistuning and localization of stresses are known problems with this. A potential problem with this is also the possibility of coalescence flutter of the modes. Even if the modes are frequency separated at zero rotational speed, the centrifugal stiffening may cause the modes to attract and even cross (or veer) at some rotational speed. In design, mode separation criteria are sometimes applied in order to minimize the risk of encountering unknown dynamic phenomena. This study is performed to better understand the dynamics of closely spaced modes with respect to risk for coalescence flutter. A reduced order aeroelastic system is then constructed that describes the interaction between the different modes. The aeroelastic couplings are then calculated for the 2 mode system. The method is general in terms of mode shapes and number of interacting modes. A parametrical study is performed in order to study how strongly the modes interact when the frequency separation is decreased and if there is a risk of destructive coalescence flutter. The investigation is performed on a high pressure ratio front stage fan blade. The tendency of the modes to interact depends on the strength of the coupling compared to the strength of the pure structural modes. The tendency towards instability was increased in cases where the stability margin was smaller of the single modes. The results can be considered to support a separation criterion of 2% for the lower. A re-evaluation should be considered if lighter blade material and increased loads are to be used.

A method is proposed for HCF-analysis that is suitable for use in early design stages of turbomachinery blades. Quantitative measures of the risk for later encountering HCF life limiting vibrations are the goal for the development. The novelty of the system is the unique and rational way all design data are processed resulting in a mode risk priority listing. The method makes extensive use of FE calculated modal analyses and simple assumptions on the modal force and damping. The modal force is taken proportional to the tangential force on the blade over the operating range. This choice is made because the tangential force is known early on from the compressor performance map, and gives a reasonable scaling with the operating point. Crossings occurring at low speed get a lower force than at high speed. The system damping used is a constant critical damping ratio. Using a modal force and damping along with the FE model forced response amplitude can be directly computed at resonance crossings inside operating envelope. The modal force calculated this way can be compared to the force amplitude needed to reach the fatigue limit in a Haigh diagram. Using the Haigh diagram this way allows modes with localized high stresses, so-called hot spots, to be highlighted. Taking the ratio of the forces gives a ranking value that can be used to compare risk. Details of the technique along with example applications to compressor blades are presented in the paper. It is found that many mode crossings can be excluded as low risk this way and that a rational way of prioritizing is achieved.

A method is proposed for the determination of the aeroelastic behavior of a system responding to mode-shapes different to the tuned in-vacuo ones, due to mistuning, mode family interaction or any other source of mode-shape perturbation. The method is based on the generation of a data base of unsteady aerodynamic forces arising from the motion of arbitrary modes and uses Least Square approximations for the prediction of any responding mode. The use of a reduced order technique allows for mistuning analyses and is also applied for the selection of a limited number of arbitrary modes. The application on a transonic compressor blade shows that the method captures well the aeroelastic properties in a wide frequency range. A discussion of the influence of the mode-shapes and frequency on the final stability response is also provided.