In the present contribution the results of two three-dimensional viscous flutter analyses for a turbine cascade, Standard Configuration no. 11, are presented. The steady state and transient flow simulations were performed using the commercially available CFD solver ANSYS CFX 13.0 and a modified version of the CFD solver TBLOCK developed by Denton which is widely used in turbomachinery industry. The flutter analyses are performed under two different flow conditions. A subsonic, attached flow case and an off-design transonic case with a separated flow region near the trailing edge and a normal shock which are both located on the suction side. For each flutter analysis, the aeroelastic solution is computed for a large number of interblade phase angles. The results of ANSYS CFX and TBLOCK are compared to one another as well as to other CFD codes and experimental data. To reduce computing time, a phase-shifted boundary condition was implemented in TBLOCK. First results are shown in comparison to ANSYS CFX and its new implemented Fourier transformation method. The results of TBLOCK and ANSYS CFX agree well with experimental results. First results applying the phase-shifted boundary condition show that this method is suitable for calculating the aerodynamic damping with less numerical effort.

During the aeromechanical design process of turbomachinery blading, one of the main goals is to improve the blade loading which may lead to a higher risk of flutter. To avoid flutter induced blade failure during operation, the final blade design has to fulfill certain aero mechanical requirements. These refer to the permitted static and dynamic stress levels as well as the aeroelastic stability constraint of flutter for the whole operating range. In this contribution, an efficient workflow for three-dimensional viscous flutter stability analyses will be presented using the three-dimensional viscous flow solver TBLOCK and the open-source software package CalculiX for FE modal analyses. For this purpose, the workflow is applied to the first compressor rotor of a state of the art gas turbine. The flutter analysis is performed for several operating points to predict an accurate flutter envelope for the whole operating range of the investigated compressor stage. To reduce the numerical effort, only the first two mode shapes are considered with respect to different shaft speeds. In addition, phase-shifted boundary conditions are applied to all flutter calculations using the traveling wave mode domain taking all possible inter-blade phase angles into account. The results of the flutter analysis show no indications for flutter within the projected operating range of the rotor and for the considered mode shapes. In conclusion, the described workflow is able to determine the critical flutter stability boundaries of the investigated compressor rotor with reasonable numerical effort.

Usually, in a turbine an uneven number of blades are selected for vane and blade rows to reduce the level of interaction forces. To consider all unsteady flow phenomena within a turbine the computation of the full annulus is required causing considerable computational cost. Transient blade row methods using few passages reduce the numerical effort significantly. Nevertheless, those approaches provide accurate results. This contribution presents three different unsteady approaches to compare the accuracy and the computational effort, using a full annulus unsteady CFD simulation as a reference. The first approach modifies the blade-to-blade ratio whereas the second method scales the circumferential flow pattern to reach spatial and temporal periodicity. Third approach is based on time-inclining method to overcome unequal blade pitches with less numerical effort. All unsteady CFD simulations are carried out for the transonic test turbine VKI BRITE EURAM using the commercial CFD solver ANSYS CFX 14.5. The resulting unsteady pressure disturbances and blade forces of the different transient blade row methods are compared to each other as well as to experimental data. Finally, the accuracy and the computational costs are discussed in more detail.